ExtremeXOS Concepts Guide 12.0 - 100262-00 Rev 01

ExtremeXOS Concepts Guide
Software Version 12.0
Extreme Networks, Inc.

3585 Monroe Street
Santa Clara, California 95051
(888) 257-3000
(408) 579-2800
http://www.extremenetworks.com
Published: 2007
Part number: 100262-00 Rev 01
AccessAdapt, Alpine, BlackDiamond, ESRP, Ethernet Everywhere, Extreme Enabled, Extreme Ethernet Everywhere,
Extreme Networks, Extreme Standby Router Protocol, Extreme Turbodrive, Extreme Velocity, ExtremeWare,
ExtremeWorks, ExtremeXOS, the Go Purple Extreme Solution, Sentriant, ServiceWatch, ScreenPlay, Summit,
SummitStack, Unified Access Architecture, Unified Access RF Manager, UniStack, Universal Port, the Extreme
Networks logo, the Alpine logo, the BlackDiamond logo, the Extreme Turbodrive logo, the Summit logos, the
Powered by ExtremeXOS logo, and the Color Purple, among others, are trademarks or registered trademarks of
Extreme Networks, Inc. or its subsidiaries in the United States and/or other countries.
Specifications are subject to change without notice.
Adobe, Flash, and Macromedia are registered trademarks of Adobe Systems Incorporated in the U.S. and/or other
countries. AutoCell is a trademark of AutoCell. Avaya is a trademark of Avaya, Inc. Internet Explorer is a
registered trademark of Microsoft Corporation. Merit is a registered trademark of Merit Network, Inc. Mozilla
Firefox is a registered trademark of the Mozilla Foundation. sFlow is a registered trademark of sFlow.org. Solaris
and Java are trademarks of Sun Microsystems, Inc. in the U.S. and other countries.
All other trademarks, registered trademarks, and service marks are property of their respective owners.
© 2007 Extreme Networks, Inc. All Rights Reserved.
2 Extreme XOS 12.0 Concepts Guide

Contents
Preface......................................................................................................................................... 29
Introduction .............................................................................................................................29
Terminology........................................................................................................................29
Conventions..............................................................................................................................29
Platform-Dependent Conventions ..........................................................................................30
Text Conventions.................................................................................................................30
Related Publications .................................................................................................................31
Using ExtremeXOS Publications Online .................................................................................31
Part 1: Using ExtremeXOS

Chapter 1: ExtremeXOS Overview ................................................................................................... 35
Platforms and Required Software Versions ...................................................................................35
Software Licenses and Feature Packs..........................................................................................37
Software Licensing Overview ................................................................................................37
Software License Levels .......................................................................................................38
Obtaining a License Voucher ................................................................................................40
Enabling and Verifying Licenses ...........................................................................................40
Security Licensing...............................................................................................................41
Feature Packs Overview .......................................................................................................41
Obtaining and Enabling Feature Packs ..................................................................................42
Displaying Software Licenses and Feature Packs ....................................................................43
Chapter 2: Accessing the Switch.................................................................................................... 45

Understanding the Command Syntax...........................................................................................45
Syntax Helper .....................................................................................................................46
Command Shortcuts ............................................................................................................47
Names ...............................................................................................................................47
Symbols .............................................................................................................................48
Limits ................................................................................................................................48
Port Numbering ........................................................................................................................48
Stand-alone Switch Numerical Ranges ..................................................................................49
Modular Switch and SummitStack Numerical Ranges .............................................................49
Stacking Port Numerical Ranges...........................................................................................49
Line-Editing Keys......................................................................................................................50
Command History......................................................................................................................50
Common Commands..................................................................................................................50
Accessing the Switch the First Time............................................................................................53
Safe Defaults Setup Method.................................................................................................53
Configuring Management Access ................................................................................................54
Account Access Levels.........................................................................................................54
Configuring the Banner ........................................................................................................55
Startup Screen and Prompt Text ...........................................................................................55
Extreme XOS 12.0 Concepts Guide 3

Contents
Default Accounts.................................................................................................................57
Creating a Management Account...........................................................................................57
Failsafe Account .................................................................................................................58
Managing Passwords .................................................................................................................59
Applying a Password to the Default Account ..........................................................................59
Applying Security to Passwords.............................................................................................60
Displaying Passwords...........................................................................................................61
Access to Both MSM Console Ports—Modular Switches Only.........................................................61
Access to an Active Node in a SummitStack ................................................................................61
Domain Name Service Client Services .........................................................................................61
Checking Basic Connectivity.......................................................................................................62
Ping...................................................................................................................................62
Traceroute ..........................................................................................................................63
Displaying Switch Information ....................................................................................................64
Chapter 3: Managing the Switch .................................................................................................... 65

Overview ..................................................................................................................................65
Understanding the ExtremeXOS Shell..........................................................................................66
Using the Console Interface .......................................................................................................66
Using the 10/100 Ethernet Management Port ..............................................................................67
Using EPICenter to Manage the Network .....................................................................................67
Authenticating Users .................................................................................................................68
RADIUS Client ....................................................................................................................68
TACACS+ ...........................................................................................................................68
Management Accounts.........................................................................................................68
Using Telnet .............................................................................................................................68
About the Telnet Client ........................................................................................................69
About the Telnet Server .......................................................................................................69
Connecting to Another Host Using Telnet...............................................................................70
Configuring Switch IP Parameters .........................................................................................70
Configuring Telnet Access to the Switch ................................................................................72
Disconnecting a Telnet Session ............................................................................................75
Using Secure Shell 2.................................................................................................................75
Using the Trivial File Transfer Protocol ........................................................................................76
Connecting to Another Host Using TFTP ................................................................................76
Understanding System Redundancy—Modular Switches and SummitStack Only .............................77
Node Election .....................................................................................................................78
Replicating Data Between Nodes ..........................................................................................79
Viewing Node Status............................................................................................................81
Understanding Hitless Failover Support—Modular Switches and SummitStack Only ........................82
Protocol Support for Hitless Failover .....................................................................................82
Platform Support for Hitless Failover.....................................................................................85
Hitless Failover Caveats .......................................................................................................86
Understanding Power Supply Management ..................................................................................87
Using Power Supplies—Modular Switches Only ......................................................................87
Using Power Supplies—Summit Family of Switches Only ........................................................90
Using Power Supplies - SummitStack Only ............................................................................91
Displaying Power Supply Information ....................................................................................91

Contents
Using the Simple Network Management Protocol .........................................................................91

Enabling and Disabling SNMPv1/v2c and SNMPv3 ................................................................92
Accessing Switch Agents......................................................................................................93
Supported MIBs ..................................................................................................................93
Configuring SNMPv1/v2c Settings ........................................................................................93
Displaying SNMP Settings....................................................................................................94
SNMPv3.............................................................................................................................94
Message Processing.............................................................................................................96
SNMPv3 Security ................................................................................................................96
SNMPv3 MIB Access Control ...............................................................................................98
SNMPv3 Notification...........................................................................................................99
Using the Simple Network Time Protocol...................................................................................102
Configuring and Using SNTP ..............................................................................................102
SNTP Example..................................................................................................................105
Chapter 4: Managing the ExtremeXOS Software............................................................................. 107

Overview of the ExtremeXOS Software .......................................................................................107
Understanding the ExtremeXOS Software.............................................................................107
Using the ExtremeXOS File System ...........................................................................................108
Moving or Renaming Files on the Switch .............................................................................109
Copying Files on the Switch ...............................................................................................110
Displaying Files on the Switch ............................................................................................111
Transferring Files to and from the Switch ............................................................................113
Deleting Files from the Switch............................................................................................115
Managing the Configuration File ...............................................................................................116
Managing ExtremeXOS Processes .............................................................................................117
Displaying Process Information...........................................................................................117
Stopping a Process............................................................................................................118
Starting a Process .............................................................................................................119
Understanding Memory Protection ............................................................................................120
Monitoring CPU Utilization.......................................................................................................121
Disabling CPU Monitoring ..................................................................................................121
Enabling CPU Monitoring ...................................................................................................121
Displaying CPU Utilization History ......................................................................................122
Chapter 5: Configuring Slots and Ports on a Switch....................................................................... 125

Configuring a Slot on a Modular or SummitStack Switch .............................................................125
Overview...........................................................................................................................125
I/O Ports on BlackDiamond 8810 MSM-G8X Module ............................................................126
I/O Ports on BlackDiamond 8806 MSM-G8X Module ............................................................127
Disabling MSM-G8X I/O Ports—BlackDiamond 8800
a-series and e-series Modules Only ...........................................................................................127
Configuring Ports on a Switch...................................................................................................128
Port Numbering ................................................................................................................128
Enabling and Disabling Switch Ports ...................................................................................129
Configuring Switch Port Speed and Duplex Setting ...............................................................130
WAN PHY OAM—BlackDiamond 10808 and Summit X450a Series Switches Only ..................134
Jumbo Frames ........................................................................................................................135
Jumbo Frames on the BlackDiamond 8800 Series Switch, SummitStack,
and Summit Family of Switches Only ..................................................................................135

Contents
Enabling Jumbo Frames.....................................................................................................136

Path MTU Discovery ..........................................................................................................137
IP Fragmentation with Jumbo Frames..................................................................................137
IP Fragmentation within a VLAN .........................................................................................138
Link Aggregation on the Switch ................................................................................................139
Link Aggregation Overview..................................................................................................139
Link Aggregation and Software-Controlled Redundant Ports—Summit Family of Switches Only .140
Dynamic Versus Static Load Sharing ...................................................................................140
Load-Sharing Algorithms....................................................................................................141
LACP—Dynamic Link Aggregation.......................................................................................143
Guidelines for Load Sharing ...............................................................................................145
Configuring Switch Load Sharing ........................................................................................148
Load-Sharing Examples .....................................................................................................150
Displaying Switch Load Sharing..........................................................................................151
Mirroring ................................................................................................................................151
Guidelines for Mirroring on the Summit Family of Switches Only............................................152
Guidelines for Mirroring on the BlackDiamond 8800 Series Switches and SummitStack Only ...153
Guidelines for Mirroring on the BlackDiamond 10808 and 12800 Series Switches Only ..........154
Mirroring Rules and Restrictions for All Switches..................................................................155
Mirroring Examples ...........................................................................................................156
Verifying the Mirroring Configuration ...................................................................................157
Extreme Discovery Protocol ......................................................................................................157
Software-Controlled Redundant Port and Smart Redundancy .......................................................158
Guidelines for Software-Controlled Redundant Ports and Port Groups .....................................159
Configuring Software-Controlled Redundant Ports.................................................................160
Verifying Software-Controlled Redundant Port Configurations.................................................160
Configuring Automatic Failover for Combination Ports—Summit X450, X450a, and
X450e Series of Switches Only .................................................................................................161
Displaying Port Configuration Information..................................................................................163
Chapter 6: Universal Port............................................................................................................. 165

Universal Port .........................................................................................................................165
Events that Trigger Profiles ................................................................................................169
Executing Static Profiles ....................................................................................................170
Handling Profile Execution Errors .......................................................................................170
Universal Port Variables .....................................................................................................171
Sample Profile Configurations ..................................................................................................172
Static Profile Example .......................................................................................................172
Configuring a Profile with QoS Support................................................................................173
Chapter 7: CLI Scripting............................................................................................................... 175

CLI Scripting Capabilities ........................................................................................................175
Scripting Variables ............................................................................................................175
Enable | Disable CLI Scripting ............................................................................................176
Local Variables .................................................................................................................176
Variable Manipulation ........................................................................................................176
Session Variables ..............................................................................................................177
Control Structures .............................................................................................................177
Operators .........................................................................................................................178
Built-In Functions .............................................................................................................179
Error Handling ..................................................................................................................179

Contents
CLI Scripting Examples ...........................................................................................................179

Sample Script to Create 100 VLAN's...................................................................................179
Chapter 8: Link Layer Discovery Protocol...................................................................................... 181

Overview ................................................................................................................................181
LLDP Packets .........................................................................................................................183
Transmitting LLDP Messages ...................................................................................................184
Receiving LLDP Messages........................................................................................................185
Managing LLDP ......................................................................................................................185
Supported TLVs ......................................................................................................................186
Mandatory TLVs ................................................................................................................189
Optional TLVs ...................................................................................................................190
Configuring LLDP....................................................................................................................195
Enabling and Disabling LLDP .............................................................................................195
Configuring the System Description TLV Advertisement.........................................................196
Configuring LLDP Timers ...................................................................................................196
Configuring SNMP for LLDP ...............................................................................................196
Configuring Optional TLV Advertisements ............................................................................197
Unconfiguring LLDP ..........................................................................................................201
Displaying LLDP Settings.........................................................................................................201
Displaying LLDP Port Configuration Information and Statistics ..............................................201
Displaying LLDP Information Detected from Neighboring Ports ..............................................202
Chapter 9: Connectivity Fault Management—BlackDiamond 10808 and 12800 Series

Switches Only ............................................................................................................................. 203
Overview of CFM Elements .......................................................................................................203
Ping and Traceroute ................................................................................................................206
Supported Instances for CFM ...................................................................................................207
Configuring CFM .....................................................................................................................207
Creating Maintenance Domains ..........................................................................................208
Creating and Associating MAs.............................................................................................209
Creating MPs and the CCM Transmission Interval .................................................................210
Executing Layer 2 Ping and Traceroute Messages .................................................................211
Displaying CFM.......................................................................................................................211
CFM Example .........................................................................................................................212
Chapter 10: Power Over Ethernet.................................................................................................. 213

Extreme Networks PoE Devices.................................................................................................213
Summary of PoE Features ........................................................................................................214
Power Checking for PoE Module ...............................................................................................215
Power Delivery ........................................................................................................................215
Enabling PoE to the Switch ................................................................................................215
Power Reserve Budget on the Summit X450e-48p Switch and Per Slot on Modular Switches ...216
PD Disconnect Precedence on the Summit X450e-48p Switch and Modular Switches .............217
Port Disconnect or Fault ....................................................................................................218
Port Power Reset...............................................................................................................218
PoE Usage Threshold.........................................................................................................218
Legacy Devices .................................................................................................................219
PoE Operator Limits ..........................................................................................................219

Contents
Configuring PoE ......................................................................................................................220

Enabling Inline Power........................................................................................................220
Reserving Power for the Summit X450e-48p Switch or a Slot on Modular Switches .................221
Setting the Disconnect Precedence on a Summit X450e-48p Switch or Modular Switches .......222
Configuring the Usage Threshold ........................................................................................223
Configuring the Switch to Detect Legacy PDs .......................................................................224
Configuring the Operator Limit ...........................................................................................224
Configuring PoE Port Labels ...............................................................................................224
Power Cycling Connected PDs ............................................................................................225
Displaying PoE Settings and Statistics ......................................................................................225
Clearing Statistics .............................................................................................................225
Displaying System Power Information..................................................................................225
Displaying Slot PoE Information on Modular Switches...........................................................226
Displaying PoE Status and Statistics on Stand-alone Switches...............................................227
Displaying Port PoE Information .........................................................................................227
Chapter 11: Status Monitoring and Statistics ................................................................................ 231

Overview of Status Monitoring ..................................................................................................231
Viewing Port Statistics .............................................................................................................231
Viewing Port Errors ..................................................................................................................232
Using the Port Monitoring Display Keys .....................................................................................234
Viewing VLAN Statistics...........................................................................................................234
Performing Switch Diagnostics .................................................................................................235
Running Diagnostics on the BlackDiamond 10808 Switch and the
BlackDiamond 8800 Series Switches..................................................................................236
Running Diagnostics on the BlackDiamond 12800 Series Switches .......................................237
Running Diagnostics on the SummitStack or Summit Family of Switches ...............................238
Observing LED Behavior During a Diagnostic Test.................................................................239
Displaying Diagnostic Test Results......................................................................................245
Using the System Health Checker .............................................................................................245
Understanding the System Health Checker—BlackDiamond 10808 and
BlackDiamond 12800 Series Switches Only ........................................................................245
Understanding the System Health Checker—BlackDiamond 8800 Series Switch Only .............246
Understanding the System Health Checker—Summit Family of Switches Only ........................247
Enabling Backplane Diagnostic Packets on the Switch—Modular Switches Only......................247
Configuring Backplane Diagnostic Packets on the Switch—Modular Switches Only ..................247
Disabling Backplane Diagnostic Packets on the Switch—Modular Switches Only .....................248
Displaying the System Health Check Setting—All Platforms ..................................................248
System Health Check Examples: Backplane Diagnostics—Modular Switches Only ...................248
Setting the System Recovery Level............................................................................................250
Configuring Software Recovery............................................................................................250
Configuring Hardware Recovery—SummitStack and Summit Family of Switches Only ..............251
Configuring Module Recovery—Modular Switches Only .........................................................253
Using ELSM ...........................................................................................................................260
About ELSM .....................................................................................................................261
ELSM Hello Messages .......................................................................................................261
ELSM Port States..............................................................................................................262
Link States .......................................................................................................................262
ELSM Link States .............................................................................................................263
ELSM Timers ....................................................................................................................264
Configuring ELSM on a Switch ...........................................................................................265

Contents
Displaying ELSM Information .............................................................................................268

Using ELSM with Layer 2 Control Protocols .........................................................................270
ELSM Configuration Example .............................................................................................270
Viewing Fan Information ..........................................................................................................271
Viewing the System Temperature ..............................................................................................272
System Temperature Output—Modular Switches and SummitStack Only ................................272
System Temperature Output—Summit Family of Switches Only .............................................272
Power Supply Temperature—Modular Switches Only.............................................................273
Fan Tray Temperature—BlackDiamond 10808 Switch Only...................................................273
Using the Event Management System/Logging ...........................................................................273
Sending Event Messages to Log Targets...............................................................................274
Filtering Events Sent to Targets ..........................................................................................275
Displaying Real-Time Log Messages ....................................................................................283
Displaying Event Logs........................................................................................................284
Uploading Event Logs ........................................................................................................284
Displaying Counts of Event Occurrences ..............................................................................284
Displaying Debug Information.............................................................................................285
Logging Configuration Changes...........................................................................................286
Using sFlow............................................................................................................................286
Licensing .........................................................................................................................287
Sampling Mechanisms.......................................................................................................288
Configuring sFlow..............................................................................................................288
Additional sFlow Configuration Options ...............................................................................291
sFlow Configuration Example..............................................................................................292
Displaying sFlow Information..............................................................................................293
Using RMON ..........................................................................................................................293
About RMON ....................................................................................................................293
Supported RMON Groups of the Switch ...............................................................................294
Configuring RMON ............................................................................................................296
Event Actions ...................................................................................................................296
Displaying RMON Information ............................................................................................297
Chapter 12: Virtual LANs ............................................................................................................. 299

Overview of Virtual LANs..........................................................................................................299
Benefits ...........................................................................................................................299
Virtual Routers and VLANs—BlackDiamond 10808 and 12800 Series Switches Only ..............300
Types of VLANs.......................................................................................................................300
Port-Based VLANs .............................................................................................................301
Tagged VLANs ..................................................................................................................303
Protocol-Based VLANs .......................................................................................................305
Precedence of Tagged Packets Over Protocol Filters .............................................................307
Default VLAN....................................................................................................................307
VLAN Names ..........................................................................................................................307
Renaming a VLAN .............................................................................................................308
Configuring VLANs on the Switch .............................................................................................308
Creating and Configuring VLANs .........................................................................................309
Enabling and Disabling VLANs ...........................................................................................309
VLAN Configuration Examples ............................................................................................310
Displaying VLAN Settings.........................................................................................................311
Displaying Protocol Information ..........................................................................................312

Contents
Chapter 13: vMAN and Tunneling ................................................................................................. 313

Overview ................................................................................................................................313
vMAN Overview .................................................................................................................313
BlackDiamond 12800 Series Switch vMAN Enhancements ...................................................315
vMANs on the BlackDiamond 8800 Series Switches, SummitStack, and the
Summit Family of Switches Only ........................................................................................316
vMANs on the BlackDiamond 10808 and
BlackDiamond 12800 Series Switches Only ........................................................................317
Licensing .........................................................................................................................318
vMAN Features .......................................................................................................................318
Inter vMAN Forwarding Using vMAN Ingress ACL .................................................................319
vMAN Flood Groups...........................................................................................................319
vMAN Learning Domain .....................................................................................................320
LAG Filtering using vMAN Egress ACL .................................................................................321
STAG Ethertype Translation................................................................................................321
MAC-in-MAC Tunneling—BlackDiamond 10808 and 12800 Series Switches Only ..................322
QoS Queue on Egress Port with vMAN packets .....................................................................325
Egress Queue on the BlackDiamond 10808 and
12800 Series Switches Only ..............................................................................................325
Egress Queue on the BlackDiamond 8800 Series Switches, SummitStack, and
the Summit Family of Switches Only ...................................................................................326
vMAN Configuration ................................................................................................................326
Guidelines for Configuring vMANs.......................................................................................326
Configuring vMANs—BlackDiamond 8800 Series Switches, SummitStack, and
Configuring vMAN—BlackDiamond 12800 Series Switches...................................................328
Configuring vMANs—BlackDiamond 10808.........................................................................329
Displaying vMAN Configurations .........................................................................................330
Configuring MAC-in-MAC Tunnels .......................................................................................330
vMAN Examples......................................................................................................................332
MAC-in-MAC Tunneling Example ........................................................................................333
BlackDiamond 10808 Switch Example ...............................................................................335
BlackDiamond 8810 Switch Example .................................................................................335
Inter vMAN Forwarding Using vMAN Ingress ACL Example ....................................................337
LAG Filtering using vMAN Egress ACL Example....................................................................339
STAG Ethertype Translation Example ..................................................................................340
Chapter 14: Web-Based Device Management................................................................................ 343

ScreenPlay Overview................................................................................................................343
Setting Up ScreenPlay.............................................................................................................343
HTTP and HTTPS Setup ....................................................................................................343
Client Setup .....................................................................................................................344
Launching ScreenPlay .......................................................................................................345
ScreenPlay Dashboard .............................................................................................................346
Configuration ..........................................................................................................................348
Ports Configuration............................................................................................................348
VLAN Configuration ...........................................................................................................349
Stacking Configuration.......................................................................................................349
SNMP Configuration ..........................................................................................................350

Contents
Statistics and Monitoring .........................................................................................................351

Event Log .........................................................................................................................352
Ports................................................................................................................................352
QoS .................................................................................................................................353
Administration ........................................................................................................................354
User Accounts ..................................................................................................................354
User Sessions ...................................................................................................................355
Chapter 15: Forwarding Database................................................................................................. 357

Overview of the FDB ................................................................................................................357
FDB Contents ...................................................................................................................357
How FDB Entries Get Added...............................................................................................358
FDB Entry Types ...............................................................................................................358
Differing FDB Table Sizes—BlackDiamond 8800 Series Switches, SummitStack,
and the Summit Family of Switches Only...................................................................................359
FDB Configuration Examples ....................................................................................................360
Adding a Permanent Static Entry ........................................................................................360
Configuring the FDB Aging Time.........................................................................................360
Clearing FDB Entries .........................................................................................................361
Displaying FDB Entries ............................................................................................................361
MAC-Based Security................................................................................................................362
Disabling MAC Address Learning ........................................................................................362
Disabling Egress Flooding ..................................................................................................363
Displaying Learning and Flooding Settings...........................................................................365
Multicast FDB with Multiport Entry—BlackDiamond 8800 Series Switches,
SummitStack, and the Summit Family of Switches Only .............................................................365
Chapter 16: Virtual Routers.......................................................................................................... 367

Virtual Routers Overview ..........................................................................................................367
Types of Virtual Routers .....................................................................................................368
Virtual Router Configuration Domain—BlackDiamond 10808 and BlackDiamond
12000 series Switches Only...............................................................................................369
Using Virtual Routers—BlackDiamond 10808 and BlackDiamond 12000 series Switches Only ......370
Creating Virtual Routers .....................................................................................................370
Configuring Ports to a Single or to Multiple Virtual Router(s) .................................................370
Adding Routing Protocols to a Virtual Router........................................................................371
Displaying Ports and Protocols............................................................................................372
Configuring the Routing Protocols and VLANs ......................................................................372
Virtual Router Configuration Example ........................................................................................373
Chapter 17: Policy Manager ........................................................................................................ 375

Policy Manager .......................................................................................................................375
Creating and Editing Policies....................................................................................................375
Using the Edit Command ...................................................................................................376
Using a Separate Machine .................................................................................................376
Checking Policies ..............................................................................................................376
Refreshing Policies............................................................................................................377
Applying Policies ....................................................................................................................377
Applying ACL Policies........................................................................................................378
Applying Routing Policies ..................................................................................................378

Contents
Chapter 18: Access Lists (ACLs)................................................................................................... 379

ACLs......................................................................................................................................380
ACL Rule Syntax .....................................................................................................................381
Matching All Egress Packets...............................................................................................382
Comments and Descriptions in ACL Policy Files ...................................................................382
Types of Rule Entries.........................................................................................................383
Match Conditions ..............................................................................................................384
Actions.............................................................................................................................384
Action Modifiers................................................................................................................384
ACL Rule Syntax Details ....................................................................................................386
IPv6 ACL Address Masks—BlackDiamond 10808 and BlackDiamond 12800
Series Switches Only .........................................................................................................391
vMAN ACLs—BlackDiamond 10808 and BlackDiamond 12800 Series Switches Only ...................392
vMAN ACL Actions ............................................................................................................392
vMAN ACL Action Modifiers ...............................................................................................393
vMAN ACL Examples—BlackDiamond 10808 and BlackDiamond 12800 Series
Switches Only ...................................................................................................................394
Layer-2 Protocol Tunneling ACLs—BlackDiamond 8800 a-Series modules and
e-Series and Summit X250e, X450a, and X450e Switches Only ..................................................394
Dynamic ACLs ........................................................................................................................395
Creating the Dynamic ACL Rule ..........................................................................................395
Configuring the ACL Rule on the Interface ...........................................................................396
Configuring ACL Priority...........................................................................................................396
ACL Evaluation Precedence—BlackDiamond 10808 and BlackDiamond 12800 Series
Switches Only .........................................................................................................................400
Rule Evaluation.................................................................................................................400
Precedence of Dynamic ACLs .............................................................................................400
Precedence of L2/L3/L4 ACL Entries on BlackDiamond10k and 12K Switches........................400
Precedence Among Interface Types.....................................................................................401
ACL Evaluation Precedence—BlackDiamond 8800 Series switches, SummitStack,
and the Summit Family Switches Only ......................................................................................401
Rule Evaluation.................................................................................................................401
Precedence of Dynamic ACLs .............................................................................................402
Precedence of L2/L3/L4 ACL Entries...................................................................................402
Precedence Among Interface Types.....................................................................................402
Redundant Rules ..............................................................................................................402
Applying ACL Policy Files ........................................................................................................402
Displaying and Clearing ACL Counters .................................................................................403
Example ACL Rule Entries .................................................................................................403
ACL Mechanisms ....................................................................................................................405
ACL Masks and Rules—BlackDiamond 8800 Series Switches and the Summit X450
Series Switches Only .........................................................................................................405
ACL Slices and Rules—BlackDiamond 8800 a-series and e-series Modules and
Summit X450a, X450e, and X250e Series Switches Only .....................................................411
Policy Based Routing...............................................................................................................421
Layer 3 Policy Based Redirect ............................................................................................421
Layer 2 Policy Based Redirect ............................................................................................422
Configuring Policy Based Routing .......................................................................................424
ACL Troubleshooting—BlackDiamond 8800, SummitStack, and Summit Family of Switches .........425

Contents
Chapter 19: Routing Policies ....................................................................................................... 427

Routing Policies......................................................................................................................427
Routing Policy File Syntax..................................................................................................427
Applying Routing Policies ..................................................................................................432
Policy Examples ................................................................................................................432
Chapter 20: Quality of Service ..................................................................................................... 437

Overview of Policy-Based Quality of Service ...............................................................................437
Applications and Types of QoS .................................................................................................438
Voice Applications.............................................................................................................438
Video Applications.............................................................................................................438
Critical Database Applications ............................................................................................439
Web Browsing Applications ................................................................................................439
File Server Applications .....................................................................................................439
Configuring QoS......................................................................................................................440
Configuring QoS on the BlackDiamond 8800 Series Switches, SummitStack, and
QoS Profiles ...........................................................................................................................441
QoS Profiles on the BlackDiamond 8800 Series Switches, SummitStack, and
QoS Profiles on the BlackDiamond 10808 and
12800 Series Switches......................................................................................................443
Traffic Groupings ....................................................................................................................444
Precedence of Traffic Groupings .........................................................................................444
ACL-Based Traffic Groupings..............................................................................................446
Explicit Class of Service (802.1p and DiffServ) Traffic Groupings ..........................................446
Physical and Logical Groupings ..........................................................................................453
Verifying QoS Configuration and Performance ............................................................................455
Monitoring Performance.....................................................................................................455
Displaying QoS Profile Information......................................................................................456
Guidelines for Configuring QoS.................................................................................................457
Metering Using ACLs—BlackDiamond 8800 Series, SummitStack, the Summit Family
of Switches, and BlackDiamond 12800 R-Series Switch Only .....................................................457
Creating the ACL Meter......................................................................................................458
Configuring the ACL Meter .................................................................................................458
Associating the Meter with an ACL......................................................................................458
Displaying Meters..............................................................................................................459
Egress Traffic Rate Limiting—BlackDiamond 8800 Series Switches, SummitStack,
and the Summit Family of Switches Only...................................................................................459
Applying Egress Bandwidth to a Port—BlackDiamond 10808 and 12800 Series Switches Only......460
Applying Egress Bandwidth to a QoS Queue...............................................................................461
Bi-Directional Rate Shaping—BlackDiamond 10808 Switch Only ................................................461
Bandwidth Settings ...........................................................................................................462
Configuring Bi-Directional Rate Shaping..............................................................................463
Hierarchical QoS—BlackDiamond 12800 R-Series Switch Only...................................................464
Overview...........................................................................................................................465
Setting the HQoS Mode .....................................................................................................466
HQoS Implementation .......................................................................................................467
Guidelines for Using Ingress-Only and Ingress and Egress HQoS ............................................472

Contents
Configuring HQoS Ingress and Egress Queues ......................................................................472

Displaying HQoS ...............................................................................................................476
HQoS Examples ................................................................................................................478
Chapter 21: Network Login .......................................................................................................... 481

Network Login Overview ...........................................................................................................481
Web-Based, MAC-Based, and 802.1x Authentication............................................................482
Multiple Supplicant Support ..............................................................................................483
Campus and ISP Modes .....................................................................................................484
Network Login and Hitless Failover—Modular Switches and SummitStack Only.......................484
Configuring Network Login .......................................................................................................485
Enabling or Disabling Network Login on the Switch ..............................................................486
Enabling or Disabling Network Login on a Specific Port ........................................................486
Configuring the Move Fail Action ........................................................................................486
Displaying Network Login Settings ......................................................................................487
Exclusions and Limitations.................................................................................................487
Authenticating Users ...............................................................................................................487
Creating User Accounts on the RADIUS Server.....................................................................488
Configuring Local Database Authentication ..........................................................................492
802.1x Authentication.............................................................................................................496
Interoperability Requirements.............................................................................................496
Enabling and Disabling 802.1x Network Login .....................................................................497
802.1x Network Login Configuration Example......................................................................497
Configuring Guest VLANs ...................................................................................................498
Post-authentication VLAN Movement ..................................................................................501
802.1x Authentication and Network Access Protection .........................................................502
Web-Based Authentication .......................................................................................................506
Enabling and Disabling Web-Based Network Login ...............................................................506
Configuring the Base URL ..................................................................................................506
Configuring the Redirect Page ............................................................................................507
Configuring Session Refresh ...............................................................................................507
Configuring Logout Privilege...............................................................................................507
Configuring the Login Page ................................................................................................507
Customizable Authentication Failure Response ....................................................................509
Web-Based Network Login Configuration Example ................................................................510
Web-Based Authentication User Login.................................................................................511
MAC-Based Authentication ......................................................................................................513
Enabling and Disabling MAC-Based Network Login ...............................................................514
Associating a MAC Address to a Specific Port ......................................................................514
Adding and Deleting MAC Addresses...................................................................................514
Displaying the MAC Address List ........................................................................................515
Secure MAC Configuration Example ....................................................................................515
MAC-Based Network Login Configuration Example................................................................516
Additional Network Login Configuration Details ..........................................................................516
Configuring Netlogin MAC-Based VLANs..............................................................................517
Configuring Dynamic VLANs for Netlogin .............................................................................519
Configuring Netlogin Port Restart........................................................................................521
Chapter 22: Security ................................................................................................................... 523

Security Overview....................................................................................................................523
Safe Defaults Mode .................................................................................................................525

Contents
MAC Security..........................................................................................................................525
Limiting Dynamic MAC Addresses.......................................................................................526
MAC Address Lockdown.....................................................................................................528
MAC Address Lockdown with Timeout .................................................................................529
DHCP Server ..........................................................................................................................533
Enabling and Disabling DHCP ............................................................................................533
Configuring the DHCP Server..............................................................................................533
Displaying DHCP Information .............................................................................................534
IP Security .............................................................................................................................535
DHCP Snooping and Trusted DHCP Server...........................................................................535
Source IP Lockdown ..........................................................................................................538
ARP Learning ...................................................................................................................539
Gratuitous ARP Protection..................................................................................................541
ARP Validation..................................................................................................................543
Denial of Service Protection .....................................................................................................544
Configuring Simulated Denial of Service Protection ..............................................................545
Configuring Denial of Service Protection ..............................................................................545
Protocol Anomaly Protection...............................................................................................546
Authenticating Users Using RADIUS or TACACS+ ......................................................................547
RADIUS ...........................................................................................................................547
Configuring RADIUS ..........................................................................................................550
TACACS+ .........................................................................................................................555
Hyptertext Transfer Protocol .....................................................................................................560
Secure Shell 2 ........................................................................................................................560
Enabling SSH2 for Inbound Switch Access ..........................................................................561
Viewing SSH2 Information .................................................................................................563
Using ACLs to Control SSH2 Access ...................................................................................563
Using SCP2 from an External SSH2 Client ..........................................................................565
Understanding the SSH2 Client Functions on the Switch ......................................................566
Using SFTP from an External SSH2 Client...........................................................................567
Secure Socket Layer ................................................................................................................569
Enabling and Disabling SSL ...............................................................................................570
Creating Certificates and Private Keys .................................................................................570
Displaying SSL Information ................................................................................................572
Chapter 23: CLEAR-Flow .............................................................................................................. 573

Overview ................................................................................................................................573
Configuring CLEAR-Flow ..........................................................................................................573
Displaying CLEAR-Flow Configuration and Activity................................................................574
Adding CLEAR-Flow Rules to ACLs ...........................................................................................574
CLEAR-Flow Rule Match Type ............................................................................................575
CLEAR-Flow Rule Match Conditions....................................................................................576
CLEAR-Flow Rule Actions ..................................................................................................582
CLEAR-Flow Rule Examples .....................................................................................................587
Count Expression Example .................................................................................................587
Delta Expression Example ..................................................................................................588
Ratio Expression Example ..................................................................................................589
Delta-Ratio Expression Example..........................................................................................590

Contents
Chapter 24: Configuring Stacked Switches ................................................................................... 591

Overview ................................................................................................................................591
Understanding SummitStack Terms ....................................................................................592
Required Software Release.................................................................................................594
Understanding SummitStack Configuration Parameters, Configuration Files,
and Port Numbering ..........................................................................................................594
Understanding Stacking Link Overcommitment ....................................................................595
About SummitStack Logging Messages................................................................................595
About QoS in Stacking.......................................................................................................596
About Power Management and Power Over Ethernet on Stacking ...........................................597
About Stacking Node Roles, Redundancy, and Failover .........................................................597
About the Failsafe Account on SummitStack Nodes..............................................................598
SummitStack Compatible Switches...........................................................................................598
SummitStack Topologies..........................................................................................................599
Ring Topology ...................................................................................................................600
Daisy Chain Topology.........................................................................................................601
Stack Depth .....................................................................................................................602
Logging into a SummitStack ....................................................................................................603
Logging in Through the Console Port ...................................................................................603
Logging in from the Management Network ...........................................................................603
Logging Into a Node From Another Node .............................................................................604
Managing Licenses on a SummitStack ......................................................................................604
Viewing Switch Licenses ....................................................................................................605
Enabling a Switch License .................................................................................................605
Restricting a Switch License Level ......................................................................................606
Upgrading Licenses ...........................................................................................................606
Configuring a New Stack..........................................................................................................607
Configuration Procedure.....................................................................................................608
Converting a Standalone Node Deployment to a Stack ................................................................609
SummitStack Configuration Tasks.............................................................................................611
Enabling the Stack ............................................................................................................611
Verifying the Configuration .................................................................................................611
Setting the Command Prompt.............................................................................................614
Configuring Slot Numbers ..................................................................................................615
Configuring Node Priority ...................................................................................................615
Assigning a MAC address for the Stack................................................................................616
Configuring Master-Capability.............................................................................................617
Configuring an Alternate Management IP Address ................................................................618
Configuring the Failsafe Account on a Stack ........................................................................619
Saving the Configuration ....................................................................................................619
Managing an Operating SummitStack........................................................................................619
Stacking LEDs ..................................................................................................................619
Configuring and Viewing Stacking Port Status ......................................................................620
Adding a Node to an Active Topology ..................................................................................620
Replacing a Node with the Same Switch Type......................................................................621
Replacing a Node with a Different Switch Type ....................................................................622
Merging Two Stacks ..........................................................................................................623
Synchronize Command Operation........................................................................................624
Upgrading ExtremeXOS on a Stack......................................................................................624
Rebooting a Stack .............................................................................................................626

Contents
Configuration Examples ...........................................................................................................626

Building and Deploying a New Stack ...................................................................................626
Joining Two Stacks............................................................................................................630
Adding a Node to a Stack...................................................................................................636
Configuring an Alternate IP Address ....................................................................................638
Configuring a MAC address ................................................................................................639
Configuring the License Level Restriction ............................................................................642
Bringing All Nodes Up to the Same License Level ................................................................642
VLAN Management in Stacking...........................................................................................643
Troubleshooting a Stack...........................................................................................................644
Managing a Dual Master Situation ......................................................................................645
Setting Traps for Stacking ..................................................................................................647
Connecting to a SummitStack with No Master......................................................................648
Rescuing a Stack That Has No Master-Capable Node............................................................648
FAQs on SummitStack.............................................................................................................650
Part 2: Using Switching and Routing Protocols

Chapter 25: Ethernet Automatic Protection Switching.................................................................... 653
Licensing ...............................................................................................................................653
Overview of the EAPS Protocol .................................................................................................654
Fast Convergence ..............................................................................................................655
EAPS and Hitless Failover—Modular Switches and SummitStack Only ...................................655
Fault Detection and Recovery ...................................................................................................656
Link Down Message Sent by a Transit Node .........................................................................657
Ring Port Down Event Sent by Hardware Layer .....................................................................658
Polling .............................................................................................................................658
Restoration Operations.......................................................................................................658
Multiple EAPS Domains...........................................................................................................659
EAPS Data VLAN Spanning Two Rings Connected by One Switch...........................................659
Multiple EAPS Domains per Ring—Spatial Reuse.................................................................660
Multiple EAPS Rings Sharing a Common Link ......................................................................661
Configuring EAPS on a Switch ..................................................................................................662
Creating and Deleting an EAPS Domain...............................................................................663
Defining the EAPS Mode of the Switch................................................................................664
Configuring EAPS Polling Timers ........................................................................................665
Configuring the Primary and Secondary Ports .......................................................................666
Configuring the EAPS Control VLAN ....................................................................................666
Adding the EAPS Protected VLANs .....................................................................................667
Enabling and Disabling Fast Convergence ............................................................................668
Enabling and Disabling an EAPS Domain.............................................................................668
Enabling and Disabling EAPS on the Switch ........................................................................668
Unconfiguring an EAPS Ring Port .......................................................................................669
Disabling EAPS Loop Protection Warning Messages ..............................................................670
Displaying EAPS Status and Counter Information .................................................................670
Configuring EAPS Shared Ports ................................................................................................675
Steady State .....................................................................................................................675
Common Link Failures .......................................................................................................676
Flushing the FDBs.............................................................................................................678
Creating and Deleting a Shared Port....................................................................................678
Defining the Mode of the Shared Port..................................................................................678

Contents
Configuring the Link ID of the Shared Port...........................................................................678

Configuring the Shared Port Segment Timer.........................................................................679
Unconfiguring an EAPS Shared Port....................................................................................679
Displaying EAPS Shared-Port Status and Counter Information ...............................................679
EAPS Shared Port Configuration Rules ......................................................................................684
EAPS Shared Port Configuration Examples ................................................................................685
Basic Configuration ...........................................................................................................685
Basic Core Configuration....................................................................................................686
Right Angle Configuration ..................................................................................................686
Combined Basic Core and Right Angle Configuration ............................................................687
Large Core and Access Rings Configuration..........................................................................687
Advanced Configuration .....................................................................................................688
Chapter 26: Spanning Tree Protocol............................................................................................. 689

Overview of the Spanning Tree Protocol.....................................................................................689
Compatibility Between 802.1D-1998 and 802.1D-2004 STP Bridges ...................................690
Spanning Tree Domains ...........................................................................................................693
Member VLANs .................................................................................................................694
STPD Modes.....................................................................................................................695
Encapsulation Modes.........................................................................................................696
STP States .......................................................................................................................697
Binding Ports....................................................................................................................698
Rapid Root Failover ...........................................................................................................700
STPD BPDU Tunneling ......................................................................................................700
STP and Hitless Failover—Modular Switches Only ................................................................702
STP Configurations..................................................................................................................703
Basic STP Configuration ....................................................................................................704
Multiple STPDs on a Port ...................................................................................................706
VLANs Spanning Multiple STPDs........................................................................................706
EMISTP Deployment Constraints ........................................................................................707
Per VLAN Spanning Tree..........................................................................................................709
STPD VLAN Mapping.........................................................................................................709
Native VLAN .....................................................................................................................709
Rapid Spanning Tree Protocol ..................................................................................................709
RSTP Concepts .................................................................................................................710
RSTP Operation ................................................................................................................713
Multiple Spanning Tree Protocol...............................................................................................721
MSTP Concepts ................................................................................................................721
MSTP Operation................................................................................................................730
STP Rules and Restrictions ......................................................................................................732
Configuring STP on the Switch .................................................................................................732
STP Configuration Examples ..............................................................................................734
Displaying STP Settings...........................................................................................................739
Chapter 27: Extreme Standby Router Protocol ............................................................................... 743

Licensing ...............................................................................................................................743
Overview of ESRP ...................................................................................................................744
ESRP Modes of Operation ..................................................................................................744
ESRP and ELRP................................................................................................................744
Reasons to Use ESRP ........................................................................................................745

Contents
ESRP Concepts.......................................................................................................................745
ESRP-Aware Switches .......................................................................................................746
Standard and Extended ESRP ............................................................................................748
ESRP Domains .................................................................................................................749
Linking ESRP Switches......................................................................................................750
ESRP and Hitless Failover—Modular Switches and SummitStack Only ...................................750
Determining the ESRP Master ..................................................................................................751
Master Switch Behavior .....................................................................................................752
Pre-Master Switch Behavior................................................................................................752
Slave Switch Behavior .......................................................................................................752
Neutral Switch Behavior ....................................................................................................752
Electing the Master Switch.................................................................................................753
ESRP Failover Time...........................................................................................................753
ESRP Election Algorithms ..................................................................................................754
Configuring an ESRP Domain on a Switch .................................................................................756
Creating and Deleting an ESRP Domain...............................................................................756
Configuring the ESRP Domain ID........................................................................................757
Adding VLANs to an ESRP Domain .....................................................................................757
Enabling and Disabling an ESRP Domain ............................................................................758
Advanced ESRP Features.........................................................................................................759
ESRP Tracking..................................................................................................................759
ESRP Port Restart .............................................................................................................762
ESRP Host Attach .............................................................................................................762
ESRP Port Weight and Don’t Count .....................................................................................763
ESRP Groups ....................................................................................................................764
Selective Forwarding .........................................................................................................765
Displaying ESRP Information ...................................................................................................767
Using ELRP with ESRP............................................................................................................767
Using ELRP with ESRP to Recover Loops ............................................................................768
Configuring ELRP..............................................................................................................768
Displaying ELRP Information..............................................................................................770
ESRP Examples ......................................................................................................................770
Single Domain Using Layer 2 and Layer 3 Redundancy.........................................................770
Multiple Domains Using Layer 2 and Layer 3 Redundancy ....................................................772
ESRP Cautions .......................................................................................................................775
Configuring ESRP and IP Multinetting.................................................................................775
ESRP and STP..................................................................................................................775
ESRP and VRRP ...............................................................................................................775
ESRP Groups and Host Attach............................................................................................775
Port Configurations and ESRP ............................................................................................775
Chapter 28: Virtual Router Redundancy Protocol........................................................................... 777

Licensing ...............................................................................................................................777
Overview of VRRP ...................................................................................................................778
VRRP and Hitless Failover—Modular Switches and SummitStack Only ...................................778
Determining the VRRP Master ..................................................................................................779
VRRP Tracking..................................................................................................................779
Electing the Master Router.................................................................................................781
Additional VRRP Highlights......................................................................................................782
VRRP Operation ......................................................................................................................783

Contents
Simple VRRP Network Configuration ...................................................................................783

Fully Redundant VRRP Network..........................................................................................784
VRRP Configuration Parameters................................................................................................785
VRRP Examples ......................................................................................................................786
Configuring the Simple VRRP Network ................................................................................786
Configuring the Fully Redundant VRRP Network...................................................................787
VRRP Cautions .......................................................................................................................788
Assigning Multiple Virtual IP Addresses...............................................................................788
VRRP and ESRP ...............................................................................................................788
Chapter 29: MultiProtocol Label Switching (MPLS) Protocol.......................................................... 789

About MPLS ...........................................................................................................................789
How MPLS Works..............................................................................................................789
MPLS Terms and Acronyms................................................................................................790
Label Switched Paths ........................................................................................................791
MPLS Data Plane ....................................................................................................................791
Label Switch Routers.........................................................................................................792
Supporting Quality of Service Features ................................................................................793
MPLS Layer ......................................................................................................................793
Routing Using LSPs ..........................................................................................................795
About MPLS Layer-2 VPNs.................................................................................................798
Configuring MPLS ...................................................................................................................799
Configuring MPLS LSR ID ..................................................................................................799
Configuring MPLS Interfaces ..............................................................................................799
Enabling MPLS .................................................................................................................800
Enabling MPLS Interfaces..................................................................................................800
Propagation of IP TTL ........................................................................................................800
Configuring Penultimate Hop Popping .................................................................................801
Configuring QoS Mappings .................................................................................................801
Mapping Dot1p to EXP Bits................................................................................................802
Resetting MPLS Configuration Parameter Values ..................................................................802
Displaying MPLS Configuration Information .........................................................................803
Configuration Example.............................................................................................................807
MPLS Configuration Constraints ...............................................................................................809
Chapter 30: Label Distribution Protocol ........................................................................................ 811

Overview of LDP......................................................................................................................811
LDP Neighbor Discovery.....................................................................................................811
Advertising Labels .............................................................................................................812
Propagating Labels ............................................................................................................812
MPLS LDP Signaling Modes.....................................................................................................812
Label Advertisement Modes................................................................................................812
Label Retention Modes ......................................................................................................813
LSP Control Modes............................................................................................................813
Configuring LDP......................................................................................................................814
Enabling LDP on the Switch...............................................................................................814
Enabling LDP on a VLAN ...................................................................................................814
Enabling LDP Loop Detection .............................................................................................815
Configuring an LDP Label Advertisement Filter ....................................................................815
Configuring LDP Session Timers .........................................................................................815

Contents
Displaying LDP Information................................................................................................816

Displaying LDP Interface Information ..................................................................................817
Displaying LDP Peer Session Information ............................................................................818
Displaying LDP Protocol Counters .......................................................................................818
Displaying LDP Labels .......................................................................................................819
Displaying LDP LSP Forwarding Database............................................................................821
Chapter 31: Configuring MPLS VPLS Layer-2 VPNs........................................................................ 823

Overview of MPLS Layer-2 VPNs...............................................................................................823
VPLS Support ...................................................................................................................823
Layer-2 VPN Service Deliminators.......................................................................................824
MPLS Pseudo Wires ..........................................................................................................824
Layer-2 VPN Domains........................................................................................................826
MAC Learning ...................................................................................................................827
Spanning Tree Protocols ....................................................................................................827
IP Protocol Considerations .................................................................................................827
VPLS Layer 2 VPN Characteristics ............................................................................................828
Configuring MPLS Layer-2 VPNs...............................................................................................828
Configuring MPLS for Establishing VPLSs............................................................................829
Creating a VPLS Domain ....................................................................................................829
Deleting a VPLS Domain ....................................................................................................829
Enabling a VPLS Domain ...................................................................................................829
Disabling a VPLS Domain...................................................................................................830
Adding a VPLS Peer ..........................................................................................................830
Deleting a VPLS Peer.........................................................................................................830
Adding a VPLS Service ......................................................................................................830
Deleting a VPLS Service.....................................................................................................831
Enabling a VPLS Service ....................................................................................................831
Disabling a VPLS Service ...................................................................................................831
Configuring VPLS Options ..................................................................................................831
Configuring VPLS MTU ......................................................................................................832
Configuring VPLS FDB Aging Timer.....................................................................................832
Unconfiguring VPLS Options ..............................................................................................832
Displaying VPLS VPN Status ..............................................................................................832
Displaying VPLS VPN Peers................................................................................................833
VPLS VPN Configuration Examples ...........................................................................................833
Basic Point-to-Point VPLS Configuration Example ................................................................833
Multipoint Full Mesh VPLS Configuration Example ...............................................................834
Chapter 32: Configuring RSVP-TE................................................................................................. 837

RSVP Elements.......................................................................................................................838
Message Types..................................................................................................................838
Reservation Styles .............................................................................................................841
Traffic Engineering..................................................................................................................842
RSVP Tunneling ................................................................................................................842
RSVP Objects ...................................................................................................................842
Establishing RSVP-TE LSPs .....................................................................................................844
RSVP-TE Implementation ........................................................................................................844
Explicit Route Path LSPs ...................................................................................................845
Route Recording ...............................................................................................................846
LSP Session Attributes ......................................................................................................846

Contents
Bandwidth Reservation ......................................................................................................846

Bandwidth Management for RSVP-TE LSPs .........................................................................847
Redundant LSPs ...............................................................................................................848
Improving LSP Scaling ......................................................................................................849
Configuring RSVP-TE...............................................................................................................851
Configuring RSVP-TE on the Switch ....................................................................................852
Configuring RSVP-TE on a VLAN.........................................................................................852
Configuring RSVP-TE Protocol Parameters ...........................................................................852
Create an RSVP-TE Path ....................................................................................................853
Configuring an Explicit Route .............................................................................................853
Reserve Bandwidth for MPLS .............................................................................................854
Create an RSVP-TE Profile .................................................................................................855
Configuring an RSVP-TE Profile ..........................................................................................855
Configuring an RSVP-TE LSP .............................................................................................856
Adding a Path to an RSVP-TE LSP ......................................................................................856
Displaying RSVP-TE LSP Configuration Information ..............................................................857
Displaying the RSVP-TE Paths............................................................................................857
Displaying the RSVP-TE Path Profile ...................................................................................857
Displaying the RSVP-TE LSP ..............................................................................................858
Chapter 33: Operational Administration and Management.............................................................. 863

Failure Detection and Alerts .....................................................................................................863
Diagnostic Information ............................................................................................................863
Chapter 34: IPv4 Unicast Routing................................................................................................. 865

Overview of IPv4 Unicast Routing .............................................................................................865
Router Interfaces ..............................................................................................................866
Populating the Routing Table .............................................................................................867
Proxy ARP ..............................................................................................................................870
ARP-Incapable Devices ......................................................................................................870
Proxy ARP Between Subnets ..............................................................................................871
Configuring IPv4 Unicast Routing .............................................................................................871
Verifying the IPv4 Unicast Routing Configuration .......................................................................871
Routing Configuration Example.................................................................................................872
IPv4 Multinetting ....................................................................................................................873
Multinetting Topology ........................................................................................................874
How Multinetting Affects Other Features .............................................................................875
Configuring IPv4 Multinetting.............................................................................................878
IP Multinetting Examples ...................................................................................................879
Configuring DHCP/BOOTP Relay ...............................................................................................879
Configuring the DHCP Relay Agent Option (Option 82) .........................................................880
Verifying the DHCP/BOOTP Relay Configuration ...................................................................881
UDP Forwarding......................................................................................................................881
Configuring UDP Forwarding ..............................................................................................881
UDP Echo Server ..............................................................................................................883
IP Broadcast Handling.............................................................................................................883
IP Broadcast Handling Details ............................................................................................883
Command-line Support for IP Broadcast Handling ................................................................884
IP Route Compression .......................................................................................................885

Contents
Chapter 35: IPv6 Unicast Routing................................................................................................. 893

Overview of IPv6 Unicast Routing .............................................................................................893
Router Interfaces ..............................................................................................................894
Specifying IPv6 Addresses .................................................................................................895
Neighbor Discovery Protocol ...............................................................................................897
Populating the Routing Table .............................................................................................899
Configuring IP Unicast Routing ................................................................................................901
Verifying the IP Unicast Routing Configuration .....................................................................902
IPv6 Forwarding Behavior ........................................................................................................902
Hardware IPv6 Unicast Forwarding Support .........................................................................902
Hardware Tunnel Support ..................................................................................................903
Routing Configuration Example.................................................................................................903
Tunnel Configuration Examples ................................................................................................905
6in4 Tunnel Configuration Example ....................................................................................905
6to4 Tunnel Configuration Example ....................................................................................907
Chapter 36: RIP........................................................................................................................... 911

Overview ................................................................................................................................911
RIP Versus OSPF ..............................................................................................................911
Advantages of RIP and OSPF..............................................................................................912
Overview of RIP ......................................................................................................................912
Routing Table ...................................................................................................................912
Split Horizon ....................................................................................................................913
Poison Reverse .................................................................................................................913
Triggered Updates .............................................................................................................913
Route Advertisement of VLANs ...........................................................................................913
RIP Version 1 Versus RIP Version 2 ....................................................................................913
Route Redistribution ...............................................................................................................914
Configuring Route Redistribution ........................................................................................914
RIP Configuration Example ......................................................................................................915
Chapter 37: RIPng ....................................................................................................................... 919

Overview ................................................................................................................................919
RIPng Versus OSPFv3........................................................................................................919
Advantages of RIPng and OSPFv3.......................................................................................920
Overview of RIPng ...................................................................................................................920
Routing Table ...................................................................................................................920
Split Horizon ....................................................................................................................921
Poison Reverse .................................................................................................................921
Triggered Updates .............................................................................................................921
Route Advertisement of VLANs ...........................................................................................921
RIPng Configuration Example ...................................................................................................922
Chapter 38: OSPF ........................................................................................................................ 925

Overview of OSPF....................................................................................................................925
Licensing .........................................................................................................................925
OSPF Edge Mode ..............................................................................................................926

Contents
Link State Database ..........................................................................................................926

Graceful OSPF Restart .......................................................................................................927
Areas ...............................................................................................................................928
Point-to-Point Support .......................................................................................................932
OSPF Timers and Authentication ........................................................................................934
Configuring OSPF....................................................................................................................934
Configuring OSPF Wait Interval...........................................................................................935
OSPF Wait Interval Parameters ...........................................................................................935
OSPF Configuration Example....................................................................................................936
Configuration for ABR1......................................................................................................937
Configuration for IR1 .........................................................................................................937
Displaying OSPF Settings.........................................................................................................938
Chapter 39: OSPFv3 .................................................................................................................... 939

Overview of OSPFv3 ................................................................................................................939
Licensing .........................................................................................................................939
Link State Database ..........................................................................................................939
Areas ...............................................................................................................................940
Link-Type Support.............................................................................................................943
OSPFv3 Timers .................................................................................................................945
OSPFv3 Configuration Example ................................................................................................946
Configuration for Router 1..................................................................................................947
Chapter 40: Border Gateway Protocol ........................................................................................... 949

Licensing ...............................................................................................................................949
Overview ................................................................................................................................950
BGP Attributes........................................................................................................................950
BGP Communities ...................................................................................................................950
BGP Features .........................................................................................................................951
Route Reflectors ...............................................................................................................951
Route Confederations ........................................................................................................953
Route Aggregation .............................................................................................................956
Using the Loopback Interface .............................................................................................956
BGP Peer Groups ..............................................................................................................956
BGP Route Flap Dampening ...............................................................................................957
BGP Route Selection .........................................................................................................959
Stripping Out Private AS Numbers from Route Updates ........................................................959
Route Redistribution .........................................................................................................960
BGP Static Network...........................................................................................................960
Graceful BGP Restart ........................................................................................................961
Chapter 41: Multicast Routing and Switching................................................................................ 963

Overview ................................................................................................................................963
Multicast Static Routes......................................................................................................963

Contents
PIM Overview....................................................................................................................964
IGMP Overview .................................................................................................................967
Configuring IP Multicast Routing Using PIM ..............................................................................969
PIM Configuration Examples ..............................................................................................970
Multicast VLAN Registration.....................................................................................................972
Basic MVR Deployment......................................................................................................973
Inter-Multicast VLAN Forwarding ........................................................................................977
MVR Configurations...........................................................................................................978
Chapter 42: IPv6 Multicast .......................................................................................................... 985

Overview ................................................................................................................................985
MLD Overview...................................................................................................................985
Chapter 43: Multicast Source Discovery Protocol (MSDP) ............................................................. 987

Overview ................................................................................................................................987
Supported Platforms..........................................................................................................988
Limitations .......................................................................................................................988
PIM Border Configuration.........................................................................................................988
MSDP Peers ...........................................................................................................................988
MSDP Default Peers ..........................................................................................................989
Peer Authentication...........................................................................................................989
Policy Filters.....................................................................................................................990
SA Request Processing ......................................................................................................990
MSDP Mesh-Groups ................................................................................................................990
SA Cache ...............................................................................................................................991
Maximum SA Cache Entry Limit .........................................................................................992
Redundancy ...........................................................................................................................992
Scaling Limits ........................................................................................................................993
SNMP MIBs ...........................................................................................................................993
Configuration for MSDP-1 ..................................................................................................994
Configuration for MSDP-2 ..................................................................................................994
Part 3: Appendixes
Appendix A: Software Upgrade and Boot Options........................................................................... 997
Downloading a New Image .......................................................................................................997
Image Filename Prefixes ....................................................................................................998
Understanding the Image Version String ..............................................................................998
Software Signatures...........................................................................................................999
Selecting a Primary or a Secondary Image ...........................................................................999
Installing a Core Image ......................................................................................................999
Installing a Modular Software Package ..............................................................................1002
Rebooting the Switch ......................................................................................................1005
Rebooting the Management Module—Modular Switches Only ..............................................1005
Understanding Hitless Upgrade—Modular Switches Only ..........................................................1006
Understanding the I/O Version Number—BlackDiamond 8800 Series Switch Only.................1006
Performing a Hitless Upgrade ...........................................................................................1007
Hitless Upgrade Examples................................................................................................1012

Contents
Saving Configuration Changes ................................................................................................1014

Viewing a Configuration ...................................................................................................1015
Returning to Factory Defaults ...........................................................................................1016
ASCII-Formatted Configuration Files .................................................................................1016
Using TFTP to Upload the Configuration..................................................................................1018
Using TFTP to Download the Configuration ..............................................................................1019
Synchronizing Nodes—Modular Switches and SummitStack Only ..............................................1020
Additional Behavior on the BlackDiamond 8800 Series Switch Only.....................................1021
Automatic Synchronization of Configuration Files ...............................................................1021
Accessing the Bootloader .......................................................................................................1022
Upgrading the BootROM—BlackDiamond 10808 Switch Only ...................................................1023
Upgrading the BootROM—Summit Family of Switches and SummitStack Only............................1023
Accessing the Bootstrap CLI on the Summit Family of Switches ..........................................1024
Upgrading the Firmware—BlackDiamond 8800 Series Switches Only.........................................1024
Upgrading the Firmware—BlackDiamond 12800 Series Switches Only.......................................1026
Accessing the Bootstrap CLI on the BlackDiamond 12800 Series Switches ..........................1026
Displaying the BootROM and Firmware Versions.......................................................................1027
Appendix B: Troubleshooting ..................................................................................................... 1029

Troubleshooting Checklists.....................................................................................................1030
Layer 1 ..........................................................................................................................1030
Layer 2 ..........................................................................................................................1030
Layer 3 ..........................................................................................................................1031
LEDs....................................................................................................................................1033
Using the Command Line Interface .........................................................................................1034
General Tips and Recommendations .................................................................................1035
MSM Prompt—Modular Switches Only ..............................................................................1036
Node Prompt—SummitStack Only ....................................................................................1037
Command Prompt ...........................................................................................................1037
Port Configuration ...........................................................................................................1037
Software License Error Messages ......................................................................................1038
VLANs............................................................................................................................1039
STP ...............................................................................................................................1039
ESRP .............................................................................................................................1040
VRRP .............................................................................................................................1041
Using Standalone ELRP to Perform Loop Tests ........................................................................1041
About Standalone ELRP...................................................................................................1042
Configuring Standalone ELRP...........................................................................................1042
Displaying Standalone ELRP Information...........................................................................1043
Using the Rescue Software Image—Modular Switches Only.......................................................1044
Obtaining the Rescue Image from a TFTP Server ................................................................1044
Obtaining the Rescue Image from an External Compact Flash Memory Card—
BlackDiamond 8800 Series Switch Only............................................................................1045
Rescuing a Node in a SummitStack ..................................................................................1047
Debug Mode .........................................................................................................................1048
Saving Debug Information ......................................................................................................1048
Enabling the Switch to Send Debug Information to the Memory Card ...................................1049
Copying Debug Information to an External Memory Card—Modular Switches Only .................1049
Copying Debug Information to a TFTP Server .....................................................................1050
Managing Debug Files .....................................................................................................1050

Contents
Evaluation Precedence for ACLs .............................................................................................1054

TOP Command......................................................................................................................1055
TFTP Server Requirements.....................................................................................................1055
System Health Check ............................................................................................................1055
Overview of the System Health Checker .............................................................................1055
Enabling and Disabling Backplane Diagnostic Packets on the Switch—
Modular Switches Only ....................................................................................................1056
Configuring Backplane Diagnostic Packets on the Switch—Modular Switches Only ................1057
System Odometer..................................................................................................................1057
Monitored Components ....................................................................................................1057
Recorded Statistics .........................................................................................................1057
Temperature Operating Range ................................................................................................1059
Unsupported Module Type—BlackDiamond 10808 Switch and BlackDiamond
8800 Series Switches Only ....................................................................................................1059
Corrupted BootROM on the BlackDiamond 8800 Series Switch .................................................1060
Inserting Powered Devices in the PoE Module—BlackDiamond 8800 Series Switch Only .............1060
Modifying the Hardware Table Hash Algorithm—BlackDiamond 8800 Series Switch,
SummitStack, and the Summit Family of Switches Only ...........................................................1060
Configuring the Hash Algorithm ........................................................................................1061
Viewing the Hash Algorithm Setting ..................................................................................1061
Untagged Frames on the 10 Gbps Module—BlackDiamond 10808 Switch Only ..........................1062
Understanding the Error Reading Diagnostics Message—Summit Family of Switches Only ...........1062
Running MSM Diagnostics from the Bootloader—BlackDiamond 10808 Switch Only...................1062
Contacting Extreme Networks Technical Support......................................................................1063
Appendix C: CNA Agent.............................................................................................................. 1065

Overview ..............................................................................................................................1065
Redundancy—Modular Switches and SummitStack Only.....................................................1066
Downloading the CNA Agent Software Module..........................................................................1066
Running the Tests .................................................................................................................1066
Configuring the CNA Agent ....................................................................................................1067
Enabling the CNA Agent ..................................................................................................1067
Connecting to the CNA Server ..........................................................................................1067
Configuring the Interface .................................................................................................1068
Clearing the Counters ......................................................................................................1068
Displaying CNA Agent Information ....................................................................................1068
Troubleshooting ..............................................................................................................1068
Appendix D: Supported Protocols, MIBs, and Standards............................................................... 1069

MIB Support Details ..............................................................................................................1073
Standard MIBs................................................................................................................1073
Extreme Networks Proprietary MIBs ..................................................................................1091
Index of Commands ................................................................................................................... 1103
Glossary ................................................................................................................................... 1111
Index ........................................................................................................................................ 1139

Contents

Preface
This preface provides an overview of this guide, describes guide conventions, and lists other
publications that might be useful.
Introduction
This guide provides the required information to configure ExtremeXOS™ software version 12.0 running
on switches from Extreme Networks®.
The guide is intended for use by network administrators who are responsible for installing and setting
up network equipment. It assumes a basic working knowledge of:
● Local area networks (LANs)
● Ethernet concepts
● Ethernet switching and bridging concepts
● Routing concepts
● Internet Protocol (IP) concepts
● Routing Information Protocol (RIP) and Open Shortest Path First (OSPF)
● Border Gateway Protocol (BGP-4) concepts
● IP multicast concepts
● Protocol Independent Multicast (PIM) concepts
● Simple Network Management Protocol (SNMP)
NOTE
If the information in the release notes shipped with your switch differs from the information in this guide, follow the
release notes.
Terminology
When features, functionality, or operation is specific to a switch family, the family name is used.
Explanations about features and operations that are the same across all product families simply refer to
the product as the “switch.”
Conventions
This section describes conventions used in the documentation:
● Platform-Dependent Conventions on page 30
● Text Conventions on page 30

Preface
Platform-Dependent Conventions
Unless otherwise noted, all information applies to all platforms supported by ExtremeXOS software,
which are the following:
● BlackDiamond® 10808 switch

● BlackDiamond 8800 series of switches
● Summit® family of switches
● BlackDiamond 12800 R-series switch
● BlackDiamond 12800 series switches
When a feature or feature implementation applies to specific platforms, the specific platform is noted in
the heading for the section describing that implementation.
Finally, minor differences in platform implementations are called out in a note, as shown below:
NOTE
This is a note.
Text Conventions
Table 1 and Table 2 list conventions that are used throughout this guide.
Table 1: Notice icons

Icon Notice Type Alerts you to...
Note Important features or instructions.
Caution Risk of personal injury, system damage, or loss of data.
Warning Risk of severe personal injury.

Related Publications
Table 2: Text conventions

Convention Description
Screen displays This typeface indicates command syntax, or represents information as it appears on
the screen.
The words “enter” When you see the word “enter” in this guide, you must type something, and then
and “type” press the Return or Enter key. Do not press the Return or Enter key when an
instruction simply says “type.”
[Key] names Key names are written with brackets, such as [Return] or [Esc].
If you must press two or more keys simultaneously, the key names are linked with a
plus sign (+). Example:
Press [Ctrl]+[Alt]+[Del].
Words in italicized type Italics emphasize a point or denote new terms at the place where they are defined in
the text. (Italics are also used when referring to publication titles.)
Related Publications
The publications related to this one are:
● ExtremeXOS release notes
● ExtremeXOS Command Reference Guide
● Extreme Networks Consolidated ExtremeXOS Hardware Installation Guide
Documentation for Extreme Networks products is available on the World Wide Web at the following
location:
http://www.extremenetworks.com/
Using ExtremeXOS Publications Online

You can access ExtremeXOS publications by downloading them from the Extreme Networks World
Wide Web location or from your ExtremeWare product CD. Publications are provided in Adobe®
Portable Document Format (PDF). Displaying or printing PDF files requires that your computer be
equipped with Adobe Reader® software, which is available free of charge from Adobe Systems
Incorporated.
The user guide PDF file provides links that connect you directly to relevant command information in
the command reference guide PDF file. This quick-referencing capability enables you to easily find
detailed information in the command reference guide for any command mentioned in the user guide.
To ensure that the quick-referencing feature functions properly:

1 Download both the user guide PDF file and the command reference guide PDF file to the same
destination directory on your computer.
2 You may open one or both PDF files. To enable cross-referenced linking between the user guide and
command reference guide; however, it is recommended that for ease of use, you keep both files open
concurrently on your computer desktop.

Preface
NOTE
If you activate a cross-referencing link from the concepts guide PDF file to the command reference PDF file when
the command reference PDF file is closed (that is, not currently open on your computer desktop), the system will
close the user guide PDF file and open the command reference PDF file. To keep both PDF files open when you
activate a cross-reference link, open both PDF files before using the link.
All of these documents are available in Adobe PDF format. You must have Acrobat Reader 5.0 or later
to properly open the documents. You must have Acrobat Reader 6.0 or later to use the cross-reference
linking feature from the ExtremeXOS Concepts Guide to the ExtremeXOS Command Reference Guide.

1 Using ExtremeXOS
1 ExtremeXOS Overview
This chapter provides an overview of the ExtremeXOS version 12.0 software, and includes the following
topics:
● Platforms and Required Software Versions on page 35

● Software Licenses and Feature Packs on page 37
Platforms and Required Software Versions

ExtremeXOS is the full-featured software operating system that is designed to run on the Extreme
Networks devices.
NOTE
See “Platform-Dependent Conventions” in the Preface for information on how to determine platform-specific
implementations or features.
ExtremeXOS supports the following platforms:

● BlackDiamond 12804 switch—ExtremeXOS 11.4.1 and higher
■ MSM-5
■ MSM-5R
● BlackDiamond 12802 switch—ExtremeXOS 12.0 and higher
■ MSM-5
■ MSM-5R
CAUTION
You cannot mix BlackDiamond 12800 R-series modules with non-R-series modules in a BlackDiamond 12800 series
chassis. If you attempt to mix these, the system returns an error message.
● BlackDiamond® 10808 switch—ExtremeXOS 10.1 and higher

■ MSM-1
■ MSM-1XL
● BlackDiamond 8810 switch (formerly known as Aspen)—ExtremeXOS 11.1 and higher
■ G48T module—ExtremeXOS 11.1 and higher
■ G48P module—ExtremeXOS 11.1 and higher
■ G24X module—ExtremeXOS 11.1 and higher
■ 10G4X module—ExtremeXOS 11.1 and higher
These modules are hereinafter referred to as BlackDiamond 8800 original modules.

ExtremeXOS Overview
● BlackDiamond 8806 switch—ExtremeXOS 11.3.1. and higher

The BlackDiamond 8806 and 8810 switch and all modules that function with both chassis are
hereinafter referred as the BlackDiamond 8800 series switch.
● MSMs for the BlackDiamond 8800 series switch
■ MSM-G8X—ExtremeXOS 11.1 and higher
■ MSM-48—ExtremeXOS 11.6 and higher
● BlackDiamond a-series modules (function with BlackDiamond 8806 and 8810 chassis)
■ G48Ta module—ExtremeXOS 11.5 and higher
■ G48Xa module—ExtremeXOS 11.5 and higher
■ 10G4Xa module—ExtremeXOS 11.6 and higher
■ 10G4Ca module—ExtremeXOS 12.0 and higher
● BlackDiamond e-series modules (function with BlackDiamond 8806 and 8810 chassis)—ExtremeXOS
11.5 and higher
■ G48Te module
■ G48Pe module
● Summit® X450 series switch
■ Summit X450-24t - ExtremeXOS 11.2 and higher
■ Summit X450-24x - ExtremeXOS 11.2 and higher
● Summit X450a series switches
■ Summit X450a-24t—ExtremeXOS 11.5 and higher
■ Summit X450a-48t—ExtremeXOS 11.5 and higher
■ Summit X450a-24t DC—ExtremeXOS 11.5 and higher
■ Summit X450a-24x—ExtremeXOS 11.6 and higher
■ Summit X450a-24x DC—ExtremeXOS 11.6 and higher
■ Summit X450a-48t DC—ExtremeXOS 11.6 and higher
■ Summit X450a-48tDC - ExtremeXOS 11.6 and higher
● Summit X450e series switches
■ Summit X450e-24p—ExtremeXOS 11.5 and higher
■ Summit X450e-48p—ExtremeXOS 11.6 and higher
■ X250e-48t—ExtremeXOS 12.0 and higher
■ X250e-48p—ExtremeXOS 12.0 and higher
■ X250e-24t—ExtremeXOS 12.0 and higher
■ X250e-24p—ExtremeXOS 12.0 and higher
■ X250e-24x—ExtremeXOS 12.0 and higher
■ X250e-24xTAA—ExtremeXOS 12.0 and higher
■ SummitStack—ExtremeXOS 12.0 and higher
The term “Summit family” refers to all Summit switches. A SummitStack is any combination of up
to eight Summit Family switches that are connected together with stacking cables and are running
ExtremeXOS 12.0 or higher.

Software Licenses and Feature Packs
NOTE
For more information on Extreme Networks switch hardware components, see the Extreme Networks Consolidated
ExtremeXOS Hardware Installation Guide.

This section discusses the software licensing and feature packs for the Extreme Networks devices, as
follows:
● Software Licensing Overview on page 37
● Software License Levels on page 38
● Obtaining a License Voucher on page 40
● Enabling and Verifying Licenses on page 40
● Security Licensing on page 41
● Feature Packs Overview on page 41
● Obtaining and Enabling Feature Packs on page 42
● Displaying Software Licenses and Feature Packs on page 43
Software Licensing Overview

All Extreme Networks products that run ExtremeXOS software offer a default software license, which
you can upgrade if your network environment requires additional functionality. There are four levels of
software licenses: Edge, Advanced Edge, Core, and Advanced Core. The Edge license includes a basic
feature set and the Core/Advanced Core license includes the highest level of functionality. Each license
level builds on the features of the license level below it. For example, the Advanced Edge license
includes all of the features in the Edge license, plus the features in the Advance Edge license. The Core
license includes all of the features in the Advanced Edge license, plus the features in the Core license,
and so on. See Table 3 for a list of software licenses available for each platform that supports
ExtremeXOS.
Table 3: Licensing requirements for ExtremeXOS version 12.0

Summit
X450e and Summit X450 BlackDiamond
X250e and X450a 8800 series BlackDiamond BlackDiamond BlackDiamond
License level switches switches switches 10808 switch 12802 switch 12804 switch
Edge Default – – – – –
Advanced Edge Requires Default Default – Default –
license
Core Not Requires a Requires a Default for Requires a Default
available license license chassis that license
upgrade upgrade ship with an upgrade
MSM-1
Advanced Core Not Not available Not available Default for Not available Not available
available chassis that
ship with an
MSM-1XL

ExtremeXOS Overview
NOTE
With ExtremeXOS software version 11.2, BGP functionality moved from the Advanced Core license to the Core
license. After version 11.2, BGP functionality is supported in the Core license and the Advanced Core license is no
longer offered.
Software keys are stored in the EEPROM of the chassis and, once enabled, persist through reboots,
software upgrades, power outages, and reconfigurations. Because the license is stored in the EEPROM
of the chassis (not the MSM card), the license persists even if you change MSM cards. The keys are
unique to the chassis or switch and are not transferable.
The licensing levels required for each feature, if any, are outlined in the discussion of each feature. If
licensing is not mentioned, the full functionality of that feature is available on every switch.
If you attempt to execute a command and you either do not have the required license or have reached
the limits defined by the current license level, the system returns one of the following messages:
Error: This command cannot be executed at the current license level.

Error: You have reached the maximum limit for this feature at this license level.
You can obtain a trial license, which allows you to use the license for either 30, 60, or 90 days; you can
downgrade trial licenses to a lower software license level during the trial period. After you enable the
trial license, the switch behaves as if all software license levels and feature packs were enabled. The trial
license key contains all the necessary information on the license level and the number of days. Once you
have a trial license for any one of these periods, you cannot extend the time of the trial. And, trial
licenses can be applied only once for each software version; if you upgrade to a different software
version, you can reapply your trial license.
Software License Levels

This section describes the licensing capabilities of the various switches.
Upgrading License on the Summit X450e Series Switch Only

The Summit X450e series switch ships with an Edge license. With ExtremeXOS 12.0, you can obtain the
Advanced Edge license for these switches. The Core and Advanced Core licenses are not available for
the Summit X450e and X250 series switches.
NOTE
Refer to the specific chapter of the ExtremeXOS Concepts Guide to determine if the Advanced Edge license is
required for some functionality. If not noted, all functionality is available, and a license is not required.
Upgrading License on the BlackDiamond 8800 Series Switch and Summit X450 and
X450a Series Switches Only
The BlackDiamond 8800 series switch and the Summit X450 and X450a series switches ship with an
Advanced Edge license. With ExtremeXOS 12.0, you can obtain the Core license for these switches. The
Core license provides additional functionality for some features, as well as Border Gateway Protocol
(BGP) functionality, on these switches.

The Advanced Core license is not available for the BlackDiamond 8800 series switch or the Summit
X450 or X450a series switches.
NOTE
Refer to the specific chapter of the ExtremeXOS Concepts Guide to determine if the Core license is required for
some functionality. If not noted, all functionality is available, and a license is not required.
The BlackDiamond 10808 Switch Only

The licensing levels on the BlackDiamond 10808 switch depend on the MSM that shipped with your
system:
● The MSM-1 ships with a Core license; you cannot upgrade to an Advanced Core license.
● The MSM-1XL ships with an Advanced Core license.
NOTE
Beginning with ExtremeXOS version 11.2, you can run BGP with a Core license on the BlackDiamond 10808 switch;
that is both the MSM1 and the MSM1XL now run BGP.
The BlackDiamond 12800 Series Switches Only

The Advanced Edge license is the default for a BlackDiamond 12802; the Core license is the default for a
BlackDiamond 12804.
Licensed and Feature Pack Features in ExtremeXOS Version 11.2 or later

The following features require an Advanced Edge license when you are using ExtremeXOS version 11.6
or higher:
● vMANTM
● VRRP
● OSPF Edge Mode
● PIM Edge Mode
The following features require a Core license when you are using ExtremeXOS version 11.2 or higher:
● EAPS full (multiple rings with multiple interconnecting points)
● PIM DM full
● PIM SM full
● OSPFv2 full (not limited to 2 adjacencies)
● OSPFv3 full (not limited to 2 adjacencies)
● BGP4
● MSDP

ExtremeXOS Overview
The following feature requires a Feature Pack license when you are using ExtremeXOS version 11.6 or
higher:
● MPLS—Available only on the BlackDiamond 10808 MSM-1XL and BlackDiamond 12800R series
switches
● MEF—Available only on the BlackDiamond 12800R series switches
Obtaining a License Voucher

You can order the desired functionality from the factory, using the appropriate model of the desired
product. If you order licensing from the factory, the license arrives in a separate package from the
switch. After the license key is installed, it should not be necessary to enter the information again.
However, Extreme Networks recommends keeping the certificate for your records.
You can obtain a regular license or a trial license, which allows you use of the license for either 30, 60 or
90 days; you cannot downgrade licenses. The software key contains all the necessary information on the
license level, whether regular or trial, and number of days for trial license.
You can upgrade the license of an existing product by purchasing a license voucher from Extreme
Networks. Contact your supplier to purchase a voucher.
The voucher contains information and instructions on obtaining a license key for the switch using the
Extreme Networks Support website at:
http://www.extremenetworks.com/support/techsupport.asp
or by phoning Extreme Networks Technical Support at:

● (800) 998-2408
● (408) 579-2826
After you have obtained a license, you cannot downgrade licenses. The license key contains all the
necessary information on the license level.
Enabling and Verifying Licenses

To enable the license, use the following command:
enable license {software} <key>
You can enable the software license key one switch at a time by using the command listed above, or you
can download a license file onto the switch and then enable the licenses through file upload. (The
license file has the extension <.xlic>.) You download the file using TFTP into the switch. The file can
contain licenses for some or all of the Extreme Networks switches that the customer owns. During file
processing, only those license keys destined for the specific switch are used to attempt enabling the
licenses. The license file is a text file that has the switch serial number, software license type, and license
key; it is removed from the switch after the licenses are enabled. To enable licenses using the license file,
use the following command:
enable license file <filename>
To verify the licensing on the switch, use the following command:
show licenses

Security Licensing
Certain additional ExtremeXOS features, such as the use of SSH2 encryption, may be under United
States export restriction control. Once the export restrictions are met, you can obtain information on
enabling these features at no charge from Extreme Networks.
The SSH2 feature is in a separate, loadable software module, which must be installed on the Extreme
Networks switches.
Obtaining a Security License

To obtain information on installing features that require export restriction, access the Extreme Networks
Support website at:
http://www.extremenetworks.com/go/security.htm
Fill out a contact form to indicate compliance or noncompliance with the export restrictions. If you are
in compliance, you will be given information that will allow you to install security features. You need
the following capabilities to ensure the process works:
● Email address
● Ability to use a Web browser to download a file
● WinZip format to uncompress a .zip file
Security Features Under License Control

ExtremeXOS software supports the SSH2 protocol, which allows the encryption of sessions between an
SSH2 client and an Extreme Networks switch, as well as the Secure Copy Protocol (SCP). The
encryption methods used are under export restriction control.
Feature Packs Overview

In addition to licenses, some Extreme Networks features are available only by obtaining a separate
feature pack. Like licenses, feature packs may be obtained from Extreme Networks at an additional cost.
Feature packs must be purchased for each switch; they are tied to a switch chassis and are not portable.
Feature packs remain in place during software upgrades and or module replacements.
MPLS Feature Pack—BlackDiamond 10808 MSM-1XL and BlackDiamond 12800R

Switches Only
MPLS support, including VPLS L2 VPNs, may be obtained from Extreme Networks at an additional
cost, as a feature pack. A separate feature pack must be purchased for each switch that will run MPLS
and VPLS.
After you enable the feature pack, the MPLS CLI commands are visible; without an enabled MPLS
feature pack, you will not see any MPLS commands on the console.

ExtremeXOS Overview
MEF Feature Pack—BlackDiamond 12800R Switches Only

ExtremeXOS version 12.0 software introduces a feature pack for bandwidth mode hierarchical Quality
of Service (QoS), which meets the standards of the Metro Ethernet Forum (MEF) 14. Strict priority mode
hierarchical QoS, introduced in ExtremeXOS 11.4 does not require that you purchase a separate feature
pack to implement.
If you do not have an enabled feature pack for bandwidth mode hierarchical QoS, you do not see any of
the associated CLI commands. Additionally, after enabling the feature pack, you must issue the
following CLI command in order to see the bandwidth mode commands:
configure traffic mode pr-and-cr
Obtaining and Enabling Feature Packs

Contact your supplier to purchase a voucher for feature packs. The voucher contains information and
instructions on obtaining a feature pack for the switch using the Extreme Networks Support website at:
http://www.extremenetworks.com/support/techsupport.asp
or by phoning Extreme Networks Technical Support at:

● (800) 998-2408
● (408) 579-2826
To activate the MPLS license, you need to know the serial number of the switch. The show version
command will display the switch serial number. The example below is from a BlackDiamond 10808
switch.
BD-10808.3 # show version
Chassis : 804300-00-09 0438F-00850 Rev 9.0

Slot-1 : 804403-00-08 0405F-00063 Rev 8.0 BootROM: 1.3.0.0 IMG: 12.0.0.6
Slot-2 : 804406-00-02 0511F-00605 Rev 2.0
Slot-3 : 804406-00-02 0511F-00602 Rev 2.0
Slot-4 :
Slot-5 :
Slot-6 :
Slot-7 :
Slot-8 :
MSM-A : 804301-00-09 0434F-00125 Rev 9.0 BootROM: 1.0.1.5 IMG: 12.0.0.6
MSM-B : 804301-00-09 0437F-00026 Rev 9.0 BootROM: 1.0.1.5 IMG: 12.0.0.6
PSUCTRL-1 : 704025-00-06 0438F-00089 Rev 6.0 BootROM: 5.3
Image : ExtremeXOS version 12.0.0.1 v1160b1 by release-manager

on Mon Oct 8 10:18:06 PDT 2006
BootROM : 1.0.1.5

When provided with the chassis serial number, Extreme Networks will generate a license key that will
activate the feature pack key. The feature pack key is a 20-digit hex number and is valid only for the
switch with that serial number. Once the feature pack key is entered, the feature is activated and can be
enabled, as shown in the following:
* BD-10808.6 # enable license 9a13-69bd-80e9-45b8-f7fd

Enabled license successfully.
The output of the show licenses command shows that the feature pack has been licensed.
Displaying Software Licenses and Feature Packs

You can display information on the type of license and feature pack on your Extreme Networks device
by using the show licenses command.
NOTE
For default settings of individual ExtremeXOS features, see individual chapters in this guide.

ExtremeXOS Overview

2 Accessing the Switch
This chapter describes the following topics:

● Understanding the Command Syntax on page 45
● Port Numbering on page 48
● Line-Editing Keys on page 50
● Command History on page 50
● Common Commands on page 50
● Accessing the Switch the First Time on page 53
● Configuring Management Access on page 54
● Managing Passwords on page 59
● Access to Both MSM Console Ports—Modular Switches Only on page 61
● Domain Name Service Client Services on page 61
● Checking Basic Connectivity on page 62
● Displaying Switch Information on page 64
Understanding the Command Syntax

This section describes the steps to take when entering a command. Refer to the sections that follow for
detailed information on using the command line interface (CLI).
ExtremeXOS command syntax is described in detail in the ExtremeXOS Command Reference Guide. Some
commands are also described in this user guide, in order to describe how to use the features of the
ExtremeXOS software. However, only a subset of commands are described here, and in some cases only
a subset of the options that a command supports. The ExtremeXOS Command Reference Guide should be
considered the definitive source for information on ExtremeXOS commands.
You may enter configuration commands at the # prompt. At the > prompt, you may enter only
monitoring commands, not configuration commands. When you log in as administrator (which has read
and write access), you see the # prompt. When you log in as user (which has only read access), you will
see the > prompt. As you are booting up, you may see the > command prompt. When the bootup
process is complete, the # prompt is displayed.
When entering a command at the prompt, ensure that you have the appropriate privilege level. Most
configuration commands require you to have the administrator privilege level. For more information on
setting CLI privilege levels, see the ExtremeXOS Command Reference Guide. To use the CLI:
1 Enter the command name.
If the command does not include a parameter or values, skip to step 3. If the command requires
more information, continue to step 2.

Accessing the Switch
2 If the command includes a parameter, enter the parameter name and values.
The value part of the command specifies how you want the parameter to be set. Values include
numerics, strings, or addresses, depending on the parameter.
3 After entering the complete command, press [Return].
NOTE
If an asterisk (*) appears in front of the command line prompt, it indicates that you have outstanding configuration
changes that have not been saved. For more information on saving configuration changes, see Appendix A, “Software
Upgrade and Boot Options.”
This section describes the following topics:

● Syntax Helper on page 46
● Command Shortcuts on page 47
● Names on page 47
● Symbols on page 48
● Limits on page 48
Syntax Helper
The CLI has a built-in syntax helper. If you are unsure of the complete syntax for a particular
command, enter as much of the command as possible and press [Tab] or [?]. The syntax helper provides
a list of options for the remainder of the command and places the cursor at the end of the command
you have entered so far, ready for the next option.
If you enter an invalid command, the syntax helper notifies you of your error and indicates where the
error is located.
If the command is one where the next option is a named component (such as a VLAN, access profile, or
route map), the syntax helper also lists any currently configured names that might be used as the next
option. In situations where this list is very long, the syntax helper lists only one line of names, followed
by an ellipses (...) to indicate that there are more names that can be displayed.
The syntax helper also provides assistance if you have entered an incorrect command.
Abbreviated Syntax
Abbreviated syntax is the shortest unambiguous allowable abbreviation of a command or parameter.
Typically, this is the first three letters of the command. If you do not enter enough letters to allow the
switch to determine which command you mean, the syntax helper provides a list of the options based
on the portion of the command you have entered.
NOTE
When using abbreviated syntax, you must enter enough characters to make the command unambiguous and
distinguishable to the switch.

Understanding the Command Syntax
Command Shortcuts
Components are typically named using the create command. When you enter a command to configure
a named component, you do not need to use the keyword of the component. For example, to create a
VLAN, enter a VLAN name:
create vlan engineering
After you have created the name for the VLAN, you can then eliminate the keyword vlan from all
other commands that require the name to be entered. For example, instead of entering the modular
switch command:
configure vlan engineering delete port 1:3,4:6
you could enter the following shortcut:
configure engineering delete port 1:3,4:6
Although it is helpful to have unique names for system components, this is not a requirement. If
ExtremeXOS encounters any ambiguity in the components within your command, it generates a
message requesting that you clarify the object you specified.
NOTE
If you use the same name across categories (for example, STPD and VLAN names), Extreme Networks recommends
that you specify the identifying keyword as well as the actual name. If you do not use the keyword, the system may
return an error message.
Names
All named components within a category of the switch configuration, such as VLAN, must have a
unique name. Names can be re-used across categories, however. Names must begin with an
alphabetical character and cannot contain any spaces. The maximum length for a name is 32 characters.
Names may contain alphanumeric characters and underscores (_) and cannot be keywords, such as
vlan, stp, and so on.
NOTE
If you use the same name across categories (for example, STPD and VLAN names), Extreme Networks recommends

Symbols
You may see a variety of symbols shown as part of the command syntax. These symbols explain how to
enter the command, and you do not type them as part of the command itself. Table 4 summarizes
command syntax symbols.
Table 4: Command syntax symbols

Symbol Description
angle brackets < > Enclose a variable or value. You must specify the variable or value. For example, in
the syntax
configure vlan <vlan> ipaddress <ipaddress>
you must supply a VLAN name for <vlan name> and an address for <ipaddress>
when entering the command. Do not type the angle brackets.
square brackets [ ] Enclose a required value or list of required arguments. One or more values or
arguments can be specified. For example, in the syntax
disable port [<port_list> | all]
you must specify either specific ports or all for all ports when entering the command.
Do not type the square brackets.
vertical bar | Separates mutually exclusive items in a list, one of which must be entered. For
example, in the syntax
configure snmp add community [readonly | readwrite]
<alphanumeric_string>
you must specify either the read or write community string in the command. Do not
type the vertical bar.
braces { } Enclose an optional value or a list of optional arguments. One or more values or
arguments can be specified. For example, in the syntax
reboot {time <month> <day> <year> <hour> <min> <sec>} {cancel}
{msm <slot_id>} {slot <slot-number> | node-address <node-
address> | stack-topology {as-standby} }
You can specify either a particular date and time combination, or the keyword
cancel to cancel a previously scheduled reboot. (In this command, if you do not
specify an argument, the command will prompt, asking if you want to reboot the
switch now.) Do not type the braces.
Limits
The command line can process up to 4500 characters, including spaces. If you attempt to enter more
than 4500 characters, the switch emits an audible “beep” and will not accept any further input. The first
4500 characters are processed, however.
Port Numbering
ExtremeXOS runs on both stand-alone and modular switches, and the port numbering scheme is
slightly different on each. This section describes the following topics:
● Stand-alone Switch Numerical Ranges on page 49
● Modular Switch and SummitStack Numerical Ranges on page 49
● Stacking Port Numerical Ranges on page 49

Port Numbering
NOTE
The keyword all acts on all possible ports; it continues on all ports even if one port in the sequence fails.
Stand-alone Switch Numerical Ranges

On Summit family switches, the port number is simply noted by the physical port number, as shown
below:
5
Separate the port numbers by a dash to enter a range of contiguous numbers, and separate the numbers
by a comma to enter a range of noncontiguous numbers:
● x-y—Specifies a contiguous series of ports on a stand-alone switch.
● x,y—Specifies a noncontiguous series of ports on a stand-alone switch.
● x-y,a,d—Specifies a contiguous series of ports and a noncontiguous series of ports on a stand-alone
switch.
Modular Switch and SummitStack Numerical Ranges

On a modular switch or SummitStack, such as the BlackDiamond 10808 switch, the port number is a
combination of the slot number and the port number. The nomenclature for the port number is as
follows:
slot:port
For example, if an I/O module that has a total of four ports is installed in slot 2 of the chassis, the
following ports are valid:
● 2:1
● 2:2
● 2:3
● 2:4
You can also use wildcard combinations (*) to specify multiple modular slot and port combinations. The
following wildcard combinations are allowed:
● slot:*—Specifies all ports on a particular I/O module.
● slot:x-slot:y—Specifies a contiguous series of ports on a particular I/O module.
● slot:x-y—Specifies a contiguous series of ports on a particular I/O module.
● slota:x-slotb:y—Specifies a contiguous series of ports that begin on one I/O module or
SummitStack node and end on another node.
Stacking Port Numerical Ranges

On a SummitStack, stacking port number is a combination of the slot number and the stacking port
number shown near the connector on the back of the Summit family switch:
slot:port

These numbers are context-specific. For example, while the front-panel port 2:1 on a Summit X450a-24t
is a 10/100/1000 Ethernet port, the stacking port 2:1 is a 10Gb port on the rear panel of the X450a-24t
that has been marked as "Stacking Port 1". When no context is given, port 2:1 refers to a front-panel port
on the Summit family switch.
The use of wildcards and ranges for stacking ports is the same as described in "Modular Switch and
SummitStack Numerical Ranges".
Line-Editing Keys
Table 5 describes the line-editing keys available using the CLI.
Table 5: Line-editing keys
Key(s) Description
Left arrow or [Ctrl] + B Moves the cursor one character to the left.
Right arrow or [Ctrl] + F Moves the cursor one character to the right.
[Ctrl] + H or Backspace Deletes character to left of cursor and shifts remainder of line to left.
Delete or [Ctrl] + D Deletes character under cursor and shifts remainder of line to left.
[Ctrl] + K Deletes characters from under cursor to end of line.
Insert Toggles on and off. When toggled on, inserts text and shifts previous text to right.
[Ctrl] + A Moves cursor to first character in line.
[Ctrl] + E Moves cursor to last character in line.
[Ctrl] + L Clears screen and movers cursor to beginning of line.
[Ctrl] + P or Up Arrow Displays previous command in command history buffer and places cursor at end of
command.
[Ctrl] + N or Down Arrow Displays next command in command history buffer and places cursor at end of
command.
[Ctrl] + U Clears all characters typed from cursor to beginning of line.
[Ctrl] + W Deletes previous word.
[Ctrl] + C Interrupts the current CLI command execution.
Command History
ExtremeXOS “remembers” the commands you enter. You can display a list of these commands by using
the following command:
history
Common Commands
Table 6 describes some of the common commands used to manage the switch. Commands specific to a
particular feature may also be described in other chapters of this guide. For a detailed description of the
commands and their options, see the ExtremeXOS Command Reference Guide.

Common Commands
Table 6: Common commands

Command Description
clear session [<sessId> | all] Terminates a Telnet or SSH2 session from the switch.
configure account [all | <name>] Configures a user account password.
Passwords can have a minimum of 0 character and can
have a maximum of 32 characters. Passwords are case-
sensitive; user names are not case sensitive.
configure banner Configures the banner string. You can enter up to 24
rows of 79-column text that is displayed before the login
prompt of each session. Press [Return] at the beginning
of a line to terminate the command and apply the
banner. To clear the banner, press [Return] at the
beginning of the first line.
configure ports <port_list> auto off Manually configures the port speed and duplex setting of
speed [10 | 100 | 1000 | 10000] one or more ports on a switch.
duplex [half | full]
configure slot <slot> module Configures a slot for a particular I/O module card.
<module_type>
NOTE: This command is available only on modular
switches.
configure ssh2 key {pregenerated} Generates the SSH2 host key.
You must install the SSH software module in addition to
the base image to run SSH.
configure sys-recovery-level [all | Configures a recovery option for instances where an
none] exception occurs in ExtremeXOS.
configure time <month> <day> <year> Configures the system date and time. The format is as
<hour> <min> <sec> follows:
mm dd yyyy hh mm ss
The time uses a 24-hour clock format. You cannot set the
year earlier than 2003 or past 2036.
configure timezone {name <tz_name>} Configures the time zone information to the configured
<GMT_offset> {autodst {name offset from GMT time. The format of GMT_offset is +/-
<dst_timezone_ID>} {<dst_offset>} minutes from GMT time. The autodst and noautodst
{begins [every <floatingday> | on options enable and disable automatic Daylight Saving
<absoluteday>] {at <time_of_day>} Time change based on the North American standard.
{ends [every <floatingday> | on
Additional options are described in the ExtremeXOS
<absoluteday>] {at <time_of_day>}}} |
Command Reference Guide.
noautodst}
configure vlan <vlan_name> ipaddress Configures an IP address and subnet mask for a VLAN.
[<ipaddress> {<ipNetmask>} | ipv6-
link-local | {eui64}
<ipv6_address_mask>]
create account [admin | user] Creates a user account. This command is available to
<account-name> {encrypted <password>} admin-level users and to users with RADIUS command
authorization. The username is between 1 and 32
characters, the password is between 0 and 32 characters.
create vlan <vlan_name> {vr <vr-name>} Creates a VLAN.
NOTE: The BlackDiamond 8800 series switch (formerly
known as Aspen), SummitStack, and the Summit family
of switches do not use the vr optional parameter.
delete account <name> Deletes a user account.
delete vlan <vlan_name> Deletes a VLAN.
disable bootp vlan [<vlan> | all] Disables BOOTP for one or more VLANs.

Table 6: Common commands (Continued)

Command Description
disable cli-config-logging Disables logging of CLI commands to the Syslog.
disable clipaging Disables pausing of the screen display when a show
command output reaches the end of the page.
disable idletimeout Disables the timer that disconnects all sessions. After
being disabled, console sessions remain open until the
switch is rebooted or until you log off. Telnet sessions
remain open until you close the Telnet client. SSH2
sessions time out after 61 minutes of activity.
disable port [<port_list> | all] Disables one or more ports on the switch.
disable ssh2 Disables SSH2 Telnet access to the switch.
disable telnet Disables Telnet access to the switch.
enable bootp vlan [<vlan> | all] Enables BOOTP for one or more VLANs.
enable cli-config-logging Enables the logging of CLI configuration commands to
the Syslog for auditing purposes. The default setting is
enabled.
enable clipaging Enables pausing of the screen display when show
command output reaches the end of the page. The
default setting is enabled.
enable idletimeout Enables a timer that disconnects all sessions (Telnet,
SSH2, and console) after 20 minutes of inactivity. The
default setting is enabled.
enable license {software} <key> Enables a particular software feature license. Specify
<license_key> as an integer.
The command unconfigure switch {all} does not
clear licensing information. This license cannot be
disabled once it is enabled on the switch.
enable ssh2 {access-profile Enables SSH2 sessions. By default, SSH2 is disabled.
[<access_profile> | none]} {port When enabled, SSH uses TCP port number 22.
<tcp_port_number>} {vr [<vr_name> |
all | default]}
enable telnet Enables Telnet access to the switch. By default, Telnet
uses TCP port number 23.
history Displays the commands entered on the switch.
show banner Displays the user-configured banner.
unconfigure switch {all} Resets all switch parameters (with the exception of
defined user accounts, and date and time information) to
the factory defaults.
If you specify the keyword all, the switch erases the
currently selected configuration image in flash memory
and reboots. As a result, all parameters are reset to
default settings.

Accessing the Switch the First Time
Accessing the Switch the First Time

With software version 11.2, ExtremeXOS adds to the security features offered on your switch.
When you take your switch from the box and set it up for the first time, you must connect to the
console to access the switch. You are prompted with an interactive script that specifically asks if you
want to disable Telnet and SNMP, so these will not be available on your switch at next reboot. This is
called the safe defaults mode.
After you connect to the console and log into the switch, the screen displays several interactive
questions that lead you through configuring the management access that you want. You disable SNMP,
or Telnet access by using the interactive script (refer to “Safe Defaults Setup Method” on page 53).
All ports are enabled in the factory default setting; you can choose to have all unconfigured ports
disabled on reboot using the interactive questions.
In addition, you can return to the safe defaults mode by issuing the following commands:
● unconfigure switch all
● configure safe-default-script
Safe Defaults Setup Method

After you connect to the console port of the switch, or after you issue the unconfigure switch all or
configure safe-default-script CLI command, the system returns the following interactive script:
This switch currently has all management methods enabled for convenience reasons.
Please answer these questions about the security settings you would like to use.
Telnet is enabled by default. Telnet is unencrypted and has been the target of
security exploits in the past.
Would you like to disable Telnet? [y/N]:
SNMP access is enabled by default. SNMP uses no encryption, SNMPv3 can be

configured to eliminate this problem.
Would you like to disable SNMP? [y/N]:
All ports are enabled by default. In some secure applications, it maybe more
desirable for the ports to be turned off.
Would you like unconfigured ports to be turned off by default? [y/N]:
Changing the default failsafe account username and password is highly

recommended. If you choose to do so, please remember the username and
password as this information cannot be recovered by Extreme Networks.
Would you like to change the failsafe account username and password
now? [y/N]:
Would you like to permit failsafe account access via the management port?
[y/N]:

Since you have chosen less secure management methods, please remember to
increase the security of your network by taking the following actions:
* change your admin password

* change your failsafe account username and password
* change your SNMP public and private strings
* consider using SNMPv3 to secure network management traffic
All the changes you make using this interactive script can be saved through switch reboots, if you save
the setting. If you want to change the management access:
● Use the configure safe-default-script command—maintains your configuration and reruns the
script.
● Use the unconfigure switch all command—resets your switch to the default factory setting and
reruns this script.
You see this interactive script only under the following conditions:
● Initial login (when you use the switch the first time)
● After unconfigure switch all CLI command
● After configure safe-default-script command
Configuring Management Access

This section discusses the following material:
● Account Access Levels on page 54
● Configuring the Banner on page 55
● Startup Screen and Prompt Text on page 55
● Default Accounts on page 57
● Creating a Management Account on page 57
● Failsafe Account on page 58
Account Access Levels

ExtremeXOS supports the following two levels of management:
● User
● Administrator
In addition to the management levels, you can optionally use an external RADIUS server to provide CLI
command authorization checking for each command. For more information on RADIUS, see Chapter
16, “Security.”
User Account
A user-level account has viewing access to all manageable parameters, with the exception of:
● User account database
● SNMP community strings

A person with a user-level account can use the ping command to test device reachability and change
the password assigned to the account name. If you have logged on with user capabilities, the command
line prompt ends with a (>) sign. For example:
BD-1.2 >
Administrator Account
A person with an administrator-level account can view and change all switch parameters. With this
level, you can also add and delete users, as well as change the password associated with any account
name (to erase the password, use the unconfigure switch all command).
The administrator can disconnect a management session that has been established by way of a Telnet
connection. If this happens, the user logged on by way of the Telnet connection is notified that the
session has been terminated.
If you have logged on with administrator capabilities, the command line prompt ends with a (#) sign.
For example:
BD-1.18 #
Configuring the Banner

You can configure a banner that displays as soon as you power-up the switch, before the login prompt.
To add a banner to your switch, use the following command:
configure banner {acknowledge)
Using the acknowledge parameter prompts the user with the following message after the banner
appears and before the login prompt:
Hit any key to accept these provisions.
To disable the acknowledgement feature, which forces the user to press a key before the login screen
displays, use the configure banner command omitting the acknowledge parameter.
Startup Screen and Prompt Text

Once you log into the switch, the system displays the startup screen, as follows:
login: admin
password: blue7
ExtremeXOS
Copyright (C) 2000-2006 Extreme Networks. All rights reserved.
Protected by US Patent Nos: 6,678,248; 6,104,700; 6,766,482; 6,618,388; 6,034,957;
6,859,438; 6,912,592; 6,954,436; 6,977,891; 6,980,550; 6,981,174; 7,003,705; 7,01
2,082.
==============================================================================
Press the <tab> or '?' key at any time for completions.

Remember to save your configuration changes.
* <switchname>.1 #

You must have an administrator-level account to change the text of the prompt. The prompt text is
taken from the SNMP sysname setting.
The number that follows the period after the switch name indicates the sequential line of the specific
command or line for this CLI session.
If an asterisk (*) appears in front of the command line prompt, it indicates that you have outstanding
configuration changes that have not been saved. For example:
* BD-1.19 #
If you have logged on with administrator capabilities, the command line prompt ends with a (#) sign.
For example:
BD-1.18 #
If you have logged on with user capabilities, the command line prompt ends with a (>) sign. For
example:
BD-1.2 >
Using the system recovery commands (refer to Chapter 5, “Status Monitoring and Statistics,” for
information on system recovery), you can configure either one or more specified slots on a modular
switch or the entire stand-alone switch to shut down in case of an error. If you have configured this
feature and a hardware error is detected, the system displays an explanatory message on the startup
screen. The message is slightly different, depending on whether you are working on a modular switch
or a stand-alone switch.
The following sample shows the startup screen if any of the slots in a modular switch are shut down as
a result of the system recovery configuration:
login: admin
password:
ExtremeXOS
6,859,438; 6,912,592; 6,954,436; 6,977,891; 6,980,550; 6,981,174; 7,003,705; 7,01
2,082.
==============================================================================

The I/O modules in the following slots are shut down: 1,3
Use the "clear sys-recovery-level" command to restore I/O modules
! BD-8810.1 #
When an exclamation point (!) appears in front of the command line prompt, it indicates that one or
more slots or the entire stand-alone switch are shut down as a result of your system recovery
configuration and a switch error. (Refer to Chapter 5, “Status Monitoring and Statistics,” for complete
information on system recovery and system health check features.)

The following sample shows the startup screen if a stand-alone switch is shut down as a result of the
system recovery configuration:
login: admin
password:
ExtremeXOS
6,859,438; 6,912,592; 6,954,436; 6,977,891; 6,980,550; 6,981,174; 7,003,705; 7,01
2,082.
==============================================================================

All switch ports have been shut down.

Use the "clear sys-recovery-level" command to restore all ports.
! SummitX450-24x.1 #
Default Accounts
By default, the switch is configured with two accounts, as shown in Table 7.
Table 7: Default accounts
Account Name Access Level
admin This user can access and change all manageable parameters. However, the user may not
delete all admin accounts.
user This user can view (but not change) all manageable parameters, with the following
exceptions:
• This user cannot view the user account database.
• This user cannot view the SNMP community strings.
To change the password on the default account, see “Applying a Password to the Default Account” on
page 59.
Creating a Management Account

The switch can have a total of 16 management accounts. You can use the default names (admin and
user), or you can create new names and passwords for the accounts. Passwords can have a minimum of
0 characters and a maximum of 32 characters.
To create a new account:

1 Log in to the switch as admin.
2 At the password prompt, press [Return], or enter the password that you have configured for the
admin account.
3 Add a new user by using the following command:
create account [admin | user] <account-name> {encrypted <password>}
If you do not want a password associated with the specified account, press [Enter] twice.

Viewing Accounts
To view the accounts that have been created, you must have administrator privileges. To see the
accounts, use the following command:
show accounts
Deleting an Account
To delete a account, you must have administrator privileges. To delete an account, use the following
command:
delete account <name>
Failsafe Account
The failsafe account is the account of last resort to access your switch. This account is never displayed
by the show accounts command, but it is always present on the switch. The failsafe account has
"admin" access level. To configure the account name and password for the failsafe account, use the
following command:
configure failsafe-account
You will be prompted for the failsafe account name and prompted twice to specify the password for the
account. For example:
BD-10808.1 # configure failsafe-account

enter failsafe user name: blue5green
enter failsafe password:
enter password again:
BD-10808.2
The failsafe account is immediately saved to NVRAM.
You need not provide the existing failsafe account information to change it.
NOTE
The information that you use to configure the failsafe account cannot be recovered by Extreme Networks. Technical
support cannot retrieve passwords or account names for this account. Protect this information carefully.
To access your switch using the failsafe account, you must connect to the serial port of the switch. You
cannot access the failsafe account through any other port.
At the switch login prompt, carefully enter the failsafe account name. If you enter an erroneous account
name, you cannot re-enter the correct name.
After you have entered the failsafe account name, you are prompted to enter the password. You will
have three tries to enter the password correctly.

Managing Passwords
Managing Passwords
When you first access the switch, you have a default account. You configure a password for your
default account. As you create other accounts (see “Creating a Management Account” on page 57), you
configure passwords for those accounts.
Beginning with ExtremeXOS version 11.2, the software allows you to apply additional security to the
passwords. You can enforce a specific format and minimum length for the password. Additionally, you
can age out the password, prevent a user from employing a previously used password, and lock users
out of the account after three consecutive failed login attempts.
Beginning with ExtremeXOS version 11.5, you can change the password to an encrypted password after
you create an account.

● Applying a Password to the Default Account on page 59
● Applying Security to Passwords on page 60
● Displaying Passwords on page 61
Applying a Password to the Default Account

Default accounts do not have passwords assigned to them. Passwords can have a minimum of 0
character and can have a maximum of 32 characters. (If you specify the format of passwords using the
configure account password-policy char-validation command, the minimum is 8 characters.)
NOTE
Passwords are case-sensitive; user names are not case-sensitive.
To add a password to the default admin account:

1 Log in to the switch using the name admin.
2 At the password prompt, press [Enter].
3 Add a default admin password of green by entering the following command:
configure account admin green
To add a password to the default user account:

1 Log in to the switch using the name user.
2 At the password prompt, press [Enter], or enter the password that you have configured for the
admin account.
3 Add a default user password by blue entering the following command:
configure account user blue
NOTE
If you forget your password while logged out of the CLI, you can use the bootloader to reinstall a default switch
configuration, which allows access to the switch without a password. Note that this process reconfigures all switch
settings back to the initial default configuration.

Applying Security to Passwords

You can increase the security of your system by enforcing password restrictions, which will make it
more difficult for unauthorized users to access your system.
You can specify that each password must include at least two characters of each of the following four
character types:
● Upper-case A-Z
● Lower-case a-z
● 0-9
● !, @, #, $, %, ^, *, (, )
To set this format for the password, use the following command:
configure account [all | <name>] password-policy char-validation [none | all-char-

groups]
You can enforce a minimum length for the password and set a maximum time limit, after which the
password will not be accepted.
To set a minimum length for the password, use the following command:
configure account [all | <name>] password-policy min-length [<num_characters> |

none]
To age out the password after a specified time, use the following command:
configure account [all | <name>] password-policy max-age [<num_days> | none]
You can block users from employing previously used passwords by issuing the command:
configure account [all | <name>] password-policy history [<num_passwords> | none]
By default, the system terminates a session after the user has three consecutive failed login attempts.
The user may then launch another session (which again would terminate after three consecutive failed
login attempts). To increase security, you can lock users out of the system entirely after three failed
consecutive login attempts. To use this feature, use the following command:
configure account [all | <name>] password-policy lockout-on-login-failures [on |

off]
NOTE
If you are not working on SSH, you can configure the number of failed logins that trigger lockout, using the
configure cli max-failed-logins <num-of-logins> command. (This command also sets the number of
failed logins that terminate the particular session.)
After the user’s account is locked out (using the configure account password-policy lockout-on-
login-failures command), it must be specifically re-enabled by an administrator. To re-enable a
locked-out account, use the following command:
clear account [all | <name>] lockout
Selecting the all option affects the setting of all existing and future new accounts.

Access to Both MSM Console Ports—Modular Switches Only
NOTE
The default admin account and failsafe accounts are never locked out, no matter how many consecutive failed login
attempts.
Displaying Passwords
To display the accounts and any applied password security, use the following command:
show accounts password-policy
You can also display which accounts may be locked out by issuing the following command:
show accounts
Access to Both MSM Console Ports—Modular Switches

Only
With ExtremeXOS version 11.2, you can access either the primary or the backup MSM regardless of
which console port you are connected to.
Use the following command:
telnet msm [a | b]
Access to an Active Node in a SummitStack

With ExtremeXOS version 12.0, you can access any active node in a SummitStack from any other active
node in a SummitStack. Use the following command:
telnet slot <slot-number>
Domain Name Service Client Services

The Domain Name Service (DNS) client in ExtremeXOS augments the following commands to allow
them to accept either IP addresses or host names:
● telnet
● download bootrom
● download image
● ping
● traceroute
● configure radius server

● configure tacacs server

● create cfm domain dns ma-level
In addition, the nslookup utility can be used to return the IP address of a hostname. (This command is
available only on the Default VR on the BlackDiamond 10808 and 12800 series switch.)
You can specify up to eight DNS servers for use by the DNS client using the following command:
configure dns-client add
You can specify a default domain for use when a host name is used without a domain. Use the
following command:
configure dns-client default-domain
For example, if you specify the domain xyz-inc.com as the default domain, then a command such as
ping accounting1 will be taken as if it had been entered ping accounting1.xyz-inc.com.
Checking Basic Connectivity
NOTE
The BlackDiamond 8800 series switch, SummitStack, and the Summit family of switches do not support user-
created virtual routers. You must use the vr option for these commands when running them on VR-Mgmt.
The switch offers the following commands for checking basic connectivity:
● ping
● traceroute
Ping
The ping command enables you to send Internet Control Message Protocol (ICMP) echo messages to a
remote IP device. The ping command is available for both the user and administrator privilege level.
Beginning with ExtremeXOS version 11.2, you can ping an IPv6 address.
The ping command syntax is:
ping {count <count> {start-size <start-size>} | continuous {start-size <start-size>} |

{start-size <start-size> {end-size <end-size>}}} {udp} {dont-fragment} {ttl <ttl>}
{tos <tos>} {interval <interval>} {vr <vrid>} {ipv4 <host> | ipv6 <host>} {from} {with
record-route}

Checking Basic Connectivity
Options for the ping command are described in Table 8.
Table 8: Ping command parameters

Parameter Description
count Specifies the number of ping requests to send.
start-size Specifies the size, in bytes, of the packet to be sent, or the starting size if
incremental packets are to be sent.
continuous Specifies that UDP or ICMP echo messages to be sent continuously. This option can
be interrupted by pressing [Ctrl] + C.
end-size Specifies an end size for packets to be sent.
udp Specifies that the ping request should use UDP instead of ICMP.
dont-fragment Sets the IP to not fragment the bit.
ttl Sets the TTL value.
tos Sets the TOS value.
interval Sets the time interval between sending out ping requests.
vr Specifies the virtual router name to use for sending out the echo message. If not
specified, VR-Default is used.
NOTE: The BlackDiamond 8800 series switch, SummitStack, and the Summit
family of switches do not support user-created VRs.
ipv4 Specifies IPv4 transport.
ipv6 Specifies IPv6 transport.
NOTE: If you are contacting an IPv6 link local address, you must specify the VLAN
you sending the message from: ping <ipv6> <link-local address>
%<vlan_name> <host>.
host Specifies a host name or IP address (either v4 or v6).
from Uses the specified source address. If not specified, the address of the transmitting
interface is used.
with record-route Sets the traceroute information.
If a ping request fails, the switch stops sending the request after three attempts. Press [Ctrl] + C to
interrupt a ping request earlier. The statistics are tabulated after the ping is interrupted or stops.
You use the ipv6 variable to ping an IPv6 host by generating an ICMPv6 echo request message and
sending the message to the specified address. If you are contacting an IPv6 link local address, you must
specify the VLAN you are sending the message from, as shown in the following example (you must
include the % sign): ping <ipv6> <link-local address> %<vlan_name> <host>.
Traceroute
The traceroute command enables you to trace the routed path between the switch and a destination
endstation. The traceroute command syntax is:
traceroute {vr <vrid>} {ipv4 <host>} {ipv6 <host>} {ttl <number>} {from <from>} {[port
<port>] | icmp}
Where:
● vr is the name of the virtual router (the BlackDiamond 8800 series switch and the Summit X450
family of switches do not support user-created VRs).
● ipv4/ipv6 is the transport.

● from uses the specified source address in the ICMP packet. If not specified, the address of the
transmitting interface is used.
● host is the host of the destination endstation. To use the hostname, you must first configure DNS.
● ttl configures the switch to trace the hops until the time-to-live has been exceeded for the switch.
● port uses the specified UDP port number.
● icmp uses ICMP echo messages to trace the routed path.
Beginning with ExtremeXOS, you can trace the route between the switch and an IPv6 address.
However, you must specify the target’s IPv6 address to use this command.
Displaying Switch Information

To display basic information about the switch, use the following command:
show switch

3 Managing the Switch

● Overview on page 65
● Understanding the ExtremeXOS Shell on page 66
● Using the Console Interface on page 66
● Using the 10/100 Ethernet Management Port on page 67
● Using EPICenter to Manage the Network on page 67
● Authenticating Users on page 68
● Using Telnet on page 68
● Using Secure Shell 2 on page 75
● Using the Trivial File Transfer Protocol on page 76
● Understanding System Redundancy—Modular Switches and SummitStack Only on page 77
● Understanding Hitless Failover Support—Modular Switches and SummitStack Only on page 82
● Understanding Power Supply Management on page 87
● Using the Simple Network Management Protocol on page 91
● Using the Simple Network Time Protocol on page 102
Overview
Using ExtremeXOS, you can manage the switch using the following methods:
● Access the command line interface (CLI) by connecting a terminal (or workstation with terminal-
emulation software) to the console port.
● Access the switch remotely using TCP/IP through one of the switch ports or through the dedicated
10/100 unshielded twisted pair (UTP) Ethernet management port. Remote access includes:
■ Telnet using the CLI interface.
■ Secure Shell (SSH2) using the CLI interface.
■ Simple Network Management Protocol (SNMP) access using EPICenter or another SNMP
manager.
● Download software updates and upgrades. For more information, see Appendix A, “Software
Upgrade and Boot Options.”
NOTE
Beginning with ExtremeXOS 11.2, the switch accepts IPv6 Telnet and SSH2 connections.

Managing the Switch
The switch supports up to the following number of concurrent user sessions:

● One console session
■ Two console sessions are available if two management modules are installed.
● Eight shell sessions
● Eight Telnet sessions
● Eight Trivial File Transfer Protocol (TFTP) sessions
● Eight SSH2 sessions
Understanding the ExtremeXOS Shell

When you log in to ExtremeXOS from a terminal, you enter the shell with a shell prompt displayed. At
the prompt, you input the commands to be executed on the switch. After the switch processes and
executes a command, the results are relayed to and displayed on your terminal.
The shell supports ANSI, VT100, and XTERM terminal emulation and adjusts to the correct terminal
type and window size. In addition, the shell supports UNIX-style page view for page-by-page
command output capability.
By default, up to eight active shell sessions can access the switch concurrently; however, you can
change the number of simultaneous, active shell sessions supported by the switch. You can configure
up to 16 active shell sessions. Configurable shell sessions include both Telnet and SSH connections (not
console CLI connections). If only eight active shell sessions can access the switch, a combination of eight
Telnet and SSH connections can access the switch even though Telnet and SSH each support eight
connections. For example, if you have six Telnet sessions and two SSH sessions, no one else can access
the switch until a connection is terminated or you access the switch via the console.
If you configure a new limit, only new incoming shell sessions are affected. If you decrease the limit
and the current number of sessions already exceeds the new maximum, the switch refuses only new
incoming connections until the number of shell session drops below the new limit. Already connected
shell sessions are not disconnected as a result of decreasing the limit.
To configure the number of shell sessions accepted by the switch, use the following command:
configure cli max-sessions
For more information about the line-editing keys that you can use with the XOS shell, see “Line-Editing
Keys” on page 50.
Using the Console Interface

The CLI built into the switch is accessible by way of the 9-pin, RS-232 port labeled console. On a
modular switch, the console port is located on the front of the Management Switch Fabric Module
(MSM). On a stand-alone switch, the console port is located on the front panel.
NOTE
For more information on the console port pinouts, see the hardware installation guide that shipped with your switch.

Using the 10/100 Ethernet Management Port
After the connection has been established, you see the switch prompt and you can log in.
Using the 10/100 Ethernet Management Port

The MSM provides a dedicated 10/100 Mbps Ethernet management port. This port provides dedicated
remote access to the switch using TCP/IP. It supports the following management methods:
● Telnet/SSH2 using the CLI interface
● SNMP access using EPICenter or another SNMP manager
The switch uses the Ethernet management port only for host operation, not for switching or routing.
The TCP/IP configuration for the management port is done using the same syntax as used for virtual
LAN (VLAN) configuration. The VLAN mgmt comes preconfigured with only the management port as
a member. The management port is a member of the virtual router VR-Mgmt.
When you configure the IP address for the VLAN mgmt, this address gets assigned to the primary
MSM. You can connect to the management port on the primary MSM for any switch configuration. The
management port on the backup MSM is available only when failover occurs. At that time, the primary
MSM relinquishes its role, the backup MSM takes over, and the VLAN mgmt on the new primary MSM
acquires the IP address of the previous primary MSM.
To configure the IP address and subnet mask for the VLAN mgmt, use the following command:
configure vlan mgmt ipaddress <ip_address>/<subnet_mask>
To configure the default gateway (you must specify VR-Mgmt for the management port and VLAN
mgmt), use the following command:
configure iproute add default <gateway> {vr <vrname>} {<metric>} {multicast-only |

unicast-only}
The following example configuration sets the management port IP address to 192.168.1.50, mask length
of 25, and configures the gateway to use 192.168.1.1:
configure vlan mgmt ipaddress 192.168.1.50/25

configure iproute add default 192.168.1.1 vr vr-mgmt
Using EPICenter to Manage the Network

EPICenter is a powerful yet easy-to-use application suite that facilitates the management of a network
of Extreme Networks switches, as well as selected third-party switches. EPICenter offers a
comprehensive set of network management tools that are easy to use from a client workstation running
EPICenter client software, or from a workstation configured with a web browser and the Java plug-in.
For more information about the EPICenter management software available from Extreme Networks, go
to http://www.extremenetworks.com
To review the EPICenter documentation, go to http://www.extremenetworks.com/services/swuserguides.asp

Managing the Switch
Authenticating Users
ExtremeXOS provides three methods to authenticate users who log in to the switch:
● RADIUS client
● TACACS+
● Local database of accounts and passwords
NOTE
You cannot configure RADIUS and TACACS+ at the same time.
RADIUS Client
Remote Authentication Dial In User Service (RADIUS, RFC 2138) is a mechanism for authenticating and
centrally administrating access to network nodes. The ExtremeXOS RADIUS client implementation
allows authentication for Telnet or console access to the switch.
For detailed information about RADIUS and configuring a RADIUS client, see Chapter 16, “Security.”
TACACS+
Terminal Access Controller Access Control System Plus (TACACS+) is a mechanism for providing
authentication, authorization, and accounting on a central server, similar in function to the RADIUS
client. The ExtremeXOS version of TACACS+ is used to authenticate prospective users who are
attempting to administer the switch. TACACS+ is used to communicate between the switch and an
authentication database.
For detailed information about TACACS+ and configuring TACACS+, see Chapter 16, “Security.”
Management Accounts
ExtremeXOS supports two levels of management accounts (local database of accounts and passwords):
User and Administrator. A user level account can view but not change all manageable parameters, with
the exception of the user account database and SNMP community strings. An administrator level
account can view and change all manageable parameters.
For detailed information about configuring management accounts, see Chapter 3, “Accessing the
Switch.”
Using Telnet
ExtremeXOS supports the Telnet Protocol based on RFC 854. Telnet allows interactive remote access to
a device and is based on a client/server model. ExtremeXOS uses Telnet to connect to other devices
from the switch (client) and to allow incoming connections for switch management using the CLI
(server).

Using Telnet
This section describes the following Telnet topics:

● About the Telnet Client on page 69
● About the Telnet Server on page 69
● Connecting to Another Host Using Telnet on page 70
● Configuring Switch IP Parameters on page 70
● Configuring Telnet Access to the Switch on page 72
● Disconnecting a Telnet Session on page 75
About the Telnet Client

Before you can start an outgoing Telnet session on the switch, you must set up the IP parameters
described in “Configuring Switch IP Parameters” on page 70. Telnet is enabled and uses VR-Mgmt by
default.
NOTE
Maximize the Telnet screen so that automatically updating screens display correctly.
If you use Telnet to establish a connection to the switch, you must specify the IP address or host name
of the device that you want to connect to. Check the user manual supplied with the Telnet facility if you
are unsure of how to do this.
After the connection is established, you see the switch prompt and you can log in.
The same is true if you use the switch to connect to another host. From the CLI, you must specify the IP
address or host name of the device that you want to connect to. If the host is accessible and you are
allowed access, you may log in.
For more information about using the Telnet client on the switch, see “Connecting to Another Host
Using Telnet” on page 70.
About the Telnet Server

Any workstation with a Telnet facility should be able to communicate with the switch over a TCP/IP
network using VT100 terminal emulation.
Up to eight active Telnet sessions can access the switch concurrently. If you enable the idle timer using
the enable idletimeout command, the Telnet connection times out after 20 minutes of inactivity by
default. If a connection to a Telnet session is lost inadvertently, the switch terminates the session within
two hours.
Beginning with ExtremeXOS 11.2, the switch accepts IPv6 connections.
For information about the Telnet server on the switch, see the following sections:
● Configuring Telnet Access to the Switch on page 72
● Disconnecting a Telnet Session on page 75

Managing the Switch
Connecting to Another Host Using Telnet

You can Telnet from the current CLI session to another host using the following command:
telnet {vr <vr_name>} [<host_name> | <remote_ip>] {<port>}
NOTE
The BlackDiamond 8800 series switch (formerly known as Aspen), SummitStack, and the Summit family of switches
do not support user-created VRs.
If the TCP port number is not specified, the Telnet session defaults to port 23. If the virtual router name
is not specified, the Telnet session defaults to VR-Mgmt. Only VT100 emulation is supported.
Beginning with ExtremeXOS 11.2, you can use Telnet to access either the primary or the backup MSM
regardless of which console port you are connected to. For more information see Chapter 3, “Accessing
the Switch.”
Configuring Switch IP Parameters

To manage the switch by way of a Telnet connection or by using an SNMP Network Manager, you
must first configure the switch IP parameters.
Using a BOOTP or DHCP Server

If you are using IP and you have a Bootstrap Protocol (BOOTP) server set up correctly on your
network, you must provide the following information to the BOOTP server:
● Switch Media Access Control (MAC) address, found on the rear label of the switch
● IP address
● Subnet address mask (optional)
The switch contains a BOOTP and Dynamic Host Configuration Protocol (DHCP) client, so if you have
a BOOTP or DHCP server in your IP network, you can have it assign IP addresses to the switch. This is
more likely to be desirable on the switch's VLAN mgmt than it is on any other VLANs.
You can enable the BOOTP or DHCP client per VLAN by using the following commands:
enable bootp vlan [<vlan> | all]

enable dhcp vlan [<vlan_name> | all]
You can disable the BOOTP or DHCP client per VLAN by using the following commands:
disable bootp vlan [<vlan> | all]

disable dhcp vlan [<vlan_name> | all]
To view the current state of the BOOTP or DHCP client, use the following command:
show dhcp-client state
The switch does not retain IP addresses assigned by BOOTP or DHCP through a power cycle, even if
the configuration has been saved. To retain the IP address through a power cycle, you must configure
the IP address of the VLAN using the CLI or Telnet.

Using Telnet
If you need the switch's MAC address to configure your BOOTP or DHCP server, you can find it on the
rear label of the switch. Note that all VLANs configured to use BOOTP or DHCP use the same MAC
address to get their IP address, so you cannot configure the BOOTP or DHCP server to assign multiple
specific IP addresses to a switch depending solely on the MAC address.
Manually Configuring the IP Settings

If you are using IP without a BOOTP server, you must enter the IP parameters for the switch in order
for the SNMP Network Manager or Telnet software to communicate with the device. To assign IP
parameters to the switch, you must perform the following tasks:
● Log in to the switch with administrator privileges using the console interface.
● Assign an IP address and subnet mask to a VLAN.
The switch comes configured with a default VLAN named default. To use Telnet or an SNMP
Network Manager, you must have at least one VLAN on the switch, and that VLAN must be
assigned an IP address and subnet mask. IP addresses are always assigned to each VLAN. The
switch can be assigned multiple IP addresses (one for each VLAN).
NOTE
For information on creating and configuring VLANs, see Chapter 6, “Virtual LANs.”
To manually configure the IP settings:

1 Connect a terminal or workstation running terminal emulation software to the console port, as
detailed in “Using the Console Interface” on page 66.
2 At your terminal, press [Return] one or more times until you see the login prompt.
3 At the login prompt, enter your user name and password. Note that they are both case-sensitive.
Ensure that you have entered a user name and password with administrator privileges.
■ If you are logging in for the first time, use the default user name admin to log in with
administrator privileges. For example:
login: admin
Administrator capabilities enable you to access all switch functions. The default user names have
no passwords assigned.
■ If you have been assigned a user name and password with administrator privileges, enter them at
the login prompt.
4 At the password prompt, enter the password and press [Return].
When you have successfully logged in to the switch, the command line prompt displays the name of
the switch.
5 Assign an IP address and subnetwork mask for the default VLAN by using the following command:
configure vlan <vlan_name> ipaddress [<ipaddress> {<ipNetmask>} | ipv6-link-local
| {eui64} <ipv6_address_mask>]
For example:
configure vlan default ipaddress 123.45.67.8 255.255.255.0
Your changes take effect immediately.

Managing the Switch
NOTE
As a general rule, when configuring any IP addresses for the switch, you can express a subnet mask by using
dotted decimal notation or by using classless inter domain routing notation (CIDR). CIDR uses a forward slash
plus the number of bits in the subnet mask. Using CIDR notation, the command identical to the previous
example is: configure vlan default ipaddress 123.45.67.8/24
6 Configure the default route for the switch using the following command:
unicast-only}
For example:
configure iproute add default 123.45.67.1
7 Save your configuration changes so that they will be in effect after the next switch reboot.
■ If you want to save your changes to the currently booted configuration, use the following
command:
save
■ ExtremeXOS allows you to select or create a configuration file name of your choice to save the
configuration to. If you want to save your changes to an existing or new configuration file, use
save configuration [<existing-config> | <new-config>]
8 When you are finished using the facility, log out of the switch by typing:
logout or quit
Configuring Telnet Access to the Switch

By default, Telnet services are enabled on the switch and all virtual routers listen for incoming Telnet
requests. Beginning with ExtremeXOS 11.2, the switch accepts IPv6 connections.
NOTE
created virtual routers.
ExtremeXOS 11.2 introduces the concept of safe defaults mode. Safe defaults mode runs an interactive
script that allows you to enable or disable SNMP, Telnet, and switch ports. When you set up your
switch for the first time, you must connect to the console port to access the switch. After logging in to
the switch, you enter safe defaults mode. Although SNMP, Telnet, and switch ports are enabled by
default, the script prompts you to confirm those settings.
If you choose to keep the default setting for Telnet—the default setting is enabled—the switch returns
the following interactive script:
Since you have chosen less secure management methods, please remember to increase
the security of your network by taking the following actions:
For more detailed information about safe defaults mode, see “Safe Defaults Setup Method” on page 53.

Using Telnet
To configure the virtual router from which you receive a Telnet request, use the following command:
configure telnet vr [all | default | <vr_name>]
To change the default TCP port number, use the following command:
configure telnet port [<portno> | default]
The range for the port number is 1 through 65535. The following TCP port numbers are reserved and
cannot be used for Telnet connections: 22, 80, and 1023. If you attempt to configure a reserved port, the
switch displays an error message.
Using ACLs to Control Telnet Access

By default, Telnet services are enabled on the switch. You can restrict Telnet access by using an access
control list (ACL) and implementing an ACL policy. You configure an ACL policy to permit or deny a
specific list of IP addresses and subnet masks for the Telnet port.
There are two methods to load ACL policies to the switch:

● Use the edit policy command to launch a VI-like editor on the switch. You can create the policy
directly on the switch.
● Use the tftp command to transfer a policy that you created using a text editor on another system to
the switch.
For more information about creating and implementing ACLs and policies, see Chapter 11, “Policy
Manager,” and Chapter 12, “Access Lists (ACLs).”
Sample ACL Policies. The following are sample policies that you can apply to restrict Telnet access.
In the following example named MyAccessProfile.pol, the switch permits connections from the subnet
10.203.133.0/24 and denies connections from all other addresses:
MyAccessProfile.pol
Entry AllowTheseSubnets {
if {
source-address 10.203.133.0 /24;
}
then
{
permit;
}
}
In the following example named MyAccessProfile.pol, the switch permits connections from the subnets
10.203.133.0/24 or 10.203.135.0/24 and denies connections from all other addresses:
MyAccessProfile.pol
if {
}
then

Managing the Switch
{
permit;
}
}
In the following example named MyAccessProfile_2.pol, the switch does not permit connections from
the subnet 10.203.133.0/24 but accepts connections from all other addresses:
MyAccessProfile_2.pol
Entry dontAllowTheseSubnets {
if {
}
then
{
deny;
}
}
Entry AllowTheRest {
If {
; #none specified
}
then
{
permit;
}
}
the subnets 10.203.133.0/24 or 10.203.135.0/24 but accepts connections from all other addresses:
if {
}
then
{
deny;
}
}
If {
; #none specified
}
then
{
permit;
}
}

Using Secure Shell 2
Configuring Telnet to Use ACL Policies. This section assumes that you have already loaded the policy on
the switch. For more information about creating and implementing ACLs and policies, see Chapter
11, “Policy Manager,” and Chapter 12, “Access Lists (ACLs).”
To configure Telnet to use an ACL policy to restrict Telnet access, use the following command:
configure telnet access-profile [<access_profile> | none]
Use the none option to remove a previously configured ACL.
Viewing Telnet Information

To display the status of Telnet, including the current TCP port, the virtual router used to establish a
Telnet session, and whether ACLs are controlling Telnet access, use the following command:
show management
Disabling and Enabling Telnet

You can choose to disable Telnet by using the following command:
disable telnet
To re-enable Telnet on the switch, use the following command:
enable telnet
You must be logged in as an administrator to configure the virtual router(s) used by Telnet and to
enable or disable Telnet.
Disconnecting a Telnet Session

A person with an administrator level account can disconnect a Telnet management session. If this
happens, the user logged in by way of the Telnet connection is notified that the session has been
terminated.
To terminate a Telnet session:

1 Log in to the switch with administrator privileges.
2 Determine the session number of the session you want to terminate by using the following
command:
show session {{detail} {<sessID>}} {history}
3 Terminate the session by using the following command:
clear session [<sessId> | all]
Using Secure Shell 2

Secure Shell 2 (SSH2) is a feature of ExtremeXOS that allows you to encrypt session data between a
network administrator using SSH2 client software and the switch. Configuration, policy, image, and
public key files may also be transferred to the switch using the Secure Copy Program 2 (SCP2).
Beginning with ExtremeXOS 11.2, you can also use SSH2 to connect to other devices from the switch.

Managing the Switch
The ExtremeXOS CLI provides a command that enables the switch to function as an SSH2 client,
sending commands to a remote system via an SSH2 session.
Beginning with ExtremeXOS 11.2, the switch accepts IPv6 connections.
Up to eight active SSH2 sessions can run on the switch concurrently. If you enable the idle timer using
the enable idletimeout command, the SSH2 connection times out after 20 minutes of inactivity by
default. If you disable the idle timer using the disable idletimeout command, the SSH2 connection
times out after 61 minutes of inactivity. If a connection to an SSH2 session is lost inadvertently, the
switch terminates the session within 61 minutes.
For detailed information about SSH2, see Chapter 16, “Security.”
Using the Trivial File Transfer Protocol

ExtremeXOS supports the Trivial File Transfer Protocol (TFTP) based on RFC 1350. TFTP is a method
used to transfer files from one network device to another. The ExtremeXOS TFTP client is a command
line application used to contact an external TFTP server on the network. For example, ExtremeXOS uses
TFTP to download software image files, switch configuration files, and ACLs from a server on the
network to the switch.
Up to eight active TFTP sessions can run on the switch concurrently.
Extreme Networks recommends using a TFTP server that supports blocksize negotiation (as described
in RFC 2348, TFTP Blocksize Option), to enable faster file downloads and larger file downloads.
For additional information about TFTP, see the following chapters:

● For information about downloading software image files, BootROM files, and switch configurations,
see Appendix A, “Software Upgrade and Boot Options.”
● For information about downloading ACL (and other) policy files, see Chapter 11, “Policy Manager.”
● For information about using TFTP to transfer files to and from the switch, see Chapter 5, “Managing
the ExtremeXOS Software.”
● For information about configuring core dump files and managing the core dump files stored on your
switch, see Appendix B, “Troubleshooting.” If configured, you can transfer core dump (debug) files
from either the internal memory card or the removable external compact flash card. You can install a
removable external compact flash card in only a modular switch.
Connecting to Another Host Using TFTP

You can TFTP from the current CLI session to another host to transfer files using the following
command:
tftp [<host-name> | <ip-address>] {-v <vr_name>} [-g | -p] [{-l [internal-memory
<local-file-internal> | memorycard <local-file-memcard> | <local-file>} {-r <remote-
file>} | {-r <remote-file>} {-l [internal-memory <local-file-internal> | memorycard
<local-file-memcard> | <local-file>]}]
NOTE
created VRs.

Understanding System Redundancy—Modular Switches and SummitStack Only
The TFTP session defaults to port 69. If you do not specify a virtual router, VR-Mgmt is used.
For example, to connect to a remote TFTP server with an IP address of 10.123.45.67 and “get” or retrieve
an ExtremeXOS configuration file named XOS1.cfg from that host, use the following command:
tftp 10.123.45.67 -g -r XOS1.cfg
When you “get” the file via TFTP, the switch saves the file to the primary MSM. If the switch detects a
backup MSM in the running state, the file is replicated to the backup MSM.
To view the files you retrieved, enter the ls command at the command prompt.
Beginning with ExtremeXOS 11.4, in addition to the tftp command, the following two commands are
available for transferring files to and from the switch:
● tftp get [<host-name> | <ip-address>] {-vr <vr_name>} [{[internal-memory <local-

file-internal> | memorycard <local-file-memcard> | <local_file>} {<remote_file>} |
{<remote_file>} {[internal-memory <local-file-internal> | memorycard <local-file-
memcard> | <local_file>]}] {force-overwrite}
NOTE
created VRs.
By default, if you transfer a file with a name that already exists on the system, the switch prompts
you to overwrite the existing file. For more information, see the tftp get command in the
ExtremeXOS Command Reference Guide.
● tftp put [<host-name> | <ip-address>] {-vr <vr_name>} [{[internal-memory <local-
memcard> | <local_file>]}]
NOTE
created VRs.
Understanding System Redundancy—Modular Switches

and SummitStack Only
If you install two MSMs or nodes in the chassis, or if you configure two master-capable nodes in a
SummitStack, one assumes the role of primary (also called "master") and the other assumes the role of
backup. The primary MSM or node provides all of the switch management functions including bringing
up and programming the I/O modules, running the bridging and routing protocols, and configuring
the switch. The primary MSM or node also synchronizes the backup MSM or node in case it needs to
take over the management functions if the primary MSM or node fails.

Managing the Switch
For SummitStack, a node can be a redundant primary node if it has been configured to be master-
capable. To configure master capability on one or all nodes in a SummitStack, use one of the following
commands:
● configure stacking [node-address <node-address> | slot <slot-number>] master-

capability [on | off]
● configure stacking redundancy [none | minimal | maximal]
Node Election
Node election is based on leader election between the MSMs installed in the chassis, or master-capable
nodes present in a SummitStack. By default, the MSM installed in slot A or the SummitStack node in
slot 1 has primary status. The Device Manager collects the node health information and forwards that
information to the Node Manager. The Node Manager then computes the quality of the node which is
later used in leader election.
When nodes exchange their health information, they determine the healthiest node. Based on the
election results obtained from all of the nodes, the healthiest node wins the election criteria.
At the end of the election process, a primary node is selected. The primary node runs the switch
management functions, and the backup node is available if the primary fails.
Determining the Primary Node

The following parameters determine the primary node:
● Node state—The node state must be STANDBY to participate in leader election and be selected as
primary. If the node is in the INIT, DOWN, or FAIL states, it cannot participate in leader election.
For more information about the node states, see “Viewing Node Status” on page 81.
● Configuration priority—This is a user assigned priority. The configured priority is compared only
after the node meets the minimum thresholds in each category for it to be healthy. Required
processes and devices must not fail.
● Software health—This represents the percent of processes available.
● Health of secondary hardware components—This represents the health of the switch components,
such as power supplies, fans, and so forth.
● Slot ID—The MSM slot where the node is installed (MSM-A or MSM-B).
Configuring the Node Priority on a Modular Switch

To configure the priority of an MSM node, use the following command:
configure node slot <slot_id> priority <node_pri>
If you do not configure any priorities, MSM-A has a higher priority than MSM-B. For the slot_id
parameter, enter A for the MSM installed in slot A or B for the MSM installed in slot B. By default, the
priority is 0 and the node priority range is 1 through 100. The higher the value, the higher the priority.

Configuring the Node Priority on a SummitStack

To configure the priority of a node in a SummitStack, use the following command:
configure stacking {node-address <node-address> | slot <slot-number>} priority

[<node-pri> | automatic]
If you do not configure any priorities, slot 1 has the highest priority, slot 2 the second highest priority,
and so forth in order of increasing slot number. Enter a number from 1 through 8 for the slot-number.
You may also use the factory assigned MAC address as the node-address value. By default the priority
is "automatic" and the node-pri value is any number between 1 and 100. The higher the value, the
higher the priority.
Relinquishing Primary Status

Before relinquishing primary status and initiating failover, review the section “Synchronizing Nodes—
Modular Switches and SummitStack Only” on page 1020 to confirm that your platform and both
installed MSMs are running software that supports the synchronize command.
You can cause the primary to failover to the backup, thereby relinquishing its primary status. To cause
the failover:
1 Use the show switch {detail} command on the primary or the backup node to confirm that the
nodes are synchronized and have identical software and switch configurations before failover. The
output displays the status of the nodes, with the primary node showing MASTER and the backup
node showing BACKUP (InSync).
A node may not be synchronized because checkpointing did not occur, incompatible software is
running on the primary and backup, or the backup is down.
● If the nodes are not synchronized and both nodes are running a version of ExtremeXOS that
supports synchronization, proceed to step 2.
● If the nodes are synchronized, proceed to step 3.
2 If the nodes are not synchronized because of incompatible software, use the synchronize command
to ensure that the backup has the same software in flash as the primary.
The synchronize command:
● Reboots the backup node to prepare it for synchronizing with the primary node
● Copies both the primary and secondary software images
● Copies both the primary and secondary configurations
● Reboots the backup node after replication is complete
After you confirm the nodes are synchronized, proceed to step 3.
3 If the nodes are synchronized, use the run msm-failover command to initiate failover from the
primary node to the backup node.
Replicating Data Between Nodes

ExtremeXOS replicates configuration and run-time information between the primary node and the
backup node so that the system can recover if the primary fails. This method of replicating data is
known as checkpointing. Checkpointing is the process of automatically copying the active state from the
primary to the backup, which allows for state recovery if the primary fails.

Managing the Switch
Replicating data consists of the following three steps:

1 Configuration synchronization—Relays current and saved configuration information from the
primary to the backup
2 Bulk checkpoint—Ensures that each individual application running on the system is synchronized
with the backup
3 Dynamic checkpoint—Checkpoints any new state changes from the primary to the backup
To monitor the checkpointing status, use the show checkpoint-data {<process>} command.
Relaying Configuration Information

To facilitate a failover from the primary node to the backup node, the primary transfers its active
configuration to the backup. Relaying configuration information is the first level of checkpointing.
During the initial switch boot-up, the primary’s configuration takes effect. During the initialization of a
node, its configuration is read from the local flash. After the primary and backup nodes have been
elected, the primary transfers its current active configuration to the backup. After the primary and
backup nodes are synchronized, any configuration change you make to the primary is relayed to the
backup and incorporated into the backup’s configuration copy.
NOTE
To ensure that all of the configuration commands in the backup’s flash are updated, issue the save command after
you make any changes. On a SummitStack, the save configuration command will normally save the primary node's
configuration file to all active nodes in the SummitStack.
If a failover occurs, the backup node continues to use the primary’s active configuration. If the backup
determines that it does not have the primary’s active configuration because a run-time synchronization
did not happen, the backup uses the configuration stored in its flash memory. Because the backup
always uses the primary’s active configuration, the active configuration remains in affect regardless of
the number of failovers.
NOTE
If you issue the reboot command before you save your configuration changes, the switch prompts you to save your
changes. To keep your configuration changes, save them before you reboot the switch.
Bulk Checkpointing
Bulk checkpointing causes the primary and backup run-time states to be synchronized. Since
ExtremeXOS runs a series of applications, an application starts checkpointing only after all of the
applications it depends on have transferred their run-time states to the backup MSM node.
After one application completes bulk checkpointing, the next application proceeds with its bulk
checkpointing.
To monitor the checkpointing status, use the show checkpoint-data {<process>} command.
To view the status of bulk checkpointing and see if the backup node is synchronized with the primary
node, use the show switch {detail} command.

Dynamic Checkpointing
After an application transfers its saved state to the backup node, dynamic checkpointing requires that
any new configuration information or state changes that occur on the primary be immediately relayed
to the backup. This ensures that the backup has the most up-to-date and accurate information.
Viewing Checkpoint Statistics

To view and check the status of one or more processes being copied from the primary to the backup
node, use the following command:
show checkpoint-data {<process>}
This command is also helpful in debugging synchronization problems that occur at run time.
This command displays, in percentages, the amount of copying completed by each process and the
traffic statistics between the process on both the primary and the backup nodes.
Viewing Node Status

ExtremeXOS allows you to view node statistical information. Each node in a modular switch, or
stackable switch in a SummitStack installed in your system is self-sufficient and runs the ExtremeXOS
management applications. By reviewing this output, you can see the general health of the system along
with other node parameters.
To view node status, use the following command:
show node {detail}
In a SummitStack, the "show stacking" command will show the node roles of active nodes.
Table 9 lists the node status collected by the switch.
Table 9: Node states

Node State Description
BACKUP In the backup state, this node becomes the primary node if the primary fails or enters the DOWN
state. The backup node also receives the checkpoint state data from the primary.
DOWN In the down state, the node is not available to participate in leader election. The node enters this
state during any user action, other than a failure, that makes the node unavailable for
management. Examples of user actions are:
• Upgrading the software
• Rebooting the system using the reboot command
• Initiating an MSM failover using the run msm-failover command
• Synchronizing the MSMs software and configuration in non-volatile storage using the
synchronize command
FAIL In the fail state, the node has failed and needs to be restarted or repaired. The node reaches this
state if the system has a hardware or software failure.
INIT In the initial state, the node is being initialized. A node stays in this state when it is coming up
and remains in this state until it has been fully initialized. Being fully initialized means that all of
the hardware has been initialized correctly and there are no diagnostic faults.
MASTER In the primary (master) state, the node is responsible for all switch management functions.

Managing the Switch
Table 9: Node states (Continued)

Node State Description
STANDBY In the standby state, leader election occurs—the primary and backup nodes are elected. The
priority of the node is only significant in the standby state.
In SummitStack, there can be more than two master-capable nodes. All such nodes that do not
get elected either Master or Backup remain in Standby state.
Understanding Hitless Failover Support—Modular

Switches and SummitStack Only
The term "hitless failover" has slightly different meaning on a modular chassis and a SummitStack. On a
modular chassis, MSMs do not directly control customer ports; such ports are directly controlled by
separate processors. However, a SummitStack node has customer ports that are under the control of its
single central processor. When a modular chassis MSM failover occurs, all of the ports in the chassis are
under the control of separate processors which can communicate with the backup MSM, so all ports
continue to function. In a SummitStack, failure of the primary node results in all ports that require that
node's processor for normal operation going down. The remaining SummitStack nodes' ports continue
to function normally. Aside from this difference, hitless failover is the same on modular chassis and
SummitStack.
As described in the section, “Understanding System Redundancy—Modular Switches and SummitStack
Only” on page 77, if you install two MSMs (nodes) in a chassis or SummitStack, one assumes the role of
primary and the other assumes the role of backup. The primary node provides all of the switch
management functions including bringing up and programming the I/O modules or other nodes in the
SummitStack, running the bridging and routing protocols, and configuring the switch. The primary
node also synchronizes the backup node in case it needs to take over the management functions if the
primary node fails.
The configuration is one of the most important pieces of information checkpointed to the backup node.
Each component of the system needs to checkpoint whatever runtime data is necessary to allow the
backup node to take over as the primary node if a failover occurs, including the protocols and the
hardware dependent layers. For more information about checkpointing data and relaying configuration
information, see “Replicating Data Between Nodes” on page 79.
Not all protocols support hitless failover; see Table 10 for a detailed list of protocols and their support.
Layer 3 forwarding tables are maintained for pre-existing flows, but subsequent behavior depends on
the routing protocols used. Static Layer 3 configurations and routes are hitless. You must configure
OSPF graceful restart for OSPF routes to be maintained, and you must configure BGP graceful restart
for BGP routes to be maintained. For more information about OSPF, see Chapter 10, “OSPF,” and for
more information about BGP, see Chapter 12, “Border Gateway Protocol.” For routing protocols that do
not support hitless failover, the new primary node removes and re-adds the routes.
Protocol Support for Hitless Failover

Table 10 summarizes the protocol support for hitless failover. Unless otherwise noted, the behavior is
the same for all modular switches.
If a protocol indicates support for hitless failover, additional information is also available in that
particular chapter. For example, for information about network login support of hitless failover, see
Chapter 15, “Network Login.”

Understanding Hitless Failover Support—Modular Switches and SummitStack Only
Table 10: Protocol support for hitless failover

Protocol Behavior Hitless
Spanning Tree Protocol STP supports hitless failover including catastrophic failure of the primary Yes
(STP) node without interruption. There should be no discernible network event
external to the switch. The protocol runs in lock step on both master and
backup nodes and the backup node is a hot spare that can take over at any
time with no impact on the network.
Ethernet Automatic The primary node replicates all EAPS BPDUs to the backup, which allows Yes
Protection Switching the backup to be aware of the state of the EAPS domain. Since both primary
(EAPS) and backup nodes receive EAPS BPDUs, each node maintains equivalent
EAPS states.
By knowing the state of the EAPS domain, the EAPS process running on the
backup node can quickly recover after a primary node failover. Although both
primary and backup nodes receive EAPS BPDUs, only the primary transmits
EAPS BPDUs to neighboring switches and actively participates in EAPS.
Extreme Standby If failover occurs on the ESRP MASTER switch, it sends a hello packet with Yes
Router Protocol (ESRP) the HOLD bit set. On receiving this packet, the ESRP SLAVE switch freezes
all further state transitions. The MASTER switch keeps sending hellos with
the HOLD bit set on every hello interval. When the MASTER is done with its
failover, it sends another hello with the HOLD bit reset. The SLAVE switch
resumes normal processing. (If no packet is received with the HOLD bit
reset, the SLAVE timeouts after a certain time interval and resumes normal
processing.)
Failover on the ESRP SLAVE switch is of no importance because it is the
SLAVE switch.
Virtual Router VRRP supports hitless failover. The primary node replicates VRRP PDUs to Yes
Redundancy Protocol the backup, which allows the primary and backup nodes to run VRRP in
(VRRP) parallel. Although both nodes receive VRRP PDUs, only the primary
transmits VRRP PDUs to neighboring switches and participates in VRRP.
Extreme Discovery EDP does not checkpoint protocol data units (PDUs) or states, so the backup No
Protocol (EDP) node does not have the neighbor’s information. If the backup node becomes
the primary node, and starts receiving PDUs, the new primary learns about
its neighbors.
Protocol Independent After a failover, all hardware and software caches are cleared and learning No
Multicast (PIM) from the hardware is restarted. This causes a traffic interruption since it is
the same as if the switch rebooted for all Layer 3 multicast traffic.
Extreme Loop Recovery If you use ELRP as a standalone tool, hitless failover support is not needed No
Protocol (ELRP) since the you initiate the loop detection.
If you use ELRP in conjunction with ESRP, ELRP does not interfere with the
hitless failover support provided by ESRP.
Although there is no hitless failover support in ELRP itself, ELRP does not
affect the network behavior if a failover occurs.
Link Layer Discovery Since LLDP is more of a tool than a protocol, there is no hitless failover No
Protocol (LLDP) support. LLDP is similar to EDP, but there is also a MIB interface to query
the information learned. After a failover, it takes 30 seconds or greater
before the MIB database is fully populated again.
Link Aggregation If the backup node becomes the primary node, there is no traffic disruption. Yes
Control Protocol (LACP)
Routing Information RIP does not support graceful restart, so the route manager deletes all RIP No
Protocol (RIP) routes 1 second after the failover occurs. This results in a traffic interruption
as well as an increase in control traffic as RIP
re-establishes its database.

Managing the Switch
Table 10: Protocol support for hitless failover (Continued)

Protocol Behavior Hitless
Routing Information RIPng does not support graceful restart, so the route manager deletes all No
Protocol next RIPng routes 1 second after the failover occurs. This results in a traffic
generation (RIPng) interruption.
In addition, after RIPng comes up on the new primary node, it relearns the
routes from its neighbors. This causes an increase in control traffic onto the
network.
Open Shortest Path If you configure OSPF graceful restart, there is no traffic interruption. Yes
First (OSPF) However, after OSPF comes up after restart, OSPF re-establishes sessions
with its neighbors and relearns Link State Advertisements (LSAs) from all of
the neighbors. This causes an increase in control traffic onto the network.
If you do not configure graceful restart, the route manager deletes all OSPF
routes 1 second after the failover occurs, which results in a traffic
interruption in addition to the increased control traffic.
Open Shortest Path OSPFv3 does not support graceful restart, so the route manager deletes all No
First v3 (OSPFv3) OSPFv3 routes 1 second after the failover occurs. This results in a traffic
interruption.
In addition, after OSPFv3 comes up on the new primary node, it relearns the
routes from its neighbors. This causes an increase in control traffic onto the
network.
Border Gateway If you configure BGP graceful restart, by default the route manager does not Yes
Protocol (BGP) delete BGP routes until 120 seconds after failover occurs. There is no traffic
interruption. However, after BGP comes up after restart, BGP re-establishes
sessions with its neighbors and relearns routes from all of them. This causes
an increase in control traffic onto the network.
If you do not configure graceful restart, the route manager deletes all BGP
routes 1 second after the failover occurs, which results in a traffic
interruption in addition to the increased control traffic.
Power over Ethernet The PoE configuration is checkpointed to the backup node. This ensures that Yes
(PoE) if the backup takes over, all ports currently powered stay powered after the
failover and the configured power policies are still in place.
This behavior is applicable only on the BlackDiamond 8800 series switch
and SummitStack.
Network Login 802.1x Authentication Yes
Authenticated clients continue to remain authenticated after failover.
However, 1 second after failover, all authenticated clients are forced to re-
authenticate themselves.
Information about unauthenticated clients is not checkpointed so any such
clients that were in the process of being authenticated at the instant of
failover must go through the authentication process again from the beginning
after failover.
Network Login MAC-Based Authentication Yes
Continued
Authenticated clients continue to remain authenticated after failover so the
failover is transparent to them. Information about unauthenticated clients is
not checkpointed so any such clients that were in the process of being
authenticated at the instant of failover must go through the authentication
process again from the beginning after failover.
In the case of MAC-Based authentication, the authentication process is very
short with only a single packet being sent to the switch so it is expected to
be transparent to the client stations.
Network Login Web-Based Authentication Yes
Continued
Web-based Netlogin users continue to be authenticated after a failover.

Understanding Hitless Failover Support—Modular Switches and SummitStack Only
Platform Support for Hitless Failover

Table 11 lists when each platform and management module supported hitless failover for a specific
protocol. If you are running an earlier version of ExtremeXOS than what is listed in the ExtremeXOS
version column, the switch does not support hitless failover for that protocol. For example, if you have
a BlackDiamond 10808 switch running ExtremeXOS 11.4, the switch does not support VRRP hitless
failover.
Remember, as described in Table 10, not all protocols support hitless failover.
Table 11: Platform support for hitless failover

Platform Management Module Protocol ExtremeXOS Version
BlackDiamond 10808 switch MSM-1 and BGP graceful restart 11.4
MSM-1XL
EAPS 11.4
ESRP 11.0
LACP 11.4
Network login 11.3
OSPF graceful restart 11.3
STP 11.0
VRRP 11.6
BlackDiamond 8800 series switch MSM-G8X BGP graceful restart 11.4
EAPS 11.4
ESRP 11.3
LACP 11.4
Network login 11.3
PoE 11.3
STP 11.3
VRRP 11.6
MSM-48 BGP graceful restart 11.6
EAPS 11.6
ESRP 11.6
LACP 11.6
Network login 11.6
PoE 11.6
STP 11.6
VRRP 11.6
SummitStack Any member of the BGP graceful restart 12.0
Summit family of
switches
(features available EAPS 12.0
depend on license
level)
ESRP 12.0
LACP 12.0

Managing the Switch
Table 11: Platform support for hitless failover (Continued)

Platform Management Module Protocol ExtremeXOS Version
Network login 12.0
STP 12.0
VRRP 12.0
BlackDiamond 12800 series switch MSM-5 and BGP graceful restart 11.4
MSM-5R
EAPS 11.4
ESRP 11.4
LACP 11.4
Network login 11.4
STP 11.4
VRRP 11.6
Hitless Failover Caveats

This section describes the caveats for hitless failover. Check the latest version of the ExtremeXOS release
notes for additional information.
Caveats for All Modular Switches

The following summary describes the hitless failover caveats for all modular switches:
● A brief traffic interruption (less than 1/2 of a second) can occur when the failover happens due to
reprogramming of the I/O modules to bypass the original primary MSM’s switch chips.
Caveats for SummitStack

The following describes the hitless failover caveats for the SummitStack:
● All customer ports and the stacking links connected to the failed primary node will go down. In the
recommended stack ring configuration, the stack becomes a daisy chain until the failed node restarts
or is replaced.
● A brief traffic interruption (less than 50 milliseconds) can occur when the traffic on the ring is
rerouted because the active topology becomes a daisy chain.
● Since the SummitStack can contain more than two master-capable nodes, it is possible to
immediately elect a new backup node. If a new backup node is elected, when the original primary
node restarts, it will become a standby node.
● To simulate the behavior of a chassis, a MAC address of one of the nodes is designated as the seed
to form a "stack MAC address". When a failover occurs, the SummitStack continues to be identified
with this address.

Understanding Power Supply Management
Caveats for the BlackDiamond 10808 Switch Only

The following summary describes the hitless failover caveats for the BlackDiamond 10808 switch and
the BlackDiamond 8800 series switch:
● There is a 50% reduction in backplane bandwidth during a failover.
Caveats for the BlackDiamond 8800 Series Switch Only

The following summary describes the hitless failover caveats for the BlackDiamond 8800 series switch:
● I/O modules not yet in the Operational state are powered off and the card state machine is restarted
to bring them to the Operational state. This results in a delay in the I/O module becoming
Operational.
Caveats for the BlackDiamond 12800 Series Switch Only

The following summary describes the hitless failover caveats for the BlackDiamond 12800 series switch:
● There is only one active link to the backplane during a failover.

This section describes how ExtremeXOS manages power consumption on the switch:
● Using Power Supplies—Modular Switches Only on page 87
● Using Power Supplies—Summit Family of Switches Only on page 90
● Displaying Power Supply Information on page 91
Using Power Supplies—Modular Switches Only

ExtremeXOS monitors and manages power consumption on the switch by periodically checking the
power supply units (PSUs) and testing them for failures. To determine the health of the PSU,
ExtremeXOS checks the voltage, current, and temperature of the PSU.
The power management capability of ExtremeXOS:

● Protects the system from overload conditions
● Monitors all installed PSUs, even installed PSUs that are disabled
● Enables and disables PSUs as required
● Powers up or down I/O modules based on available power and required power resources
● Logs power resource changes, including power budget, total available power, redundancy, and so
on
● Detects and isolates faulty PSUs
The switch includes two power supply controllers that collect data from the installed PSUs and report
the results to the MSM modules. When you first power on the switch, the power supply controllers
enable a PSU. As part of the power management function, the power controller disables the PSU if an
unsafe condition arises. For more information about the power supply controller, see the Extreme
Networks Consolidated XOS Hardware Installation Guide.

Managing the Switch
If you have a BlackDiamond 8800 Power over Ethernet (PoE) G48P module installed in the
BlackDiamond 8800 series switch, there are specific power budget requirements and configurations
associated with PoE that are not described in this section. For more detailed information about PoE, see
Chapter 4, “Power Over Ethernet.”
ExtremeXOS 11.6 introduces support for the 600/900 W AC PSU for the BlackDiamond 8806 switch.
You can mix existing 700/1200 W AC PSUs and 600/900 W AC PSUs in the same chassis; however, you
must be running ExtremeXOS 11.6 to support the 600/900 W AC PSUs. If you install the 600/900 W AC
PSU in a chassis other than the BlackDiamond 8806, ExtremeXOS provides enough power to boot-up
the chassis, display a warning message in the log, and disable the PSU. If this occurs, you see a message
similar to the following:
<Warn:HAL.Sys.Warning>MSM-A:Power supply in slot 6 is not supported and is being

disabled.
When a combination of 700/1200 W AC PSUs and 600/900 W AC PSUs are powered on in the same
BlackDiamond 8806 chassis, all 700/1200 W AC PSUs are budgeted “down” to match the lower
powered 600/900 W AC output values to avoid PSU shutdown. For more information about the 600/
900 W AC PSU, see the Extreme Networks Consolidated XOS Hardware Installation Guide.
This section describes the following power management topics:

● Initial System Boot-Up on page 88
● Power Redundancy on page 89
● Power Management Guidelines on page 89
● Overriding Automatic Power Supply Management on page 90
Initial System Boot-Up

When ExtremeXOS boots up, it reads and analyzes the installed I/O modules. ExtremeXOS considers
the I/O modules for power up from the lowest numbered slot to the highest numbered slot, based on
their power requirements and the available system power. If the system does not have enough power,
some I/O modules are not powered up. For example, ExtremeXOS:
● Collects information about the PSUs installed to determine how many are running and how much
power each can supply.
● Checks for PSU failures.
● Calculates the number of I/O modules to power up based on the available power budget and the
power requirements of each I/O module, including PoE requirements for the BlackDiamond 8800
PoE I/O module.
● Reserves the amount of power required to power up a second MSM if only one MSM is installed.
● Reserves the amount of power required to power all fans and chassis components.
● Calculates the current power surplus or shortfall.
● Logs and sends SNMP traps for transitions in the overall system power status, including whether
the available amount of power is:
■ Redundant or N+1—Power from a single PSU can be lost and no I/O modules are powered
down.
■ Sufficient, but not redundant—Power from a single PSU is lost, and one or more I/O modules
are powered down.
■ Insufficient—One or more modules are not powered up due to a shortfall of available power.

By reading the PSU information, ExtremeXOS determines the power status and the total amount of
power available to the system. The total power available determines which I/O modules can be
powered up.
Power Redundancy
In simple terms, power redundancy (N+1) protects the system from shutting down. With redundancy, if
the output of one PSU is lost for any reason, the system remains fully powered. In this scenario, N is
the minimum number of power supplies needed to keep the system fully powered and the system has
N+1 PSUs powered.
If the system power status is not redundant, the removal of one PSU, the loss of power to one PSU, or a
degradation of input voltage results in insufficient power to keep all of the I/O modules powered up. If
there is not enough power, ExtremeXOS powers down the I/O modules from the highest numbered slot
to the lowest numbered slot until the switch has enough power to continue operation.
If you install or provide power to a new PSU, I/O modules powered down due to earlier insufficient
power are considered for power up from the lowest slot number to the highest slot number, based on
the I/O module’s power requirements.
Whenever the system experiences a change in power redundancy, including a change in the total
available power, degraded input voltage, or a return to redundant power, the switch sends messages to
the syslog.
Power Management Guidelines

The following list describes some key issues to remember when identifying your power needs and
installing PSUs:
● If you disable a slot, the I/O module installed in that slot is always powered down regardless of the
number of PSUs installed.
● If a switch has PSUs with a mix of both 220V AC and 110V AC inputs, ExtremeXOS maximizes
system power by automatically taking one of two possible actions:
■ If all PSUs are enabled then all PSUs must be budgeted at 110V AC to prevent overload of PSUs
with 110V AC inputs.
OR
■ If the PSUs with 110V AC inputs are disabled, then the PSUs with 220V AC inputs can be
budgeted with a higher output per PSU.
ExtremeXOS computes the total available power using both methods and automatically uses the PSU
configuration that provides the greatest amount of power to the switch. Table 12 lists combinations
where ExtremeXOS maximizes system power by disabling the PSUs with 110V AC inputs.
Table 12: PSU combinations where 110V PSUs are disabled

Number of PSUs with Number of PSUs with
220V AC Inputs 110V AC Inputs
2 1
3 1
3 2
4 1
4 2

Managing the Switch
Table 12: PSU combinations where 110V PSUs are disabled (Continued)
Number of PSUs with Number of PSUs with
220V AC Inputs 110V AC Inputs
5 1
For all other combinations of 220V AC and 110V AC PSUs, ExtremeXOS maximizes system power
by enabling all PSUs and budgeting each PSU at 110V AC.
BlackDiamond 8806 switch only—When a combination of 700/1200 W AC PSUs and 600/900 W
AC PSUs are powered on in the same BlackDiamond 8806 chassis, all 700/1200 W AC PSUs are
budgeted “down” to match the lower powered 600/900 W AC output values to avoid PSU
shutdown.
NOTE
If you are running ExtremeXOS 11.2 or 11.1 and mix PSUs with 110V and 220V AC inputs, the switch budgets
all PSUs as if they have 110V AC inputs.
Overriding Automatic Power Supply Management

You can override automatic power supply management to enable a PSU with 110V AC inputs that
ExtremeXOS disables if the need arises, such as for a planned maintenance of 220V AC circuits. If the
combination of AC inputs represents one of those listed in Table 12, you can turn on a disabled PSU
using the following command:
configure power supply <ps_num> on
NOTE
If you override automatic power supply management, you may reduce the available power and cause one or more I/O
modules to power down.
To resume using automatic power supply management on a PSU, use the configure power supply
<ps_num> auto command. The setting for each PSU is stored as part of the switch configuration.
To display power supply status and power budget information, use the show power and show power
budget commands.
Using Power Supplies—Summit Family of Switches Only

On the Summit family of switches, ExtremeXOS reports when the PSU has power or has failed. The
Summit family of switches support an internal power supply with a range of 90V to 240V AC power as
well as an external redundant power supply. The Extreme Networks External Power System (EPS)
allows you to add a redundant power supply to the Summit family of switches to protect against a
power supply failure. The EPS consists of a tray or module that holds the EPS power supplies.
On non-PoE Summit switches, if you experience an internal PSU failure and do not have an external
PSU installed, the switch powers down. If you experience a PSU failure and have an external PSU
installed, the switch uses the external PSU to maintain power to the switch.

Using the Simple Network Management Protocol
On PoE Summit switches, there are specific power budget requirements and configurations associated
with PoE that are not described in this section. The PoE Summit switches respond to internal and
external PSU failures based on your PoE configurations. For more information about configuring PoE
on the Summit X450e-24p and X450e-48p switches, see Chapter 4, “Power Over Ethernet.”
For more information about the Summit X450 family of switches and the EPS, see the Extreme Networks
Consolidated XOS Hardware Installation Guide.
Using Power Supplies - SummitStack Only

Since the nodes have their own power supplies and since they cannot be shared, management is the
same as it is for standalone Summit family switches. The only difference is that the power management
commands have been centralized so that they can be issued from the primary node.
Displaying Power Supply Information

To display the status of the currently installed power supplies on all switches, use the following
command:
show power {<ps_num>} {detail}
On modular switches, the following commands provide additional power supply information.
To view the system power status and the amount of available and required power, use the following
command:
show power budget
To display the status of the currently installed power supply controllers on modular switches, use the
following command:
show power controller {<num>}

Any network manager program running the Simple Network Management Protocol (SNMP) can
manage the switch, provided the Management Information Base (MIB) is installed correctly on the
management station. Each network manager program provides its own user interface to the
management facilities.
Note, when using a network manager program to create a VLAN, Extreme Networks does not support
the SNMP create and wait operation. To create a VLAN with SNMP, use the create and go operation.
The following sections describe how to get started if you want to use an SNMP manager. It assumes
you are already familiar with SNMP management. If not, refer to the following publication:
The Simple Book

by Marshall T. Rose
ISBN 0-13-8121611-9
Published by Prentice Hall.

Managing the Switch
This section describes the following SNMP topics:

● Enabling and Disabling SNMPv1/v2c and SNMPv3 on page 92
● Accessing Switch Agents on page 93
● Supported MIBs on page 93
● Configuring SNMPv1/v2c Settings on page 93
● Displaying SNMP Settings on page 94
● SNMPv3 on page 94
● Message Processing on page 96
● SNMPv3 Security on page 96
● SNMPv3 MIB Access Control on page 98
● SNMPv3 Notification on page 99
Enabling and Disabling SNMPv1/v2c and SNMPv3

ExtremeXOS can concurrently support SNMPv1/v2c and SNMPv3. The default is both types of SNMP
enabled. Network managers can access the device with either SNMPv1/v2c methods or SNMPv3. To
enable concurrent support, use the following command:
enable snmp access
To prevent any type of SNMP access, use the following command:
disable snmp access
To prevent access using SNMPv1/v2c methods and allow access using SNMPv3 methods only, use the
following commands:
enable snmp access

disable snmp access {snmp-v1v2c}
The switch cannot be configured to simultaneously allow SNMPv1/v2c access and prevent SNMPv3
access.
Most of the commands that support SNMPv1/v2c use the keyword snmp; most of the commands that
support SNMPv3 use the keyword snmpv3.
After a switch reboot, all slots must be in the “Operational” state before SNMP can manage and access
the slots. To verify the current state of the slot, use the show slot command.
Understanding Safe Defaults Mode and SNMP

ExtremeXOS 11.2 introduces the concept of safe defaults mode. Safe defaults mode runs an interactive
script that allows you to enable or disable SNMP, Telnet, and switch ports. When you set up your
switch for the first time, you must connect to the console port to access the switch. After logging in to
the switch, you enter safe defaults mode. Although SNMP, Telnet, and switch ports are enabled by
default, the script prompts you to confirm those settings.

If you choose to keep the default setting for SNMP—the default setting is enabled—the switch returns
the following interactive script:
Accessing Switch Agents

To access the SNMP agent residing in the switch, at least one VLAN must have an assigned IP address.
By default, SNMP access and SNMPv1/v2c traps are enabled. SNMP access and SNMP traps can be
disabled and enabled independently—you can disable SNMP access but still allow SNMP traps to be
sent, or vice versa.
Supported MIBs
In addition to private MIBs, the switch supports the standard MIBs listed in Appendix D, “Supported
Protocols, MIBs, and Standards.”
Configuring SNMPv1/v2c Settings

The following SNMPv1/v2c parameters can be configured on the switch:
● Authorized trap receivers—An authorized trap receiver can be one or more network management
stations on your network. The switch sends SNMPv1/v2c traps to all configured trap receivers. You
can specify a community string and UDP port individually for each trap receiver. All community
strings must also be added to the switch using the configure snmp add community command.
To configure a trap receiver on a switch, use the following command:
configure snmp add trapreceiver <ip_address> community [[hex <hex_community_name>]
| <community_name>] {port <port_number>} {from <src_ip_address>} {mode <trap_mode>
[enhanced | standard]}
To delete a trap receiver on a switch, use the following command:
configure snmp delete trapreceiver
Entries in the trap receiver list can also be created, modified, and deleted using the RMON2
trapDestTable MIB table, as described in RFC 2021.
● SNMP access control—This feature allows the administrator to restrict SNMP access by using the
access control list (ACL) and implementing an ACL policy. The administrator can configure an ACL
policy to either permit or deny specific list of IP address and subnet masks. There are four
subcommands for enacting access control:
■ To configure SNMP to use an ACL policy, use the following command:
configure snmp access-profile <profile_name>
By default, SNMP supports the read/write option.

Managing the Switch
■ To configure SNMP to remove a previously configured ACL policy, use the following command:
configure snmp access-profile none
■ To configure SNMP to use an ACL policy and support the read-only option, use the following
command:
configure snmp access-profile <profile_name> readonly
■ To configure SNMP to use an ACL policy and support the read/write option explicitly, use the
following command:
configure snmp access-profile <profile_name> readwrite
● Community strings—The community strings allow a simple method of authentication between the
switch and the remote network manager. There are two types of community strings on the switch:
■ Read community strings provide read-only access to the switch. The default read-only
community string is public.
■ Read-write community strings provide read- and-write access to the switch. The default read-
write community string is private.
● System contact (optional)—The system contact is a text field that enables you to enter the name of
the person(s) responsible for managing the switch.
● System name (optional)—The system name enables you to enter a name that you have assigned to
this switch. The default name is the model name of the switch (for example, BD-1.2).
● System location (optional)—Using the system location field, you can enter the location of the switch.
Displaying SNMP Settings

To display the SNMP settings configured on the switch, use the following command:
show management
This command displays the following information:

● Enable/disable state for Telnet and SNMP access
● Login statistics
■ Enable/disable state for idle timeouts
■ Maximum number of CLI sessions
● SNMP community strings
● SNMP trap receiver list
● SNMP trap receiver source IP address
● SNMP statistics counter
● Enable/disable state for Remote Monitoring (RMON)
SNMPv3
SNMPv3 is an enhanced standard for SNMP that improves the security and privacy of SNMP access to
managed devices and provides sophisticated control of access to the device MIB. The prior standard
versions of SNMP, SNMPv1, and SNMPv2c, provided no privacy and little security.

The following six RFCs provide the foundation for the Extreme Networks implementation of SNMPv3:
● RFC 2570, Introduction to version 3 of the Internet-standard Network Management Framework, provides an
overview of SNMPv3.
● RFC 2571, An Architecture for Describing SNMP Management Frameworks, talks about SNMP
architecture, especially the architecture for security and administration.
● RFC 2572, Message Processing and Dispatching for the Simple Network Management Protocol (SNMP),
talks about the message processing models and dispatching that can be a part of an SNMP engine.
● RFC 2573, SNMPv3 Applications, talks about the different types of applications that can be associated
with an SNMPv3 engine.
● RFC 2574, The User-Based Security Model for Version 3 of the Simple Network Management Protocol
(SNMPv3), describes the User-Based Security Model (USM).
● RFC 2575, View-based Access Control Model (VACM) for the Simple Network Management Protocol
(SNMP), talks about VACM as a way to access the MIB.
The SNMPv3 standards for network management were primarily driven by the need for greater
security and access control. The new standards use a modular design and model management
information by cleanly defining a message processing (MP) subsystem, a security subsystem, and an
access control subsystem.
The MP subsystem helps identify the MP model to be used when processing a received Protocol Data
Unit (PDU), which are the packets used by SNMP for communication. The MP layer helps in
implementing a multilingual agent, so that various versions of SNMP can coexist simultaneously in the
same network.
The security subsystem features the use of various authentication and privacy protocols with various
timeliness checking and engine clock synchronization schemes. SNMPv3 is designed to be secure
against:
● Modification of information, where an in-transit message is altered
● Masquerades, where an unauthorized entity assumes the identity of an authorized entity
● Message stream modification, where packets are delayed and/or replayed
● Disclosure, where packet exchanges are sniffed (examined) and information is learned about the
contents
The access control subsystem provides the ability to configure whether access to a managed object in a
local MIB is allowed for a remote principal. The access control scheme allows you to define access
policies based on MIB views, groups, and multiple security levels.
In addition, the SNMPv3 target and notification MIBs provide a more procedural approach for
generating and filtering of notifications.
SNMPv3 objects are stored in non-volatile memory unless specifically assigned to volatile storage.
Objects defined as permanent cannot be deleted.
NOTE
In SNMPv3, many objects can be identified by a human-readable string or by a string of hexadecimal octets. In
many commands, you can use either a character string, or a colon-separated string of hexadecimal octets to specify
objects. To indicate hexadecimal octets, use the keyword hex in the command.

Managing the Switch
Message Processing
A particular network manager may require messages that conform to a particular version of SNMP. The
choice of the SNMPv1, SNMPv2c, or SNMPv3 MP model can be configured for each network manager
as its target address is configured. The selection of the MP model is configured with the mp-model
keyword in the following command:
configure snmpv3 add target-params [[hex <hex_param_name>] | <param_name>] user [[hex
<hex_user_name>] | <user_name>] mp-model [snmpv1 | snmpv2c | snmpv3] sec-model [snmpv1
| snmpv2c | usm] {sec-level [noauth | authnopriv | priv]} {volatile}
SNMPv3 Security
In SNMPv3 the User-Based Security Model (USM) for SNMP was introduced. USM deals with security
related aspects like authentication, encryption of SNMP messages, and defining users and their various
access security levels. This standard also encompasses protection against message delay and message
replay.
USM Timeliness Mechanisms

An Extreme Networks switch has one SNMPv3 engine, identified by its snmpEngineID. The first four
octets are fixed to 80:00:07:7C, which represents the Extreme Networks vendor ID. By default, the
additional octets for the snmpEngineID are generated from the device MAC address.
Every SNMPv3 engine necessarily maintains two objects: SNMPEngineBoots, which is the number of
reboots the agent has experienced and SNMPEngineTime, which is the local time since the engine reboot.
The engine has a local copy of these objects and the latestReceivedEngineTime for every authoritative
engine it wants to communicate with. Comparing these objects with the values received in messages
and then applying certain rules to decide upon the message validity accomplish protection against
message delay or message replay.
In a chassis, the snmpEngineID is generated using the MAC address of the MSM with which the switch
boots first. In a SummitStack, the MAC address chosen for the snmpEngineID is the configured stack
MAC address.
The snmpEngineID can be configured from the command line, but when the snmpEngineID is changed,
default users revert back to their original passwords/keys, and non-default users are reset to the
security level of no authorization, no privacy. To set the snmpEngineID, use the following command:
configure snmpv3 engine-id <hex_engine_id>
SNMPEngineBoots can also be configured from the command line. SNMPEngineBoots can be set to any
desired value but will latch on its maximum, 2147483647. To set the SNMPEngineBoots, use the
following command:
configure snmpv3 engine-boots <(1-2147483647)>
Users, Groups, and Security

SNMPv3 controls access and security using the concepts of users, groups, security models, and security
levels.

Users. Users are created by specifying a user name. Depending on whether the user will be using
authentication and/or privacy, you would also specify an authentication protocol (MD5 or SHA) with
password or key, and/or privacy (DES) password or key. To create a user, use the following command:
configure snmpv3 add user [[hex <hex_user_name>] | <user_name>] {authentication [md5 |

sha] [hex <hex_auth_password> | <auth_password>]} {privacy [hex <hex_priv_password> |
<priv_password>]} {volatile}
A number of default, permanent users are initially available. The default user names are: admin, initial,
initialmd5, initialsha, initialmd5Priv, initialshaPriv. The default password for admin is password. For the
other default users, the default password is the user name.
To display information about a user, or all users, use the following command:
show snmpv3 user {[[hex <hex_user_name>] | <user_name>]}
To delete a user, use the following command:
configure snmpv3 delete user [all-non-defaults | [[hex <hex_user_name>] |

<user_name>]]
NOTE
The SNMPv3 specifications describe the concept of a security name. In the ExtremeXOS implementation, the user
name and security name are identical. In this manual, both terms are used to refer to the same thing.
Groups. Groups are used to manage access for the MIB. You use groups to define the security model,
the security level, and the portion of the MIB that members of the group can read or write. To
underscore the access function of groups, groups are defined using the following command:
configure snmpv3 add access [[hex <hex_group_name>] | <group_name>] {sec-model [snmpv1
| snmpv2c | usm]} {sec-level [noauth | authnopriv | priv]} {read-view [[hex
<hex_read_view_name>] | <read_view_name>]} {write-view [[hex <hex_write_view_name>]] |
<write_view_name>]} {notify-view [[hex <hex_notify_view_name]] | <notify_view_name>]}
{volatile}
The security model and security level are discussed in “Security Models and Levels” on page 98. The
view names associated with a group define a subset of the MIB (subtree) that can be accessed by
members of the group. The read view defines the subtree that can be read, write view defines the
subtree that can be written to, and notify view defines the subtree that notifications can originate from.
MIB views are discussed in “SNMPv3 MIB Access Control” on page 98.
A number of default (permanent) groups are already defined. These groups are: admin, initial, v1v2c_ro,
v1v2c_rw. To display information about the access configuration of a group or all groups, use the
following command:
show snmpv3 access {[[hex <hex_group_name>] | <group_name>]}
Users are associated with groups using the following command:
configure snmpv3 add group [[hex <hex_group_name>] | <group_name>] user [[hex

<hex_user_name>] | <user_name>] {sec-model [snmpv1| snmpv2c | usm]} {volatile}
To show which users are associated with a group, use the following command:
show snmpv3 group {[[hex <hex_group_name>] | <group_name>] {user [[hex

<hex_user_name>] | <user_name>]}}

Managing the Switch
To delete a group, use the following command:
configure snmpv3 delete access [all-non-defaults | {[[hex <hex_group_name>] |

<group_name>] {sec-model [snmpv1 | snmpv2c | usm] sec-level [noauth | authnopriv |
priv]}}]
When you delete a group, you do not remove the association between the group and users of the
group. To delete the association between a user and a group, use the following command:
configure snmpv3 delete group {[[hex <hex_group_name>] | <group_name>]} user [all-non-

defaults | {[[hex <hex_user_name>] | <user_name>] {sec-model [snmpv1|snmpv2c|usm]}}]
Security Models and Levels. For compatibility, SNMPv3 supports three security models:
● SNMPv1—no security
● SNMPv2c—community strings based security
● SNMPv3—USM security
The default is USM. You can select the security model based on the network manager in your network.
The three security levels supported by USM are:

● noAuthnoPriv—No authentication, no privacy. This is the case with existing SNMPv1/v2c agents.
● AuthnoPriv—Authentication, no privacy. Messages are tested only for authentication.
● AuthPriv—Authentication, privacy. This represents the highest level of security and requires every
message exchange to pass the authentication and encryption tests.
When a user is created, an authentication method is selected, and the authentication and privacy
passwords or keys are entered.
When MD5 authentication is specified, HMAC-MD5-96 is used to achieve authentication with a 16-octet
key, which generates an 128-bit authorization code. This authorization code is inserted in
msgAuthenticationParameters field of SNMPv3 PDUs when the security level is specified as either
AuthnoPriv or AuthPriv. Specifying SHA authentication uses the HMAC-SHA protocol with a 20-octet
key for authentication.
For privacy, a 16-octet key is provided as input to DES-CBS encryption protocol, which generates an
encrypted PDU to be transmitted. DES uses bytes 1-7 to make a 56 bit key. This key (encrypted itself) is
placed in msgPrivacyParameters of SNMPv3 PDUs when the security level is specified as AuthPriv.
SNMPv3 MIB Access Control

SNMPv3 provides a fine-grained mechanism for defining which parts of the MIB can be accessed. This
is referred to as the View-Based Access Control Model (VACM).
MIB views represent the basic building blocks of VACM. They are used to define a subset of the
information in the MIB. Access to read, to write, and to generate notifications is based on the
relationship between a MIB view and an access group. The users of the access group can then read,
write, or receive notifications from the part of the MIB defined in the MIB view as configured in the
access group.
A view name, a MIB subtree/mask, and an inclusion or exclusion define every MIB view. For example,
there is a System group defined under the MIB-2 tree. The Object Identifier (OID) for MIB-2 is 1.3.6.1.2,
and the System group is defined as MIB-2.1.1, or directly as 1.3.6.1.2.1.1.

To define a MIB view which includes only the System group, use the following subtree/mask
combination:
1.3.6.1.2.1.1/1.1.1.1.1.1.1.0
The mask can also be expressed in hex notation (this is used for the ExtremeXOS CLI):
1.3.6.1.2.1.1/fe
To define a view that includes the entire MIB-2, use the following subtree/mask:
1.3.6.1.2.1.1/1.1.1.1.1.0.0.0
which, in the CLI, is:
1.3.6.1.2.1.1/f8
When you create the MIB view, you can choose to include the MIB subtree/mask or to exclude the MIB
subtree/mask. To create a MIB view, use the following command:
configure snmpv3 add mib-view [[hex <hex_view_name>] | <view_name>] subtree

<object_identifier> {/<subtree_mask>} {type [included | excluded]} {volatile}
After the view has been created, you can repeatedly use the configure snmpv3 add mib-view
command to include and/or exclude MIB subtree/mask combinations to precisely define the items you
want to control access to.
In addition to the user-created MIB views, there are three default views. These default views are of
storage type permanent and cannot be deleted, but they can be modified. The default views are:
defaultUserView, defaultAdminView, and defaultNotifyView. To show MIB views, use the following
command:
show snmpv3 mib-view {[[hex <hex_view_name>] | <view_name>] {subtree

<object_identifier>}}
To delete a MIB view, use the following command:
configure snmpv3 delete mib-view [all-non-defaults | {[[hex <hex_view_name>] |

<view_name>] {subtree <object_identifier>}}]
MIB views that are used by security groups cannot be deleted.
SNMPv3 Notification
SNMPv3 can use either SNMPv1 traps or SNMPv2c notifications to send information from an agent to
the network manager. The terms trap and notification are used interchangeably in this context.
Notifications are messages sent from an agent to the network manager, typically in response to some
state change on the agent system. With SNMPv3, you can define precisely which traps you want sent,
to which receiver by defining filter profiles to use for the notification receivers.
To configure notifications, you configure a target address for the target that receives the notification, a
target parameters name, and a list of notification tags. The target parameters specify the security and
MP models to use for the notifications to the target. The target parameters name also points to the filter
profile used to filter the notifications. Finally, the notification tags are added to a notification table so
that any target addresses using that tag will receive notifications.

Managing the Switch
Target Addresses
A target address is similar to the earlier concept of a trap receiver. To configure a target address, use
configure snmpv3 add target-addr [[hex <hex_addr_name] | <addr_name>] param [[hex
<hex_param_name] | <param_name>] ipaddress [[<ip_address> {<netmask>}] | <ip_address>]
{transport-port <port_number> {from <src_ip_address>} {tag-list <tag_list>} {volatile}
In configuring the target address you supply an address name that identifies the target address, a
parameters name that indicates the MP model and security for the messages sent to that target address,
and the IP address and port for the receiver. The parameters name also is used to indicate the filter
profile used for notifications. The target parameters is discussed in “Target Parameters” next.
The from option sets the source IP address in the notification packets.
The tag-list option allows you to associate a list of tags with the target address. The tag defaultNotify
is set by default. Tags are discussed in the section “Notification Tags”.
To display target addresses, use the following command:
show snmpv3 target-addr {[[hex <hex_addr_name>] | <addr_name>]}
To delete a single target address or all target addresses, use the following command:
configure snmpv3 delete target-addr [{[[hex <hex_addr_name>] | <addr_name>]} | all]
Target Parameters
Target parameters specify the MP model, security model, security level, and user name (security name)
used for messages sent to the target address. See “Message Processing” on page 96 and “Users, Groups,
and Security” on page 96 for more details on these topics. In addition, the target parameter name used
for a target address points to a filter profile used to filter notifications. When you specify a filter profile,
you associate it with a parameter name, so you must create different target parameter names if you use
different filters for different target addresses.
To create a target parameter name and to set the message processing and security settings associated
with it, use the following command:
configure snmpv3 add target-params [[hex <hex_param_name>] | <param_name>] user [[hex

<hex_user_name>] | <user_name>] mp-model [snmpv1 | snmpv2c | snmpv3] sec-model [snmpv1
| snmpv2c | usm] {sec-level [noauth | authnopriv | priv]} {volatile}
To display the options associated with a target parameters name or all target parameters names, use the
following command:
show snmpv3 target-params {[[hex <hex_target_params>] | <target_params>]}
To delete one or all the target parameters, use the following command:
configure snmpv3 delete target-params [{[[hex <hex_param_name>] | <param_name>]} |

all]

Filter Profiles and Filters

A filter profile is a collection of filters that specifies which notifications should be sent to a target
address. A filter is defined by a MIB subtree and mask and by whether that subtree and mask is
included or excluded from notification.
When you create a filter profile, you are associating only a filter profile name with a target parameter
name. The filters that make up the profile are created and associated with the profile using a different
command.
To create a filter profile, use the following command:
configure snmpv3 add filter-profile [[hex <hex_profile_name>] | <profile_name>] param

[[hex <hex_param_name>]] | <param_name>] {volatile}
After the profile name has been created, you associate filters with it using the following command:
configure snmpv3 add filter [[hex <hex_profile_name>] | <profile_name>] subtree

<object_identifier> {/<subtree_mask>} type [included | excluded] {volatile}
The MIB subtree and mask are discussed in “SNMPv3 MIB Access Control” on page 98, as filters are
closely related to MIB views. You can add filters together, including and excluding different subtrees of
the MIB until your filter meets your needs.
To display the association between parameter names and filter profiles, use the following command:
show snmpv3 filter-profile {[[hex <hex_profile_name>] | <profile_name>]} {param [[hex

<hex_param_name>] | <param_name>]}
To display the filters that belong a filter profile, use the following command:
show snmpv3 filter {[[hex <hex_profile_name>] | <profile_name>] {{subtree}

<object_identifier>}
To delete a filter or all filters from a filter profile, use the following command:
configure snmpv3 delete filter [all | [[hex <hex_profile_name>] | <profile_name>]

{subtree <object_identifier>}]]
To remove the association of a filter profile or all filter profiles with a parameter name, use the
following command:
configure snmpv3 delete filter-profile [all |[[hex <hex_profile_name>] |

<profile_name>] {param [[hex <hex_param_name>] | <param_name>}]]
Notification Tags
When you create a target address, either you associate a list of notification tags with the target or by
default, the defaultNotify tag is associated with the target. When the system generates notifications, only
those targets associated with tags currently in the standard MIB table, called snmpNotifyTable, are
notified.
To add an entry to the table, use the following command:
configure snmpv3 add notify [[hex <hex_notify_name>] | <notify_name>] tag [[hex

<hex_tag>] | <tag>] {volatile}

Managing the Switch
Any targets associated with tags in the snmpNotifyTable are notified, based on the filter profile
associated with the target.
To display the notifications that are set, use the following command:
show snmpv3 notify {[[hex <hex_notify_name>] | <notify_name>]}
To delete an entry from the snmpNotifyTable, use the following command:
configure snmpv3 delete notify [{[[hex <hex_notify_name>] | <notify_name>]} | all-non-

defaults]
You cannot delete the default entry from the table, so any targets configured with the defaultNotify tag
will always receive notifications consistent with any filter profile specified.
Configuring Notifications
Because the target parameters name points to a number of objects used for notifications, configure the
target parameter name entry first. You can then configure the target address, filter profiles and filters,
and any necessary notification tags.
Using the Simple Network Time Protocol

ExtremeXOS supports the client portion of the Simple Network Time Protocol (SNTP) Version 3 based
on RFC1769. SNTP can be used by the switch to update and synchronize its internal clock from a
Network Time Protocol (NTP) server. After SNTP has been enabled, the switch sends out a periodic
query to the indicated NTP server, or the switch listens to broadcast NTP updates. In addition, the
switch supports the configured setting for Greenwich Mean time (GMT) offset and the use of Daylight
Saving Time.
Configuring and Using SNTP

To use SNTP:
1 Identify the host(s) that are configured as NTP server(s). Additionally, identify the preferred method
for obtaining NTP updates. The options are for the NTP server to send out broadcasts or for
switches using NTP to query the NTP server(s) directly. A combination of both methods is possible.
You must identify the method that should be used for the switch being configured.
2 Configure the Greenwich Mean Time (GMT) offset and Daylight Saving Time preference. The
command syntax to configure GMT offset and usage of Daylight Saving Time is as follows:
configure timezone {name <tz_name>} <GMT_offset>
{autodst {name <dst_timezone_ID>} {<dst_offset>}
{begins [every <floatingday> | on <absoluteday>] {at <time_of_day_hour>
<time_of_day_minutes>}
{ends [every <floatingday> | on <absoluteday>] {at <time_of_day_hour>
<time_of_day_minutes>}}}
By default, Daylight Saving Time is assumed to begin on the first Sunday in April at 2:00 AM, and
end the last Sunday in October at 2:00 AM and to be offset from standard time by one hour. If this is

the case in your time zone, you can set up automatic daylight savings adjustment with the
command:
configure timezone <GMT_offset> autodst
If your time zone uses starting and ending dates and times that differ from the default, you can
specify the starting and ending date and time in terms of a floating day, as follows:
configure timezone name MET 60 autodst name MDT begins every last sunday march at
1 30 ends every last sunday october at 1 30
You can also specify a specific date and time, as shown in the following command:
configure timezone name NZST 720 autodst name NZDT 60 begins every first sunday
october at 2 00 ends on 3 16 2004 at 2 00
The optional time zone IDs are used to identify the time zone in display commands such as show
switch {detail}.
Table 13 describes the command options in detail.
Table 13: Time zone configuration command options
tz_name Specifies an optional name for this timezone specification. May be up to six characters in
length. The default is an empty string.
GMT_offset Specifies a Greenwich Mean Time (GMT) offset, in + or - minutes.
autodst Enables automatic Daylight Savings Time.
dst_timezone_ID Specifies an optional name for this Daylight Savings Time specification. May be up to six
characters in length. The default is an empty string.
dst_offset Specifies an offset from standard time, in minutes. Value is in the range of 1 to 60.
Default is 60 minutes.
floatingday Specifies the day, week, and month of the year to begin or end Daylight Savings Time
each year. Format is:
<week> <day> <month> where:
• <week> is specified as [first | second | third | fourth | last]
• <day> is specified as [sunday | monday | tuesday | wednesday | thursday | friday |
saturday]
• <month> is specified as [january | february | march | april | may | june | july | august |
september | october | november | december]
Default for beginning is first sunday april; default for ending is last sunday october.
absoluteday Specifies a specific day of a specific year on which to begin or end DST. Format is:
<month> <day> <year> where:
• <month> is specified as 1-12
• <day> is specified as 1-31
• <year> is specified as 1970 - 2035
The year must be the same for the begin and end dates.
time_of_day_hour Specifies the time of day to begin or end Daylight Savings Time. May be specified as an
hour (0-23). Default is 2.
time_of_day_minut Specify the minute to begin or end Daylight Savings Time. May be specified as a minute
es (0-59).
noautodst Disables automatic Daylight Savings Time.
Automatic Daylight Savings Time changes can be enabled or disabled. The default setting is enabled.
To disable automatic Daylight Savings Time, use the command:
configure timezone {name <tz_name>} <GMT_offset> noautodst

Managing the Switch
3 Enable the SNTP client using the following command:

enable sntp-client
After SNTP has been enabled, the switch sends out a periodic query to the NTP servers defined in
step 4 (if configured) or listens to broadcast NTP updates from the network. The network time
information is automatically saved into the onboard real-time clock.
4 If you would like this switch to use a directed query to the NTP server, configure the switch to use
the NTP server(s). If the switch listens to NTP broadcasts, skip this step. To configure the switch to
use a directed query, use the following command:
configure sntp-client [primary | secondary] <host-name-or-ip> {vr <vr_name>}
NTP queries are first sent to the primary server. If the primary server does not respond within 1
second, or if it is not synchronized, the switch queries the secondary server (if one is configured). If
the switch cannot obtain the time, it restarts the query process. Otherwise, the switch waits for the
sntp-client update interval before querying again.
5 Optionally, the interval for which the SNTP client updates the real-time clock of the switch can be
changed using the following command:
configure sntp-client update-interval <update-interval>
The default sntp-client update-interval value is 64 seconds.
6 You can verify the configuration using the following commands:
■ show sntp-client
This command provides configuration and statistics associated with SNTP and its connectivity to
the NTP server.
■ show switch {detail}
This command indicates the GMT offset, the Daylight Savings Time configuration and status, and
the current local time.
NTP updates are distributed using GMT time. To properly display the local time in logs and other time-
stamp information, the switch should be configured with the appropriate offset to GMT based on
geographical location. Table 14 lists GMT offsets.
Table 14: Greenwich Mean Time offsets

GMT
Offset in GMT Offset
Hours in Minutes Common Time Zone References Cities
+0:00 +0 GMT - Greenwich Mean London, England; Dublin, Ireland;
Edinburgh, Scotland; Lisbon, Portugal;
UT or UTC - Universal (Coordinated)
Reykjavik, Iceland; Casablanca, Morocco
WET - Western European
-1:00 -60 WAT - West Africa Cape Verde Islands
-2:00 -120 AT - Azores Azores
-3:00 -180 Brasilia, Brazil; Buenos Aires, Argentina;
Georgetown, Guyana
-4:00 -240 AST - Atlantic Standard Caracas; La Paz
-5:00 -300 EST - Eastern Standard Bogota, Columbia; Lima, Peru; New York,
NY, Trevor City, MI USA
-6:00 -360 CST - Central Standard Mexico City, Mexico
-7:00 -420 MST - Mountain Standard Saskatchewan, Canada
-8:00 -480 PST - Pacific Standard Los Angeles, CA, Santa Clara, CA,
Seattle, WA USA

Table 14: Greenwich Mean Time offsets (Continued)

GMT
Offset in GMT Offset
Hours in Minutes Common Time Zone References Cities
-9:00 -540 YST - Yukon Standard
-10:00 -600 AHST - Alaska-Hawaii Standard
CAT - Central Alaska
HST - Hawaii Standard
-11:00 -660 NT - Nome
-12:00 -720 IDLW - International Date Line West
+1:00 +60 CET - Central European Paris France; Berlin, Germany;
Amsterdam, The Netherlands; Brussels,
FWT - French Winter
Belgium; Vienna, Austria; Madrid, Spain;
MET - Middle European Rome, Italy; Bern, Switzerland;
Stockholm, Sweden; Oslo, Norway
MEWT - Middle European Winter
SWT - Swedish Winter
+ 2:00 +120 EET - Eastern European, Russia Zone 1 Athens, Greece; Helsinki, Finland;
Istanbul, Turkey; Jerusalem, Israel;
Harare, Zimbabwe
+3:00 +180 BT - Baghdad, Russia Zone 2 Kuwait; Nairobi, Kenya; Riyadh, Saudi
Arabia; Moscow, Russia; Tehran, Iran
+4:00 +240 ZP4 - Russia Zone 3 Abu Dhabi, UAE; Muscat; Tblisi;
Volgograd; Kabul
+5:00 +300 ZP5 - Russia Zone 4
+5:30 +330 IST - India Standard Time New Delhi, Pune, Allahabad, India
+6:00 +360 ZP6 - Russia Zone 5
+7:00 +420 WAST - West Australian Standard
+8:00 +480 CCT - China Coast, Russia Zone 7
+9:00 +540 JST - Japan Standard, Russia Zone 8
+10:00 +600 EAST - East Australian Standard
GST - Guam Standard
Russia Zone 9
+11:00 +660
+12:00 +720 IDLE - International Date Line East Wellington, New Zealand; Fiji, Marshall
Islands
NZST - New Zealand Standard
NZT - New Zealand
SNTP Example
In this example, the switch queries a specific NTP server and a backup NTP server. The switch is
located in Cupertino, California, and an update occurs every 20 minutes. The commands to configure
the switch are as follows:
configure timezone -480 autodst
configure sntp-client update-interval 1200
enable sntp-client
configure sntp-client primary 10.0.1.1
configure sntp-client secondary 10.0.1.2

Managing the Switch

4 Managing the ExtremeXOS Software

● Overview of the ExtremeXOS Software on page 107
● Using the ExtremeXOS File System on page 108
● Managing the Configuration File on page 116
● Managing ExtremeXOS Processes on page 117
● Understanding Memory Protection on page 120
● Monitoring CPU Utilization on page 121
Overview of the ExtremeXOS Software

The ExtremeXOS software platform is a distributed software architecture. The distributed architecture
consists of separate binary images organized into discreet software modules with messaging between
them. The software and system infrastructure subsystem form the basic framework of how the
ExtremeXOS applications interact with each other, including the system startup sequence, memory
allocation, and error events handling. Redundancy and data replication is a built-in mechanism of
ExtremeXOS. The system infrastructure provides basic redundancy support and libraries for all of the
ExtremeXOS applications.
Understanding the ExtremeXOS Software
NOTE
For information about downloading and upgrading a new software image, saving configuration changes, and
upgrading the BootROM, see Appendix A, “Software Upgrade and Boot Options.”
Like any advanced operating system, ExtremeXOS gives you the tools to manage your switch and
create your network configurations. With the introduction of ExtremeXOS, the following enhancements
and functionality have been added to the switch operating system:
● File system administration
● Configuration file management
● Process control
● Memory protection
● CPU monitoring
File system administration—With the enhanced file system, you can move, copy, and delete files from
the switch. The file system structure allows you to keep, save, rename, and maintain multiple copies of
configuration files on the switch. In addition, you can manage other entities of the switch such as
policies and access control lists (ACLs).

Managing the ExtremeXOS Software
Configuration file management—With the enhanced configuration file management, you can oversee
and manage multiple configuration files on your switch. In addition, you can upload, download,
modify, and name configuration files used by the switch.
Process control—With process control, you can stop and start processes, restart failed processes, and
update the software for a specific process or set of processes.
Memory protection—With memory protection, each function can be bundled into a single application
module running as a memory protected process under real-time scheduling. In essence, ExtremeXOS
protects each process from every other process in the system. If one process experiences a memory
fault, that process cannot affect the memory space of another process.
CPU monitoring—With CPU monitoring, you can monitor CPU utilization for Management Switch
Fabric Modules (MSMs) or Summit family switches whether or not the switches are included in a
SummitStack, and the individual processes running on the switch. Monitoring the workload of the CPU
allows you to troubleshoot and identify suspect processes.
The following sections describe in more detail how to manage the ExtremeXOS software.
Using the ExtremeXOS File System

The file system in ExtremeXOS is the structure by which files are organized, stored, and named. The
switch can store multiple user-defined configuration and policy files, each with its own name.
Using a series of commands, you can manage the files on your system. For example, you can rename or
copy a configuration file on the switch, display a comprehensive list of the configuration and policy
files on the switch, or delete a policy file from the switch.
NOTE
Filenames are case-sensitive. For information on filename restrictions, refer to the specific command in the
You can also download configuration and policy files from the switch to a network Trivial File Transfer
Protocol (TFTP) server using TFTP. For detailed information about downloading switch configurations,
see Appendix A, “Software Upgrade and Boot Options.” For detailed information about downloading
policies and ACLs, see Chapter 11, “Policy Manager.”
With guidance from Extreme Networks Technical Support personnel, you can configure the switch to
capture core dump files, which contain debugging information that is useful in troubleshooting
situations. For more information about configuring core dump files and managing the core dump files
stored on your switch, see Appendix B, “Troubleshooting.”
This section describes the following file management topics:

● Moving or Renaming Files on the Switch on page 109
● Copying Files on the Switch on page 110
● Displaying Files on the Switch on page 111
● Transferring Files to and from the Switch on page 113
● Deleting Files from the Switch on page 115

Moving or Renaming Files on the Switch

To move or rename an existing configuration, policy, or if configured, core dump file in the system, use
mv [internal-memory <old-name-internal> internal-memory <new-name-internal> |
internal-memory <old-name-internal> memorycard <new-name-memorycard> | memorycard
<old-name-memorycard> memorycard <new-name-memorycard> | memorycard <new-name-
memorycard> <new-name> | <old-name> memorycard <new-name-memorycard> | <old-name>
<new-name>]
Where the following is true:

● internal-memory—Specifies the internal memory card. Specify internal-memory if you configured
core dumps and are sending debug files to the internal memory.
● old-name-internal—Specifies the current name of the core dump file located on the internal
memory card.
● new-name-internal—Specifies the new name of the core dump file located on the internal memory
card.
● memorycard—Specifies the removable external compact flash memory card. (This parameter is
available only on modular switches.)
● old-name-memorycard—Specifies the current name of the file located on the external compact flash
memory card. Depending on your switch configuration, you can have configuration, policy, or core
dump files stored in this card. (This parameter is available only on modular switches.)
● new-name-memorycard—Specifies the new name of the file located on the external compact flash
memory card. (This parameter is available only on modular switches.)
● old-name—Specifies the current name of the configuration or policy file.
● new-name—Specifies the new name of the configuration or policy file.
XML-formatted configuration files have a .cfg file extension. The switch only runs .cfg files. ASCII-
formatted configuration files have an .xsf file extension. See “ASCII-Formatted Configuration Files” on
page 1016 for more information. Policy files have a .pol file extension.
When you rename a file, make sure the renamed file uses the same file extension as the original file. If
you change the file extensions, the file may be unrecognized by the system. For example, if you have an
existing configuration file named test.cfg, the new filename must include the .cfg file extension.
When you rename a file on the switch, a message similar to the following appears:
Rename config test.cfg to config megtest.cfg on switch? (y/n)
Enter y to rename the file on your system. Enter n to cancel this process and keep the existing filename.
If you attempt to rename an active configuration file (the configuration currently selected the boot the
switch), the switch displays an error similar to the following:
Error: Cannot rename current selected active configuration.
For more information about configuring core dump files and managing the core dump files stored on
your switch, see Appendix B, “Troubleshooting.”

Modular Switches and SummitStack Only

This command also replicates the action from the primary node to the backup node. For example, if you
rename a file on the primary node, the same file on the backup node is renamed.
For the memorycard option, this command can move files between the external memory card and the
switch. If you use the memorycard option for both the old-name and the new-name, this command only
renames a file on the external memory card.
Examples
The following example renames the configuration file named Test.cfg to Final.cfg:
mv Test.cfg Final.cfg
On a modular switch, the following command moves the configuration file named test1.cfg from the
switch to the external memory card:
mv test1.cfg memorycard test1.cfg
Copying Files on the Switch

The copy function allows you to make a copy of an existing file before you alter or edit the file. By
making a copy, you can easily go back to the original file if needed.
To copy an existing configuration or policy file on your switch, use the following command:
cp [internal-memory <old-name-internal> internal-memory <new-name-internal> |

<old-name-memorycard> memorycard <new-name-memorycard> | memorycard <old-name-
<new-name>]

● internal-memory—Specifies the internal memory card. Specify internal-memory if you configured
core dumps and are sending debug files to the internal memory.
● old-name-internal—Specifies the name of the core dump file located on the internal memory card
that you want to copy.
● new-name-internal—Specifies the name of the newly copied core dump file located on the internal
memory card.
● old-name-memorycard—Specifies the name of the file located on the external compact flash memory
card that you want to copy. Depending on your switch configuration, you can have configuration,
policy, or core dump files stored in this card. (This parameter is available only on modular switches.)
● new-name-memorycard—Specifies the name of the newly copied file located on the external compact
flash memory card. (This parameter is available only on modular switches.)
● old-name—Specifies the name of the configuration or policy file that you want to copy.
● new-name—Specifies the name of the copied configuration or policy file.

XML-formatted configuration files have a .cfg file extension. The switch only runs .cfg files. ASCII-
formatted configuration files have an .xsf file extension. See “ASCII-Formatted Configuration Files” on
page 1016 for more information. Policy files have a .pol file extension.
When you copy a configuration or policy file from the system, make sure you specify the appropriate
file extension. For example, if you want to copy a policy file, specify the filename and .pol.
When you copy a file on a the switch, a message similar to the following appears:
Copy config test.cfg to config test1.cfg on switch? (y/n)
Enter y to copy the file. Enter n to cancel this process and not copy the file.
If you enter y, the switch copies the file with the new name and keeps a backup of the original file with
the original name. After the switch copies the file, use the ls command to display a complete list of
files.

copy a file on the primary node, the same file is copied to the backup node.
For the memorycard option, the source and/or destination is the memorycard. You must mount the
memory card for this operation to succeed. This command copies a file from the switch to the external
memory card or a file already on the card. If you copy a file from the switch to the external memory
card, and the new filename is identical to the source file, you do not need to re-enter the filename.
Example
The following example copies an existing configuration file named test.cfg and names the copied
configuration file test_rev2.cfg:
cp test.cfg test_rev2.cfg
On a modular switch, the following command makes a copy of a configuration file named primary.cfg
from the switch to the external memory card with the same name, primary.cfg:
cp primary.cfg memorycard
Displaying Files on the Switch

To display a list of the configuration, policy, or if configured, core dump files stored on your switch,
ls {internal-memory | memorycard}


● internal-memory—Lists the core dump files that are present and saved in the internal memory
card.
If the switch is not configured to save debug files or has not saved any debug files, no files are
displayed. For more information about configuring core dump files and managing the core dump
files stored on your switch, see Appendix B, “Troubleshooting.”
● memorycard—Lists all files that are stored in the external compact flash memory card. (This
parameter is available only on modular switches.)
If you do not specify a parameter, this command lists all of the files stored on your switch.
Output from this command includes the file size, date and time the file was last modified, and the file
name.
Example
The following command displays all of the configuration and policy files stored on your switch:
ls
The following is sample output from this command:
total 424
-rw-r--r-- 1 root root 50 Jul 30 14:19 hugh.pol
-rw-r--r-- 1 root root 94256 Jul 23 14:26 hughtest.cfg
-rw-r--r-- 1 root root 100980 Sep 23 09:16 megtest.cfg
-rw-r--r-- 1 root root 35 Jun 29 06:42 newpolicy.pol
-rw-r--r-- 1 root root 100980 Sep 23 09:17 primary.cfg
-rw-r--r-- 1 root root 94256 Jun 30 17:10 roytest.cfg
On a modular switch, the following command displays all of the configuration and policy files stored
on the external memory card:
ls memorycard
-rwxr-xr-x 1 root 0 15401865 Mar 30 00:03 bd10K-11.2.0.13.xos

-rwxr-xr-x 1 root 0 10 Mar 31 09:41 test-1.pol
-rwxr-xr-x 1 root 0 10 Apr 4 09:15 test.pol
-rwxr-xr-x 1 root 0 10 Mar 31 09:41 test_1.pol
-rwxr-xr-x 1 root 0 223599 Mar 31 10:02 v11_1_3.cfg

Transferring Files to and from the Switch

TFTP allows you to transfer files to and from the switch, internal memory card, and on a modular
switch, the external memory card. This section describes the commands used to transfer files to and
from the switch.
To transfer a configuration or policy file from a TFTP server, internal memory card, or external memory
card to the switch, use the tftp and tftp get commands:
● tftp [<host-name> | <ip-address>] {-v <vr_name>} [-g | -p] [{-l [internal-memory
<local-file-internal> | memorycard <local-file-memcard> | <local-file>} {-r
<remote-file>} | {-r <remote-file>} {-l [internal-memory <local-file-internal> |
memorycard <local-file-memcard> | <local-file>]}]

● host-name—Specifies the name of the remote host on the network.
● ip-address—Specifies the IP address of the TFTP server on the network.
● vr_name—Specifies the name of the virtual router.
NOTE
created VRs.
● -g—Gets the specified file from the TFTP server and copies it to the local host. (This parameter is
available only on the tftp command.)
● get—Gets the specified file from the TFTP server and copies it to the local host. (This is part of the
tftp get command.)
● internal-memory—Specifies the internal memory card.
● local-file-internal—Specifies the name of the core dump file located on the internal memory
card.
● local-file-memcard—Specifies the name of the file on the external compact flash memory card.
(This parameter is available only on modular switches.)
● local-file—Specifies the name of the file (configuration file, policy file) on the local host.
● remote-file—Specifies the name of the file on the remote host.
● force-overwrite—Specifies the switch to automatically overwrite an existing file. (This parameter
is available only on the tftp get command.)
NOTE
By default, if you transfer a file with a name that already exists on the system, the switch prompts you to
overwrite the existing file. For more information, see the tftp get command in the ExtremeXOS Command
Reference Guide.

To transfer a configuration or policy file from the switch to a TFTP server, internal memory card, or
external memory card, use the tftp and tftp put commands:

● host-name—Specifies the name of the remote host on the network.
● ip-address—Specifies the IP address of the TFTP server on the network.
NOTE
created VRs.
● -p—Puts the specified file from the local host and copies it to the TFTP server. (This parameter is
● put—Puts the specified file from the local host and copies it to the TFTP server. (This is part of the
tftp put command.)
card.
● local-file-memcard—Specifies the name of the file on the external compact flash memory card.
(This parameter is available only on modular switches.)
For more information about TFTP, see Chapter 4, “Managing the Switch.” For detailed information
about downloading software image files, BootROM files, and switch configurations, see Appendix
A, “Software Upgrade and Boot Options.” For more information about configuring core dump files and
managing the core dump files stored on your switch, see Appendix B, “Troubleshooting.”
Modular Switches Only

For the memorycard option, this command transfers an existing file to or from the external compact
flash memory card.

Example
The following example uses the tftp command to download the configuration file named XOS1.cfg
from the TFTP server:
tftp 10.123.45.67 -g -r XOS1.cfg
The following example uses the tftp get command to download the configuration file from the TFTP
server:
tftp get 10.123.45.67 XOS1.cfg
The following example uses the tftp put command to upload the configuration file from the switch to
the TFTP server:
tftp put 10.123.45.67 XOS1.cfg
NOTE
On a modular switch, you can transfer files to and from the switch and an installed external compact flash memory
card.
Deleting Files from the Switch

To delete a configuration, policy, or if configured, core dump file from your system, use the following
command:
rm {internal-memory | memorycard} <file-name>

● file-name—Specifies the name of the configuration or policy file to delete.
When you delete a configuration or policy file from the system, make sure you specify the appropriate
file extension. For example, if you want to delete a policy file, specify the filename and .pol. After you
delete a file, it is unavailable to the system.
When you delete a file from the switch, a message similar to the following appears:
Remove testpolicy.pol from switch? (y/n)
Enter y to remove the file from your system. Enter n to cancel the process and keep the file on your
system.
If you attempt to delete an active configuration file (the configuration currently selected the boot the
switch), the switch displays an error similar to the following:
Error: Cannot remove current selected active configuration.


delete a file on the primary node, the same file on the backup node is deleted.
Modular Switches only. For the memorycard option, this command removes/deletes an existing file on
the external memory card.
Example
The following example removes the policy file named newpolicy.pol from the system:
rm newpolicy.pol
On a modular switch with an external memory card installed, the following command removes the
policy file named test.pol from the external memory card:
rm memorycard test.pol
Managing the Configuration File

The configuration is the customized set of parameters that you have selected to run on the switch.
Table 15 describes some of the key areas of configuration file management in ExtremeXOS.
Table 15: Configuration file management

Task Behavior
Configuration file database ExtremeXOS supports saving a configuration file into any named file and
supports more than two saved configurations.
For example, you can download a configuration file from a network TFTP
server and save that file as primary, secondary, or with a user-defined name.
You also select where to save the configuration: primary or secondary
partition, or another space.
The file names primary and secondary exist for backward compatibility with
ExtremeWare.
Downloading configuration files ExtremeXOS uses the tftp and tftp get commands to download
configuration files from the network TFTP server to the switch.
For more information about downloading configuration files, see “Using TFTP
to Download the Configuration” on page 1019.
Uploading configuration files ExtremeXOS uses the tftp and tftp put commands to upload
configuration files from the switch to the network TFTP server.
For more information about uploading configuration files, see “Using TFTP to
Upload the Configuration” on page 1018.
Managing configuration files, The following commands allow you to manage configuration files:
including listing, copying,
• ls—Lists all of the configuration files in the system
deleting, and renaming
• cp—Makes a copy of an existing configuration file in the system
• rm—Removes/deletes an existing configuration file from the system
• mv—Renames an existing configuration file
Configuration file type ExtremeXOS configuration files are saved in Extensible Markup Language
(XML) format. Use the show configuration command to view on the CLI
your currently running switch configuration.

Managing ExtremeXOS Processes
Table 15: Configuration file management (Continued)

Task Behavior
ASCII-formatted configuration Beginning with ExtremeXOS 11.4, you can upload your current configuration
file in ASCII format to a network TFTP server. The uploaded ASCII file retains the
CLI format.
To view your configuration in ASCII format, save the configuration with the
.xsf file extension (known as the XOS CLI script file). This saves the XML-
based configuration in an ASCII format readable by a text editor.
ExtremeXOS uses the upload configuration command to upload the
ASCII-formatted configuration file from the switch to the network TFTP server.
ExtremeXOS uses the tftp and tftp get commands to download
configuration files from the network TFTP server to the switch.
For more information about ASCII-formatted configuration files, see “ASCII-
Formatted Configuration Files” on page 1016.
XML configuration mode Indicated by (xml) at the front of the switch prompt. Do not use. Use the
command disable xml-mode to disable this mode.
Displaying configuration files You can also see a complete list of configuration files by entering the ls
command followed by the Tab key.
For more information about saving, uploading, and downloading configuration files, see “Saving
Configuration Changes” on page 1014.

ExtremeXOS consists of a number of cooperating processes running on the switch. With process control,
under certain conditions, you can stop and start processes, restart failed processes, examine information
about the processes, and update the software for a specific process or set of processes.

● Displaying Process Information on page 117
● Stopping a Process on page 118
● Starting a Process on page 119
Displaying Process Information

To display information about the processes in the system, use the following command:
show process {<name>} {detail} {description} {slot <slotid>}

● name—Specifies the name of the process.
● detail—Specifies more detailed process information, including memory usage statistics, process ID
information, and process statistics.

● description—Describes the name of all of the processes or the specified process running on the
switch.
● slotid—On a Modular chassis, specifies the slot number of the MSM. A specifies the MSM installed
in slot A. B specifies the MSM installed in slot B. On a SummitStack, specifies the target node's slot
number. The number is a value from 1 to 8. (This parameter is available only on modular switches
and SummitStack.)
The show process and show process slot <slotid> commands display the following information
in a tabular format:
● Card—The name of the module where the process is running (modular switches only).
● Process Name—The name of the process.
● Version—The version number of the process. Options are:
■ Version number—A series of numbers that identify the version number of the process. This is
helpful to ensure that you have version-compatible processes and if you experience a problem.
■ Not Started—The process has not been started. This can be caused by not having the appropriate
license or for not starting the process.
● Restart—The number of times the process has been restarted. This number increments by one each
time a process stops and restarts.
● State—The current state of the process. Options are:
■ No License—The process requires a license level that you do not have. For example, you have not
upgraded to that license, or the license is not available for your platform.
■ Ready—The process is running.
■ Stopped—The process has been stopped.
● Start Time—The current start time of the process. Options are:
■ Day/Month/Date/Time/Year—The date and time the process began. If a process terminates and
restarts, the start time is also updated.
■ Not Started—The process has not been started. This can be caused by not having the appropriate
license or for not starting the process.
If you specify the detail keyword, more specific and detailed process information is displayed. The
show process detail and show process slot <slotid> detail commands display the following
information in a multi-tabular format:
● Detailed process information
● Memory usage configurations
● Recovery policies
● Process statistics
● Resource usage
Stopping a Process
If recommended by Extreme Networks Technical Support personnel, you can stop a running process.
To stop a running process, use the following command:
terminate process <name> [forceful | graceful] {msm <slot>}
In a SummitStack:
terminate process <name> [forceful | graceful] {slot <slot>}


● forceful—Specifies that the software quickly terminate a process. Unlike the graceful option, the
process is immediately shutdown without any of the normal process cleanup.
● graceful—Specifies that the process shutdown gracefully by closing all opened connections,
notifying peers on the network, and other types of process cleanup.
● slot—For a modular chassis, specifies the slot number of the MSM. A specifies the MSM installed in
slot A. B specifies the MSM installed in slot B. On a SummitStack, specifies the target node's slot
number. The number is a value from 1 to 8. (This parameter is available only on modular switches
and SummitStack.)
NOTE
Do not terminate a process that was installed since the last reboot unless you have saved your configuration. If you
have installed a software module and you terminate the newly installed process without saving your configuration,
your module may not be loaded when you attempt to restart the process with the start process command.
To preserve a process’s configuration during a terminate and (re)start cycle, save your switch configuration before
terminating the process. Do not save the configuration during the process terminate and re(start) cycle. If you save
the configuration after terminating a process, and before the process (re)starts, the configuration for that process is
lost.
You can also use a single command to stop and restart a running process during a software upgrade on
the switch. By using the single command, there is less process disruption and it takes less time to stop
and restart the process. To stop and restart a process during a software upgrade, use the following
command:
restart process [class <cname> | <name> {msm <slot>}]

● cname—Specifies that the software terminates and restarts all instances of the process associated with
a specific routing protocol on all VRs.
Starting a Process
To start a process, use the following command:
start process <name> {msm <slot>}
In a SummitStack:
start process <name> {slot <slot>}

slot A. B specifies the MSM installed in slot B. On a SummitStack, specifies the slot number of the
target node. The number is a value from 1 to 8. (This parameter is available only on modular
switches and SummitStack.)

You are unable to start a process that is already running. If you try to start a currently running process,
for example telnetd, an error message similar to the following appears:
Error: Process telnetd already exists!
As described in the section, “Stopping a Process” on page 118, you can use a single command, rather
than multiple commands, to stop and restart a running process. To stop and restart a process during a
software upgrade, use the following command:
In a SummitStack:
restart process [class <cname> | <name> {slot <slot>}]
For more detailed information, see the previous section or the ExtremeXOS Command Reference
Guide.omm
Understanding Memory Protection

ExtremeXOS provides memory management capabilities. With ExtremeXOS, each process runs in a
protected memory space. This infrastructure prevents one process from overwriting or corrupting the
memory space of another process. For example, if one process experiences a loop condition, is under
some type of attack, or is experiencing some type of problem, that process cannot take over or overwrite
another processes’ memory space.
Memory protection increases the robustness of the system. By isolating and having separate memory
space for each individual process, you can more easily identify the process or processes that experience
a problem.
To display the current system memory and that of the specified process, use the following command:
show memory process <name> {slot <slotid>}

● slot—On a modular chassis, specifies the slot number of the MSM. A specifies the MSM installed in
The show memory process command displays the following information in a tabular format:
● System memory information (both total and free)
● Current memory used by the individual processes
The current memory statistics for the individual process also includes the following:
● The module (MSM A or MSM B) and the slot number of the MSM (modular switches only)
● The name of the process
You can also use the show memory {slot [slotid | a | b]} command to view the system memory
and the memory used by the individual processes, even for all processes on all MSMs installed in
modular switches. The slot parameter is available only on modular switches and SummitStack.

Monitoring CPU Utilization
In general, the free memory count for an MSM or Summit family switch decreases when one or more
running processes experiences an increase in memory usage. If you have not made any system
configuration changes, and you observe a continued decrease in free memory, this might indicate a
memory leak.
The information from these commands may be useful for your technical support representative if you
experience a problem.

You can monitor the CPU utilization and history for all of the processes running on the switch. By
viewing this history on a regular basis, you can see trends emerging and identify processes with peak
utilization. Monitoring the workload of the CPU allows you to troubleshoot and identify suspect
processes before they become a problem. By default, the switch monitors CPU utilization every 20
seconds. In addition, if CPU utilization of a process exceeds 60% of the regular operating basis, the
switch logs an error message specifying the process name and the current CPU utilization for the
process.
Disabling CPU Monitoring

To disable CPU monitoring, use the following command:
disable cpu-monitoring
This command disables CPU monitoring on the switch; however, it does not clear the monitoring
interval. Therefore, if you altered the monitoring interval, this command does not return the monitoring
interval to 20 seconds. The next time you enable CPU monitoring, the switch uses the existing
configured interval.
Enabling CPU Monitoring

To enable CPU monitoring, use the following command:
enable cpu-monitoring {interval <seconds>} {threshold <percent>}

● seconds—Specifies the monitoring interval. The default interval is 20 seconds, and the range is 5 to
60 seconds. Extreme Networks recommends the default setting for most network environments. If
you enter a number lower than 20 seconds, CPU utilization may increase.
● threshold—Specifies the CPU threshold value. CPU usage is measured in percentages. The default
is 60%, and the range is 0% to 100%.
By default, CPU monitoring is enabled and occurs every 20 seconds. The default CPU threshold value is
60%.

Displaying CPU Utilization History

To display the CPU utilization history of one or more processes, use the following command:
show cpu-monitoring {process <name>} {slot <slotid>}

Output from this command includes the following information:

● Card—The location (MSM A or MSM B) where the process is running on a modular switch.
● Process—The name of the process.
● Range of time (5 seconds, 10 seconds, and so forth)—The CPU utilization history of the process or
the system. The CPU utilization history goes back only 1 hour.
● Total User/System CPU Usage—The amount of time recorded in seconds that the process spends
occupying CPU resources. The values are cumulative meaning that the values are displayed as long
as the system is running. You can use this information for debugging purposes to see where the
process spends the most amount of time: physical memory or virtual memory.
The following is sample truncated output from a modular switch:
show cpu-monitoring
CPU Utilization Statistics - Monitored every 5 seconds

-------------------------------------------------------------------------------
Card Process 5 10 30 1 5 30 1 Max Total

secs secs secs min mins mins hour User/System
util util util util util util util util CPU Usage
(%) (%) (%) (%) (%) (%) (%) (%) (secs)
-------------------------------------------------------------------------------
MSM-A System 0.0 0.0 0.1 0.0 0.0 0.0 0.0 0.9
MSM-B System 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
MSM-A GNSS_cpuif 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
MSM-A GNSS_ctrlif 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
MSM-A GNSS_esmi 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
MSM-A GNSS_fabric 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
MSM-A GNSS_mac_10g 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
MSM-A GNSS_pbusmux 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
MSM-A GNSS_pktengine 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
MSM-A GNSS_pktif 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
MSM-A GNSS_switch 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
MSM-A aaa 0.0 0.0 0.0 0.0 0.0 0.0 0.0 8.4 0.82 0.56
MSM-A acl 0.0 0.0 0.0 0.0 0.0 0.0 0.0 7.5 0.37 0.33
MSM-A bgp 0.0 0.0 0.0 0.0 0.0 0.0 0.0 5.2 0.27 0.42
MSM-A cfgmgr 0.0 0.9 0.3 3.7 1.2 1.2 1.3 27.3 7.70 7.84
MSM-A cli 0.0 0.0 0.0 48.3 9.6 2.5 2.1 48.3 0.51 0.37
MSM-A devmgr 0.0 0.0 0.0 0.9 0.3 0.2 0.2 17.1 2.22 2.50
MSM-A dirser 0.0 0.0 0.0 0.0 0.0 0.0 0.0 9.5 0.0 0.0

MSM-A dosprotect 0.0 0.0 0.0 0.0 0.0 0.0 0.0 3.8 0.20 0.26
MSM-A eaps 1.9 0.9 0.4 0.0 0.0 0.0 0.0 8.4 2.40 1.40
MSM-A edp 0.0 0.0 0.0 0.0 0.0 0.0 0.0 10.2 0.99 0.47
MSM-A elrp 0.0 0.0 0.0 0.0 0.0 0.0 0.0 8.4 0.44 0.28
MSM-A ems 0.0 0.0 0.0 0.0 0.0 0.0 0.0 12.2 1.1 1.16
MSM-A epm 0.0 0.0 0.0 0.9 0.1 0.2 0.2 4.7 2.6 4.18
MSM-A esrp 0.0 0.0 0.0 0.0 0.0 0.0 0.0 7.5 0.44 0.36
MSM-A etmon 0.9 0.4 0.6 1.2 1.1 1.0 1.0 23.3 21.84 7.24
...
The following is sample truncated output from a Summit family switch:
CPU Utilization Statistics - Monitored every 25 seconds

-----------------------------------------------------------------------
Process 5 10 30 1 5 30 1 Max Total

secs secs secs min mins mins hour User/System
util util util util util util util util CPU Usage
(%) (%) (%) (%) (%) (%) (%) (%) (secs)
-----------------------------------------------------------------------
System n/a n/a 0.0 0.9 0.1 0.2 0.5 34.6

aaa n/a n/a 0.0 0.0 0.0 0.0 0.0 1.8 1.72 0.78
acl n/a n/a 0.0 0.0 0.0 0.0 0.0 0.0 0.40 0.24
bgp n/a n/a 0.0 0.0 0.0 0.0 0.0 12.6 11.18 2.21
cfgmgr n/a n/a 0.0 0.0 0.0 0.0 0.8 39.8 4743.92 3575.79
cli n/a n/a 0.0 0.0 0.0 0.0 0.0 0.0 0.59 0.42
devmgr n/a n/a 0.0 0.0 0.0 0.0 0.0 19.5 74.44 24.52
dirser n/a n/a 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
dosprotect n/a n/a 0.0 0.0 0.0 0.0 0.0 0.0 0.8 0.12
eaps n/a n/a 0.0 0.0 0.0 0.0 0.1 5.5 36.40 15.41
edp n/a n/a 0.0 0.0 0.0 0.0 0.0 11.1 10.92 3.97
elrp n/a n/a 0.0 0.0 0.0 0.0 0.0 0.0 0.49 0.44
ems n/a n/a 0.0 0.0 0.0 0.0 0.0 0.0 1.19 1.29
epm n/a n/a 0.0 0.0 0.0 0.0 0.0 30.7 48.74 32.93
esrp n/a n/a 0.0 0.0 0.0 0.0 0.0 2.7 0.82 0.45
etmon n/a n/a 0.0 0.0 0.0 0.0 0.5 30.5 4865.78 873.87
...


5 Configuring Slots and Ports on a Switch

● Configuring a Slot on a Modular or SummitStack Switch on page 125
● Disabling MSM-G8X I/O Ports—BlackDiamond 8800 a-series and e-series Modules Only on page 127
● Configuring Ports on a Switch on page 128
● Jumbo Frames on page 135
● Link Aggregation on the Switch on page 139
● Mirroring on page 151
● Extreme Discovery Protocol on page 157
● Software-Controlled Redundant Port and Smart Redundancy on page 158
● Configuring Automatic Failover for Combination Ports—Summit X450, X450a, and X450e Series of
Switches Only on page 161
● Displaying Port Configuration Information on page 163
Configuring a Slot on a Modular or SummitStack Switch

This section describes configuring slots on a modular switch, which are the BlackDiamond 10808
switch, the BlackDiamond 12800 series switches, the SummitStack, and the BlackDiamond 8800 series
switch. This section describes the following topics:
● I/O Ports on BlackDiamond 8810 MSM-G8X Module on page 126
● I/O Ports on BlackDiamond 8806 MSM-G8X Module on page 127
Overview
In a SummitStack, a slot number is assigned to a node through configuration and stored in the node's
NVRAM. It only takes effect when the node restarts. In the following descriptions, the term "inserted
into a slot" in a SummitStack means that the node has become active, and because of its configured slot
value it appears to be present in a slot when the "show slot" command is run. The relationship of a
node and a slot does not change if the SummitStack is rewired. The term "module" refers to a Summit
family switch that may be present in the stack as an active node.
If a slot has not been configured for a particular type of module, then any type of module is accepted in
that slot, and a default port and VLAN configuration is automatically generated.
After any port on the module has been configured (for example, a VLAN association, a VLAN tag
configuration, or port parameters), all the port information and the module type for that slot must be
saved to non-volatile storage. Otherwise, if the modular switch or SummitStack is rebooted or the
module is removed from the slot, the port, VLAN, and module configuration information is not saved.

Configuring Slots and Ports on a Switch
NOTE
For information on saving the configuration, see Appendix A, “Software Upgrade and Boot Options.”
You configure the modular switch or a SummitStack with the type of input/output (I/O) module that is
installed in each slot. To do this, use the following command:
configure slot <slot> module <module_type>
You can also preconfigure the slot before inserting the module. This allows you to begin configuring the
module and ports before installing the module in the chassis or activating the related node in the
SummitStack.
If a slot is configured for one type of module, and a different type of module is inserted, the inserted
module is put into a mismatch state and is not brought online. To use the new module type in a slot,
the slot configuration must be cleared or configured for the new module type. To clear the slot of a
previously assigned module type, use the following command:
clear slot <slot>
All configuration information related to the slot and the ports on the module is erased. If a module is
present when you issue this command, the module is reset to default settings.
To display information about a particular slot, use the following command:

show slot {<slot>} {detail}
Information displayed includes:

● Module type, part number and serial number
● Current state (power down, operational, diagnostic, mismatch)
● Port information
If no slot is specified, information for all slots is displayed.
All slots on the modular switches are enabled by default. To disable a slot, use the following CLI
command:
disable slot
To re-enable slot, use the following CLI command:
enable slot
I/O Ports on BlackDiamond 8810 MSM-G8X Module

On the BlackDiamond 8810 switch, the MSM-G8X module also has eight 1 Gbps fiber SFP GBIC data, or
I/O, ports. You configure these ports exactly as you do any other ports on the switch.
Additionally, one slot on the BlackDiamond 8810 switch is dedicated to MSM-G8X use—slot A, or slot
5. Slot B, or slot 6, is a dual-purpose slot; it can be used for a secondary MSM-G8X or for a module
consisting solely of data, or I/O, ports.
The primary MSM-G8X must be in slot A in the BlackDiamond 8810 switch, which is referred to as slot
5 when working with the data ports. If you have a secondary MSM-G8X, that one goes into slot B,

Disabling MSM-G8X I/O Ports—BlackDiamond 8800 a-series and e-series Modules Only
which is slot 6 when you work with the data ports. So, when you work with the data ports on the
MSM-G8X, you specify slot 5 if you have one MSM-G8X, and slot 5 or 6 if you have two MSMs in the
switch.
When you issue any slot commands specifying a slot that contains an MSM-G8X (slot 5 with one
MSM-G8X and slots 5 and 6 with two MSMs) on the BlackDiamond 8810 switch, those commands affect
only the data ports on that slot; the MSMs remain unaffected. When you issue most msm commands on
this switch, those commands affect only the MSM-G8X host CPU subsystem; the I/O ports remain
unaffected. The sole exception is that the reboot msm command reboots both the MSM-G8X and the I/O
ports on that module.
I/O Ports on BlackDiamond 8806 MSM-G8X Module

On the BlackDiamond 8806 switch, the MSM-G8X module also has eight 1 Gbps fiber SFP GBIC data, or
I/O, ports. You configure these ports exactly as you do any other ports on the switch.
Additionally, one slot on the BlackDiamond 8806 switch is dedicated to MSM-G8X use—slot A, or slot
3. Slot B, or slot 4, is a dual-purpose slot; it can be used for a secondary MSM-G8X or for a module
consisting solely of data, or I/O, ports.
The primary MSM-G8X must be in slot A in the BlackDiamond 8806 switch, which is referred to as slot
3 when working with the data ports. If you have a secondary MSM-G8X, that one goes into slot B,
which is slot 4 when you work with the data ports. So, when you work with the data ports on the
MSM-G8X, you specify slot 3 if you have one MSM-G8X, and slot 3 or 4 if you have two MSMs in the
switch.
When you issue any slot commands specifying a slot that contains an MSM-G8X (slot 3 with one
MSM-G8X and slots 3 and 4 with two MSMs) on the BlackDiamond 8806 switch, those commands affect
only the data ports on that slot; the MSMs remain unaffected. When you issue most msm commands on
this switch, those commands affect only the MSM-G8X host CPU subsystem; the I/O ports remain
unaffected. The sole exception is that the reboot msm command reboots both the MSM-G8X and the I/O
ports on that module.
Disabling MSM-G8X I/O Ports—BlackDiamond 8800

a-series and e-series Modules Only
Beginning with ExtremeXOS version 11.5, the BlackDiamond 8800 series switch introduces new
modules—the BD8800 a-series and e-series modules—with increased capabilities over the BD8800
original modules. The BlackDiamond 8800 a-series and e-series modules are as follows:
● G48Te module
● G48Pe module
● G48Ta module
● G48Xa module
● 10G4Xa module
You may want to run a BlackDiamond 8800 series chassis with all a-series and e-series modules. To
obtain the full functionality of the BlackDiamond 8800 a-series and e-series modules, you must disable

the I/O ports on the MSM-G8X. Use the following commands to disable the I/O ports on the MSM-G8X
in the BlackDiamond 8800 chassis:
unconfigure slot <slot>

disable slot <slot> {offline}
where the slot number is the one containing the MSM-G8X (slot 3 and/or 4 for the BlackDiamond 8806
chassis; slot 5 and/or 6 for the BlackDiamond 8810 chassis). If you are running a chassis with two
MSMs, issue the commands twice, once with each MSM-G8X slot.
NOTE
The MSM will continue to work; only the I/O ports on the module are disabled.
When you are running a BlackDiamond 8800 series switch with both original modules and
BlackDiamond a-series or e-series modules, the switch returns an error message when you attempt to
execute certain commands. (Refer to the specific feature in the ExtremeXOS Concepts Guide to see the
messages.) Once you remove the original modules and unconfigure and disable those slots, issue the
above commands to disable the I/O ports on the MSM.
Also, if you are running ExtremeXOS software that is earlier than version 11.5 and you have any of the
BlackDiamond a-series or e-series modules in the chassis, those slots display as ‘Empty’ when you use
the show slot command.
Configuring Ports on a Switch
NOTE
Beginning with ExtremeXOS version 11.2, a port can belong to multiple virtual routers (VRs). See Chapter
10, “Virtual Routers,” for more information on VRs.
This section describes the following topics of configuring ports on a switch:

● Port Numbering on page 128
● Enabling and Disabling Switch Ports on page 129
● Configuring Switch Port Speed and Duplex Setting on page 130
● WAN PHY OAM—BlackDiamond 10808 and Summit X450a Series Switches Only on page 134
Port Numbering
ExtremeXOS runs on both stand-alone and modular switches, and the port numbering scheme is
slightly different on each. This section describes the following topics:
● Stand-alone Switch Numerical Ranges on page 129
● Modular Switch and SummitStack Numerical Ranges on page 129

Stand-alone Switch Numerical Ranges

On a stand-alone switch, such as the Summit family of switches, the port number is simply noted by the
physical port number, as shown below:
5
Separate the port numbers by a dash to enter a range of contiguous numbers, and separate the numbers
by a comma to enter a range of noncontiguous numbers:
● x-y—Specifies a contiguous series of ports on a stand-alone switch
● x,y—Specifies a non contiguous series of ports on a stand-alone switch
● x-y,a,d—Specifies a contiguous series of ports and a series of noncontagious ports on a stand-alone
switch
Modular Switch and SummitStack Numerical Ranges

On a modular switch and SummitStack, such as the BlackDiamond 10808 switch, the port number is a
combination of the slot number and the port number. The nomenclature for the port number is as
follows:
slot:port
For example, if an I/O module that has a total of four ports is installed in slot 2 of the chassis, the
following ports are valid:
● 2:1
● 2:2
● 2:3
● 2:4
You can also use wildcard combinations (*) to specify multiple modular slot and port combinations. The
following wildcard combinations are allowed:
● slot:*—Specifies all ports on a particular I/O module or stack node
● slot:x-slot:y—Specifies a contiguous series of ports on multiple I/O modules or stack nodes
● slot:x-y—Specifies a contiguous series of ports on a particular I/O module or stack node
● slota:x-slotb:y—Specifies a contiguous series of ports that begin on one I/O module or stack
node and end on another I/O module or stack node
Enabling and Disabling Switch Ports

By default, all ports are enabled. To enable or disable one or more ports on a switch, use the following
commands:
enable port [<port_list> | all]
disable port [<port_list> | all]
For example, to disable slot 7, ports 3, 5, and 12 through 15 on a modular switch or SummitStack, use
disable port 7:3,7:5,7:12-7:15

Beginning with ExtremeXOS version 11.6, you have the flexibility to receive or not to receive SNMP
trap messages when a port transitions between up and down. To receive these SNMP trap messages,
enable snmp traps port-up-down ports [<port_list> | all]
To stop receiving these messages, use the following command:
disable snmp traps port-up-down ports [<port_list> | all]
Refer to “Displaying Port Configuration Information” for information on displaying link status.
NOTE
You can choose to boot the BlackDiamond 12800 series switches with the ports disabled. This information is saved
in NVRAM and is not saved in the “.cfg” configuration files. Use the commands configure switch ports
initial-mode disabled to disable and configure switch ports initial-mode enabled to enable,
and show switch to display the configuration.
Configuring Switch Port Speed and Duplex Setting
NOTE
Refer to “Displaying Port Configuration Information” for information on displaying port speed, duplex,
autonegotiation, and flow control settings.
ExtremeXOS supports the following port types:

● 10 Gbps ports
● 10/100/1000 Mbps copper ports
● 10/100/1000 Mbps copper ports with Power over Ethernet (PoE)—only on the G48P and G48Pe
modules installed in the BlackDiamond 8800 series switch and the Summit X450e-24p, X450e-48p,
X250e-24p, and X250e-48p switches
● 1 Gbps small form factor pluggable (SFP) gigabit Ethernet interface converter (GBIC) fiber ports
● 100 FX GBICs, which must have their speed configured to 100 Mbps
● 100/1000 FX/LX SFP GBIC ports—only on the Black Diamond 8800 series switch, the BlackDiamond
12800 series switch, SummitStack, and the Summit family of switches
● Wide area network (WAN) PHY port—only on the BlackDiamond 10808 and 12800 series switches
and the Summit X450a and X450e series switches
● 10/100 Mbps copper ports with Power over Ethernet (PoE) ports for the Summit X250e series
switches.
Autonegotiation determines the port speed and duplex setting for each port (except 10 Gbps ports). You
can manually configure the duplex setting and the speed of 10/100/1000 Mbps ports.
The 10/100/1000 Mbps ports can connect to either 10BASE-T, 100BASE-T, or 1000BASE-T networks. By
default, the ports autonegotiate port speed. You can also configure each port for a particular speed
(either 10 Mbps or 100 Mbps).

NOTE
With autonegotiation turned off, you cannot set the speed to 1000 Mbps.
In general, SFP gigabit Ethernet ports are statically set to 1 Gbps, and their speed cannot be modified.
However, there are two GBICs supported by Extreme that can have a configured speed:
● 100 FX GBICs, which must have their speed configured to 100 Mbps
● 100FX/1000LX GBICs, which can be configured at either speed (available only on the Black
Diamond 8800 series switch, the BlackDiamond 12800 series switch, and the Summit family of
switches)
The 10 Gbps ports always run at full duplex and 10 Gbps.
To configure port speed and duplex setting, use the following command:
configure ports <port_list> auto off speed [10 | 100 | 1000 | 10000] duplex [half |
full]
To configure the system to autonegotiate, use the following command:

configure ports <port_list> auto on {speed [10 | 100 | 1000 | 10000]} {duplex [full |
half]}
ExtremeXOS does not support turning off autonegotiation on the management port.
Table 16 lists the support for autonegotiation, speed, and duplex setting for the various types of ports.
Table 16: Support for autonegotiation on various ports

Port Autonegotiation Speed Duplex
10 Gbps Off 10000 Mbps Full duplex
1 Gbps fiber SFP GBIC On (default)
Off 1000 Mbps Full duplex
100 FX GBIC On (default) 100 Mbps Full duplex
Off
100/1000 Mbps FX/LX On (default) 100 Mbps
SFP GBIC Off Full duplex
1000 Mbps
10/100/1000 Mbps On (default) 10 Mbps Full/half duplex
Off 100 Mbps Full/half duplex
10/100 Mbps On (default) 10 Mbps Full/half duplex
Off 100 Mbps Full/half duplex
Flow control on Gigabit Ethernet ports is enabled or disabled as part of autonegotiation (see IEEE
802.3x). If autonegotiation is set to Off on the ports, flow control is disabled. When autonegotiation is
turned On, flow control is enabled.
With Extreme Networks devices, the 1 Gbps ports and the 10 Gbps ports implement flow control as
follows:
● 1 Gbps ports
■ Autonegotiation enabled
- Advertise support for pause frames

- Respond to pause frames

- Do not transmit pause frames
■ Autonegotiation disabled
- Do not advertise support for pause frames
- Do not respond to pause frames
● 10 Gbps ports on modules for the BlackDiamond 10808 and 12800 series chassis:
■ Autonegotiation always disabled
- Transmit pause frames
● 10 Gbps ports for the Summit X450, X450a, and X450e series switches, SummitStack, and on modules
for the BlackDiamond 8800 series switch:
■ Autonegotiation always disabled
Turning Off Autonegotiation on a Gigabit Ethernet Port

In certain interoperability situations, you may need to turn autonegotiation off on a fiber gigabit
Ethernet port. Although a gigabit Ethernet port runs only at full duplex, you must specify the duplex
setting.
The following example turns autonegotiation off for port 1 (a 1 Gbps Ethernet port) on a module
located in slot 1 of a modular switch:
configure ports 1:1 auto off speed 1000 duplex full
The 10 Gbps ports do not autonegotiate; they always run at full duplex and 10 Gbps speed.
Running Link Fault Signal

Beginning with ExtremeXOS 11.1, the 10 Gbps ports support the Link Fault Signal (LFS) function. This
function, which is always enabled, monitors the 10 Gbps ports and indicates either a remote fault or a
local fault. The system then stops transmitting or receiving traffic from that link. After the fault has
been alleviated, the system puts the link back up and the traffic automatically resumes.
The Extreme Networks implementation of LFS conforms to the IEEE standard 802.3ae-2002.
NOTE
On the BlackDiamond 10808 switch, the 10 Gbps module must have the serial number 804405-00-09 or higher to
support LFS. To display the serial number of the module, use the show slot <slot_number> command. (All
the modules on the BlackDiamond 8800 series switch support LFS.)

Although the physical link remains up, all Layer 2 and above traffic stops. The system sends LinkDown
and LinkUp traps when these events occur. Additionally, the system writes one or more information
messages to the syslog, as shown in the following example:
09/09/2004 14:59:08.03 <Info:vlan.dbg.info> MSM-A: Port 4:3 link up at

10 Gbps speed and full-duplex
09/09/2004 14:59:08.02 <Info:hal.sys.info> MSM-A: 4:3 - remote fault
recovered.
09/09/2004 14:59:05.56 <Info:vlan.dbg.info> MSM-A: Port 4:3 link down

due to remote fault
09/09/2004 14:59:05.56 <Info:hal.sys.info> MSM-A: 4:3 - remote fault.
09/09/2004 15:14:12.22 <Info:hal.sys.info> MSM-A: 4:3 - local fault

recovered.
09/09/2004 15:14:11.35 <Info:vlan.dbg.info> MSM-A: Port 4:3 link up at
10 Gbps speed and full-duplex

due to local fault
09/09/2004 15:13:33.56 <Info:hal.sys.info> MSM-A: 4:3 - local fault.
due to local fault
NOTE
A link down or up event may trigger Spanning Tree Protocol topology changes or transitions.
Turning off Autopolarity—Summit Family of Switches, SummitStack, and

BlackDiamond 8800 Series Switch Only
(Autopolarity detection is always on for the BlackDiamond 10808 and 12800 series switches.)
The autopolarity feature allows the system to detect and respond to the Ethernet cable type (straight-
through or crossover cable) used to make the connection to the switch port. This feature applies to only
the 10/100/1000 BASE-T ports on the switch.
When the autopolarity feature is enabled, the system causes the Ethernet link to come up regardless of
the cable type connected to the port. When the autopolarity feature is disabled, you need a crossover
cable to connect other networking equipment and a straight-through cable to connect to endstations.
The autopolarity feature is enabled by default.
To disable or enable autopolarity detection, use the following command:
configure ports [<port_list> | all] auto-polarity [off | on]

● port_list—Specifies one or more ports on the switch
● all—Specifies all of the ports on the switch
● off—Disables the autopolarity detection feature on the specified ports
● on—Enables the autopolarity detection feature on the specified ports

Under certain conditions, you might opt to turn autopolarity off on one or more ports. The following
example turns autopolarity off for ports 5 to 7 on a Summit X450 series switch:
configure ports 5-7 auto-polarity off
When autopolarity is disabled on one or more Ethernet ports, you can verify that status by using the
command:
show ports information detail
WAN PHY OAM—BlackDiamond 10808 and Summit X450a Series

Switches Only
Beginning with Extreme XOS 11.6 version software, you can configure WAN PHY OAM on the LW
XENPAK module ports on the BlackDiamond 10808 and Summit X450a series switches whether or not
included in a SummitStack. The LW XENPAK provides an interface connection between a 10G Ethernet
and a 10G SONET/SDH network from a 10G Ethernet equipment port.
The WAN-PHY OAM feature is a subset of the SONET/SDH overhead function. The WAN PHY
interface is defined in IEEE 802.3ae; ExtremeXOS 11.6 supports the LW XENPAK only. The WAN-PHY
feature is available on LW XENPAK ports on the BlackDiamond 10808 switch the Summit X450a series
switches, and beginning in ExtremeXOS 12.0 on the BlackDiamond 12800 series switch.
Configuring WAN PHY OAM Parameters

The following are configurable WAN PHY OAM parameters:
● Framing—either SONET or SDH; default is SONET.

● Clock source—either internal or line; default is line.
● J0 section trace string—16-character string; the default is 15 NULL characters.
● J1 path trace string—16-character string; the default is 15 NULL characters
● Loopback—line, internal, or off; the default is off
To set the framing, use the following command:
configure ports <port_list> wan-phy framing [sonet | sdh]
To choose the clock source, use the following command:
configure ports <port_list> wan-phy clocking [line | internal]
To set a section trace ID, use the following command:
configure ports <port_list> wan-phy trace-section <id_string>
To set a path trace ID, use the following command:
configure ports <port_list> wan-phy trace-path <id_string>
To set a WAN PHY port to loopback, use the following command:
configure ports <port_list> wan-phy loopback [line | off]

Jumbo Frames
Displaying WAN PHY OAM Information

You display information on the WAN PHY ports using the following commands:
show ports {mgmt | <port_list>} information {detail}

show ports <port_list> wan-phy configuration
show ports <port_list> wan-phy errors {no-refresh}
show ports <port_list> wan-phy events {no-refresh}
show ports <port_list> wan-phy overhead {no-refresh}
Jumbo Frames
Jumbo frames are Ethernet frames that are larger than 1522 bytes, including four bytes used for the cyclic
redundancy check (CRC). Extreme products support switching and routing of jumbo frames at wire-
speed on all ports. The configuration for jumbo frames is saved across reboots of the switch.
Jumbo frames are used between endstations that support larger frame sizes for more efficient transfers
of bulk data. Both endstations involved in the transfer must be capable of supporting jumbo frames.
The switch only performs IP fragmentation, or participates in maximum transmission unit (MTU)
negotiation on behalf of devices that support jumbo frames.
You need jumbo frames when running the Extreme Networks vMAN implementation. When you are
working on the BlackDiamond 10808 or a BlackDiamond 12800 series switch, the switch enables jumbo
frames when you configure vMANs. If you are working on the BlackDiamond 8800 series switch,
SummitStack, and the Summit family of switches, you can enable and disable jumbo frames on
individual ports before configuring vMANs. For more information on configuring vMANs, refer to
Chapter 6, “Virtual LANs.”
Refer to “Displaying Port Configuration Information” for information on displaying jumbo frame status.
Jumbo Frames on the BlackDiamond 8800 Series Switch,

SummitStack, and Summit Family of Switches Only
Beginning in ExtremeXOS 11.6, you can enable jumbo frames per port on the following switches:
● BlackDiamond a-series and e-series modules
● Summit X450a and X450e series switches and beginning with ExtremeXOS 12.0, the Summit X250e
series switches.
If you attempt to enable jumbo frames per port on the original BlackDiamond 8800 modules, the system
returns the following message:
Error: You must enable jumbo-frame on all MSM-G8X I/O ports and G48T, G48P, G24x, and
10G4X ports globally. Use ‘enable jumbo-frame ports all’.
If you attempt to enable jumbo frames per port on the Summit X450 series switch, the system returns
the following error message:
Error: You must enable jumbo-frame on all S450-24X and S450-24T ports globally. Use
'enable jumbo-frame port all'.

If you attempt to enable jumbo frames per port on a BlackDiamond a-series or e-series module that co-
exists in the chassis with an original BlackDiamond 8800 module, the system returns the following
message:
Error: You must first enable jumbo-frame on all MSM-G8X I/O ports and G48T, G48P,
G24X, and 10G4X port globally. Use ‘enable jumbo-frame ports all’ and the ‘disable
jumbo-frame port <port_list>’ on any other port.
The following information applies to jumbo frames on the original BlackDiamond 8800 series modules
and the Summit X450 series switches:
● The original BlackDiamond 8800 series modules and the Summit X450 series switches support jumbo
frames on the entire switch; you cannot enable or disable jumbo frames per port.
The original BlackDiamond 8800 series module and the Summit X450 series switches enable or
disable jumbo frames on the entire switch; jumbo frames are either enabled or disabled on every
port on the switch. The system returns an error message if you attempt to enter specified ports.
● To enable jumbo frame support on the original BlackDiamond 8800 series modules or the Summit
X450 series switches, use the following command:
enable jumbo-frame ports all
After you issue this command, any new modules you add to the switch will also have jumbo frames
enabled.
● When you configure vMANs on the BlackDiamond 8800 series switch, ,SummitStack, and the
Summit family of switches, you can enable or disable jumbo frames for individual ports before
configuring the vMANs.
Enabling Jumbo Frames
NOTE
Some network interface cards (NICs) have a configured maximum MTU size that does not include the additional 4
bytes of CRC. Ensure that the NIC maximum MTU size is at or below the maximum MTU size configured on the
switch. Frames that are larger than the MTU size configured on the switch are dropped at the ingress port.
To enable jumbo frame support, enable jumbo frames on the desired ports. To set the maximum jumbo
frame size, use the following command:
configure jumbo-frame-size <framesize>
The jumbo frame size range is 1523 to 9216. This value describes the maximum size of the frame in
transit (on the wire), and includes 4 bytes of CRC plus another 4 bytes if 802.1Q tagging is being used.
Set the MTU size for the VLAN by using the following command:
configure ip-mtu <mtu> vlan <vlan_name>
Next, enable support on the physical ports that will carry jumbo frames using the following command:
enable jumbo-frame ports [all | <port_list>]

Jumbo Frames
Path MTU Discovery
NOTE
The original BlackDiamond 8800 series modules and the Summit X450 series switches do not support the router
specification for path MTU discovery.
Beginning with ExtremeXOS 11.6, the BlackDiamond a-series and e-series modules and the Summit
X450a, X450e, and X250e series switches, whether or not included in a SummitStack, support path MTU
discovery.
Using path MTU discovery, a source host assumes that the path MTU is the MTU of the first hop
(which is known). The host sends all datagrams on that path with the “don’t fragment” (DF) bit set,
which restricts fragmentation. If any of the datagrams must be fragmented by an Extreme switch along
the path, the Extreme switch discards the datagrams and returns an ICMP Destination Unreachable
message to the sending host, with a code meaning “fragmentation needed and DF set”. When the
source host receives the message (sometimes called a “Datagram Too Big” message), the source host
reduces its assumed path MTU and retransmits the datagrams.
The path MTU discovery process ends when one of the following is true:
● The source host sets the path MTU low enough that its datagrams can be delivered without
fragmentation.
● The source host does not set the DF bit in the datagram headers.
If it is willing to have datagrams fragmented, a source host can choose not to set the DF bit in datagram
headers. Normally, the host continues to set DF in all datagrams, so that if the route changes and the
new path MTU is lower, the host can perform path MTU discovery again.
IP Fragmentation with Jumbo Frames
NOTE
The original BlackDiamond 8800 series modules and the Summit X450 series switches do not support
fragmentation of any IP packets they forward.
Beginning with ExtremeXOS 11.6, The BlackDiamond a-series and e-series modules and the Summit
X450a and X450e series switches support fragmentation of IP packets. Beginning with ExtremeXOS 12.0,
the Summit X250e series switches support fragmentation of IP packets. The above support is included
whether or not the switches are present in a SummitStack.
ExtremeXOS supports the fragmenting of IP packets. If an IP packet originates in a local network that
allows large packets and those packets traverse a network that limits packets to a smaller size, the
packets are fragmented instead of discarded.
This feature is designed to be used in conjunction with jumbo frames. Frames that are fragmented are
not processed at wire-speed within the switch fabric.

NOTE
Jumbo frame-to-jumbo frame fragmentation is not supported. Only jumbo frame-to-normal frame fragmentation is
supported.
To configure VLANs for IP fragmentation:

1 Enable jumbo frames on the incoming port.
NOTE
If you are working with the original BlackDiamond 8800 series modules or a Summit X450 series switch, you
must enable jumbo frames for the entire switch.
2 Add the port to a VLAN.

3 Assign an IP address to the VLAN.
4 Enable ipforwarding on the VLAN.
5 Set the MTU size for the VLAN, using the following command:
configure ip-mtu <mtu> vlan <vlan_name>
The ip-mtu value ranges between 1500 and 9216, with 1500 the default.
NOTE
To set the MTU size greater than 1500, all ports in the VLAN must have jumbo frames enabled.
IP Fragmentation within a VLAN

ExtremeXOS supports IP fragmentation within a VLAN. This feature does not require you to configure
the MTU size. To use IP fragmentation within a VLAN:
1 Enable jumbo frames on the incoming port.
2 Add the port to a VLAN.
3 Assign an IP address to the VLAN.
4 Enable ipforwarding on the VLAN.
If you leave the MTU size configured to the default value, when you enable jumbo frame support on a
port on the VLAN you will receive a warning that the ip-mtu size for the VLAN is not set at maximum
jumbo frame size. You can ignore this warning if you want IP fragmentation within the VLAN, only.
However, if you do not use jumbo frames, IP fragmentation can only be used for traffic that stays
within the same VLAN. For traffic that is sent to other VLANs, to use IP fragmentation, all ports in the
VLAN must be configured for jumbo frame support.
NOTE
IP Fragmentation within a VLAN does not apply to the Summit X450e and X450a series, the Summit X250e series
switches (whether or not included in a SummitStack), and the BlackDiamond 8800 a-series and e-series modules.
The platforms which currently support Fragmentation do so only for layer-3 forwarding.

Link Aggregation on the Switch

The link aggregation (also known as load sharing) feature allows you to increase bandwidth and
availability by using a group of ports to carry traffic in parallel between switches. Load sharing, link
aggregation, and trunking are terms that have been used interchangeably in Extreme Networks
documentation to refer to the same feature, which allows multiple physical ports to be aggregated into
one logical port, or link aggregation group (LAG). Refer to IEEE 802.3ad for more information on this
feature. The advantages to link aggregation include an increase in bandwidth and link redundancy.

● Link Aggregation Overview on page 139
● Link Aggregation and Software-Controlled Redundant Ports—Summit Family of Switches Only on
page 140
● Dynamic Versus Static Load Sharing on page 140
● Load-Sharing Algorithms on page 141
● LACP—Dynamic Link Aggregation on page 143
● Guidelines for Load Sharing on page 145
● Configuring Switch Load Sharing on page 148
● Load-Sharing Examples on page 150
● Displaying Switch Load Sharing on page 151
Link Aggregation Overview
NOTE
All ports in a LAG must be running at the same speed and duplex setting. Each port can belong to only one LAG.
Load sharing allows the switch to use multiple ports as a single logical port, or LAG. For example,
VLANs see the LAG as a single logical port. And, although you can only reference the master port of a
LAG to a Spanning Tree Domain (STPD), all the ports of the LAG actually belong to the specified STPD.
Most load-sharing algorithms guarantee packet sequencing between clients.
Link aggregation, or load sharing, is disabled by default.
If a port in a load-sharing group (or LAG) fails, traffic is redistributed to the remaining ports in the
LAG. If the failed port becomes active again, traffic is redistributed to include that port.
NOTE
Load sharing must be enabled on both ends of the link, or a network loop may result.
Link aggregation is most useful when:

● The egress bandwidth of traffic exceeds the capacity of a single link.
● Multiple links are used for network resiliency.
In both situations, the aggregation of separate physical links into a single logical link multiplies total
link bandwidth in addition to providing resiliency against individual link failures.

In modular switches, ExtremeXOS supports LAGs across multiple modules, so resiliency is also
provided against individual module failures.
The software supports control protocols across the LAGs, both static and dynamic. If you add the
protocols (for example, EAPS, ESRP, and so forth) to the port and then create a LAG on that port, you
may experience a slight interruption in the protocol operation. To seamlessly add or delete bandwidth
when running control protocols, Extreme Networks recommends that you create a LAG consisting of
only one port. Then add your protocols to that port and add other ports as needed.
vMAN ports can belong to LAGs. If any port in the LAG is enabled for vMAN, all ports in the group
are automatically enabled to handle jumbo size frames on the BlackDiamond 10808 and 12800 series
switches; you must enable jumbo frames on the BlackDiamond 8800 series switch, SummitStack, and
the Summit family of switches. Also, vMAN is automatically enabled on all ports of the untagged LAG.
NOTE
Beginning with ExtremeXOS 11.4, you can use vMAN ACLs to configure load sharing on a vMAN. See Chapter
12, “Access Lists (ACLs),” for complete information on vMAN ACLs.
You can run the Link Layer Discovery Protocol (LLDP) on ports in a LAG.
Link Aggregation and Software-Controlled Redundant Ports—

Summit Family of Switches Only
If you are configuring software-controlled redundant ports and link aggregation together, the following
rules apply:
● Only the master logical port can be a either a primary or redundant port.
● You must unconfigure the software-controlled redundant ports before either configuring or
unconfiguring load sharing.
● The entire LAG must go down before the software-controlled redundant port takes effect.
Dynamic Versus Static Load Sharing

Beginning with 11.3, ExtremeXOS software supports two broad categories of load sharing, or link
aggregation:
● Dynamic load sharing—Dynamic load sharing is a grouping of ports that use the Link Aggregation
Control Protocol (LACP) to dynamically determine if link aggregation is possible and then to
automatically configure the aggregation. LACP is part of the IEEE 802.3ad standard and allows the
switch to dynamically reconfigure the link aggregation groups (LAGs). The LAG is enabled only
when LACP detects that the remote device is also using LACP and is able to join the LAG.
● Static load sharing—Static load sharing is a grouping of ports specifically configured to load share.
The switch ports at each end must be specifically configured as part of a load-sharing group.
NOTE
The platform-related load-sharing algorithms apply to LACP (as well as static load sharing).

Load-Sharing Algorithms
With ExtremeXOS version 11.2, you can use IPv6 addresses in the load-sharing algorithms.
Load-sharing, or link aggregation, algorithms allow you to select the distribution technique used by the
LAG to determine the output port selection. Algorithm selection is not intended for use in predictive
traffic engineering.
The ExtremeXOS software supports static load sharing, which is a grouping of ports specifically
configured to load share. Beginning with ExtremeXOS version 11.3, the software also supports dynamic
load sharing. (See “LACP—Dynamic Link Aggregation” on page 143 for complete information on
LACP.) The switch ports at each end must be configured as part of a load-sharing group. Additionally,
you can choose the load-sharing algorithm used by the group.
The dynamic load-sharing feature, or Link Aggregation Control Protocol (LACP), is standards-based
and compatible with most third-party switches. Static load sharing is also compatible with most third-
party switches.
NOTE
Always reference the master logical port of the load-sharing group when configuring or viewing VLANs. VLANs
configured to use other ports in the LAG will have those ports deleted from the VLAN when link aggregation is
enabled.
Link Aggregation Algorithm on the Summit 450a, X450e and X250e Series Switches
NOTE
You cannot configure port-based load sharing on the Summit X450a, X450e and X250e series switches (whether or
not included in a SummitStack).
Address-based load sharing. When you configure address-based load sharing, the switch examines a
specific place in the packet to determine which egress port to use for forwarding traffic:
● For Layer 2 load sharing, the switch uses the MAC source address and destination address.
● For Layer 3/Layer 4 load sharing, the switch uses the IP source and destination address and the
Layer 4 port.
You control the field examined by the switch for address-based load sharing when the load-sharing
group is created by using the following command:
enable sharing grouping
If the packet is not IP, the switch applies the Layer 2 algorithm, which is the default setting.
Link Aggregation Algorithm on the BlackDiamond 8800 Series Switch, SummitStack,

and the Summit X450 Series Switches
NOTE
You cannot configure port-based load sharing on the BlackDiamond 8800 series switch, SummitStack, or the
Summit X450 series switches.

● For Layer 3 load sharing, the switch uses the IP source address and destination address.
You control the field examined by the switch for address-based load sharing when the load-sharing
group is created by using the following command:
enable sharing grouping
If the packet is not IP, the switch applies the Layer 2 algorithm, which is the default setting.
Link Aggregation Algorithms on the BlackDiamond 10808 and 12800 series Switches
You can configure one of two load-sharing, or link aggregation, algorithms on the BlackDiamond 10808
and 12800 series switches, as follows:
● Port-based—Uses the ingress port to determine which physical port in the load-sharing group is
used to forward traffic out of the switch.
● Address-based—Uses addressing information to determine which physical port in the load-sharing
group to use to forward traffic out of the switch. Addressing information is based on the packet
protocol, as follows:
■ IP packets—Uses the source and destination MAC and IP addresses and the TCP port number.
■ All other packets—Uses the source and destination MAC address.
If you do not explicitly select an algorithm, the port-based scheme is used. However, the address-based
algorithm has a more even distribution and is the recommended choice.
● For Layer 3 load sharing, the switch uses the IP source address and destination address.
● For Layer 4 load sharing, the switch using the TCP source and destination port number.
You can control the field examined by the switch for address-based load sharing by using the following
command:
configure sharing address-based
In this command, CHK SUM indicates that the switch should examine the IP check sum. Examining the
IP check sum in addition to the other parameters produces a random traffic pattern on the egress of the
load-sharing links because the IP check sum includes the packet length, which is likely to change from
packet to packet.
NOTE
Those variables with CHK_SUM apply only to IPv4 packets.
This feature is available only for the address-based load-sharing algorithm. The selected address-based
algorithm is applied to the entire switch, to all the load-sharing groups configured as address-based.
Layer 2 is the default setting.

The master port of the load-sharing group can be the monitor port for port-mirroring.
LACP—Dynamic Link Aggregation
NOTE
Beginning in ExtremeXOS version 11.4, LACP fails over hitlessly in the event of a failover to a duplicate MSM in a
modular switch.
Beginning with ExtremeXOS version 11.3, you can run the Link Aggregation Control Protocol (LACP)
on Extreme Networks devices. LACP enables dynamic load sharing and hot standby for link
aggregation links, in accordance with the IEEE 802.3ad standard. All third-party devices supporting
LACP run with Extreme Networks devices.
The addition of LACP provides the following enhancements to static load sharing, or link aggregation:
● Automatic configuration
● Rapid configuration and reconfiguration
● Deterministic behavior
● Low risk of duplication or misordering
After you enable load-sharing, the LACP protocol is enabled by default. You configure dynamic link
aggregation by first assigning a primary, or logical, port to the group, or LAG and then specifying the
other ports you want in the LAG.
LACP, using an automatically generated key, determines which links can aggregate. Each link can
belong to only one LAG. LACP determines which links are available. The communicating systems
negotiate priority for controlling the actions of the entire trunk (LAG), using LACP, based on the lowest
system MAC number. You can override this automatic prioritization by configuring the system priority
for each LAG.
After you enable and configure LACP, the system sends PDUs (LACPDUs) on the LAG ports. The
LACPDUs inform the remote system of the identity of the sending system, the automatically generated
key of the link, and the desired aggregation capabilities of the link. If a key from a particular system on
a given link matches a key from that system on another link, those links are aggregatable. After the
remote system exchanges LACPDUs with the LAG, the system determines the status of the ports and
whether to send traffic on which ports.
Among those ports deemed aggregatable by LACP, the system uses those ports with the lowest port
number as active ports; the remaining ports aggregatable to that LAG are put into standby status.
Should an active link fail, the standby ports become active, also according to the lowest port number.
(See “Configuring LACP” on page 148 for the number of active and standby LACP links supported per
platform.)
All ports configured in a LAG begin in an unselected state. Based on the LACPDUs exchanged with the
remote link, those ports that have a matching key are moved into a selected state. If there is no matching
key, the ports in the LAG remain in the unselected state.
However if more ports in the LAG are selected than the aggregator can handle because of the system
hardware, those ports that fall out of the hardware’s capability are moved into standby state. The lowest
numbered ports are the first to be automatically added to the aggregator; the rest go to standby. As the

name implies, these ports are available to join the aggregator if one of the selected ports should go
offline.
Beginning with ExtremeXOS version 11.4, you can configure the port priority to ensure the order that
ports join the aggregator. However, that port must first be added to the LAG before you can configure
the LACP settings. Again, if more than one port is configured with the same priority, the lowest-
numbered port joins the aggregator first.
After the ports in the LAG move into the selected state, LACP uses the mux portion of the protocol to
determine which ports join the aggregator and can collect and distribute traffic. A few seconds after a
port is selected, it moves into the mux state of waiting, and then into the mux state of attached. The
attached ports then send their own LACP sync messages announcing that they are ready to receive
traffic.
The protocol keeps sending and receiving LACPDUs until both sides of the link have echoed back each
other’s information; the ends of the link are then considered synchronized. After the sync messages
match up on each end, that port is moved into the aggregator (into the mux state of collecting-
distributing) and is able to collect and distribute traffic.
The protocol then enables the aggregated link for traffic and monitors the status of the links for changes
that may require reconfiguration. For example, if one of the links in a LAG goes down and there are
standby links in that LAG, LACP automatically moves the standby port into selected mode and that
port begins collecting and distributing traffic.
The marker protocol portion of LACP ensures that all traffic on a link has been received in the order in
which it was sent and is used when links must be dynamically moved between aggregation groups. The
Extreme Networks LACP implementation responds to marker frames but does not initiate these frames.
NOTE
Always verify the LACP configuration by issuing the show ports sharing command; look for the ports specified
as being in the aggregator. You can also display the aggregator count by issuing the show lacp lag command.
Beginning with ExtremeXOS 11.4, you can configure additional parameters for the LACP protocol and
the system sends certain SNMP traps in conjunction with LACP. The system sends a trap when a
member port is added to or deleted from an aggregator.
The system now detects and blocks loopbacks; that is, the system does not allow a pair of ports that are
in the same LAG but are connected to one another by the same link to select the same aggregator. If a
loopback condition exists between two ports, they cannot aggregate. Ports with the same MAC address
and the same admin key cannot aggregate; ports with the same MAC address and a different admin key
can belong to the same LAG.
The system sends an error message if a LAG port is configured and up but still not attached to the
aggregator or in operation within 60 seconds. Use the show lacp member-port <port> detail
command to display the churn on both sides of the link. If the Churn value is shown as True in the
display, check your LACP configuration. The issue may be either on your end or on the partner link,
but you should check your configuration. The display shows as True until the aggregator forms, when
it changes to display as False.
A LAG port moves to expired and then to the defaulted state when it fails to receive an LACPDU from
its partner for a specified time. You can configure this timeout value as long, which is 90 seconds, or
short, which is 3 seconds; the default is long. (In ExtremeXOS version 11.3, the timeout value is not

configurable and is set as long, or 90 seconds.) Use the show lacp lag <group-id> detail command
to display the timeout value for the LAG.
There are two LACP activity modes: active and passive. In LACP active mode, the switch periodically
sends LACPDUs; in passive mode, the switch sends LACPDUs only when it receives one from the other
end of the link. The default is active mode. (In ExtremeXOS version 11.3, the mode is not configurable;
it is always active mode.) Use the show lacp lag <group-id> detail command to display the LACP
mode for the LAG.
NOTE
One side of the link must be in active mode in order to pass traffic. If you configure your side in the passive mode,
ensure that the partner link is in LACP active mode.
A LAG port moves into a defaulted state after the timeout value expires with no LACPDUs received for
the other side of the link. You can configure whether you want this defaulted LAG port removed from
the aggregator or added back into the aggregator. If you configure the LAG to remove ports that move
into the default state, those ports are removed from the aggregator and the port state is set to
unselected. The default configuration for defaulted ports is to be removed, or deleted, from the
aggregator. (In ExtremeXOS version 11.3, defaulted ports in the LAG are always removed from the
aggregator; this is not configurable.)
NOTE
To force the LACP trunk to behave like a static sharing trunk, use the configure sharing lacp defaulted-
state-action command to add ports to the aggregator.
If you configure the LAG to add the defaulted port into the aggregator, the system takes inventory of
the number of ports currently in the aggregator. If there are fewer ports in the aggregator than the
maximum number allowed, the system adds the defaulted port to the aggregator (port set to selected
and collecting-distributing). If the aggregator has the maximum ports, the system adds the defaulted
port to the standby list (port set to standby). Use the show lacp lag <group-id> {detail} command
to display the defaulted action set for the LAG.
NOTE
If the defaulted port is assigned to standby, that port automatically has a lower priority than any other port in the
LAG (including those already in standby).
Guidelines for Load Sharing

This sections discusses the guidelines for load sharing on various platforms.
Configuring Load Sharing on the Summit Family of Switches

The following rules apply to load sharing on the Summit family of switches:
● One static LAG can contain up to 8 ports.
● The maximum number of LAGs is 128.
● You can configure only the address-based load-sharing algorithm.

● The default load-sharing algorithm is L2 address-based aggregation.

● Broadcast, multicast, or unknown unicast packets are transmitted differently depending on the
device you are using:
■ On the Summit X450 series switches, these packets are transmitted on a single port of a LAG.
■ On the Summit X450a, X450e and X250e series switches, these packets are distributed across all
members of a LAG.
● You can use Layer 3 information and Layer 4 port information in the algorithm on the following
devices:
■ Summit X450, X450e and X250 series switches.
NOTE
See “Configuring LACP” on page 148 for the maximum number of links, selected and standby, per LACP.
Configuring Load Sharing on SummitStack and the BlackDiamond 8800 Series Switch
The following rules apply to load sharing on the BlackDiamond 8800 series switch:
● You can configure only the address-based load-sharing algorithm.
● The default load-sharing algorithm is L2 address-based aggregation.
● Broadcast, multicast, or unknown unicast packets are transmitted differently depending on the
device you are using:
■ On the BlackDiamond 8800 original modules or Summit X450 Series switches, these packets are
transmitted on a single port of a LAG.
■ On the BlackDiamond 8800 a-series and e-series modules, Summit X450a, X450e, and X250e
switches, these packets are distributed across all members of a LAG.
● Beginning with ExtremeXOS version 11.5, the maximum number of LAGs is 128 unless a 10G4X
module is in use in the chassis or a Summit X450 series switch is in the SummitStack, in which case
the maximum number of LAGs is 32.
If you attempt to configure more than 32 LAGs on a BlackDiamond 8800 chassis that contains an
10G4X module or has a slot configured for the 10G4X module, or if you attempt to configure more
than 32 LAGs on a SummitStack that contains a Summit X450 series switch, the system displays the
following error message:
Error: Slot <slot_number> can support a maximum of 32 trunks
If you want to configure more than 32 LAGs on this chassis, it is not necessary to remove the 10G4X
module; however, you must both unconfigure the slot holding the 10G4X module and disable the slot
holding that module or remove the 10G4X module. Use the following commands to unconfigure the
slot and disable that slot:
unconfigure ports wan-phy and disable slot
If you want to configure more than 32 LAGs on this SummitStack, you must unconfigure the slot
occupied by the Summit X450 series switch and remove the switch from the stack.

If you attempt to insert a 10G4X module into a BlackDiamond 8800 chassis configured for more than
32 LAGs (or attempt to insert a Summit X450 series module into a SummitStack), the module fails
the Init state with log error messages at the warning level similar to the following:
04/07/1921 23:52:28.29 <Warn:DM.Warning> MSM-A: Slot-8 FAILED (1) Error Max Load Share
Groups Exceeded(-48) from HAL on CardExec INIT(5) for slot
04/07/1921 23:52:28.29 <Warn:DM.Warning> MSM-A: Error Max Load Share Groups Exceeded(-
48) from HAL on CardExec INIT(5) for slot 8
Once you configure more than 32 LAGs on a BlackDiamond 8800 chassis or SummitStack, the 10G4X
module or Summit X450 switch will not initialize even if you reduce the number of LAGs to 32; you
must reboot the system or SummitStack first. The system logs an error message at the error level
similar to the following when you must reduce the system (even if you reduced the number of LAGs
to 32):
04/07/1921 23:52:28.29 <Erro:HAL.Card.Error> MSM-A: Slot 8 is not supported when more
than 32 load share groups have previously been configured. A system reboot is
required to clear this condition.
● You can use Layer 3 information and Layer 4 port information in the algorithm on the following
devices:
■ BlackDiamond 8800 a-series and e-series modules.
■ Summit X450a, X450e, or X250e series switches..
NOTE
Configuring Load Sharing on the BlackDiamond 10808 and 12800 series Switches
The following rules apply to load sharing on the BlackDiamond 10808 and 12800 series switches:
● The maximum number of LAGs is 128.
● You cannot specify L2 or L3 address-based load sharing.
● The default load-sharing algorithm is port-based aggregation.
NOTE
Load Sharing Rules and Restrictions for All Switches

Additionally, the following rules apply to load sharing on all switches:
● The ports in the LAG do not need to be contiguous.
● A LAG that spans multiple modules must use ports that have the same maximum bandwidth
capability, with one exception—you can mix media type on 1 Gbps ports.

Configuring Switch Load Sharing
NOTE
See “Guidelines for Load Sharing” on page 145” for specific information on load sharing for each specific device.
To set up a switch for load sharing, or link aggregation, among ports, you must create a load-sharing
group of ports, also known as a link aggregation group (LAG). The first port in the load-sharing group
is configured to be the “master” logical port. This is the reference port used in configuration commands
and serves as the LAG group ID. It can be thought of as the logical port representing the entire port
group.
All the ports in a load-sharing group must have the same exact configuration, including
autonegotiation, duplex setting, ESRP host attach or don’t-count, and so on. All the ports in a load-
sharing group must also be of the same bandwidth class.
To define a load-sharing group, or LAG, you assign a group of ports to a single, logical port number.
To enable or disable a load-sharing group, use the following commands:
enable sharing <port> grouping <port_list> {algorithm [port-based | address-based

{L2 | L3 | L3_L4}]} {lacp}
disable sharing <port>
Adding and Deleting Ports in a Load-Sharing Group

Ports can be added or deleted dynamically in a load-sharing group, or LAG. To add or delete ports
from a load-sharing group, use the following commands:
configure sharing <port> add ports <port_list>
configure sharing <port> delete ports <port_list>
NOTE
Configuring LACP
NOTE
Extreme Networks does not recommend enabling LACP and ELSM on the same port. See Chapter 5, “Status
Monitoring and Statistics,” for information on ELSM.
To configure LACP, or dynamic link aggregation, you must, again, first create a LAG. The first port in
the LAG serves as the logical port for the LAG. This is the reference port used in configuration
commands. It can be thought of as the logical port representing the entire port group, and it serves as
the LAG Group ID.

To create a LAG for LACP:

1 Create a LAG, using the following command:
enable sharing <port> grouping <port_list> {algorithm [port-based | address-based
{L2 | L3 | L3_L4}]} {lacp}
The port you assign using the first parameter becomes the logical port for the link aggregation group
and the LAG Group ID when using LACP. This logical port must also be included in the port list of
the grouping itself.
2 If you want to override the default prioritization in LACP for a specified LAG, use the following
command:
configure sharing <port> lacp system-priority <priority>
This step is optional; LACP handles prioritization using system MAC addresses.
3 Add or delete ports to the LAG as desired, using the following command:
configure sharing <port> add ports <port_list>
4 If you want to override the ports selection for joining the LAG by configuring a priority for a port
within a LAG, issue the following command:
configure lacp member-port <port> priority <port_priority>
5 If you want to change the expiry timer, use the following command:
configure sharing <port> lacp timeout [long | short]
The default value for the timeout is long, or 90 seconds.
6 If you want to change the activity mode, use the following command:
configure sharing <port> lacp activity-mode [active | passive]
The default value for the activity mode is active.
7 If you want to configure the action the switch takes for defaulted LAG ports, use the following
command:
configure sharing <port> lacp defaulted-state-action [add | delete]
The default value for defaulted LAG ports is delete the default ports.
NOTE
Always verify the LACP configuration by issuing the show ports sharing command; look for the ports listed as
being in the aggregator.
Configuring LACP on BlackDiamond 8800 series switch, SummitStack, and Summit family of switches. The
following rules apply to the number of LACP ports in one LAG on the BlackDiamond 8800 series
switch and on the Summit family of switches:
● Up to 16 link per LAG

● Up to 8 selected links and 8 standby links
● Only address-based LACP algorithm
Configuring LACP on BLackDiamond 10808 and 12800 series switches. The following rules apply to the
number of LACP ports in one LAG on the BlackDiamond 10808 and 12800 series switches:
● Up to 32 link per LAG

● Up to 16 selected links and 16 standby links
● Cannot specify L2 or L3 address-based LACP algorithm

Load-Sharing Examples
This section provides examples of how to define load sharing, or link aggregation, on stand-alone and
modular switches, as well has defining dynamic link aggregation.
Load Sharing on a Stand-alone Switch

The following example defines a static load-sharing group that contains ports 9 through 12, and uses
the first port in the group as the master logical port 9:
enable sharing 9 grouping 9-12
In this example, logical port 9 represents physical ports 9 through 12.
When using load sharing, you should always reference the master logical port of the load-sharing group
(port 9 in the previous example) when configuring or viewing VLANs; the logical port serves as the
LAG Group ID. VLANs configured to use other ports in the load-sharing group will have those ports
deleted from the VLAN when load sharing becomes enabled.
Cross-Module Load Sharing on a Modular Switch or SummitStack

The following example defines a static load-sharing group on modular switches that contains ports 9
through 12 on slot 3, ports 7 through 10 on slot 5, and uses the port 9 in the slot 3 group as the primary
logical port, or LAG Group ID:
enable sharing 5:7 grouping 3:9-3:12, 5:7-5:10
In this example, logical port 5:7 represents physical ports 3:9 through 3:12 and 5:7 through 5:10.
When using load sharing, you should always reference the LAG Group ID of the load-sharing group
(port 3:9 in the previous example) when configuring or viewing VLANs. VLANs configured to use
other ports in the load-sharing group will have those ports deleted from the VLAN when load sharing
becomes enabled.
Address-based load sharing can also span modules.
Single-Module Load Sharing on a Modular Switch or SummitStack

The following example defines a static load-sharing, or link aggregation, group that contains ports 9
through 12 on slot 3 and uses the first port as the master logical port 9, or LAG group ID:
enable sharing 3:9 grouping 3:9-3:12
In this example, logical port 3:9 represents physical ports 3:9 through 3:12.
LACP Example
The following configuration example:
● Creates a dynamic LAG with the logical port (LAG Group ID) of 10 that contains ports 10 through
12.
● Sets the system priority for that LAG to 3.

Mirroring
● Adds port 5 to the LAG.

enable sharing 10 grouping 10-12 lacp
configure sharing 10 lacp system-priority 3
configure sharing 10 add port 5
Displaying Switch Load Sharing

Beginning with ExtremeXOS version 11.3, you can use either static or dynamic load sharing (link
aggregation). LACP is used to create, configure, and display dynamic link aggregation. In the link
aggregation displays, the types are shown by the following aggregation controls:
● Static link aggregation—static
● Dynamic link aggregation—LACP
To verify your configuration, use the following command:
show ports sharing
To verify LACP configuration, use the following command:
show lacp
To display information for the specified LAG, use the following command:
show lacp lag <group-id> {detail}
To display LACP information for a specific port that is a member of a LAG, use the following
command:
show lacp member-port <port> {detail}
Refer to “Displaying Port Configuration Information” for information on displaying summary load-
sharing information.
To clear the counters, use the following command:
clear lacp counters
Beginning with ExtremeXOS version 11.4, you can display the LCAP counters for all member ports in
the system. To display the LACP counters, use the following command:
show lacp counters
Mirroring
NOTE
Beginning with ExtremeXOS software version 11.2, you can accomplish port mirroring using ACLs or
CLEAR-Flow. See Chapter 12, “Access Lists (ACLs),” for more information on ACLs and Chapter 17, “CLEAR-Flow,”
for more information on CLEAR-Flow.

Mirroring configures the switch to copy all traffic associated with one or more ports, VLANs, or virtual
ports. A virtual port is a combination of a VLAN and a port. The monitor port or ports can then be
connected to a network analyzer or RMON probe for packet analysis. The system uses a traffic filter
that copies a group of traffic to the monitor port(s). You can have only one monitor port or port list on
the switch.
NOTE
The mirroring filter limits discussed in this chapter do not apply when you are working with Sentriant devices.
Up to 16 mirroring filters and 1 monitor port or 1 monitor port list can be configured. A monitor port
list may contain up to 16 ports.
NOTE
On the BlackDiamond 8800 series switch, SummitStack, and the Summit family of switches, you can mirror up to
16 VLANs on a given port.
Mirroring is disabled by default.
NOTE
Frames that contain errors are not mirrored.
Guidelines for Mirroring on the Summit Family of Switches Only

The traffic filter on the Summit family of switches can be defined based on one of the following criteria:
● Physical port—All data that traverses the port, regardless of VLAN configuration, is copied to the
monitor port(s). You can specify which traffic the port mirrors:
■ Ingress—Mirrors traffic received at the port.
■ Egress—Mirrors traffic sent from the port.
■ Ingress and egress—Mirrors traffic either received at the port or sent from the port.
(If you omit the optional parameters, all traffic is forwarded; the default for port-based mirroring
is ingress and egress).
● VLAN—All data to a particular VLAN, regardless of the physical port configuration, is copied to the
monitor port(s).
● Virtual port—All data specific to a VLAN on a specific port is copied to the monitor port(s).
● EXOS supports up to 16 mirror filters where each filter can be a port, a VLAN, or a port + VLAN.
● Only traffic ingressing a VLAN can be monitored; you cannot specify ingressing or egressing traffic
when mirroring VLAN traffic and a virtual port filter.
● When routing between VLANs, ingress mirrored traffic is presented to the monitor port(s) as
modified for routing. This is the default behavior and the behavior when you use the command,
configure mirroring mode standard. When you use the command, configure mirroring mode
enhanced, ingress traffic is mirrored as it is received (on the wire).
● In standard mode (see the command, configure mirroring mode), even if you select ingress and
egress traffic, the packet is mirrored only the first time it matches a mirror filter and is not mirrored

Mirroring
on subsequent configured filters. In enhanced mode, packets which match both an ingress filter and
an egress filter will result in two packets egressing the monitor port or ports.
● You cannot include the monitor port, monitor port list, or loop-back port in a load sharing group.
● You cannot run sFlow and mirroring on the same switch on the Summit X450 series switches. If you
attempt to enable mirroring on these devices on a port that is already enabled for sFlow, the switch
returns the following message:
Mirroring is not compatible with SFlow. Mirroring is not enabled!
You can run mirroring and sFlow on the same device when you are running one of the following:
■ Summit X450a or X450e or X250e series switch
● Tagged and untagged traffic is mirrored as below:
■ With a monitor port or ports on all Summit series switches, all traffic egressing the monitor port
or ports is tagged. Even if some untagged ports send mirrored traffic to the monitor port or ports,
that traffic also egresses the monitor port or ports tagged with the internal VLAN ID.
■ With a monitor port or ports on all Summit series switches, all traffic ingressing the monitor port
or ports is tagged only if the ingress packet is tagged. If the packet arrived at the ingress port as
untagged, the packet egress the monitor port or ports as untagged.
● You may see a packet mirrored twice. This occurs only if both the ingress mirrored port and the
monitor port or ports are on the same one-half of the module and the egress mirrored port is either
on the other one-half of that module or on another module
Guidelines for Mirroring on the BlackDiamond 8800 Series

Switches and SummitStack Only
The traffic filter on the BlackDiamond 8800 series switch or SummitStack can be defined based on one
of the following criteria:
monitor port(s). You can specify which traffic the port mirrors:
■ Ingress—Mirrors traffic received at the port.
■ Egress—Mirrors traffic sent from the port.
■ Ingress and egress—Mirrors traffic either received at the port or sent from the port.
(If you omit the optional parameters, all traffic is forwarded; the default for port-based mirroring
is ingress and egress).
● VLAN—All data to a particular VLAN, regardless of the physical port configuration, is copied to the
monitor port(s).
● EXOS supports up to 16 mirror filters where each filter can be a port, a VLAN, or a port + VLAN.
● Only traffic ingressing a VLAN can be monitored; you cannot specify ingressing or egressing traffic
when mirroring VLAN traffic.
● When routing between VLANs, ingress mirrored traffic is presented to the monitor port(s) as
modified for routing. This is the default behavior and the behavior when you use the command
configure mirroring mode standard. When you use the command configure mirroring mode
enhanced, ingress traffic is mirrored as it is received (on the wire).
● In standard mode (see the command configure mirroring mode), even if you select ingress and
egress traffic, the packet is mirrored only the first time it matches a mirror filter and is not mirrored

on subsequent configured filters. In enhanced mode, packets which match both an ingress filter and
an egress filter will result in two packets egressing the monitor port or ports.
● You cannot include the monitor port or ports for the BlackDiamond 8800 series switch or
SummitStack in a load-sharing group.
● You cannot run sFlow and mirroring on the same switch for using the original BlackDiamond 8800
modules or Summit X450 series switches in a SummitStack. If you attempt to enable mirroring on
these devices on a port that is already enabled for sFlow, the switch returns the following message:
Mirroring is not compatible with SFlow. Mirroring is not enabled!
You can run mirroring and sFlow on the same device when you are running one of the following:
■ BlackDiamond a-series and e-series modules in a BlackDiamond 8800 chassis
■ Summit X450a, X450e, and X250e series switches in a SummitStack
● Tagged and untagged traffic is mirrored slightly differently depending on the module that the
mirrored port and the monitor port or ports are on:
■ With a monitor port(s) on a BlackDiamond 8800 original module or a Summit X450 series switch
in a SummitStack, all traffic egressing the monitor port(s) is tagged (regardless of what module
the ingressing port is on). Even if some untagged ports send mirrored traffic to the monitor
port(s), that traffic also egresses the monitor port(s) tagged with the internal VLAN ID.
■ With a monitor port or ports on a BlackDiamond 8800 a-series or e-series module or a Summit
X450a, X450e, or X250e series switch in a SummitStack, the mirrored packet is tagged only if the
ingress packet is tagged (regardless of what module the ingressing port is on). If the packet
arrived at the ingress port as untagged, the packet egress the monitor port(s) as untagged.
■ Egress mirrored packets are always transmitted as tagged on the monitor port(s).
● With the BlackDiamond 8800 a-series and e-series modules or Summit X450a, X450e, or X250e series
switches in a SummitStack, you may see a packet mirrored twice. This occurs only if both the ingress
mirrored port and the monitor port or ports are on the same one-half of the module and the egress
mirrored port is either on the other one-half of that module or on another module.
Guidelines for Mirroring on the BlackDiamond 10808 and 12800

Series Switches Only
The traffic filter on the BlackDiamond 10808 and 12800 series switches can be defined based on one of
the following criteria:
monitor port(s).
● VLAN—All data to and from a particular VLAN, regardless of the physical port configuration, is
copied to the monitor port(s).
● The master port of the load-sharing group can be the monitor port for port-mirroring.
The monitor port or ports transmits tagged or untagged frames, according to the way you configured
the monitor port(s). This feature allows you to mirror multiple ports or VLANs to a monitor port, while
preserving the ability of a single protocol analyzer to track and differentiate traffic within a broadcast
domain (VLAN) and across broadcast domains (for example, across VLANs when routing).

Mirroring
NOTE
The monitor port or ports on the BlackDiamond 10808 switch must be explicitly configured for tagged or untagged
frames beginning with ExtremeXOS version 11.0.
The traffic egressing the monitor port(s) can be either tagged or untagged. If the mirroring is enabled as
tagged on the monitor port, all traffic egressing the monitor port is tagged. In this case, even if some
untagged ports send mirrored traffic to the monitor port(s), that traffic also egresses the monitor port(s)
as tagged. And, if mirroring is enabled as untagged on the monitor port(s), all traffic egressing the
monitor port is untagged, including mirrored tagged packets.
When you upgrade to ExtremeXOS 11.0 on the BlackDiamond 10808 switches, all restored mirroring
configurations are tagged on the monitor ports.
Mirroring Rules and Restrictions for All Switches

This section summarizes the rules and restrictions for configuring mirroring:
● When you disable mirroring, all the filters are unconfigured.
● To change monitor ports, you must first remove all the filters.
● You cannot mirror the monitor port.
● The mirroring configuration is removed when you:
■ Delete a VLAN (for all VLAN-based filters).
■ Delete a port from a VLAN (for all VLAN-, port-based filters).
■ Unconfigure a slot (for all port-based filters on that slot).
● Any mirrored port can also be enabled for load sharing (or link aggregation); however, each
individual port of the load-sharing group must be explicitly configured for mirroring.
● The monitor port is automatically removed from all VLANs; you cannot add it to a VLAN.
● The mirroring filters are not confined to a single module; they can have ports that span multiple
modules.
● You cannot use the management port at all in mirroring configurations.
● You cannot run ELSM and mirroring on the same port. If you attempt to enable mirroring on a port
that is already enabled for ELSM, the switch returns a message similar to the following:
Error: Port mirroring cannot enabled on an ELSM enabled port 1.
● With one-to-many mirroring, you need to enable jumbo frame support in the mirror-to port, if you
need mirror tagged packets of length 1519 to 1522.
● The loopback port is dedicated for mirroring and hence cannot be used for other configuration and
that is indicated through glowing LED.
● One-to-many mirroring uses vMAN functionality internally. Please refer to “vMANs on the
BlackDiamond 8800 Series Switches, SummitStack, and the Summit Family of Switches Only” on
page 316 for more information on limitations of configuring vMAN and VLAN on the same physical
port. On SummitX450-24t and SummitX450-24x switches, if your configuration has ports belonging
to multiple VLANs as tagged, and if you want to configure one-to-many mirroring for any of those
VLANs, please ensure that the vMAN ethertype is set to 0x8100 before the mirroring configuration.

Mirroring Examples
Mirroring is disabled by default. To enable mirroring in single port, the following command can be
used:
enable mirroring to port <port-no>
To enable mirroring on multiple ports, use the following command:
enable mirroring to port-list <port-list> loopback-port <port>
The port-list is a list of monitor ports which will transmit identical copies of mirrored packets. The
loopback-port is an otherwise unused port required when mirroring to a port-list. The loopback-port is
not available for switching user data traffic.
To disable mirroring, use the following command:
disable mirroring
NOTE
When you change the mirroring configuration, the switch stops sending egress packets from the monitor port until
the change is complete. The ingress mirroring traffic to the monitor port and regular traffic are not affected.
BlackDiamond 8800 Series Switch, SummitStack, and the Summit Family of Switches
Only
The following example selects slot 3, port 4 on a modular switch or SummitStack as the monitor port
and sends all traffic received at slot 6, port 5 to the monitor port:
enable mirroring to port 3:4
configure mirroring add port 6:5 ingress
The following example selects slot 3, port 4 on a modular switch or SummitStack as the monitor port
and sends all traffic sent from slot 6, port 5 to the monitor port:
enable mirroring to port 3:4

configure mirroring add port 6:5 egress
The following example selects port 4 on a standalone switch as the monitor port and sends all traffic
ingressing the VLAN red to the monitor port:
enable mirroring to port 4

configure mirroring add vlan red
The following example selects port 4 on a standalone switch as the monitor port and sends all traffic
ingressing the VLAN red on port 5 to the monitor port:
enable mirroring to port 4

configure mirroring add vlan red port 5
The following example selects ports 5, 6, and 7 on slot 2 on a modular switch or SummitStack as the
monitor ports and sends all traffic received at slot 6, port 5 to the monitor ports. Slot 3, port 1 is an
unused port selected as the loopback port.
enable mirroring to port-list 2:5-2:7 loopback-port 3:1

configure mirroring add port 6:5 ingress

Extreme Discovery Protocol
BlackDiamond 10808 and 12800 series Switches Only

The following example selects slot 7, port 3 as the untagged monitor port, and sends all traffic coming
into or out of a modular switch on slot 7, port 1 to the monitor port:
enable mirroring to port 7:3 untagged
configure mirroring add port 7:1
The following example sends all traffic coming into or out of the system on slot 8, port 1 and the VLAN
default to the untagged monitor port, which is slot 7, port 3:
enable mirroring to port 7:3 untagged

configure mirroring add port 8:1 vlan default
The following example selects slot 2, port 24 and slot 7, port 3 as the tagged monitor ports, and sends
all traffic coming into or out of a modular switch on slot 7, port 1 to the monitor ports. Slot2, port 28 is
an unused port selected as the loopback port.
enable mirroring to port-list 2:24, 7:3 loopback-port 2:28 tagged

configure mirroring add port 7:1
Verifying the Mirroring Configuration

The screen output resulting from the show mirroring command lists the ports that are involved in
mirroring and which is the monitor port. The display differs slightly depending on the platform. To
display the LACP configuration, use the show lacp command; again, the display differs slightly
depending on the platform.
Extreme Discovery Protocol

The Extreme Discovery Protocol (EDP) is used to gather information about neighbor Extreme Networks
switches. EDP is used by the switches to exchange topology information. Information communicated
using EDP includes:
● Switch MAC address (switch ID)
● Switch software version information
● Switch IP address
● Switch VLAN IP information
● Switch port number
● Switch configuration data: duplex and speed
EDP is enabled on all ports by default. EDP enabled ports advertise information about the Extreme
Networks switch to other switches on the interface and receives advertisements from other Extreme
Networks switches. Information about other Extreme Networks switches is discarded after a timeout
interval is reached without receiving another advertisement.
To disable EDP on one or more ports, use the following command:
disable edp ports [<ports> | all]

To enable EDP on specified ports, use the following command:
enable edp ports [<ports> | all]
To clear EDP counters on the switch, use the following command:
clear counters edp
This command clears the following counters for EDP protocol data units (PDUs) sent and received per
EDP port:
● Switch PDUs transmitted
● VLAN PDUs transmitted
● Transmit PDUs with errors
● Switch PDUs received
● VLAN PDUs received
● Received PDUs with errors
To view EDP port information on the switch, use the following command:
show edp
Additionally, you view EDP information by using the following command:
show edp port <ports> detail
To configure the advertisement interval and the timeout interval, use the following command:
configure edp advertisment-interval <timer> holddown-interval <timeout>
Refer to “Displaying Port Configuration Information” for information on displaying EDP status.
Software-Controlled Redundant Port and

Smart Redundancy
Using the software-controlled redundant port feature you can back up a specified Ethernet port
(primary) with a redundant, dedicated Ethernet port; both ports are on the same switch. If the primary
port fails, the switch will establish a link on the redundant port and the redundant port becomes active.
Only one side of the link must be configured as redundant because the redundant port link is held in
standby state on both sides of the link. This feature provides very fast path or network redundancy.
NOTE
You cannot have any Layer 2 protocols configured on any of the VLANs that are present on the ports.
Smart Redundancy is a feature that allows control over how the failover from a redundant port to the
primary port is managed. If this feature is enabled, which is the default setting, the switch attempts to
revert to the primary port as soon as it can be recovered. If the feature is disabled, the switch attempts
only to recover the primary port to active if the redundant port fails.

Software-Controlled Redundant Port and Smart Redundancy
A typical configuration of software-controlled redundant ports is a dual-homed implementation

(Figure 1). This example maintains connectivity only if the link between switch A and switch B remains
open; that link is outside the scope of the software-controlled port redundancy on switch C.
Figure 1: Dual-homed implementation for switch C
Switch A Switch B
Primary Redundant
Link Link
Switch C
XOS002
In normal operation, the primary port is active and the software redundant switch (switch C in
Figure 1) blocks the redundant port for all traffic, thereby avoiding a loop in the network. If the switch
detects that the primary port is down, the switch unblocks the redundant port and allows traffic to flow
through that redundant port.
NOTE
The primary and redundant ports must have identical VLAN membership.
You configure the software-controlled redundant port feature either to have the redundant link always
physically up but logically blocked or to have the link always physically down. The default value is to
have the link physically down, or Off.
By default, Smart Redundancy is always enabled. If you enable Smart Redundancy, the switch
automatically fails over to the redundant port and returns traffic to the primary port after connectivity
is restored on that port. If you do not want the automatic restoration of the primary link when it
becomes active, disable Smart Redundancy.
Guidelines for Software-Controlled Redundant Ports and Port

Groups
Software-controlled redundant ports and port groups have the following limitations:
● You cannot have any Layer 2 protocols configured on any of the VLANs that are present on the
ports. (You will see an error message if you attempt to configure software redundant ports on ports
with VLANs running Layer 2 protocols.)
● The primary and redundant ports must have identical VLAN membership.
● The master port is the only port of a load-sharing group that can be configured as either a primary
or redundant port. Also, all ports on the load-sharing group must fail before the software-controlled
redundancy is triggered.
● You must disable the software redundancy on the master port before enabling or disabling load
sharing.

● You can configure only one redundant port for each primary port.
● Recovery may be limited by FDB aging on the neighboring switch for unidirectional traffic. For bi-
directional traffic, the recovery is immediate.
NOTE
On the BlackDiamond 10808 switch, on 10 Gbps modules with a serial number lower than 804405-00-09, the
software redundant port feature cover only those failures where both the TX and RX paths fail. If a single strand of
fiber is pulled on these ports, the software redundant port cannot correctly recover from the failure.To display the
serial number of the module, use the show slot <slot_number> command. (All the modules on the
BlackDiamond 8800 series switch have this serial number or higher.)
Configuring Software-Controlled Redundant Ports

When provisioning software-controlled redundant ports, configure only one side of the link as
redundant. In Figure 1 only the ports on switch C would be configured as redundant.
NOTE
To enable the software-controlled redundant port feature, the primary and redundant ports must have identical VLAN
membership.
To configure a software-controlled redundant port, use the following command:
configure ports <primaryPort> redundant <secondaryPort> {link [on | off]}
The first port specified is the primary port. The second port specified is the redundant port.
To unconfigure a software-controlled redundant port, use the following command and enter the
primary port(s):
unconfigure ports <port_list> redundant
To configure the switch for the Smart Redundancy feature, use the following command:
enable smartredundancy <port_list>
To disable the Smart Redundancy feature, use the following command:
disable smartredundancy <port_list>
Verifying Software-Controlled Redundant Port Configurations

You can verify the software-controlled redundant port configuration by issuing a variety of CLI
commands.
To display the redundant ports as well as which are active or members of load-sharing groups, use the
following command:
show ports redundant

Configuring Automatic Failover for Combination Ports—Summit X450, X450a, and X450e Series of Switches Only
To display information on which ports are primary and redundant software-controlled redundancy
ports, use the following command:
Refer to “Displaying Port Configuration Information” for more information on the show ports
information command.
Configuring Automatic Failover for Combination Ports—

Summit X450, X450a, and X450e Series of Switches
Only
The Summit X450, X450a, and X450e series has gigabit Ethernet ports; you configure automatic failover
using the four combination ports. These ports are called combination ports because either the fiber port
or the copper port is active, but they are never active concurrently. These ports, also called redundant
ports, are shared PHY copper and fiber ports.
If you plan to use the automatic failover feature, ensure that port settings are set correctly for
autonegotiation. Summit X450 series switch ports do not advertise or support flow control frames.
NOTE
You may experience a brief episode of the link going down and recovering during the failover.
To display the port type currently used as well as the preferred media setting, use the following
command:
Refer to “Displaying Port Configuration Information” for more information on the show ports
information command.
There are four ports on the Summit X450, X450a, and X450e series of switches that are designed as
combination ports for uplink redundancy. When sharing ports, only the fiber medium or only the
copper medium can be active at one time. If the copper medium goes down while transmitting packets,
the fiber medium activates and becomes the primary link; and vice-versa.
See Figure 2 for a diagram of these combination ports on the Summit X450-24t switch and Figure 3 for a
diagram of these combination ports on the Summit X450-24x switch (both switches have ports 1 to 4 as
the combination ports). If copper medium 1 goes down while transmitting packets, fiber medium 1
activates and becomes the primary link, and vice-versa.

Figure 2: Redundancy cabling on the Summit X450-24t switch
1 3 1 3
2 4 2 4
S450_004
Figure 3: Redundancy cabling for the Summit X450-24x switch
1 3 1 3
2 4 2 4
S450_005
The switch determines whether the port uses the primary or redundant media based upon the order in
which the connectors are inserted into the switch. When the switch senses a mini-GBIC and a copper
connector are inserted, the switch enables the uplink redundancy feature. As an example on the Summit
X450-24t switch (which has ports 1 to 4 as the combination ports), if you insert mini-GBICs into fiber
port 1 and fiber port 3 first and then connect copper ports 1 and 3, the switch assigns ports 1 and 3 as
redundant ports.
Hardware determines when a link is lost and swaps the primary and redundant ports to maintain
stability. After a failover occurs, the switch keeps or sticks with the current port assignment until there
is another failure or until a user changes the assignment using the CLI. To change the uplink failover
assignment, use the following command:
configure ports <port_list> preferred-medium [copper | fiber] {force}
The default preferred-medium is fiber. If you use the force option, it disables automatic failover. If you
force the preferred-medium to fiber and the fiber link goes away, the copper link is not used, even if
available.

Displaying Port Configuration Information
Displaying Port Configuration Information

You display summary port configuration information using the show ports {mgmt | <port_list>}
configuration {no-refresh} and show ports {mgmt | <port_list>} information {detail},
commands.
The show ports configuration command shows you either summary configuration information on
all the ports, or more detailed configuration information on specific ports. If you specify the no-
refresh parameter, the system displays a snapshot of the data at the time you issue the command.
The show ports information command shows you either summary information on all the ports, or
more detailed information on specific ports. The output from the command differs very slightly
depending on the platform you are using.
Beginning with ExtremeXOS software version 11.3 you can display real-time port utilization
information, by issuing the following command:
show ports {mgmt | <port_list> | stack-ports <stacking-port-list>} utilization

{bandwidth | bytes | packets}
When you use a parameter (packets, byte, or bandwidth) with the above command, the display for the
specified type shows a snapshot per port when you issued the command.


6 Universal Port
● Universal Port on page 165

■ Events that Trigger Profiles on page 169
■ Handling Profile Execution Errors on page 170
■ Universal Port Variables on page 171
● Sample Profile Configurations on page 172
■ Static Profile Example on page 172
■ Configuring a Profile with QoS Support on page 173
Universal Port
Universal Port is a flexible framework that allows the switch to take direct action based on events. The
focus is on edge ports, with the switch applying dynamic profiles (such as security and QoS policy)
based on user login, and PoE configuration based on LLDP device discovery.
NOTE
Beginning in ExtremeXOS Release 12.0, the Universal Port feature is available with an Edge license. In earlier
releases, an Advanced Edge license is required.
Universal Port management allows auto-configuration of the switch based on dynamic network events.
This is achieved by a combination of scripting and dynamic command execution capabilities.
The responsibility for assigning dynamic policies is moved to the RADIUS server via creation of
security profiles in switches that can be selected via VSA's based on various trigger events. This closes
the gap between the time the user is authenticated and allowed on the network and the time the
dynamic ACL's are deployed on the switch. In addition it provides a more flexible framework that can
respond to a larger number of trigger events such as discovery or time-based policies.
The CLI-based scripting capability in ExtremeXOS allows users to create policy templates with variables
for the match criteria. This allows you to re-use templates with different address ranges and
parameters. For example, two software development projects might have the same logical set of rules,
but different address ranges, and both could use the same policy template.
NOTE
In the CLI, “upm” is used as an abbreviation for Universal Port management.
There are two types of profiles you can configure in the system: Static and Dynamic. Profiles are
configured and run via the CLI. You use CLI commands to tell the switch to run a static profile.

Universal Port
Dynamic profiles are triggered by a defined set of events, which provide the required set of arguments
to help enforce the policy determined by that profile.
The term "profile" is distinct from the term "policy" because a policy is only one particular application
of a profile. A profile is a set of commands that cause a port configuration change.
Typically, static profiles are not event driven and do not change frequently. They are not specific to or
dependent on the specific device or user connected to a port. Changes made by static profiles are
persistent, i.e. they are saved in the configuration and thus are preserved across system reboots. Use
these profiles to implement scripts that parameterize and simplify complex configuration tasks like
Netlogin.
Static profiles can be remotely implemented by scripts running locally on the switch.
Static profiles allow all CLI commands to be executed.
By default, Universal Port profiles run in non-persistent mode. Profiles that run via event triggers are
Dynamic, which means the Universal Port prepends "configure cli mode non-persistent" to the script.
This will configure the subsequent commands that are in the profile to be run. The dynamic profile
configuration will not be restored across reboots on the system.
Dynamic profiles are applied based on the occurrence of a dynamic profile trigger. These triggers fall
into two categories: profile activation and profile deactivation triggers. A profile is run when a trigger
event occurs on a port.
There is no automatic rollback of dynamic profiles. You can roll back the configuration to any previous
state by saving information in variables that are retrievable for accomplishing the rollback.
There is no profile hierarchy. You should not configure conflicting profiles that might create different
results, based on the sequence of events.
Profiles can be configured to respond to either user events (such as NetLogin) or device events, such as
LLDP.
Device Profiles
Device profiles are applied when specific devices are detected or un-detected. They may also be
triggered by timers.
To test a profile, use the following command:
>>run upm profile <profile-name> {event <event-name> } {variables

<variable-string>}
If the variables keyword is not present, but an events variable is specified, the user will be prompted
for various environment variables appropriate for the event, including the VSA string for user
authentication. The variables are not validated for correct syntax. To see the profile execution history,
use the show upm history CLI command.
The following parameters will be configurable on the switch ports when the device profiles execute:
● LLDP Parameters
■ VLAN Name
■ Port VLAN ID
■ Power Conservation Mode

Universal Port
■ Avaya File Server

■ Avaya Call server
■ 802.1Q Framing
LLDP must be enabled on the port.
User or Security Profiles

User (security) profiles are applied dynamically when a user is authenticated or un-authenticated by
any of the following techniques:
● MAC based Netlogin

● 802.1x based Netlogin
● Web based Netlogin
The user profile can be applied to a port or a port-list.
Universal Port profiles can be timer-triggered.
User profiles applied to specific users will not be configurable through the CLI. User profiles that need
to be applied will be configurable through the RADIUS Server as is the requirement for any net-login or
CLI user.
The same profile will be applied at logon and logoff. The profile will take different actions based on the
event that triggered it. There is an option to specify a different profile for LOGOFF
The User profile name and a few parameters are configurable on the RADIUS Server using a new VSA:
EXTREME-SECURITY-PROFILE. The RADIUS dictionary value for the EXTREME-SECURITY-PROFILE
is 212.
The VSA syntax is shown below:
<profile-name> <var1>=<value1>;<var2>=<value2>;…
For Example:
EXTREME-SECURITY-PROFILE= "p1 QOS=\"QP8\";LOGOFF-PROFILE=P2;"
The string shown above in italicized font after the first equal sign (=) forms the VSA string that the
RADIUS server will send to the switch. The RADIUS server will send its dictionary value to the switch
instead of the string “EXTREME-SECURITY-PROFILE.”
These variables are available to the profile at run-time as arguments that can be used for configuration.
If a "LOGOFF-PROFILE" variable is not present, then UP management will launch the
"PROFILENAME" profile when the user is logged off (the same profile that was executed when the user
logged in). If a "LOGOFF-PROFILE" variable is present, but the UP management is not configured for
the event, then an error message is logged and nothing is done.
A user or security policy is applied when any of the following occurs:
● It is received as a RADIUS attribute on successful authentication of a client.

● A CLI command is executed to apply it.
● A timer event occurs that was probably added as part of user authentication.

Universal Port
Sample RADIUS Server Configuration

The sample RADIUS files shown below are taken from FreeRADIUS Version 1.1.2, built on Aug. 7, 2006.
These configurations are for an Avaya phone implementation.
Sample RADIUS Server Dictionary file with Extreme Networks Vendor Specific Attributes (VSAs):
VENDOR Extreme 1916
ATTRIBUTE Extreme-CLI-Authorization 201 integer Extreme

ATTRIBUTE Extreme-Shell-Command 202 string Extreme
ATTRIBUTE Extreme-Netlogin-Vlan 203 string Extreme
ATTRIBUTE Extreme-Netlogin-Url 204 string Extreme
ATTRIBUTE Extreme-Netlogin-Url-Desc 205 string Extreme
ATTRIBUTE Extreme-Netlogin-Only 206 integer Extreme
ATTRIBUTE Extreme-User-Location 208 string Extreme
ATTRIBUTE Extreme-Netlogin-Vlan-Tag 209 integer Extreme
ATTRIBUTE Extreme-Netlogin-Extended-Vlan 211 string Extreme
ATTRIBUTE Extreme-Security-Profile 212 string Extreme
VALUE Extreme-CLI-Authorization Disabled 0

VALUE Extreme-CLI-Authorization Enabled 1
VALUE Extreme-Netlogin-Only Disabled 0
VALUE Extreme-Netlogin-Only Enabled 1
# End of Dictionary
The RADIUS "users" file, which contains all the login parameters, must match the authentication type of
the Avaya users.
NOTE
The "users" file is case-sensitive and punctuation is very important for FreeRADIUS.
Sample RADIUS users file for 802.1x authentication:

<username> Auth-Type := EAP, User-Password == "12345"
Session-Timeout = 60,
Termination-Action = 1,
Extreme-Security-Profile = "user-auth LOGOFF-PROFILE=avaya-
remove;qos=\"QP1\";",
Extreme-Netlogin-Vlan = voice-avaya
Sample RADIUS users file for MAC based authentication: .

00040D9D12AF Auth-Type := Local, User-Password == "00040D9D12AF"
Termination-Action = 1,
Extreme-Security-Profile = "user-auth LOGOFF-PROFILE=avaya
remove;qos=\"QP1\";",
Extreme-Netlogin-Vlan = voice-avaya
NOTE
The Extreme-Security-Profile and Extreme-Netlogin-Vlan lines must correspond to an existing profile and VLAN on
the Extreme switch.

Universal Port
Multiple Profiles on the Same Port

You can configure multiple user profiles on a port. For instance, you might have a user-authenticated
profile and a user-unauthenticated profile.
For device events, you can configure only one profile on a port.
Events that Trigger Profiles

Table 17 shows the system triggers that might lead to the execution of a particular profile.
Table 17: Events that Trigger Profiles

Trigger Condition
Device-Detect Specific device was detected by the system.
Device-Undetect Specific device is no longer present. This could also be triggered by a timeout. This
allows the restoration of port properties to a known state.
User-Authenticated Specified user authenticated.
User-Unauthenticated Specified authenticated user has been unauthenticated.
This allows the restoration of port properties to a known state.
Timer-AT Timer scheduled to occur AT a specified time has occurred.
Timer-AFTER Timer scheduled to occur AFTER specified interval has occurred.
User-Request This profile was triggered remotely via the CLI.
The following example configures a profile to an event:
create upm profile sample1

configure port $EVENT.USER_PORT qosprofile qp1
.
configure upm event user-authenticated profile sample1 ports 1
show upm profile (to see the above binding)
Table 18 shows the CLI commands that can be executed in the dynamically applied profiles.
Note that commands executed dynamically are not saved across system reboots.
Table 18: Configuration Commands That Can Be Executed in Non-Persistent Mode

CLI Commands
ACL Commands
Dynamic ACL syntax will allow the application of all ACLs
configure access-list add <rule-name> [after
<rule> |before <rule>| first| last] ports
<portlist> [ingress |egress]
configure access-list delete <rule-name> ports
<portlist> [ingress|egress]

Universal Port
Table 18: Configuration Commands That Can Be Executed in Non-Persistent Mode (Continued)
CLI Commands
LLDP
configure lldp ports <portlist> [advertise|don't-
advertise]...
Port
[enable|disable] port <portlist>
[enable|disable]jumbo-frame ports <portlist>
PoE Budget
Enable PoE
config inline-power priority [critical|high|low]
ports <portlist>
unconfig inline-power priority ports <portlist>
PoE Limits
config inline-power operator-limit <mwatts> ports
<portlist>
VLAN
config vlan <vlan> add port <portlist> tagged
config vlan <vlan> add port <portlist> untagged
config ip-mtu <mtu> vlan <vlan-name>
QOS/Rate-limiting
802.1p priority assignment to traffic on a port
config port <portlist> qosprofile <qosprofile>
Show Commands
All show commands can be executed in non-persistent mode.
Executing Static Profiles

Profiles can be run from the command line interface, by configuring the system to run as it would when
the trigger events happen. This facility is provided to allow you to test how the system behaves when
the actual events happen. But the actual configuration is applied to the switch when the profile is run.
Handling Profile Execution Errors

To conserve resources, the switch stores only the last execution log for the profile that resulted in an
error.
Use the following command to see a tabular display showing the complete history of the last 100
profiles run:
show upm history
Use the detail keyword to display the actual executions that have happened when the profile was run.

Universal Port
Use the following command to display a specific execution that was run:
show upm history exec-id <number>
Select the exec-id number from the list in the tabular display.
Universal Port Variables

This section defines the information that will be available to any profile on execution, based on the
event that triggered the profile or CLI session.
You must enable CLI scripting before using these variables or executing a script.
Common Variables
Table 19 shows the variables that are always available for use by any script. These variables are setup
for use before a script or profile is executed.
Table 19: Common Variables

Variable Syntax Definition
$STATUS Status of last command execution.
$CLI.USER UserName who is executing this CLI.
$CLI.SESSION_TYPE Type of session of the user.
$EVENT.NAME This is the event that triggered this profile. See Table 17 for a list of triggers.
$EVENT.TIME Time this event occurred. The time will be in seconds since epoch.
$EVENT.TIMER_TYPE PERIODIC or NON_PERIODIC.
$EVENT.TIMER_NAME Name of the timer that the Universal Port is invoking.
$EVENT.TIMER_DELTA Time difference when the timer fired and when the actual shell was run in seconds.
$EVENT.PROFILE Name of the profile that is being run currently.
User Profile Variables

Table 20 shows the variables available to user profiles.
Table 20: User Profile Variables

$EVENT.USERNAME Name of user authenticated. This would be a string with the MAC address for MAC-
based user-login
$EVENT.NUMUSERS Authenticated supplicants on this port after this event occurred
$EVENT.USER_MAC MAC address of the user
$EVENT.USER_PORT Port associated with this event
$EVENT.USER_VLAN Vlan associated with this event
$EVENT.USER_IP IP address of the user if applicable, else blank

Universal Port
Device Profile Variables

Table 21 shows the variables available to Device Profiles.
Table 21: Device Profile Variables

$EVENT.DEVICE Device identification string
Possible values for EVENT.DEVICE are: AVAYA_PHONE, GEN_TEL_PHONE, ROUTER,
BRIDGE, REPEATER, WLAN_ACCESS_PT, DOCSIS_CABLE_SER, STATION_ONLY
and OTHER.
These strings correspond to the devices that the LLDP application recognizes and
reports to the Universal Port management application.
$EVENT.DEVICE_IP The IP address of the device (if available). Blank if not available.
$EVENT.DEVICE_MAC The MAC address of the device (if available). Blank if not available.
$EVENT.DEVICE_POWER The power of the device in milliwatts (if available). Blank if not available.
$EVENT.DEVICE_MANUF The manufacturer of the device.
ACTURER_NAME
$EVENT.DEVICE_MODEL Model name of the device
_NAME
Inside the profile script, you can create and use variables of your own. You are allowed to use and
modify the standard variables, but these changes to existing variables will be limited to the execution
context of the script.
Persistence Modes
To configure persistent command execution use the following command:
configure cli mode persistent
To configure non-persistent command execution use the following command:
configure cli mode non-persistent
The CLI mode will be implicitly set to the non-persistent mode when executing any dynamic profile.
Both modes allow the user to execute commands.
Sample Profile Configurations
Static Profile Example

The following configuration creates a profile and runs it statically:
* BD-10808.4 # Create upm profile p1
Enable port 1:1
.
* BD-10808.4 #run upm profile p1
* BD-10808.4 # show upm history exec 8006

Sample Profile Configurations
UPM Profile: p1
Event: User Request , Time run: 2006-10-18 11:56:15
Execution Identifier: 8006 Execution Status: Pass
Execution Information:
1 # enable cli scripting
2 # set var EVENT.NAME USER-REQUEST
3 # set var EVENT.TIME 1161172575
4 # set var EVENT.PROFILE p1
5 # enable por 1:1
Configuring a Profile with QoS Support

The example below can be used with a Summit X450 series switch that supports QoS profiles qp1 and
qp8. When the user or phone logs in with a particular MAC address, the script configures the QoS
profile configured by the user in the RADIUS server for the "USER-AUTHENTICATED" event. In this
example, the user set the QoS profile to be qp8.
Below is the upm profile configuration for the above example:
Create upm profile p1
set var z1 $uppercase($EVENT.USER_MAC)

set var z2 $uppercase(00:04:0d:9d:12:a9)
#show var z1
#show var z2
if ($match($EVENT.NAME, USER-AUTHENTICATED) == 0) then
if ($match($z1, $z2) == 0) then
configure port $EVENT.USER_PORT qosprofile $QOS
endif
endif
.
You must configure the netlogin, RADIUS server, and Universal Port management on the switch as part
of the user log-in authentication process. For the example below, the user is the MAC address of the
phone.
00040D9D12A9 Auth-Type := local, User-Password == "test"

Extreme-security-profile = "p1 QOS=\"QP8\";LOGOFF-PROFILE=p2;VLAN=\"voice-
test\";"

Universal Port

7 CLI Scripting
● CLI Scripting Capabilities on page 175

■ Scripting Variables on page 175
■ Enable | Disable CLI Scripting on page 176
■ Local Variables on page 176
■ Variable Manipulation on page 176
■ Session Variables on page 177
■ Control Structures on page 177
■ Operators on page 178
■ Built-In Functions on page 179
■ Error Handling on page 179
● CLI Scripting Examples on page 179
■ Sample Script to Create 100 VLAN's on page 179
CLI Scripting Capabilities

The CLI is enhanced to support the following features required to implement UP management:
● Scripting Variables on page 175
● Enable | Disable CLI Scripting on page 176
● Local Variables on page 176
● Variable Manipulation on page 176
● Session Variables on page 177
● Control Structures on page 177
● Operators on page 178
● Built-In Functions on page 179
● Error Handling on page 179
Scripting Variables
This section defines the information that will be available to any profile on execution, based on the
event that triggered the profile or CLI session.
NOTE
You must enable CLI scripting before using these variables or executing a script.

CLI Scripting
Common Variables
Table 22 shows the variables that are always available for use by any script. These variables are setup
for use before a script or profile is executed.
Table 22: Common Variables

$STATUS Status of last command execution.
$CLI.USER UserName who is executing this CLI.
$CLI.SESSION_TYPE Type of session of the user.
Enable | Disable CLI Scripting

The following command will allow the user to control scripting at the switch. Scripting will be disabled
by default for an interactive session.
[enable | disable] cli scripting {permanent}
You can view the various modes that have been set (scripting, persistent mode and scripting error
mode) using the following command:
show management
Local Variables
The CLI is enhanced to support variables. Some internal local variables are always available.
The format of a local variable (case insensitive) is $VARNAME.The variable name length is limited to 32
characters.
Variable Manipulation
To create and set the variable to the desired value, use the following command:
set var <varname> <expression>
To configure variable manipulations, use the following command:
set var <varname> ($v1 <operation> $v2)
Valid operations for numeric variables are: + - * / %
Examples:
● set var x 100
● set var x ($x + 2)
● set var y ($x - 100)
A variable can be referenced as $X or $(X).
If the variable name X contains special characters such as +-/*, then the variable needs to be enclosed in
parentheses.
● set var z ($(x) + 100)
If a variable already exists, it will be overwritten. No error message will be displayed.

CLI Scripting Capabilities
Only the set var CLI command supports expression evaluation.

To display all variables or a specified variables, use the following command:
show var {<varname>}
Session Variables
The profiles need some information repository to save the current profile state to be able to restore the
profile when the user logs off or the device times out. The LOAD VAR, SAVE VAR and DELETE VAR
commands accomplish variable management. Up to five variables can be imported or exported at a
time.
The syntax for these commands is:
load var key <key> [<var1> <var2> …]
save var key <key> [<var1> <var2> …]
delete var key <key>
The variables saved by the SAVE VAR command are saved using the specified key and are retrievable
and restored in the context that this profile was applied. They will be available to rollback events like
user-unauthenticate and device-undetect.
The key option allows you to save the data for this unique key. You can then retrieve this data based on
this key. You are responsible for generating unique keys. The system has a limited memory allocation to
store these variables.
The load var command allows you to import the appropriate set of variables.
The delete var command destroys the variables created.
Control Structures
The CLI is enhanced to introduce new control structures that allow the use of variables available to a
profile. The following control structures are supported. The expression must be enclosed in parentheses.
Conditional Execution
IF (<expression>) THEN
<statements>
ELSE
<statements>
ENDIF
Loop While Condition is TRUE

WHILE (<expression>) DO
<statements>
ENDWHILE
Nesting is supported up to five levels. The Ctrl-C key combination can be used to break out of any
While loop(s).
The operators mentioned in the “Operators” section below can be used in an expression in the set var
command or in an IF or WHILE condition.

CLI Scripting
If there is incorrect nesting of an IF condition or WHILE loops, an error message appears. If a user tries
to type more than five WHILE loops or five IF conditions, an error message appears. Breaking out of
any number of while(s) always clears all the while(s) .
Comments can be inserted by using the number sign (#).
Operators
The valid operators are listed below, grouped in decreasing order of precedence:
Table 23: Operators

Operand Action Comments
- Unary minus None of these operands may be applied to string operands, and
the bit-wise NOT operand may be applied only to integers.
+ Unary plus
~ Bit-wise NOT
! Logical NOT
* Multiply None of these operands may be applied to string operands, and
the remainder operand may be applied only to integers. The
/ Divide
remainder will always have the same sign as the divisor and an
% Remainder absolute value smaller than the divisor.
+ Add These operands are valid for any numeric operations.
- Subtract
<< Left shift These operands are valid for integer operands only. A right shift
always propagates the sign bit.
>> Right shift
< Boolean less Each operator produces 1 if the condition is true, 0 otherwise.
These operators may be applied to strings as well as numeric
> Boolean greater
operands, in which case string comparison is used.
<= Boolean less than or
equal
>= Boolean greater than or
equal
== Boolean equal Each operator produces a zero or one result. These operators
are valid for all operand types.
!= Boolean not equal
& Bit-wise AND This operator is valid for integer operands only.
^ Bit-wise exclusive OR This operator is valid for integer operands only.
| Bit-wise OR This operator is valid for integer operands only.
&& Logical AND This operator produces a result of 1 if both operands are non-
zero. Otherwise, it produces a result of 0. This operator is valid
for numeric operands only (integers or floating-point).
|| Logical OR This operator produces a result of 0 if both operands are zero.
Otherwise, it produces a result of 1.
This operator is valid for numeric operands only (integers or
floating-point).
x?y:z If-then-else (as in the C If x evaluates to non-zero, then the result is the value of y.
programming language) Otherwise the result is the value of z. The x operand must have
a numeric value.

CLI Scripting Examples
Built-In Functions
Table 24 shows the built-in functions:
Table 24: Built-in Functions

Syntax Function
$MATCH(string 1, string Compares the two strings s1 and s2. Returns 0 if string1 matches string2.
2)
It returns -1,0, or 1, depending on whether string1 is less than, equal to, or greater
than string2.
$UPPERCASE( string ) Returns the string uppercased.
Error Handling
The following modes will be set to control the error handling at the switch.
configure cli mode scripting abort-on-error
configure cli mode scripting ignore-error
The default error handling behavior is to ignore errors.
The user can change the options within the scripts.
CLI Scripting Examples
Sample Script to Create 100 VLAN's

The following script creates 100 VLANS with IP Addresses from 10.1.1.1/16 to 10.100.1.1/16:
enable cli scripting
Set var count 1
while ($count < 101) do
Create vlan v$count
configure vlan v$count ipaddress 10.$(count).1.1/16
set var count ($count + 1)
endwhile
show vlan
NOTE
You can use the command load script <script_name> to execute a script.

CLI Scripting

8 Link Layer Discovery Protocol

● LLDP Packets on page 183
● Transmitting LLDP Messages on page 184
● Receiving LLDP Messages on page 185
● Managing LLDP on page 185
● Supported TLVs on page 186
● Configuring LLDP on page 195
● Displaying LLDP Settings on page 201
Overview
Beginning with ExtremeXOS version 11.2, the software supports the Link Layer Discovery Protocol
(LLDP). LLDP is a Layer 2 protocol (IEEE standard 802.1ab) that is used to determine the capabilities of
devices such as repeaters, bridges, access points, routers, and wireless stations. ExtremeXOS version
11.2 LLDP support enables devices to advertise their capabilities and media-specific configuration
information and to learn the same information from the devices connected to it.
The information is represented in Type Length Value (TLV) format for each data item. The 802.1ab
specification provides detailed TLV information. The TLV information is contained and transmitted in
an LLDP protocol data unit (LLDPDU). Certain TLVs are mandatory and are always sent after LLDP is
enabled; other TLVs are optionally configured. LLDP defines a set of common advertisement messages,
a protocol for transmitting the advertisements, and a method for storing the information contained in
received advertisements. Beginning with ExtremeXOS version 11.5, the switch can receive and record
certain TLVs but not transmit these TLVs; they are TLVs originating from the power over Ethernet
(PoE) powered device (PD) connected to a port and certain inventory management TLVs.
LLDP provides a standard method of discovering and representing the physical network connections of
a given network management domain. LLDP works concurrently with Extreme Discovery Protocol
(EDP); it also works independently, you do not have to run EDP to use LLDP. The LLDP neighbor
discovery protocol allows you to discover and maintain accurate network topologies in a multivendor
environment.
The information distributed using LLDP is stored by its recipients in a standard Management
Information Base (MIB), making it possible for the information to be accessed by a Network
Management System (NMS) using a management protocol such as the Simple Network Management
Protocol (SNMP).
LLDP transmits periodic advertisements containing device information and media-specific configuration
information to neighbors attached to the same network. LLDP agents cannot solicit information from
other agents by way of this protocol. Beginning with ExtremeXOS version 11.5, the switch can transmit
and receive LLDP media endpoint discovery (MED) TLVs. Once enabled, the LLDP MED TLVs
messages are sent only after a neighbor is detected sending out LLDP MED TLVs; the LLDP MED TLVs

Link Layer Discovery Protocol
are transmitted only after the switch receives an LLDP MED TLV from a neighbor. For this reason, two
connected switches will never exchange LLDP MED TLVs.
NOTE
Network connectivity devices wait to detect LLDP MED TLVs from endpoints before they send out LLDP MED TLVs;
so 2 network connectivity devices will not exchange LLDP MED messages.
The TLV format with link layer control frames is used to communicate with other LLDP agents. LLDP
agents also receive link layer control frames, extract the information from TLVs, and store them in
LLDP MIB objects.
If the information values from the device change at any time, the LLDP agent is notified. The agent then
sends an update with the new values, which is referred to as a triggered update. If the information for
multiple elements changes in a short period, the changes are bundled together and sent as a single
update to reduce network load.
You configure LLDP per port, and each port can store received information for a maximum of four
neighbors.
NOTE
LLDP runs with link aggregation.
Beginning with ExtremeXOS version 11.5, the device can also support the following types of LLDP
TLVs:
● Avaya-Extreme Networks proprietary TLVs
● LLDP media endpoint discovery (MED) TLVs
The software supports several TLVs that are proprietary to Avaya and Extreme Networks (avaya-
extreme TLVs). These TLVs primarily advertise and receive information for Avaya voice over IP (VoIP)
telephones. Some of these TLVs primarily concern the PD; the PD receives these TLVs, but does not
transmit them. (Refer to Table 26 for a listing of the proprietary TLVs that are only received by the
switch.) These proprietary LLDPs are transmitted and received as soon as you enable LLDP and
configure the specified TLVs.
LLDP MED TLVs are sent only after the device detects a neighbor transmitting LLDP MED TLVs; and
the LLDP MED TLVs must be configured and enabled prior to the detection. You must enable the
LLDP-MED capabilities TLV before configuring and enabling any other LLDP MED TLVs. Likewise,
when disabling the LLDP MED TLVs, you must disable the LLDP-MED capabilities TLVs only after you
have disabled all other LLDP MED TLVs.
The LLDP MED protocol extension introduces a new feature called MED fast start, which is
automatically enabled when the LLDP MED capabilities TLV is enabled. When a new MED-capable
device is detected, the detecting switch sends out an LLDPDU each 1 second for the configured number
of times (called the repeat count). By default, the switch sends out the LLDPDU each 1 second 3 times;
you can change this repeat count 10 between 1 and 0 times. Once the repeat count is reached, the
configured transmit interval value is used between LLDPDUs. Use the following command to configure
the repeat count:
configure lldp med fast-start repeat-count <count>

LLDP Packets
NOTE
The fast-start feature is automatically enabled, at the default level of 3, when you enable the LLDP MED capabilities
TLV on the port.
You must enable SNMP traps separately for the LLDP MED traps; they are disabled by default. To
enable the LLDP MED SNMP traps, issue the following command:
enable snmp traps lldp-med {ports [all | <port_list>]}
In addition, the switch can receive, but not transmit, the LLDP MED inventory management TLVs.
(Refer to Table 26 for a listing of these inventory management TLVs.)
LLDP Packets
You configure the device to transmit messages, to receive messages, or both.
LLDP is enabled and configured per port.
Multiple advertisements messages (or TLVs) are transmitted in one LAN packet, the LLDPDU
(Figure 4). The LLDP packet contains the destination multicast address, the source MAC address, the
LLDP EtherType, the LLDPDU data, and a frame check sequence (FCS). The LLDP multicast address is
defined as 01:80:C2:00:00:0E, and the EtherType is defined as 0x88CC.
Figure 4: LLDP packet format

DA SA Ethertype Data + Pad
LLDP_Multicast Source MAC 88-CC LLDPDU FCS

Address Address
6 6 2 1500 4 Octets
XOS005
The following characteristics apply to LLDP packets:

● They are IEEE 802.3 Ethernet frames.
● The frames are sent as untagged frames.
● The frames are sent with a link-local-assigned multicast address as destination address.
● The Spanning Tree Protocol (STP) state of the port does not affect the transmission of LLDP frames.
The length of the packet cannot exceed 1500 bytes. As you add TLVs, you increase the length of the
LLDP frame. When you reach 1500 bytes, the remaining TLVs are dropped. Extreme Networks
recommends that you advertise information regarding only one or two VLANs on the LLDP port, to
avoid dropped TLVs.
If the system drops TLVs because of exceeded length, the system logs a message to the EMS and the
show lldp statistics commands shows this information under the Tx Length Exceeded field.

NOTE
The LLDPDU has a maximum of 1500 bytes, even with jumbo frames enabled. TLVs that exceed this limit are
dropped.
Transmitting LLDP Messages

In transmit mode, the Extreme Networks switch periodically sends out an untagged LLDPDU frame
that contains the mandatory LLDP TLVs as well as the configured optional TLVs. The LLDP agent
running on the Extreme Networks switch passes serially through the list of ports that are enabled for
LLDP and periodically transmits an LLDP frame containing the mandatory TLVs and any configured
optional TLVs. The mandatory TLVs and the system description TLV are automatically transmitted
after you enable LLDP.
The following information, when configured, can be sent at regular intervals:

● Chassis ID (mandatory)
● Port ID (mandatory)
● Time-to-live (mandatory)
● Port description
● System name
● System description (sent by default)
● System capabilities
● Management address
● 802.1-specific information
■ VLAN name
■ Port VLAN ID
■ Port and protocol VLAN ID
● 802.3-specific information
■ MAC/PHY
■ Power via MDI
■ Link aggregation
■ Maximum frame size
● Avaya-Extreme Networks proprietary information
■ Power conservation request
■ Call server
■ File server
■ 802.1Q framing information
● MED extensions (Once enabled, these are sent only when the switch detects a neighbor on the port
that transmits at least one MED TLV)
■ MED capabilities
■ Network policy
■ Location ID
■ Extended information on Power via MDI

Receiving LLDP Messages
This information is obtained from memory objects such as standard MIBs or from system management
information.
Receiving LLDP Messages

The LLDP agent running on an Extreme Networks switch receives LLDPDUs, parses the messages, and
stores the information in a remote device database. Unrecognized TLVs are also stored in the remote
device database, in order of TLV type. The information is purged after the configured timeout interval,
unless it is refreshed by the remote LLDP agent.
You access the messages from the neighbors with SNMP or the CLI. To access this information with the
CLI, use the show lldp neighbors detailed command. (You must use the detailed variable to
display this information.)
Each port can store LLDP information from a maximum of four neighbors.
Beginning with ExtremeXOS version 11.5, the software receives several TLVs that it does not transmit,
as follows:
● Avaya-Extreme Networks proprietary information

■ PD conservation level support (includes the PD’s current conservation level, typical power value,
and maximum power value, as well as power conservation levels available to that PD)
■ Endpoint IP address (including the mask and gateway addresses)
■ Converged Network Analyzer (CNA) server IP address
● Inventory management LLDP MED TLVs:
■ Hardware revision
■ Firmware revision
■ Software revision
■ Serial number
■ Manufacturer name
■ Model name
■ Asset ID
Managing LLDP
LLDP can work in tandem with EDP. LLDP is disabled by default, and EDP is enabled by default. LLDP
information is transmitted periodically and stored for a finite period. You access the information using
SNMP. A port configured to receive LLDP messages can store information for up to four neighbors.
You manage LLDP using the CLI and SNMP. (Refer to ExtremeXOS Command Reference Guide for
complete information on configuring, managing, and displaying LLDP.)
The LLDP MED TLVs begin transmission only after detecting LLDP MED TLVs transmitted by a
neighbor. After you enable LLDP, you can set a variety of time periods for the transmission and storage
of the LLDP messages (or you can use the default values), as follows:
● Reinitialization period (default is 2 seconds)

● Delay between LLDP transmissions (default is 2 seconds)—applies to triggered updates, or updates

that are initiated by a change in the topology
● Transmit interval (default is 30 seconds)—applies to messages sent periodically as part of protocol
● Time-to-live (TTL) value (default is 2 minutes)—time that the information remains in the recipient’s
LLDP database
NOTE
Once the LLDP MED TLVs begin transmitting (after detecting LLDP MED TLVs from a connected endpoint), those
TLVs are also controlled by these timers.
Each time a device receives an LLDP advertisement packet, the device stores the information and
initializes a timer that is compared to the TTL value of the packet. If the timer reaches the TTL value,
the LLDP agent deletes the stored information. This action ensures that only valid information is stored
in the LLDP agent.
After you enable LLDP, you can enable the LLDP-specific SNMP traps; the traps are disabled by
default. After you enable the LLDP-specific traps, the systems send all LLDP traps to the configured
trap receivers. You configure the period between the system sending SNMP notifications; the default
interval is 5 seconds. LLDP configurations are saved across reboots when you issue the save
configuration command.
The system logs EMS messages regarding LLDP, including when optional TLVs exceeding the 1500-byte
limit are dropped and more than 4 neighbors are detected on a port.
When both IEEE 802.1x and LLDP are enabled on the same port, LLDP packets are not sent until one or
more clients authenticate a port. Also, incoming LLDP packets are only accepted if one or more clients
are authenticated.
You can configure an optional TLV to advertise or not to advertise the device’s management address
information to the port’s neighbors. With ExtremeXOS, when enabled, this TLV sends out the IPv4
address configured on the management VLAN. If you have not configured an IPv4 address on the
management VLAN, the software advertises the system’s MAC address. LLDP does not send out IPv6
addresses in this field.
Supported TLVs
The TLVs are contained in the LLDPDU portion of the LLDP packet, and the LLDPDU cannot exceed
1500 bytes. Some TLVs are mandatory according to the 802.1ab standard, and the rest are optional. The
mandatory and system description TLVs are included by default as soon as you enable LLDP. The
system description TLV is enabled by default on the ExtremeXOS LLDP implementation. Additionally
some TLVs can be repeated in one LLDP.
NOTE
To avoid exceeding the 1500-byte limit, Extreme Networks recommends sending information on only one or two
VLANs on the LLDP port. Any TLVs that exceed the limit are dropped.

Supported TLVs
The following TLVs are enabled by default when LLDP transmit is enabled on a port:
● Chassis ID
● Port ID
● Time to live
● System description
● End-of-LLDP PDU
All of these TLVs that are sent by default are mandatory for the protocol and cannot be disabled, except
the system description. You can configure the system not to advertise the system description when
LLDP is enabled; the other four TLVs cannot be configured not to advertise. Table 25 lists all the defined
TLVs, if they are included by default after you enable LLDP, if they can be configured, if they are
mandatory or optional, and if you can repeat that TLV in one LLDP packet.
NOTE
Refer to ExtremeXOS Command Reference Guide for complete information on configuring LLDP using the CLI.
Table 25: Available TLVs for transmission

Included by
Name default User configurable Repeatable Comments
Chassis ID X Mandatory TLV
Port ID X Mandatory TLV
Time to live (TTL) X Mandatory TLV
Port description X
System name X
System description X X
System capabilities X
Management address X X ExtremeXOS sends only
1 TLV
VLAN name X X
Port VLAN ID X
Port and protocol VLAN ID X X
Protocol identity X Not supported
MAC/PHY configuration/status X
Power via MDI X
Link aggregation X
Maximum frame size X
PoE conservation level request X Avaya-Extreme
Networks proprietary
TLV
Call server X Avaya-Extreme
TLV
File server X Avaya-Extreme
TLV

Table 25: Available TLVs for transmission (Continued)

Included by
Name default User configurable Repeatable Comments
802.1Q framing X Avaya-Extreme
TLV
LLDP MED capabilities X Must be enabled before
any other MED TLV,
and must be disabled
after all other MED
TLVs
MED TLVs transmit
only after detecting a
neighbor transmitting
MED TLVs
Network policy X X Content cannot be
configured by SNMP
MED TLVs transmit
MED TLVs
Location ID X MED TLVs transmit
MED TLVs
Extended power via MDI X Can be enabled only on
a PoE-capable port
MED TLVs transmit
MED TLVs
End-of-LLDP PDU X Mandatory TLV
NOTE
Refer to ExtremeXOS Command Reference Guide for complete information on configuring LLDP using the CLI.
Table 26 lists the TLVs that the switch can receive, but not transmit. To receive any of these TLVs, the
port must be enabled for LLDP. After you enable LLDP receiving on the switch, all TLVs are received
(even if the LLDP MED capabilities TLV is not enabled). To display these received messages, use the
show lldp neighbor detailed CLI command.
Table 26: Available TLVs for reception

Name Type Comments
PoE Conservation level support Avaya-Extreme Sent by PD to advertise current power consumption
level and current conservation level including typical
power value, maximum power value, and available
conservation power levels
IP phone address Avaya-Extreme Advertises IP address and mask, as well as default
gateway address
CNA server Avaya-Extreme Advertises IP address of Converged Network Analyzer
(CNA)

Supported TLVs
Table 26: Available TLVs for reception (Continued)

Name Type Comments
Hardware revision MED
Firmware revision MED
Software revision MED
Serial number MED
Manufacturer name MED
Model name MED
Asset ID MED
Mandatory TLVs
This section describes the following mandatory TLVs, which are automatically enabled after you enable
LLDP on a port:
● Chassis ID TLV on page 189
● Port ID TLV on page 189
● TTL TLV on page 189
● End-of-LLDPDU TLV on page 190
Chassis ID TLV
This mandatory TLV is sent by default after you enable LLDP on the port. It is not configurable.
The ExtremeXOS software uses the system’s MAC address to uniquely identify the device. EDP also
uses this to identify the device.
Port ID TLV
This mandatory TLV is sent by default after you enable LLDP on the port; you cannot configure this
TLV. The port ID TLV is used to uniquely identify the port within the device.
The software uses the ifName object for this TLV, so it is the port number on stand-alone switches and
the combination of slot and port number on modular switches.
TTL TLV
The TTL TLV is mandatory, sent by default after LLDP is enabled, and nonconfigurable. This TLV
indicates how long the record should be maintained in the LLDP database. The default value is 120
seconds (or 2 minutes).
A value of 0 in the TTL TLV means the client is shutting down and that record should be deleted from
the database. When you disable an LLDP port, the triggered update LLDPU from that port contains a
TTL TLV of 0.
The TTL TLV is mandatory and is sent by default after LLDP is enabled. Although, technically, you do
not configure the TTL TLV, you can configure the transmit hold value, which is used to calculate the

TTL TLV. (See “Configuring LLDP Timers” on page 196 for more information on transmit hold value
and TTL.)
End-of-LLDPDU TLV
The end-of-LLDPDU TLV marks the end of the data. The system automatically adds this TLV to the
LLDPDU after you enable LLDP.
Optional TLVs
All the optional TLVs are configurable using the CLI and/or SNMP.
This section describes the optional TLVs, under the following categories:
● Standards-based TLVs on page 190
● Avaya-Extreme TLVs on page 193
● LLDP MED TLVs on page 193
Standards-based TLVs
NOTE
The system description TLV is automatically enabled after you enable LLDP and is always sent as part of the
LLDPDU. Although this TLV is not mandatory according to the standard, the ExtremeXOS software includes this TLV
in all LLDPDUs by default; you can configure the system not to advertise this TLV.
This section describes the following optional standards-based TLVs:

● Port description TLV on page 190
● System name TLV on page 191
● System description TLV on page 191
● System capabilities TLV on page 191
● Management address TLV on page 191
● VLAN name TLV on page 191
● Port VLAN ID TLV on page 191
● Port and protocol VLAN ID TLV on page 192
● MAC/PHY configuration/status TLV on page 192
● Power via MDI TLV on page 192
● Link aggregation TLV on page 192
● Maximum frame size TLV on page 192
Port description TLV. You configure this TLV to be advertised or not advertised. The port description
TLV contains the ifDescr object, which is the ASCII string you entered using the configure ports
display-string command. If you have not configured this parameter, the TLV carries an empty string.

Supported TLVs
System name TLV. You configure this TLV to be advertised or not advertised. The system name TLV
contains the device’s configured system name, if previously configured using SNMP. This is the
sysName as defined in RFC 3418, which you define using the configure snmp sysname command.
System description TLV. This is the only TLV that is enabled by default but not mandatory according to
the standard. The ExtremeXOS implementation sends this TLV, by default, whenever you enable LLDP
on a port. You can disable sending this TLV after you enable LLDP; but, by default, the system sends
this TLV.
When enabled, the system sends the image information (from the show version command) in the
system description TLV:
ExtremeXOS version 11.2.0.12 v1120b12 by release-manager

on Fri Mar 18 16:01:08 PST 2005
System capabilities TLV. You configure this TLV to be advertised or not advertised. The system
capabilities TLV indicates the device’s capabilities and which of these are enabled.
The ExtremeXOS software advertises bridge and router capabilities. When configured to advertise the
system capabilities, Extreme Networks devices advertise bridging capabilities. After at least one VLAN
on the device has IP forwarding enabled, the system automatically advertises router capabilities.
Management address TLV. You configure this TLV to be advertised or not advertised. The management
address TLV supplies the management entity for the device.
ExtremeXOS advertises only one management TLV. That management TLV is the IP address of the
management VLAN. If the management VLAN does not have an assigned IP address, the management
address TLV advertises the system’s MAC address. LLDP does not recognize IPv6 addresses in this
field.
VLAN name TLV. You configure this TLV to be advertised or not advertised. This TLV can be repeated
several times within one LLDPDU.
The ExtremeXOS software allows you to advertise VLAN name information to neighboring devices.
This TLV associates a VLAN name to the IEEE 802.1Q tag assigned to that VLAN.
You can enable this TLV for tagged and untagged VLANs. When you enable this TLV for tagged
VLANs, the TLV advertises the IEEE 802.1Q tag for that VLAN. (For untagged VLANs, the internal tag
is advertised.) You can specify exactly which VLANs to advertise.
By default, after you configure this TLV, the system sends all VLAN names on the port. However, each
VLAN name requires 32 bits and the LLDPDU cannot exceed 1500 bytes, so you should configure the
port to advertise only the specified VLANs.
Port VLAN ID TLV. You configure this TLV to be advertised or not advertised. The port VLAN ID
advertises the untagged VLAN on that port. Thus, only one port VLAN ID TLV can exist in the
LLDPDU.
If you configure this TLV and there is no untagged VLAN on the particular port, this TLV is not
included in the LLDPDU.

Port and protocol VLAN ID TLV. You configure this TLV to be advertised or not advertised. This TLV can
be repeated several times within one LLDPDU.
When configured, this TLV allows the port to advertise VLANs and whether the port supports protocol-
based VLANs or not. If no protocol-based VLANs are configured on the port, the TLV still advertises
the port’s capability and sets the VLAN ID value to 0.
As Extreme Networks devices are always capable of supporting protocol-based VLANs, after you
configure this TLV, the system always advertises support for this type of VLAN.
By default, after you configure this TLV, the system sends information for all VLANs on the port.
However, as VLAN TLV requires space and the LLDPDU cannot exceed 1500 bytes, you should
configure the port to advertise only specified VLANs.
MAC/PHY configuration/status TLV. You configure this TLV to be advertised or not advertised. After
configured, this TLV advertises autonegotiation and physical layer capabilities of the port. The system
adds information about the speed rate, duplex setting, bit rate, physical interface, and autonegotiation
support and status.
Power via MDI TLV . You configure this TLV to be advertised or not advertised. When enabled, this TLV
is included in the LLDPDU only for those ports that support supplying power over Ethernet (PoE).
This TLV allows network management to advertise and discover the power-via-MDI capabilities of the
sending 802.3 LAN station. The device type field contains a binary value that represents whether an
LLDP-MED device transmitting the LLDPDU is a power sourcing entity (PSE) or power device (PD), as
listed in Table 27.
Table 27: Power management TLV device information

Value Power source
0 PSE device
1 PD device
2-3 Reserved
Additional PoE information is advertised as well, including the power status, power class, and pin pairs
used to supply power.
(Refer to for “Avaya-Extreme TLVs” on page 193 and “LLDP MED TLVs” on page 193 more information
on power-related TLVs.)
Link aggregation TLV. You configure this TLV to be advertised or not advertised. When enabled, this
TLV advertises information on the port’s load-sharing (link aggregation) capabilities and status.
Maximum frame size TLV. You configure this TLV to be advertised or not advertised. This TLV allows
the port to advertise its maximum supported frame size to its neighbors.
When jumbo frames are not enabled on the specified port, the TLV reports a value of 1518 after you
configure it to advertise. If jumbo frames are enabled, the TLV inserts the configured value for the
jumbo frames.

Supported TLVs
Avaya-Extreme TLVs
This section describes the following optional proprietary Avaya-Extreme Networks TLVs that you can
configure the switch to transmit:
● PoE conservation level request TLV on page 193
● Call server TLV on page 193
● File server TLV on page 193
● 802.1Q framing TLV on page 193
NOTE
You display the values for these TLVs using the show lldp neighbors detailed command.
PoE conservation level request TLV. You configure this TLV to advertise or not advertise a requested
conservation level. This enables the PSE device to request the connected PD to go into a certain power
conservation level or request the PD to go to the maximum conservation level. By default, the requested
conservation value on this proprietary LLDP TLV is 0, which is no power conservation. This LLDP TLV
is sent out only on PoE-capable Ethernet ports.
You change this level temporarily using a network station or SNMP with the MIB; this change is not
saved across a reboot.
Call server TLV. You configure this TLV to advertise or not advertise up to 8 call servers. This TLV
allows the exchange of information between an IP phone and a network connectivity device about the
reachability of the call server for the IP phone connected to the respective port of the switch. You can
send a maximum of 8 call server addresses in a single TLV. The Avaya phone uses this addressing
information after it receives the TLV from the switch.
File server TLV. This TLV allows the exchange of information between an IP phone and the network
connectivity device concerning the reachability of the file server for the IP phone connected to the
respective port of the switch. You can advertise up to 4 file server addresses in a single TLV. The Avaya
phone uses this addressing information after it receives the TLV from the switch.
802.1Q framing TLV. Use this TLV to exchange information about Layer 2 priority tagging between a
connectivity device and an IP phone. The Avaya phone uses this addressing information after it
receives the TLV from the switch.
This TLV works as an extension of the LLDP network policy TLV. For this TLV to function, you must
enable both:
● LLDP MED capabilities TLV (see “LLDP MED capabilities TLV” on page 194)
● LLDP MED network policy TLV (see “Network policy TLV” on page 194)
LLDP MED TLVs

This section describes the optional LLDP media endpoint discovery (MED) TLVs that you can configure
the switch to transmit.

NOTE
You must configure the LLDP MED capabilities TLV before any of the other MED TLVs can be enabled. Also, this
TLV must be se to no-advertise after all other MED TLVs are set to no-advertise.
The switch sends all MED TLVs only after it detects a MED-capable device on the port. The switch does
not automatically send any MED TLVs after it is enabled; the switch must first detect a MED-capable
device on the port.
Network connectivity devices wait for LLDP MED TLVs from endpoints before they send out LLDP
MED TLVs; so two network connectivity devices will not exchange LLDP MED messages.
The following LLEP MED extension TLVs can be transmitted by the switch:
● LLDP MED capabilities TLV on page 194
● Network policy TLV on page 194
● Location identification TLV on page 194
● Extended power-via-MDI TLV on page 194
NOTE
You display the values for these TLVs using the show lldp neighbors detailed command.
LLDP MED capabilities TLV. This TLV allows LLDP MED network connectivity devices to determine that
specified endpoints support LLDP MED, and if so, to discover which LLDP MED TLVs the particular
endpoint device supports and what device class is belongs to.
This TLV must be enabled before any of the other LLDP MED TLVs can be enabled.
Network policy TLV. You configure this MED TLV to allow both network connectivity devices and
endpoint devices to advertise VLAN configuration and associated Layer 2 and Layer 3 attributes that
apply for a specific set of applications on that port.
You configure this TLV per port/VLAN. Each application can exist only once on each port. You can
configure a maximum of 8 TLVs, each with its own DSCP value and/or priority tag. This TLV tells the
endpoint the specific VLAN to use for the specific application.
Location identification TLV. You configure this TLV to advertise or not advertise a maximum of three
different location identifiers, each with a different format, as follows:
● Coordinate based, using a 16-byte hexadecimal string
● Civic-based, using a hexadecimal string with a minimum of 6 bytes
● ECS ELIN, using a numerical string with a range of 10 to 25 characters.
Extended power-via-MDI TLV. Use this TLV to advertise fine-grained power requirement details,
including the power status of the PD and the port. You can enable this TLV only on PoE-capable ports;
the switch returns an error message if you attempt to transmit this LLDP TLV over a non-PoE-capable
port.

Configuring LLDP
Configuring LLDP
You configure LLDP per port. To configure LLDP:
1 Enable LLDP on the desired port(s).
2 If desired, configure the system not to advertise the system description TLV.
3 If you want to change any default values, configure the following values:
a Reinitialize period
b Transmit interval
c Transmit delay
d Transmit hold
4 Enable the SNMP traps and configure the notification interval.
5 Configure any optional TLV advertisements, including the proprietary Avaya-Extreme TLVs, that
you want included in the LLDPDU.
6 If you want to send or receive MED extension TLVs, configure the LLDP MED capabilities TLV.
7 If you want to change the default value of 3 for the fast-start feature for LLDP MED, configure the
LLDP MED fast-start TLVs.
8 If you want SNMP traps for the LLDP MED extension TLVs, enable these traps.
This section describes how to configure LLDP using the CLI. Refer to ExtremeXOS Command Reference
Guide for complete information on configuring LLDP. You can also reference the IEEE 892.1ab standard.
Enabling and Disabling LLDP

LLDP is disabled on all ports by default. When you enable LLDP on the ports, you select whether the
ports will only transmit LLDP messages, only receive the messages, or both transmit and receive LLDP
messages.
To enable LLDP, use the following command:
enable lldp ports [all | <port_list>] {receive-only | transmit-only}
After you enable LLDP, the following TLVs are automatically added to the LLDPDU:
● Chassis ID
● Port ID
● TTL
● System description
● End of LLDPDU
All of these, except the system description, are mandated by the 802.1ab standard. Similarly, none of
these, except the system description, can be configured to advertise or not to advertise.
To disable LLDP, use the following command:
disable lldp ports [all | <port_list>] {receive-only | transmit-only}

Configuring the System Description TLV Advertisement

If you have not configured the system description using SNMP sysName before enabling LLDP, the
system sends the following information in the system description TLV:
ExtremeXOS version 11.2.0.12 v1120b12 by release-manager
on Fri Mar 18 16:01:08 PST 2005
To disable the default advertisement of the system description, use the following command:
configure lldp ports [all | <port_list>] no-advertise system-description
Configuring LLDP Timers

After you enable LLDP, the timer values assume the default values. However, if you want to change
any of these default values, use the CLI to configure the relevant timer.
NOTE
The LLDP timers apply to the entire device and are not configurable by port.
When LLDP is disabled or if the link goes down, LLDP is reinitialized. The reinitialize delay is the
number of seconds the port waits to restart LLDP state machine; the default is 2 seconds.
To change the default reinitialize delay period, use the following command:
configure lldp reinitialize-delay <seconds>
LLDP messages are transmitted at a set interval; this interval has a default value of every 30 seconds.
To change this default value, use the following command:
configure lldp transmit-interval <seconds>
The time between triggered update LLDP messages is referred to as the transmit delay, and the default
value is 2 seconds. You can change the default transmit delay value to a specified number of seconds or
to be automatically calculated by multiplying the transmit interval by 0.25. To change the value for the
transmit delay, use the following command:
configure lldp transmit-delay [ auto | <seconds>]
Each LLDP message contains a TTL value. The receiving LLDP agent discards all LLDP messages that
surpass the TTL value; the default value is 120 seconds.
The TTL is calculated by multiplying the transmit interval value and the transmit hold value; the
default transmit hold value is 4. To change the default transmit hold value, use the following command:
configure lldp transmit-hold <hold>
Configuring SNMP for LLDP

You can send SNMP traps regarding LLDP; the software supports the LLDP MIB. By default, SNMP
LLDP traps are disabled on all ports; to enable LLDP SNMP traps, use the following command:
enable snmp traps lldp {ports [all | <port_list>]}

Configuring LLDP
The traps are only sent for those ports that are both enabled for LLDP and have LLDP traps enabled.
To disable the LLDP SNMP traps, use the following command:
disable snmp traps lldp {ports [all | <port_list>]}
The default value for the interval between SNMP LLDP trap notifications is 5 seconds. To change this
interval for the entire switch for LLDP traps, use the following command:
configure lldp snmp-notification-interval <seconds>
NOTE
If you want to send traps for LLDP MED, you must configure it separately. Use the enable snmp traps lldp-
med {ports [all | <port_list>]} command to enable these traps.
Configuring Optional TLV Advertisements

By default, all optional TLVs are not added to the LLDPDU, or not advertised.
You can add optional TLVs to the LLDPDU but be aware that the total LLDPDU cannot exceed 1500
bytes, including the mandatory TLVs. Any optional added TLVs that exceed the 1500-byte limit are
dropped. You can see if you have dropped TLVs from your LLDPDU by referring to the EMS log or by
issuing the show lldp statistics command.
NOTE
Extreme Networks recommends that you advertise only one or two VLANS on specified ports to avoid dropping TLVs
from the LLDPDU.
This section describes the following types of optional TLVs:

● Configuring Standards-based Optional TLVs on page 197
● Configuring Proprietary Avaya-Extreme Optional TLVs on page 199
● Configuring LLDP MED Optional TLVs on page 200
Configuring Standards-based Optional TLVs

You configure LLDP ports to advertise any of the following optional TLVs:
● Port description TLV
● System name TLV
● System capabilities TLV
● Management address TLV
● VLAN name TLV (repeatable TLVs)
● Port VLAN ID TLV
● Port and protocol VLAN ID TLV (repeatable TLVs)
● MAC/PHY configuration/status TLV
● Power via MDI TLV

● Link aggregation TLV

● Maximum frame size TLV
Refer to “Standards-based TLVs” on page 190 for complete information on each optional TLV.
To advertise the optional port description information, use the following command:
configure lldp ports [all | <port_list>] [advertise | no-advertise] port-description
To advertise the system name, use the following command:
configure lldp ports [all | <port_list>] [advertise | no-advertise] system-name
To advertise the system capabilities, use the following command:
configure lldp ports [all | <port_list>] [advertise | no-advertise] system-

capabilities
To advertise the IP address of the management VLAN (or the system MAC address if IP is not
configured), use the following command:
configure lldp ports [all | <port_list>] [advertise | no-advertise] management-address
You can advertise more than one VLAN name per LLDP-enabled port. To do so, add one optional
VLAN name TLV for each VLAN you want to advertise. If you do not specify VLAN names, the system
sends an advertisement for all VLANs on the port.
To advertise VLAN names, use the following command:
configure lldp ports [all | <port_list>] [advertise | no-advertise] vendor-specific

dot1 vlan-name {vlan [all | <vlan_name>]}
NOTE
The total LLPDU size is 1500 bytes; any TLVs after that limit are dropped.
You can advertise the untagged, port-based VLAN for the LLDP-enabled port using the port VLAN ID
TLV. To configure the port VLAN ID TLV, use the following command:

dot1 port-vlan-ID
You can advertise more than one protocol-based VLAN per LLDP-enabled port. To do so, add one
optional port and protocol VLAN ID TLV for each VLAN you want to advertise. To advertise these
VLANs, use the following command:

dot1 port-protocol-vlan-ID {vlan [all | <vlan_name>]}
NOTE
The total LLPDU size is 1500 bytes; any TLVs after that limit are dropped.

Configuring LLDP
You can advertise the speed capabilities, autonegotiation support and status and physical interface of
the LLDP-enabled port using the MAC/PHY configuration/status TLV. To advertise this information,

dot3 mac-phy
Configure the power via MDI TLV to advertise the PoE capabilities of the LLDP-enabled port. To
advertise the PoE capabilities and status, use the following command:

dot3 power-via-mdi
You advertise the load-sharing capabilities and status of the LLDP-enabled port by configuring the link
aggregation TLV. To advertise load-sharing capabilities, use the following command:

dot3 link-aggregation
You advertise the maximum frame size available on the LLDP-enabled port using the maximum frame
size TLV. To advertise the maximum frame size, use the following command:

dot3 max-frame-size
Configuring Proprietary Avaya-Extreme Optional TLVs

You configure LLDP ports to advertise any of the following optional proprietary Avaya-Extreme
Networks TLVs:
● PoE conservation level request TLV
● Call server TLV
● File server TLV
● 802.1Q framing TLV
Refer to “Avaya-Extreme TLVs” on page 193 for complete information on each optional TLV.
To advertise a request for power conservation to the connected PDs, use the proprietary Avaya-Extreme
Networks PoE conservation level request TLV with the following command:

avaya-extreme poe-conservation-request
To advertise up to 8 call servers, use the proprietary Avaya-Extreme Networks call server TLV with the
following command:

avaya-extreme call-server <ip_address_1> {<ip_address_2> {<ip_address_3>
{<ip_address_4> {<ip_address_5> {<ip_address_6> {<ip_address_7> {ip_address_8>}}}}}}}

To advertise up to 4 file servers, use the proprietary Avaya-Extreme Networks file server TLV with the
following command:

avaya-extreme file-server <ip_address_1> {<ip_address_2> {<ip_address_3>
{<ip_address_4>}}}
To advertise the Layer 2 priority tagging information, use the proprietary Avaya-Extreme Networks
802.1Q framing TLV with the following command:

avaya-extreme dot1q-framing [tagged | untagged | auto]
NOTE
For this command to work, you must have previously enabled both the configure lldp ports vendor-
specific med capabilities and the configure lldp ports vendor-specific med policy
application commands. (See “Configuring LLDP MED Optional TLVs” for complete information.)
Configuring LLDP MED Optional TLVs

After you enable an LLDP MED TLV, the switch waits until it detects a MED-capable device before it
begins transmitting the configured LLDP MED TLVs; the switch does not transmit the MED TLVs as
soon as they are enabled, it must first detect an MED-capable device. Because network connectivity
devices wait to detect LLDP MED TLVs from endpoints before they send out LLDP MED TLVs 2
network connectivity devices will not exchange LLDP MED messages.
To receive SNMP traps on the LLDP MED, you must enable these separately from the other LLDP
traps. (See “Configuring SNMP for LLDP” on page 196).
You must configure the LLDP MED capabilities TLV before you configure any other LLDP MED TLVs.
Finally, the fast-start feature allows you to increase the learning speed of the switch for LLDP MED
TLVs. The fast-start feature is automatically enabled once you enable the LLDP MED capabilities TLV;
you can change the configuration from the default setting of 3.
Refer to “LLDP MED TLVs” on page 193 for complete information on each optional TLV.
This section describes configuring the following LLDP MED TLVs:

● LLDP MED capabilities TLV
● LLDP fast-start TLV
● Network policy TLV
● Location identification TLV
● Extended power-via-MDI TLV
To enable configuration and transmission of any other LLDP MED TLV and to determine the LLDP
MED capabilities of endpoint devices, use the following command:

med capabilities
To configure the LLDP fast-start feature, use the following command:
configure lldp med fast-start repeat-count <count>

Displaying LLDP Settings
To advertise VLAN as associated Layer 2 and Layer 3 attributes for a specified application, use the
network policy TLV with following command:

med policy application [voice | voice-signaling |guest-voice | guest-voice-signaling |
softphone-voice | video-conferencing | streaming-video | video-signaling] vlan
<vlan_name> dscp <dscp_value> {priority-tagged}
To advertise location information, use the following command:

med location-identification [coordinate-based <hex_value> | civic-based <hex_value> |
ecs-elin <elin>]
To advertise power requirement details, use the extended power-via-MDI TLV with the following
command:

med power-via-mdi
Unconfiguring LLDP
To unconfigure LLD, use the following command:
unconfigure lldp
This command only returns the LLDP timers to default values; LLDP remains enabled, and all the
configured TLVs are still advertised.
To leave LLDP enabled, but reset the advertised TLVs to the five default TLVs, use the following
command, and specify the affected ports:
unconfigure lldp port [all | <port_list>]
Displaying LLDP Settings

The system displays information on the LLDP status and statistical counters of the ports, as well as
about the LLDP advertisements received and stored by the system. You can display information on the
LLDP port configuration and on the LLDP neighbors detected on the port.
NOTE
Refer to ExtremeXOS Command Reference Guide for complete information on displaying LLDP settings.
Displaying LLDP Port Configuration Information and Statistics

To display LLDP port configuration information, use the show lldp command; to display detailed
LLDP information, add the detailed option.
To display the statistical counters related to the LLDP port, use the show lldp statistics command.

Displaying LLDP Information Detected from Neighboring Ports

To display information from LLDP neighbors detected on the port, use the show lldp neighbors
command. The following is sample output from the show lldp neighbors command. You must use
the detailed option to display information on the proprietary Avaya-Extreme Networks TLVs and the
LLDP MED TLVs.

9 Connectivity Fault Management—BlackDiamond
10808 and 12800 Series Switches Only
Connectivity Fault Management (CFM), discussed in the emerging IEEE 802.1ag specification, allows
you to detect, verify, and isolate connectivity failures in virtual bridged LANs. Part of this specification
is a toolset to manually check connectivity, which is sometimes referred to as Layer 2 ping.
NOTE
The ExtremeXOS implementation of CFM is based on draft 4.1 of the IEEE 802.1ag standard.
There is no direct interaction between CFM and other Layer 2 protocols; however, blocked Spanning
Tree Protocol (STP) ports are taken into consideration when forwarding CFM messages.

● Overview of CFM Elements on page 203
● Ping and Traceroute on page 206
● Supported Instances for CFM on page 207
● Configuring CFM on page 207
● Displaying CFM on page 211
● CFM Example on page 212
Overview of CFM Elements
NOTE
Extreme Networks uses proprietary values for the MAC addresses and Ethernet types for CFM.
You create hierarchical networks, or domains, and test connectivity within that domain by sending
Layer 2 messages, known as Connectivity Check Messages (CCMs).

Connectivity Fault Management—BlackDiamond 10808 and 12800 Series Switches Only
Figure 5 shows an example of hierarchical CFM domains.
Figure 5: CFM hierarchical domains example
Customer ISP ISP Operator ISP ISP Customer
MA level 0 4 4 5 4 4 0
CFMs
EX_116A
NOTE
The arrows in Figure 5 indicate the span that CCM messages take, not the direction. (See Table 28 for more
information on spans for CCM messages.)
To achieve this hierarchical connectivity testing, you create and configure the following entities:
● Maintenance domains, or domains
● Maintenance association (MA) level; a unique hierarchical numeric value for each domain
● Maintenance associations (MAs)
● Maintenance points (MPs), which are the following
■ Maintenance end points (MEPs), which are one of the following types:
- Inward-facing MEPs
- Outward-facing MEPs
■ Maintenance intermediate points (MIPs)
■ CFM filter function (CFF)
You must have at least one MP on an intermediate switch in your domain. Ensure that you map and
configure all ports in your domain carefully, especially the inward-facing MEPs and the outward-facing
MEPs. If these are incorrectly configured, the CCMs are sent in the wrong direction in your network,
and you will not be able to test the connectivity within the domain.
You can have up to eight domains on an Extreme Networks switch. A domain is the network or part of
the network for which faults are to be managed; it is that section where you are monitoring Layer 2
connectivity. A domain is intended to be fully connected internally.
NOTE
Domains may cross VR boundaries; domains not virtual router-aware.

Overview of CFM Elements
You assign each domain an MA level, which function in a hierarchy for forwarding CFM messages. The
MA levels are from 0 to 7. The lowest number is superior in the CFM hierarchy.
The IEEE standard 801.2ag recommends assigning different MA levels to different domains for different
network users, as follows:
● 0 to 2 for end users
● 3 and 4 for Internet service providers (ISPs)
● 5 to 7 for operators (entities carrying the information for the ISPs)
All CFM messages with a superior MA level (numerically lower) pass throughout domains with an
inferior MA level (numerically higher). CFM messages with an inferior MA level are not forwarded to
domains with a superior MA level. Refer to Table 28 for an illustration of domains with hierarchical MA
levels.
Table 28: MA levels and recommended use

MA Level 0 1 2 3 4 5 6 7
Use Customer Service provider Operator
Superiority Most < ----- Superior / Inferior ----- > Most
superior inferior
Within a given domain, you associate maintenance associations (MAs). Extreme Networks’
implementation of CFM associates MAs with VLANs. All of the ports in that VLAN are now in that
MA and its associated domain. In general, you should configure one MIP on each intermediate switch
in the domain and a MEP on every edge switch.
Each MA associates with one VLAN, and each VLAN associates with one MA. Thus, you can have a
total of 4096 MAs on a switch (each associated with 1 of 8 possible domains). The MA is unique within
the domain. You can configure up to 4096 MAs in any 1 domain. One switch can have 8 domains, 128
ports, 4096 associations (see “Supported Instances for CFM” on page 207 for supported CFM elements).
If you attempt to put 1 MA into 2 domains, the system returns an error message.
NOTE
You cannot associate the Management VLAN with an MA or a domain.
NOTE
If you configured MAC-in-MAC on a port or have an untagged vMAN on that port, you cannot configure CFM on that
port.
You assign the MPs to ports: inward-facing MEPs, outward-facing MEPs, MIPs, and CFFs. These
various MPs filter or forward the CFM messages to test the connectivity of your network.

Each configured MEP periodically sends out a Layer 2 multicast CCM message. The destination MAC
address in the CCM frame is from a multicast MAC address range that is reserved for CFM messages
Each MEP must have a MEP ID that is unique within the MA. The MEPs send the CCM messages
differently, depending on the configuration, as follows:
● The outward-facing MEPs sends out a single CCM message.
● The inward-facing MEPs potentially sends the CCM message to all ports on the VLAN (MA)—
except the sending port—depending on the MPs configured on the outgoing ports.
NOTE
Ensure that you configured the inward-facing and outward-facing MEPs correctly, or the CCMs will flow in the wrong
through the domain and not allow connectivity testing.
MIPs define intermediate points within a domain. MIPs relay the CCM messages to the next MIP or
MEP in the domain. MIPs look at the CCM and pass it along.
You configure the time interval for each MEP to send a CCM. Extreme Networks recommends setting
this interval for at least 1 second. Each MEP also makes a note of what port and what time it received a
CCM. This information is stored in the CCM database. The CCM database holds up to 64 entries per
MEP; after that, additional CCM messages are discarded.
Each CCM has a time-to-live (TTL) value also noted for that message. This TTL interval is 3.5 times the
CCM transmission interval you configured on the switch that is originating the CCM. After the TTL
expires, the connectivity is considered broken, and the system sends a message to the log. One
important result of the continual transmission of CCM frames is that the MAC address of the
originating MEP is known to all MPs in the association.
NOTE
You must configure the following EMS filter to capture the messages on the log: configure log filter
DefaultFilter add events cfm.rMEPExp. See Chapter for Chapter 5, “Status Monitoring and Statistics,” for
complete information on configuring EMS messages.
The MA values are from 0 to 7; in the hierarchy, the MA level of 0 is highest and 7 is lowest. CFFs
configured on a port block all inward and outward CCM messages with an equal or inferior MA level
(higher MA value of 0 to 7) than the MA level of the domain that CFF is in. A CFF port passes every
CFM packet that has a superior MA level (lower MA value of 0 to 7) in both directions.
Not all combinations of MPS are allowed on the same port within an MA; only the following
combinations can be on the same port within an MA:
● Inward-facing MEP and MIP
● Inward-facing MEP and outward-facing MEP
● Outward-facing MEP and MIP
Ping and Traceroute

When the operator sees a connectivity fault message from CFM in the system log, they can send a
loopback message (LBM) or a link trace message (LTM). These are also referred to as a Layer 2 ping or

Supported Instances for CFM
a traceroute message. You can send with an LBM or an LTM only from a MEP (either inward-facing or
outward-facing).
You can only send a ping from a MEP, and you ping to the unique system MAC address on the switch
you are testing connectivity to. The operator sends out a unicast LBM, and the first MIP or MEP on the
switch with the destination MAC address matching the embedded MAC address replies with an LBR.
You can only send a traceroute (or LTM) from a MEP. You send the traceroute to the unique system
MAC address on the switch to which you are testing connectivity. The system sends out an LTM to the
special multicast address. Each MIP in the path passes the frame only in the direction of the path and
sends a link trace reply (LTR) message back to the originating with information that the LTM passed.
The traceroute command displays every MIP along the path (see traceroute mac port).
Supported Instances for CFM

Table 29 displays the CFM support in ExtremeXOS software version 11.4.
Table 29: ExtremeXOS version 11.4 CFM support
Item Limit Notes
Domains 8 Per switch; one for each MA level
Associations (MAs) 4096 Per switch
Inward-facing MEPs 32 Per switch
Outward-facing MEPs 32 Per switch
MIPs 32 Per switch
Total CFM ports 128 Per switch; total number of all ports for all
VLANs assigned to an MA (see show cfm
command for ports configured for CFM)
Entries in the CCM database 64 Number of remote end points stored in a CCM
database on each MEP; 65 end points per MEP
(additional CCMs discarded after this limit is
reached)
There is no limit on the CFFs. Additionally, ExtremeXOS 11.4 does not support CFM traps, MIBs, or
alarm suppression.
NOTE
The total number of CFM ports is a guideline, not a limit enforced by the system.
Configuring CFM
To configure CFM, you create a maintenance domain and assign it a unique MA level. Next, you
associate MAs with the specified domain and assign MPs within that MA. Optionally, you can
configure the transmission interval for the CCMs.

If an MEP fails to receive a CCM before the last advertised TTL value expires, the system logs a
message. After the network administrator sees the system log message, he can send a Layer 2 ping and/
or a traceroute message to isolate the fault.
NOTE
Enhanced ACLs are automatically applied when you configure CFM using the CLI commands. These ACLs are
displayed when you issue the command show access-list dynamic.

● Creating Maintenance Domains on page 208
● Creating and Associating MAs on page 209
● Creating MPs and the CCM Transmission Interval on page 210
● Executing Layer 2 Ping and Traceroute Messages on page 211
Creating Maintenance Domains

You create maintenance domains (MDs), or domains, and assign a unique MA level at that time.
Available MA levels are numbered from 0 to 7. Lower numerical values are superior MA levels in the
CFM hierarchy; CFFs block all CCMs with an inferior MA level. Each switch can have a total of eight
domains; each with a unique MA level.
You can name domains using any one of the following five formats:
● ITU-T M.1400 carrier code
Use an alphanumeric character string with a maximum of 100 characters.
● ISO 3166 country code
● Domain name server (DNS) name
● MAC address plus 2-octet integer
Use a MAC address and a 2-octet integer. The display format is XX.XX.XX.XX.XX.XX.YYY, where X
is the MAC address, and Y is the 2-octet integer. For example, a domain name in this format using
123 as the 16-bit unsigned integer appears as follows: 00:11:22:33:44:55.123.
● RFC 2685 VPN ID
Use the 24-bit virtual private network (VPN) Organizational Unique Identifier (OUI) and a 32-bit
VPN index. The display format is XX:XX:XX.Y, where X is the VPN OUI and Y is the VPN index. For
example a domain name in this format using the VPN OUI and an index value of 55 appears as
follows: 00:11:22:.55.
NOTE
Whatever convention you choose, you must use that same format throughout the entire domain.
The CFM messages carry the domain name, so the name and naming format must be identical to be
understood throughout the domain. You can, however, use different naming conventions on different
domains on one switch (up to eight domains allowed on one switch).

Configuring CFM
To create a domain and assign an MA level using the ITU-T M.1400 carrier code convention, use the
following CLI command:
create cfm domain m1400 <name> ma-level <level>
To create a domain and assign an MA level using the ISO 3166 country code convention, use the
following CLI command:
create cfm domain iso3166 <name> ma-level <level>
To create a domain and assign an MA level using the DNS convention, use the following CLI
command:
create cfm domain dns <name> ma-level <level>
To create a domain and assign an MA level using the MAC address convention, use the following CLI
command:
create cfm domain mac <mac> <int> ma-level <level>
To create a domain and assign an MA level using the RFC 2685 VPN ID convention, use the following
CLI command:
create cfm domain vpn-id oui <oui> index <index> ma-level <level>
Although you assign an MA level to the domain when you create that domain, you can change the MA
level on an existing domain by using the following command:
configure cfm domain <domain_name> ma-level <level>
To delete a domain, use the following command:
delete cfm domain <domain>
Creating and Associating MAs

Within a given domain, you associate maintenance associations (MAs). Extreme Networks’
implementation of CFM associates MAs with VLANs. All of the ports in that VLAN are now in that
MA and its associated domain.
Each MA associates with one VLAN, and each VLAN associates with one MA. Each MA can be in only
1 domain; you can configure more than one MAs in any 1 domain but 1 MA cannot be in 2 domains.
Like the domains, ExtremeXOS supports multiple formats for naming the MA. The following formats
are supported for naming the MAs:
● Character string
● 2-octet integer
● RFC 2685 VPN ID
To add an MA to a domain using the character string format, use the following command:
configure cfm domain <domain_name> add association string <name> vlan <vlan_name>

To add an MA to a domain using the 2-octet integer format, use the following command:
configure cfm domain <domain_name> add association integer <int> vlan <vlan_name>
To add an MA to a domain using the RFC 2685 VPN ID format, use the following command:
configure cfm domain <domain_name> add association vpn-id oui <oui> index <index>
vlan <vlan_name>
To delete an MA from a domain, use the following command:
configure cfm domain <domain_name> delete association <association_name>
Creating MPs and the CCM Transmission Interval

Within an MA, you configure the following MPs:
● Maintenance end points (MEPs), which are one of the following types:
■ Inward-facing MEPs—transmit CCMs and maintain CCM database
■ Outward-facing MEPs—transmit CCMs and maintain CCM database
● Maintenance intermediate points (MIPs)—pass CCMs through
● CFM filter functions (CFFs)—block all CCMs with an inferior MA level
Each MEP must have an ID that is unique for that MEP throughout the MA.
Not all combinations of MPS are allowed on the same port within an MA; only the following
combinations can be on the same port within an MA:
● inward-facing MEP and MIP
● inward-facing MEP and outward-facing MEP
● outward-facing MEP and MIP
NOTE
Ensure that you configured the inward-facing and outward-facing MEPs correctly, or the CCMs will flow in the wrong
through the domain and not allow connectivity testing.
To configure inward-facing and outward-facing MEPs and its unique MEP ID, use the following
command:
configure cfm domain <domain_name> association <association_name> ports <port_list>

add end-point [inward-facing | outward-facing ] <mepid>
To change the MEP ID on an existing MEP, use the following command:
configure cfm domain <domain-name> association <association_name> ports <port_list>

end-point [inward-facing | outward-facing] mepid <mepid>
To delete inward-facing and outward-facing MEPs, use the following command:

delete end-point [inward-facing | outward-facing ]

Displaying CFM
To configure a MIP, use the following command:

add intermediate-point
To delete a MIP, use the following command:

delete intermediate-point
To configure a CFF, use the following command:

add filter-function
To delete a CFF, issue the following command:

delete filter-function
To configure the transmission interval for the MEP to send CCMs, issue the following command:

end-point [inward-facing | outward-facing] transmit-interval <interval>
Executing Layer 2 Ping and Traceroute Messages

If the system logs a missed CCM message, the operator can use Layer 2 ping and traceroute messages
to isolate the fault. (See “Ping and Traceroute” on page 206 for information on how each MP handles
these messages.)
NOTE
You must have all the CFM parameters configured on your network before issuing the ping and traceroute messages.
To send a Layer 2 ping, use the following command:
ping mac <mac> port <port> {domain} <domain_name> {association} <association_name>
To send a Link Trace Message (LTM) and receive information on the path, use the following command:
traceroute mac <mac> {inward-facing-end-point} port <port> {domain} <domain_name>

{association} <association_name> {ttl <ttl>}
Displaying CFM
To verify your CFM configuration, you can display the current CFM configuration using the show cfm
command. The information this command displays includes the total ports configured for CFM, the
domain names and MA levels, the MAs and associated VLANs, and the inward-facing and outward-
facing MEPs.

To display the CCM database for each MEP, use the show cfm detail command. To display statistics
on the CFM functioning on the switch, use the show cfm statistics command.
To clear these counters, use the clear counters command.
CFM Example
As shown in Figure 6, this examples assumes a simple network; this example assumes that CFM is
configured on the access switches, as well as the necessary VLANs configured with the ports added.
Figure 6: CFM configuration example
1:3
1:1 1:1 1:2 1:1
Switch 1 Switch 2 Switch 3
= Inward MEP
= Outward MEP
= Intermediate point
EX_111A
To configure switch 1 for this example, use the following CLI commands:
create cfm domain dns corp ma-level 1
configure cfm domain corp add association string core vlan corenet
configure cfm domain corp association core ports 1:1 add end-point outward-facing 101
configure cfm domain corp association core ports 1:1 add intermediate-point
configure cfm domain corp association core ports 1:2 add intermediate-point
configure cfm domain corp association core ports 1:3 add end-point inward-facing 102
configure cfm domain corp association core ports 1:1 add end-point outward-facing 103

10 Power Over Ethernet
Power over Ethernet (PoE) is an effective method of supplying 48 VDC power to certain types of
powered devices (PDs) through Category 5 or Category 3 twisted pair Ethernet cables. PDs include
wireless access points, IP telephones, laptop computers, web cameras, and other devices. With PoE, a
single Ethernet cable supplies power and the data connection, reducing costs associated with separate
power cabling and supply.
Beginning with ExtremeXOS version 11.3, the system supports hitless failover for PoE in a system with
two Master Switch Modules (MSMs). Hitless failover means that if the primary MSM fails over to the
backup MSM, all port currently powered will maintain power after the failover and all the power
configurations remain active.
Beginning with ExtremeXOS version 12.0, similar failover support is available for SummitStack. In a
SummitStack, power is maintained across a failover on all PoE ports of non-primary nodes but is lost on
all PoE ports of the failed primary node. Power management is local to each PoE switch in a
SummitStack. The commands to manage the power can be issued centrally on the SummitStack from
the primary node.
NOTE
The BlackDiamond 8800 series switch was formerly known as Aspen.

● Extreme Networks PoE Devices on page 213
● Summary of PoE Features on page 214
● Power Checking for PoE Module on page 215
● Power Delivery on page 215
● Configuring PoE on page 220
● Displaying PoE Settings and Statistics on page 225
Extreme Networks PoE Devices

The following lists the Extreme Networks devices that support PoE and the minimum required
software:
● G48P module for the BlackDiamond 8800 series switch—ExtremeXOS 11.1 and higher
● G48Pe module for the BlackDiamond 8800 series switch—ExtremeXOS 11.5 and higher
● Summit X450e-24p switch—ExtremeXOS 11.5 and higher
● Summit X450e-48p switch—ExtremeXOS 11.6 and higher
● Summit X250e-24p switch - ExtremeXOS 12.0 and higher
● Summit X250e-48p switch - ExtremeXOS 12.0 and higher

Power Over Ethernet
The Summit X450e-48p switch is offers an optional external power supply chassis (EPS-C) than accepts
up to three power modules (EPS-600LS). Table 30 describes the EPS-C/EPS-600LS options for the
X450e-48p.
Table 30: EPS-C/EPS-600LS Options for the X450e-48p

Number of EPS-
600LS Module X450e-48p PoE Capability
0 Internal PSU 370 watt PoE power
1 Redundant 370 watt PoE power
2 Full 740 watt PoE power. No PoE redundancy
3 Full 740 watt PoE power. 2:1 power module redundancy
When using the EPS-C with 0 or 1 EPS-600LS modules, 370 watts of PoE power is available for the
X450e-48p PDs. E.g. 15.4 watts to 24 ports, 7.7 watts to 48 ports or any combination where the total PoE
power does not exceed the 370 watt PSU capacity.
If the total system demands exceed the available power on the Summit X450e-48p switch, the user can
specify the port priorities and port disconnect precedence.
When using the EPS-C with 2 or 3 EPS-600LS modules, 740 watts of PoE power is available. E.g. 15.4
watts to 48 ports.
When disconnecting an EPS-C from the X450e-48p, the user should turn off the EPS-600LS one module
at a time. This allows an uninterrupted PoE transition from full power to redundant half power to
internal power. Suddenly removing an EPS-C with 2 or 3 EPS-600LS modules providing power will
cause an interruption of PoE power to the PDs.
NOTE
Refer to the Extreme Networks Consolidated Hardware Guide for complete information on power availability using the
Summit X450e-48p switch in conjunction with the EPS-600 PSU.
Summary of PoE Features

The PoE feature supports the following PoE features:
● Configuration and control of the power distribution for PoE at the system, slot, and port levels
● Real-time discovery and classification of IEEE 802.3af-compliant PDs and many legacy devices
● Monitor and control of port PoE fault conditions including exceeding configured class limits and
power limits and short-circuit detection
● Support for configuring and monitoring PoE status at the system, slot, and port levels
● Management of an over-subscribed power budget
● Port LED control for indicating the link state
● Support for hitless failover in a chassis with two MSMs
For detailed information on using the PoE commands to configure, manage, and display PoE settings,
refer to the ExtremeXOS Command Reference Guide.

Power Checking for PoE Module
Power Checking for PoE Module

PoE modules require more power than other I/O modules. When a chassis containing a PoE module is
booted or a new PoE module is inserted, the power drain is calculated. Before the PoE module is
powered up, the chassis calculates the power budget and powers up the PoE module only if there is
enough power. The chassis powers up as many I/O modules as possible with lower-numbered slots
having priority.
NOTE
If your chassis has an inline power module and there is not enough power to supply the configured inline power for
the slot, that slot will not power on; the slot will not function in data-only mode without enough power for inline
power.
If a PoE module is inserted into a chassis, the chassis calculates the power budget and only powers up
the PoE module if there is enough power. Installed modules are not affected. However, if you reboot
the chassis, power checking proceeds as described in the previous paragraph. If there is now enough
power, I/O modules that were not powered up previously are powered up.
If you lose power or the overall available power decreases, the system removes power to the I/O
modules beginning with the highest numbered slots until enough power is available. Inline power
reserved for a slot that is not used cannot be used by other PoE slots (inline power is not shared among
PoE modules).
Before you install your PoE module, consult your sales team to determine the required power budget.
Power Delivery
This section describes how the system provides power to the PDs.
Enabling PoE to the Switch

You enable or disable inline power to the entire switch, or per slot or per port.
If you are working on a BlackDiamond 8000 family of switches chassis, you must reserve power for
each PoE slot. By default, 50 watts of inline power is provided to each slot. If you are working with a
Summit X450e-48p switch whether or not included in a SummitStack, you reserve power for the entire
switch, depending on the configuration of the optional EPS-600 PSUs. (Refer to “Power Reserve Budget
on the Summit X450e-48p Switch and Per Slot on Modular Switches” for information on reserving
power on these devices.)
To enable inline power to the switch, slot, or port, use the following commands:
enable inline-power
To disable inline power to the switch, use the following command:
disable inline-power
Disabling inline power removes power immediately to all connected PDs. The default value is enabled.

Power Over Ethernet
Power Reserve Budget on the Summit X450e-48p Switch and Per

Slot on Modular Switches
On modular switches, the power budget is provided on a per slot basis, not switchwide. You reserve
power for each slot, or PoE module. Power reserved for a specific PoE module cannot be used by any
other slot regardless of how much power is actually consumed on the specified slot. The default power
budget reserved for each PoE module is 50 W. The minimum power you can assign to a slot is 37 W, or
0 W if the slot is disabled. The maximum possible for each slot is 768 W.
For the Summit X450e-24p and X450e-48p, the power budget is set by the capability of the power
supplies connected. For each of these Summit switches, the internal PSU is capable of 370 watts of PoE
power.
For the Summit X450e-48p switch, the PoE capability is increased to 740 watts when operating with the
optional EPS-C with 2 or 3 EPS-600LS modules installed. (refer to the Extreme Networks Consolidated
Hardware Guide for complete information on power availability with this optional unit.)
To reduce the chances of ports fluctuating between powered and non-powered states, newly inserted
PDs are not powered when the actual delivered power for the module or switch is within
approximately 19 W of the configured inline power budget for that slot. However, actual aggregate
power can be delivered up to the configured inline power budget for the slot or switch (for example,
when delivered power from ports increases or when the configured inline power budget for the slot is
reduced).
NOTE
Extreme Networks recommends that, when using a modular switch, you fully populate a single PoE module with PDs
until the power usage is just below the usage threshold, instead of spacing PDs evenly across PoE modules.
If you disable a slot with a PoE module, the reserved power budget remains with that slot until you
unconfigure or reconfigure the power budget. Also, you can reconfigure the reserved power budget for
a PoE module without disabling the device first; you can reconfigure dynamically. These settings are
preserved across reboots and other power-cycling conditions.
The total of all reserved slot power budgets cannot be larger than the total available power to the
switch. If the base module power requirements plus the reserved PoE power for all modules exceeds
the unallocated power in the system, the lowest numbered slots have priority in getting power and one
or more modules in higher-numbered slots will be powered down.
NOTE
On modular switches, PoE modules are not powered-up at all, even in data-only mode, if the reserved PoE power
cannot be allocated to that slot.
To reset the reserved power budget for a slot to the default value of 50 W, use the following command:
unconfigure inline-power budget slot <slot>

Power Delivery
PD Disconnect Precedence on the Summit X450e-48p Switch and

Modular Switches
After a PD is discovered and powered on a Summit X450e-48p switch or a modular PoE switch, the
actual power drain is continuously measured. If the usage for power by PDs is within 19 W of the
reserved power budget for the PoE switch or module, the system begins denying power to PDs.
To supply power to all PDs, you can reconfigure the reserved power budget for the switch or slot, so
that enough power is available to power all PDs. You reconfigure the reserved power budget
dynamically; you do not have to disable the device to reconfigure the power budget.
You configure the switch to handle a request for power that exceeds the power budget situation in one
of two ways, called the disconnect precedence:
● Disconnect PDs according to the configured PoE port priority for each PD.
● Deny power to the next PD requesting power, regardless of that port’s PoE priority.
On modular switches, this is a switchwide configuration that applies to each slot; you cannot configure
this disconnect precedence per slot.
The default value is deny-port. So, if you do not change the default value and the switch’s or slot’s
power is exceeded, the next PD requesting power is not connected (even if that port has a higher
configured PoE port priority than those ports already receiving power). When you configure the deny-
port value, the switch disregards the configured PoE port priority and port numbering.
When the switch is configured for lowest-priority mode, PDs are denied power based on the individual
port’s configured PoE priority. If the next PD requesting power is of a higher configured PoE priority
than an already powered port, the lower-priority port is disconnected and the higher-priority port is
powered.
To configure the disconnect precedence for the switch, use the following command:
configure inline-power disconnect-precedence [deny-port | lowest-priority]
To reset the disconnect precedence value to the default value of deny port to the switch, use the
following command:
unconfigure inline-power disconnect-precedence
PoE Port Priority on the Summit X450-48p Switch and Modular Switches
On the Summit X450e-48p switch and modular switches, you can configure the PoE priority for each
port as low, high, or critical; the default value is low. If you configure the disconnect precedence of the
switch as lowest priority, the switch disconnects those PDs with lower PoE port priorities when the
reserved switch or slot power budget is exceeded; the system continues supplying power to PDs with
higher PoE port priorities.
To set the PoE port priority, use the following command:
configure inline-power priority [critical | high | low] ports <port_list>
To reset the PoE priority of the ports to the default value of low, use the following command:
unconfigure inline-power priority ports [all | <port_list>]

Power Over Ethernet
If several PDs have the same configured PoE port priority, the priority is determined by the port
number. The highest port number has the lowest PoE priority.
The switch withdraws power (or disconnects) those ports with the highest port number (s). That is, the
highest port number is the lowest PoE priority.
Port Disconnect or Fault

On modular PoE switches, when a port is disconnected, the power is removed from that port and can
be used only by ports on the same slot. The power from the disconnected port is not redistributed to
any other slot.
On all PoE devices, when a port enters a fault state because of a class violation or if you set the operator
limit lower than the amount requested by the PD, the system removes power from that port. The power
removed is, again, available only to other ports on the same slot or stand-alone switch; it cannot be
redistributed to other slots on modular switches. The port stays in the fault state until you disable that
port, or disconnect the attached PD, or reconfigure the operator limit to be high enough to satisfy the
PD requirements.
To display the status of PoE ports, including disconnected or faulted ports, use the following command:
show inline-power info ports
When a port is disconnected or otherwise moves into a fault state, SNMP generates an event (after you
configure SNMP and a log message is created).
Port Power Reset

You can set ports to experience a power-down, discover, power-up cycle.
On the Summit X450e-48p switch and modular PoE switches, this power-cycling occurs without
returning the power to the slot’s reserved power budget. This function allows you to reset PDs without
losing their claim to the reserved power budget.
To power cycle specified ports, use the following commands:
reset inline-power ports <port_list>
Ports are immediately depowered and repowered, maintaining current power allocations on modular
switches.
PoE Usage Threshold

The system generates an SNMP event when any slot or stand-alone switch has consumed a specified
percentage of that slot’s reserved power budget or of the entire power for the stand-alone switch. The
default value is 70%; you can configure this threshold to generate events from 1% to 99% consumption
of the reserved power budget. You can also configure the system to log an Event Management System
(EMS) message when the usage threshold is crossed (refer to Chapter 5, “Status Monitoring and
Statistics,” for more information on EMS). On modular switches, this threshold percentage is set to be
the same for each PoE slot; you cannot configure it differently for each PoE module.

Power Delivery
On modular switches, although the threshold percentage of measured to budgeted power applies to all
PoE modules, the threshold measurement applies only to the percentage per slot of measured power to
budgeted power use; it does not apply to the amount of power used switchwide.
To configure the threshold percentage of budgeted power used on a slot or the total power on a stand-
alone switch that causes the system to generate an SNMP event and EMS message, use the following
command:
configure inline-power usage-threshold <threshold>
To reset the threshold that causes the system to generate an SNMP event and EMS message per slot to
70% for measured power compared to budgeted power, use the following command:
unconfigure inline-power usage-threshold
Legacy Devices
ExtremeXOS software allows the use of non-standard PDs with the switch. These are PDs that do not
comply with the IEEE 802.3af standard.
The system detects non-standard PDs using a capacitance measurement. You must enable the switch to
detect legacy devices; the default value is disabled. You configure the detection of legacy PoE devices
per slot.
Detecting a PD through capacitance is used only if the following two conditions are both met:
● Legacy PD detection is enabled.
● The system unsuccessfully attempted to discover the PD using the standard resistance measurement
method.
To enable the switch to use legacy PDs on a modular switch, use the following command:
enable inline-power legacy slot <slot>
To enable the switch to use legacy PDs on a stand-alone switch, use the following command:
enable inline-power legacy
To disable the non-standard power detection method that allows the switch to use legacy PDs on a
modular switch, use the following command:
disable inline-power legacy slot <slot>
To disable the non-standard power detection method that allows the switch to use legacy PDs on a
stand-alone switch, use the following command:
disable inline-power legacy
PoE Operator Limits

You set the power limit that a PD can draw on the specified ports. The range is 3000 to 16800 mW, and
the default value is 15400 mW.

Power Over Ethernet
You set the operator limit on specified ports, which limits how much power a PD can draw from that
port by using the following command:
configure inline-power operator-limit <milliwatts> ports [all |<port_list>]
If the measured power for a specified port exceeds the port’s operator limit, the power is withdrawn
from that port and the port moves into a fault state.
To reset the power limit allowed for PDs to the default value of 15.4 W per port, use the following
command:
unconfigure inline-power operator-limit ports [all |<port_list>]
If you attempt to set an operator-limit outside the accepted range, the system returns an error message.
Configuring PoE
PoE supports a full set of configuration and monitoring commands that allow you configure, manage,
and display PoE settings at the system, slot, and port level. Refer to the ExtremeXOS Command Reference
Guide for complete information on using the CLI commands.
To enable inline power, or PoE, you must have a powered switch or chassis and module.
NOTE
On a module switch, if your chassis has an inline power module and there is not enough power to supply a slot, that
slot will not power on; the slot will not function in data-only mode without enough power for inline power.
To configure inline power, or PoE, you must accomplish the following tasks:
● Enable inline power to the system, slot, and/or port.
● On modular switches and the Summit X450e-48p switch, reserve power to the switch or slot, using a
power budget.
● On modular switches and the Summit X450e-48p switch, configure the disconnect precedence for the
PDs in the case of excessive power demands.
● Configure the threshold for initiating system alarms on power usage.
Additionally, you can configure the switch to use legacy PDs, apply specified PoE limits to ports, apply
labels to PoE ports, and configure the switch to allow you to reset a PD without losing its power
allocation.
Enabling Inline Power

You enable inline power to the switch, slot, or port using the following commands:
enable inline-power
enable inline-power slot <slot>
enable inline-power ports [all | <port_list>]

Configuring PoE
NOTE
On modular switches, if your chassis has an inline power module and there is not enough power to supply a slot,
that slot will not power on; the slot will not function in data-only mode without enough power for inline power.
To disable inline power to the switch, slot (on modular switches), or port, use the following commands:
disable inline-power
disable inline-power slot <slot>
disable inline-power ports [all | <port_list>]
Disabling the inline power to a PD immediately removes power from the PD.
To display the configuration for inline power, use the following command:
show inline-power
Reserving Power for the Summit X450e-48p Switch or a Slot on

Modular Switches
On modular PoE switches, you reserve power for a given slot. The power reserved for a given slot
cannot be used by any other PoE slots, even if the assigned power is not entirely used. To reallocate
power among the slots, you must reconfigure each slot for the power budget you want; the power is
not dynamically reallocated among PoE modules.
On the Summit X450e-48p switch (whether or not included in a SummitStack), which operates with the
optional EPS-600 PSU, the power budget is set by the capability of the power supplies connected. For
the X450e-48p, the internal PSU is capable of 370 watts of PoE power. (Refer to the Extreme Networks
Consolidated Hardware Guide for complete information on power availability with this optional unit.)
You do not have to disable the PoE devices to reconfigure the power budgets.
To set the budgeted power reserved for all PDs on a Summit X450e-48p switch on a slot, use the
following command:
configure inline-power budget <num_watts> {slot <slot>}
On modular switches, the default power budget is 50 W per slot, and the maximum is 768 W. The
minimum reserved power budget you can configure is 37 W for an enabled slot. If inline power on the
slot is disabled, you can configure a power budget of 0.
NOTE
Extreme Networks recommends that you fully populate a single PoE module with PDs until the power usage is just
below the usage threshold, instead of spacing PDs evenly across PoE modules.
To reset the power budget for a PoE module to the default value of 50 W, use the following command:
unconfigure inline-power budget slot <slot>
To display the reserved power budget for the PoE modules, use the following command:
show inline-power slot <slot>

Power Over Ethernet
To display the power budget for the X450e-48p, use the following command:
show inline-power
Setting the Disconnect Precedence on a Summit X450e-48p

Switch or Modular Switches
NOTE
The switch generates an SNMP event if a PD goes offline, and the port’s state moves from Power to Searching. You
must configure SNMP to generate this event.
When the actual power used by the PDs on a switch or slot exceeds the power budgeted for that switch
or slot, the switch refuses power to PDs. There are two methods used by the switch to refuse power to
PDs, and whichever method is in place applies to all PoE slots in the switch. This is called the
disconnect precedence method, and you configure one method for the entire switch.
The available disconnect precedence methods are:

● Deny port
● Lowest priority
The default value is deny port. Using this method, the switch simply denies power to the next PD
requesting power from the slot, regardless of that port’s PoE priority or port number.
Using the lowest priority method of disconnect precedence, the switch disconnects the PDs connected to
ports configured with lower PoE priorities. (Refer to “Configuring the PoE Port Priority on the Summit
X450e-48p Switch and Modular Switches” for information on port priorities.)
When several ports have the same PoE priority, the lower port numbers have higher PoE priorities.
That is, the switch withdraws power (or disconnects) those ports with the highest port number(s).
The system keeps dropping ports, using the algorithm you selected with the disconnect ports command,
until the measured inline power for the slot is lower than the reserved inline power.
To configure the disconnect precedence for the switch, use the following command:
configure inline-power disconnect-precedence [deny-port | lowest-priority]
To return the disconnect precedence to the default value of deny port, use the following command:
unconfigure inline-power disconnect-precedence
To display the currently configured disconnect precedence, use the following command:
show inline-power
PDs are not powered when the actual delivered power for the switch or module is within
approximately 19 W of the configured inline power budget for that switch or slot. However, actual
aggregate power can be delivered up to the configured inline power budget for the switch or slot (for
example, when delivered power from ports increases or when the configured inline power budget for
the slot is reduced).

Configuring PoE
Configuring the PoE Port Priority on the Summit X450e-48p Switch and Modular
Switches
You can configure the PoE port priority to be low, high, or critical. The default value is low.
If you configure the disconnect precedence as lowest priority and the PDs request power in excess of
the switch’s or slot’s reserved power budget, the system allocates power to those ports with the highest
priorities first.
If several ports have the same PoE priority, the lower port numbers have higher PoE priorities. That is,
the switch withdraws power (or disconnects) those ports with the highest port number(s).
To configure PoE port priority, use the following command:
configure inline-power priority [critical | high | low] ports <port_list>
To reset the port priority to the default value of low, use the following command:
unconfigure inline-power priority ports [all | <port_list>]
To display the PoE port priorities, use the following command:
show inline-power configuration ports <port_list>
Configuring the Usage Threshold

The system generates an SNMP event after a preset percentage of the reserved power for any slot or
total power for a stand-alone switch is actually used by a connected PD. This preset percentage is called
the usage threshold and is the percentage of the measured power to the budgeted power for each slot
or total power for a stand-alone switch.
On modular switches, although the percentage of used to budgeted power is measured by each PoE
module, you set the threshold for sending the event for the entire switch. That is, after any PoE module
passes the configured threshold, the system sends an event.
The default value for this usage threshold is 70%. You can configure the usage threshold to be any
integer between 1% and 99%.
To configure the usage threshold, use the following command:
configure inline-power usage-threshold <threshold>
To reset the usage threshold to 70%, use the following command:
unconfigure inline-power usage-threshold
To display the currently configured usage threshold, use the following command:
show inline-power

Power Over Ethernet
Configuring the Switch to Detect Legacy PDs

The PoE device can detect non-standard, legacy PDs, which do not conform to the IEEE 802.3af
standard, using a capacitance measurement. However, you must specifically enable the switch to detect
these non-standard PDs; the default value for this detection method is disabled.
This configuration applies to the entire switch; you cannot configure the detection method per slot.
The switch detects PDs through capacitance only if both of the following conditions are met:
● The legacy detection method is enabled.
● The switch unsuccessfully attempted to discover the PD using the standard resistance measurement
method.
To enable the switch to detect legacy, non-standard PDs, use the following command:
enable inline-power legacy slot <slot>
To reset the switch to the default value, which does not detect legacy PDs, use the following command:
disable inline-power legacy slot <slot>
To display the status of legacy detection, use the following command:
show inline-power
Configuring the Operator Limit

You configure the maximum amount of power that the specified port can deliver to the connected PD,
in milliwatts (mW). The default value is 15400 mW, and the range is 3000 to 16800 mW.
If the operator limit for a specified port is less than the power drawn by the legacy PD, the legacy PD is
denied power.
To configure the operator limit, use the following command:
configure inline-power operator-limit <milliwatts> ports [all |<port_list>]
To reset the operator limit to the default value of 15.4 W, use the following command:
unconfigure inline-power operator-limit ports [all |<port_list>]
To display the current operator limit on each port, use the following command:
Configuring PoE Port Labels

You can assign labels to a single or group of PoE ports using a string of up to 15 characters. To assign a
label to PoE ports, use the following command:
configure inline-power label <string> ports <port_list>
To rename a port or to return it to a blank label, reissue the command.

Displaying PoE Settings and Statistics
To display the PoE port labels, use the following command:

Power Cycling Connected PDs

To power cycle a connected PD without losing the power allocated to its port, use the following
command:
reset inline-power ports <port_list>

You can display the PoE status, configuration, and statistics for the system, slot, and port levels.
Clearing Statistics
You can clear the PoE statistics for specified ports or for all ports. To clear the statistics and reset the
counters to 0, use the following command:
clear inline-power stats ports [all | <port_list>]
Displaying System Power Information

You can display the status of the inline power for the system and, for additional information, display
the power budget of the switch.
Displaying System PoE Status

To display the PoE status for the switch, use the following command:
show inline-power
The command provides status for the following areas:

● Configured inline power status—The status of the inline power for the switch: enabled or disabled.
● System power surplus—The surplus amount of power on the system, in watts, available for
budgeting.
● Redundant power surplus—The amount of power on the system, in watts, available for budgeting if
one power supply is lost.
● System power usage threshold—The configured power usage threshold for each slot, shown as a
percentage of budgeted power. After this threshold has been passed on any slot, the system sends an
SNMP event and logs a message.
● Disconnect precedence—The method of denying power to PDs if the budgeted power on any slot is
exceeded.
● Legacy mode—The status of the legacy mode, which allows detection of non-standard PDs.

Power Over Ethernet
The output indicates the following inline power status information for each slot:
● Inline power status—The status of inline power. The status conditions are:
■ Enabled
■ Disabled
● Firmware status—The operational status of the slot. The status conditions are:
■ Operational
■ Not operational
■ Disabled
■ Subsystem failure
■ Card not present
■ Slot disabled
● Budgeted power—The amount of inline power, in watts, that is reserved and available to the slot.
● Measured power—The amount of power, in watts, that is currently being used by the slot.
Displaying System Power Data

Additionally, you can view the distribution of power, as well as currently required and allocated
power, on the entire modular switch including the power supplies by using the following command:
show power budget
Displaying Slot PoE Information on Modular Switches

You can display PoE status and statistics per slot.
Displaying Slot PoE Status

To display PoE status for each slot, use the following command:
show inline-power slot <slot>
The command provides the following information:

● Inline power status—The status of inline power. The status conditions are:
■ Enabled
■ Disabled
● Firmware status—The operational status of the slot. The status conditions are:
■ Operational
■ Not operational
■ Disabled
■ Slot disabled
● Budgeted power—The amount of power, in watts, that is available to the slot.
● Measured power—The amount of power, in watts, that is currently being used by the slot.

Displaying Slot PoE Statistics on Modular Switches

To display the PoE statistics for each slot, use the following command:
show inline-power stats slot <slot>

● Firmware status—Displays the firmware state:
■ Operational
■ Not operational
■ Disabled
■ Slot disabled
● Firmware revision—Displays the revision number of the PoE firmware
● Total ports powered—Displays the number of ports powered on specified slot
● Total ports awaiting power—Displays the number of remaining ports in the slot that are not
powered
● Total ports faulted—Displays the number of ports in a fault state
● Total ports disabled—Displays the number of ports in a disabled state
Displaying PoE Status and Statistics on Stand-alone Switches

To display the PoE statistics for the switch, use the following command:
show inline-power stats

● Firmware status—Displays the firmware state:
■ Operational
■ Not operational
■ Disabled
● Firmware revision—Displays the revision number of the PoE firmware
● Total ports powered—Displays the number of ports powered on specified slot
● Total ports awaiting power—Displays the number of remaining ports in the slot that are not
powered
● Total ports faulted—Displays the number of ports in a fault state
● Total ports disabled—Displays the number of ports in a disabled state
Displaying Port PoE Information

You can display PoE configuration, status, and statistics per port.

Power Over Ethernet
Displaying Port PoE Configuration

To display PoE configuration for each port, use the following command:
This command provides the following information:

● Config—Indicates whether the port is enabled to provide inline power:
■ Enabled: The port can provide inline power.
■ Disabled: The port cannot provide inline power.
● Operator Limit—Displays the configured limit, in milliwatts, for inline power on the port.
● Label—Displays a text string, if any, associated with the port (15 characters maximum).
Displaying Port PoE Status

To display the PoE status per port, use the following command:
show inline-power info {detail} ports <port_list>
This command provides the following information:

● State—Displays the port power state:
■ Disabled
■ Searching
■ Delivering
■ Faulted
■ Disconnected
■ Other
■ Denied
● PD’s power class—Displays the class type of the connected PD:
■ “-----”: disabled or searching
■ “class0”: class 0 device
● Volts—Displays the measured voltage. A value from 0 to 2 is valid for ports that are in a searching
or discovered state.
● Curr—Displays the measured current, in milliamperes, drawn by the PD.
● Power—Displays the measured power, in watts, supplied to the PD.
● Fault—Displays the fault value:
■ None
■ UV/OV fault
■ UV/OV spike
■ Over current
■ Overload
■ Undefined

■ Underload
■ HW fault
■ Discovery resistance fail
■ Operator limit violation
■ Disconnect
■ Discovery resistance, A2D failure
■ Classify, A2D failure
■ Sample, A2D failure
■ Device fault, A2D failure
■ Force on error
The detail command lists all inline power information for the selected ports. Detail output displays the
following information:
● Configured Admin State
● Inline Power State
● MIB Detect Status
● Label
● Operator Limit
● PD Class
● Max Allowed Power
● Measured Power
● Line Voltage
● Current
● Fault Status
● Detailed Status
● Priority
Displaying Port PoE Statistics

To display the PoE statistics for each port, use the following command:
show inline-power stats ports <port_list>

● State—Displays the port power state:
■ Disabled
■ Searching
■ Delivering
■ Faulted
■ Disconnected
■ Other
■ Denied
● PD’s power class—Displays the class type of the connected PD:
■ “-----”: disabled or searching

Power Over Ethernet

● Absent—Displays the number of times the port was disconnected
● InvSig—Displays the number of times the port had an invalid signature
● Denied—Displays the number of times the port was denied
● Over-current—Displays the number of times the port entered an overcurrent state
● Short—Displays the number of times the port entered undercurrent state

11 Status Monitoring and Statistics

● Overview of Status Monitoring on page 231
● Viewing Port Statistics on page 231
● Viewing Port Errors on page 232
● Using the Port Monitoring Display Keys on page 234
● Viewing VLAN Statistics on page 234
● Performing Switch Diagnostics on page 235
● Using the System Health Checker on page 245
● Setting the System Recovery Level on page 250
● Using ELSM on page 261
● Viewing Fan Information on page 271
● Viewing the System Temperature on page 272
● Using the Event Management System/Logging on page 274
● Using sFlow on page 286
● Using RMON on page 293
Viewing statistics on a regular basis allows you to see how well your network is performing. If you
keep simple daily records, you can see trends emerging and notice problems arising before they cause
major network faults. In this way, statistics can help you get the best out of your network.
Overview of Status Monitoring

The status monitoring facility provides information about the switch. This information may be useful
for your technical support representative if you have a problem. ExtremeXOS includes many command
line interface (CLI) show commands that display information about different switch functions and
facilities.
NOTE
For more information about show commands for a specific ExtremeXOS feature, see the appropriate chapter in this
guide.
Viewing Port Statistics

ExtremeXOS provides a facility for viewing port statistical information. The summary information lists
values for the current counter for each port on each operational module in the system. Beginning with
ExtremeXOS 11.3, the switch automatically refreshes the display (this is the default behavior).

Status Monitoring and Statistics
You can also display a snapshot of the real-time port statistics at the time you issue the command and
view the output in a page-by-page mode (this was the default behavior in ExtremeXOS 11.2 and
earlier). This setting is not saved; therefore, you must specify the no-refresh parameter each time you
want a snapshot of the port statistics.
Values are displayed to nine digits of accuracy.
To view port statistics, use the following command:
show ports {stack-ports <stacking-port-list> | <port_list>} statistics {no-refresh}
The switch collects the following port statistical information:

● Link Status—The current status of the link. Options are:
■ Active (A)—The link is present at this port.
■ Ready (R)—The port is ready to accept a link.
■ Loopback (L)—The port is configured for WANPHY loopback.
■ Not Present (NP)—The port is configured, but the module is not installed in the slot (modular
switches only).
● Transmitted Packet Count (Tx Pkt Count)—The number of packets that have been successfully
transmitted by the port.
● Transmitted Byte Count (Tx Byte Count)—The total number of data bytes successfully transmitted
by the port.
● Received Packet Count (Rx Pkt Count)—The total number of good packets that have been received
by the port.
● Received Byte Count (RX Byte Count)—The total number of bytes that were received by the port,
including bad or lost frames. This number includes bytes contained in the Frame Check Sequence
(FCS), but excludes bytes in the preamble.
● Received Broadcast (RX Bcast)—The total number of frames received by the port that are addressed
to a broadcast address.
● Received Multicast (RX Mcast)—The total number of frames received by the port that are addressed
to a multicast address.
You can also view port statistics for SummitStack stacking ports using the following command:
show ports stack-ports {<stacking-port-list>} statistics {no-refresh}
Viewing Port Errors

The switch keeps track of errors for each port. Beginning with ExtremeXOS 11.3, the switch
automatically refreshes the display (this is the default behavior).
You can also display a snapshot of the port errors at the time you issue the command and view the
output in a page-by-page mode (this was the default behavior in ExtremeXOS 11.2 and earlier). This
setting is not saved; therefore, you must specify the no-refresh parameter each time you want a
snapshot of the port errors.
To view port transmit errors, use the following command:

show ports {stack-ports <stacking-port-list> | <port_list>} txerrors {no-refresh}

Viewing Port Errors
The switch collects the following port transmit error information:

● Port Number—The number of the port.
■ Loopback (L)—The port is configured for WANPHY loopback.
switches only).
● Transmit Collisions (TX Coll)—The total number of collisions seen by the port, regardless of whether
a device connected to the port participated in any of the collisions.
● Transmit Late Collisions (TX Late Coll)—The total number of collisions that have occurred after the
port’s transmit window has expired.
● Transmit Deferred Frames (TX Deferred)—The total number of frames that were transmitted by the
port after the first transmission attempt was deferred by other network traffic.
● Transmit Errored Frames (TX Errors)—The total number of frames that were not completely
transmitted by the port because of network errors (such as late collisions or excessive collisions).
● Transmit Lost Frames (TX Lost)—The total number of transmit frames that do not get completely
transmitted because of buffer problems (FIFO underflow).
● Transmit Parity Frames (TX Parity)—The bit summation has a parity mismatch.
To view port receive errors, use the following command:
show ports {<port_list> | stack-ports <stacking-port-list>} rxerrors {no-refresh}
The switch collects the following port receive error information:

● Port Number
switches only).
■ Loopback (L)—The port is in Loopback mode.
● Receive Bad CRC Frames (RX CRC)—The total number of frames received by the port that were of
the correct length but contained a bad FCS value.
● Receive Oversize Frames (RX Over)—The total number of good frames received by the port greater
than the supported maximum length of 1,522 bytes.
● Receive Undersize Frames (RX Under)—The total number of frames received by the port that were
less than 64 bytes long.
● Receive Fragmented Frames (RX Frag)—The total number of frames received by the port that were
of incorrect length and contained a bad FCS value.
● Receive Jabber Frames (RX Jabber)—The total number of frames received by the port that were
greater than the support maximum length and had a Cyclic Redundancy Check (CRC) error.
● Receive Alignment Errors (RX Align)—The total number of frames received by the port with a CRC
error and not containing an integral number of octets.
● Receive Frames Lost (RX Lost)—The total number of frames received by the port that were lost
because of buffer overflow in the switch.

For SummitStack stacking ports, you can also view transmit and receive errors with the following
commands:
show ports stack-ports {<stacking-port-list>} txerrors {no-refresh}
show ports stack-ports {<stacking-port-list>} rxerrors {no-refresh}
Information displayed is identical to the details displayed for non-stacking ports.
Using the Port Monitoring Display Keys

Table 31 describes the keys used to control the displays that appear if you use any of the show ports
commands without specifying the no-refresh parameter (this is the default behavior).
Table 31: Port monitoring display keys with auto-refresh enabled

Key(s) Description
U Displays the previous page of ports.
D Displays the next page of ports.
[Esc] Exits from the screen.
0 Clears all counters.
[Space] Cycles through the following screens:
• Packets per second
• Bytes per second
• Percentage of bandwidth
NOTE: Available only using the show ports utilization
command.
Table 32 describes the keys used to control the displays that appear if you use any of the show ports
commands and specify the no-refresh parameter.
Table 32: Port monitoring displays keys with auto-refresh disabled

Key Description
Q Exits from the screen.
[Space] Displays the next page of ports.
Viewing VLAN Statistics

Beginning with ExtremeXOS 12.0, ExtremeXOS provides the facility for viewing VLAN statistics at the
port level and at the VLAN level on BlackDiamond 12800 series switches.
To configure the switch to start counting VLAN statistics, use the following commands:
clear counters
configure port [<port_list> | all} monitor vlan <vlan name>
Up to four VLANs can be monitored on the same port by issuing the command up to four times.

Performing Switch Diagnostics
To view VLAN statistics at the port level, use the following command:
show ports {port_list} vlan statistics {no-refresh}
The switch collects and displays the following statistics:
● Port—The designated port.

● VLAN—The associated VLANs.
● Rx Frames Count—The total number of frames successfully received by the port.
● Rx Byte Count—The total number of bytes that were received by the port.
● Tx Total Frames—The total number of frames that were transmitted by the port.
● Tx Byte Count—The total number of bytes that were transmitted by the port.
To view VLAN statistics at the VLAN level, use the following command:
show vlan <vlan_name> statistics {no-refresh}
The switch collects and displays the following statistics:
● VLAN—The designated VLAN.

● Rx Frames Count—The total number of frames successfully received by the port.
● Rx Byte Count—The total number of bytes that were received by the port.
● Tx Total Frames—The total number of frames that were transmitted by the port.
● Tx Byte Count—The total number of bytes that were transmitted by the port.
To stop counting VLAN statistics use the following command:
unconfigure ports {<port_list> | all} monitor vlan <vlan name>

The switch provides a facility for running normal or extended diagnostics. In simple terms, a normal
routine performs a simple ASIC and packet loopback test on all ports, and an extended routine
performs extensive ASIC, ASIC-memory, and packet loopback tests. By running and viewing the results
from diagnostic tests, you can troubleshoot and resolve network issues.
On the BlackDiamond 10808 switch and the BlackDiamond 8800 series switch, you run the diagnostic
routine on Input/Output (I/O) modules or Management Switch Fabric Modules (MSMs) without
affecting the operation of the rest of the system.
On the BlackDiamond 12800 series switches, you run the diagnostic routine on all installed I/O
modules and the primary MSM. The switch does not test the backup MSM on the BlackDiamond 12804.
Running diagnostics affects system operation.
On the Summit family of switches, you run the diagnostic routine on the switch or on the stacking
ports. Running the switch or stacking port diagnostic routine affects system operation; the switch is
unavailable during the diagnostic test.

When you run diagnostics on an I/O module, an MSM, or the Summit family of switches, the switch
verifies that the:
● Registers can be written to and read from correctly.
● Memory addresses are accessed correctly.
● Application-Specific Integrated Circuit (ASICs) and Central Processing Unit (CPUs) operate as
required.
● Data and control fabric connectivity is active (modular switches only).
● External ports can send and receive packets.
● Sensors, hardware controllers, and LEDs are working correctly.
NOTE
Before running slot diagnostics on a modular switch, you must have at least one MSM installed in the chassis.
When you run diagnostics on the SummitStack stacking ports, the switch completes a hardware test to
ensure that the stacking ports are operational.
The remainder of this section describes the following topics:

● Running Diagnostics on the BlackDiamond 10808 Switch and the BlackDiamond 8800 Series
Switches on page 236
● Running Diagnostics on the BlackDiamond 12800 Series Switches on page 237
● Running Diagnostics on the SummitStack or Summit Family of Switches on page 238
● Observing LED Behavior During a Diagnostic Test on page 239
● Displaying Diagnostic Test Results on page 245
Running Diagnostics on the BlackDiamond 10808 Switch and the

BlackDiamond 8800 Series Switches
If you run the diagnostic routine on an I/O module, that module is taken offline while the diagnostic
test is performed. Traffic to and from the ports on that I/O module is temporarily unavailable. When
the diagnostic test is complete, the I/O module is reset and becomes operational again.
If you run diagnostics on an MSM, that module is taken offline while the diagnostics test is performed.
When the diagnostic test is complete, the MSM reboots and becomes operational again.
If you run diagnostics on the primary MSM, the backup MSM assumes the role of the primary and
takes over switch operation. After the MSM completes the diagnostic routine and reboots, you can
initiate failover from the new primary MSM to the original primary MSM. Before initiating failover,
confirm that both MSMs are synchronized using the show switch command. If the MSMs are
synchronized, initiate failover using the run msm-failover command. For more detailed information
about system redundancy and MSM failover, see “Understanding System Redundancy—Modular
Switches and SummitStack Only” on page 77.
Run diagnostics on one MSM at a time. After you run the diagnostic routine on the first MSM, use the
show switch command to confirm that both MSMs are up, running, and synchronized before running
diagnostics on the second MSM.
After the switch runs the diagnostic routine, test results are saved in the module’s EEPROM and
messages are logged to the syslog.

To run diagnostics on I/O or MSM modules, use the following command:
run diagnostics [extended | normal] {slot [<slot> | A | B]}

● extended—Takes the switch fabric and ports offline and performs extensive ASIC, ASIC-memory,
and packet loopback tests. Extended diagnostic tests take a maximum of 15 minutes. The CPU is not
tested. Console access is available during extended diagnostics.
If you have a Power over Ethernet (PoE) module installed, the switch also performs an extended PoE
test, which tests the functionality of the inline power adapter.
● normal—Takes the switch fabric and ports offline and performs a simple ASIC and packet loopback
test on all ports.
● <slot>—Specifies the slot number of an I/O module. When the diagnostics test is complete, the
system attempts to bring the I/O module back online.
NOTE
BlackDiamond 8810 switch —If you run diagnostics on slots 5 and 6 with an MSM installed in those slots, the
diagnostic routine tests the I/O subsystem of the MSM.
BlackDiamond 8806 switch—if you run diagnostics on slots 3 and 4 with an MSM installed in those slots, the
diagnostic routine tests the I/O subsystem of the MSM.
BlackDiamond 8800 series switch—To run diagnostics on the management portion of the master MSM, specify
slot A or B.
● A | B—Specifies the slot letter of the primary MSM. The diagnostic routine is performed when the
system reboots. Both switch fabric and management ports are taken offline during diagnostics.
BlackDiamond 8800 Series Switch Only—Before running diagnostics on a module, you can use the
disable slot <slot> offline command to force the module to enter the offline state which takes the
switch fabric and ports offline. If you run diagnostics on a module that is not offline, the switch
automatically takes the switch fabric and ports offline when you use the run diagnostics [extended
| normal | stack-port] {slot [<slot> | A | B]} command.
After the diagnostic routine has finished, use the enable slot <slot> command to bring the module
back online and operational.
Running Diagnostics on the BlackDiamond 12800 Series

Switches
If you run the diagnostic routine on the BlackDiamond 12800 series switches, the diagnostic routine
occurs on all of the installed I/O modules and the primary MSM. If a backup MSM is present, the
switch does not test the backup MSM.
Running diagnostics affects system operation. During a diagnostic routine, the system can be offline
from 5 to 25 minutes depending on the number of modules installed and the type of diagnostic test
performed.
When the diagnostic routine runs on the I/O modules, traffic to and from the ports on the I/O modules
is unavailable. When the diagnostic test is complete, the I/O modules are reset.

When the diagnostic routine runs on the primary MSM, that module is taken offline. If you have a
backup MSM installed, it enters the reset state and remains in this state until the primary MSM finishes
the diagnostic tests. Upon completion of the diagnostic tests, the primary MSM reboots the backup
MSM and then reboots itself.
After the switch runs the diagnostic routine, test results are saved in the each module’s EEPROM and
messages are logged to the syslog.
To run diagnostics on all of the installed I/O modules and the primary MSM, use the following
command:
run diagnostics [extended | normal]

● extended—Takes the switch fabric and ports offline and performs extensive ASIC, ASIC-memory,
and packet loopback tests. Extended diagnostic tests take a maximum of 25 minutes. The CPU is not
tested. Console access is available during extended diagnostics.
● normal—Takes the switch fabric and ports offline and performs a simple ASIC and packet loopback
test on all ports. Normal diagnostic tests take a minimum of 5 minutes.
To run diagnostics on the backup MSM, you must failover to the backup MSM, thereby relinquishing
primary MSM status to the backup. After failover, you can run diagnostics on the new primary MSM.
Remember, the diagnostic routine runs on all of the installed I/O modules in addition to the new
primary MSM. For more detailed information about system redundancy and MSM failover, see
“Understanding System Redundancy—Modular Switches and SummitStack Only” on page 77.
Running Diagnostics on the SummitStack or Summit Family of

Switches
In a SummitStack, diagnostics can not be run. You need to disable stacking on the switch to be tested,
and reboot the switch before logging in. Then run the diagnostics. Once the diagnostics is complete, the
switch reboots again. Log in, enable stacking mode, and reboot the switch again. Upon reboot, the
switch rejoins the stack. You can then use the "show diagnostics" command to see the last diagnostic
result of any or all switches in the SummitStack.
If you run the diagnostic routine on the Summit family of switches, the switch reboots and then
performs the diagnostic test. During the test, traffic to and from the ports on the switch is temporarily
unavailable. When the diagnostic test is complete, the switch reboots and becomes operational again.
To run the diagnostic routine on the stack ports, you need a dedicated stacking cable that connects stack
port 1 to stack port 2, which are located at the rear of the switch. The stacking cable is available from
Extreme Networks. The switch performs a hardware test to confirm that the stack ports are operational;
traffic to and from the ports on the switch is temporarily unavailable. This Bit Error Rate Test (BERT)
provides an analysis of the number of bits transmitted in error.
After the switch runs the diagnostic routine, test results saved to the switch’s EEPROM and messages
are logged to the syslog.
To run diagnostics on the Summit family of switches, use the following command:
run diagnostics [extended | normal | stack-port]


● extended—Reboots the switch and performs extensive ASIC, ASIC-memory, and packet loopback
tests. Extended diagnostic tests take a maximum of 5 minutes. The CPU is not tested.
● normal—Reboots the switch and performs a simple ASIC and packet loopback test on all ports.
● stack-port—Performs a BERT on the stacking ports and reboots the switch.
Observing LED Behavior During a Diagnostic Test

Whether you run a diagnostic test on an I/O module, an MSM, or the Summit family of switches, LED
activity occurs during and immediately following the test. The LED behavior described in this section
relates only to the behavior associated with a diagnostic test. For more detailed information about all of
the I/O module, MSM, and switch LEDs, see the Extreme Networks Consolidated XOS Hardware
Installation Guide.
I/O Module LED Behavior—BlackDiamond 10808 Switch

Table 33 describes the BlackDiamond 10808 switch I/O module LED behavior during a diagnostic test.
Table 33: BlackDiamond 10808 switch I/O module LED behavior

LED Color Indicates
DIAG Amber blinking Diagnostic test in progress.
Amber Diagnostic failure has occurred.
Status Amber blinking Configuration error, code version error, diagnostic failure, or other severe
module error.
After the I/O module completes the diagnostic test, or the diagnostic test is terminated, the DIAG and
the Status LEDs are reset. During normal operation, the DIAG LED is off and the Status LED blinks
green.
MSM LED Behavior—BlackDiamond 10808 Switch

Table 34 describes the BlackDiamond 10808 switch MSM LED behavior during a diagnostic test.
Table 34: BlackDiamond 10808 switch MSM LED behavior

LED Color Indicates
SYS Amber blinking Diagnostic test in progress.
After the MSM completes the diagnostic test, or the diagnostic test is terminated, the SYS LED is reset.
During normal operation, the status LED blinks green.

I/O Module LED Behavior—BlackDiamond 8800 Series Switch

Table 35 describes the BlackDiamond 8800 series switch I/O module LED behavior during a diagnostic
test.
Table 35: BlackDiamond 8800 series switch I/O module LED behavior
LED Color Indicates
Green Diagnostic test has passed.
Stat Amber blinking Configuration error, code version error, diagnostic failure, or other severe
module error.
Off Diagnostic test in progress, or diagnostic failure has occurred.
After the I/O module completes the diagnostic test, or the diagnostic test is terminated, the DIAG and
green.
MSM LED Behavior—BlackDiamond 8800 Series Switch

This section describes the MSM behavior during a diagnostic test.
LED behavior during a diagnostict test on the primary MSM. Table 36 describes the BlackDiamond 8800
series switch MSM-48 LED behavior during a diagnostic test on the primary MSM.
Table 36: BlackDiamond 8800 series switch MSM-48 LED behavior during diagnostic test on primary
MSM
MSM LED Color Indicates
Primary ERR Off Depending on the situation, this state indicates:
• Diagnostic test in progress on the primary MSM.
• Diagnostic test has passed.
• Diagnostic failure has occurred.
ENV Off Depending on the situation, this state indicates:
Amber blinking Diagnostic test is in progress on the primary MSM.
Mstr/Diag Green/Off Diagnostic failure has occurred.
Off/Green Depending the situation, this state indicates:
Sys/Stat Off/Off Depending on the situation, this state indicates:
Amber/Green Diagnostic failure has occurred.
blinking

Table 36: BlackDiamond 8800 series switch MSM-48 LED behavior during diagnostic test on primary
MSM (Continued)
Backup ERR Off Depending on the situation, this state indicates:
Mstr/Diag Off/Off Diagnostic failure has occurred.
Green/Green Diagnostic test in progress on the primary MSM.
Green/Off Diagnostic test has passed.
Sys/Stat Off/Green blinking Diagnostic test has passed.
Off/Off Diagnostic test in progress on the primary MSM.
Amber/Green Diagnostic failure has occurred.
blinking
Table 37 describes the BlackDiamond 8800 series switch MSM-G8X LED behavior during a diagnostic
test on the primary MSM.
Table 37: BlackDiamond 8800 series switch MSM-G8X LED behavior during diagnostic test on
primary MSM
Primary ERR Off Depending on the situation, this state indicates:
Amber blinking Diagnostic test is in progress on the primary MSM.
Mstr Greenf Diagnostic failure has occurred.
Off Depending the situation, this state indicates:
Sys Off Depending on the situation, this state indicates:

primary MSM (Continued)
Mstr Off Diagnostic failure has occurred.
Green Diagnostic test in progress on the primary MSM.
Green Diagnostic test has passed.
Sys Off Diagnostic test has passed.
Off Diagnostic test in progress on the primary MSM.
LED behavior during a diagnostict test on the backup MSM. Table 38 describes the BlackDiamond 8800
series switch MSM-48 LED behavior during a diagnostic test on the backup MSM.
Table 38: BlackDiamond 8800 series switch MSM-48 LED behavior during diagnostic test on backup
MSM
• Diagnostic test in progress on the backup MSM.
Mstr/Diag Off/Green Depending on the situation, this state indicates:
Sys/Stat Off/Green Diagnostic test in progress on the backup MSM.
Off/Off Diagnostic test has passed.
Primary ERR Amber Depending on the situation, this state indicates:

Table 38: BlackDiamond 8800 series switch MSM-48 LED behavior during diagnostic test on backup
MSM (Continued)
Mstr/Diag Green/Off Depending on the situation, this state indicates:
Sys/Stat Off/Green blinking Depending on the situation, this state indicates:
Table 39 describes the BlackDiamond 8800 series switch MSM-G8X LED behavior during a diagnostic
test on the backup MSM.
backup MSM
Mstr Off Depending on the situation, this state indicates:
Syst Off Diagnostic test in progress on the backup MSM.
Off Diagnostic test has passed.
Primary ERR Amber Depending on the situation, this state indicates:
Mstr Green Depending on the situation, this state indicates:
Sys Off Depending on the situation, this state indicates:

I/O Module LED Behavior—BlackDiamond 12800 Series Switches

Table 40 describes the I/O module LED behavior during a diagnostic test for the R-series and non R-
series I/O modules installed in the BlackDiamond 12800 series switches.
NOTE
You cannot mix R-series and non R-series MSM or I/O modules in the chassis.
Table 40: BlackDiamond 12800 series switch I/O module LED behavior
LED Color Indicates
Status Amber blinking Configuration error, code version error, diagnostic failure, or other severe
module error.
After the I/O modules complete the diagnostic test, or the diagnostic test is terminated, the DIAG and
green. If you start another diagnostic test, the LED returns to blinking amber.
MSM LED Behavior—BlackDiamond 12800 Series Switches

Table 41 describes the MSM LED behavior during a diagnostic test for the R-series and non R-series
MSM modules installed in the BlackDiamond 12800 series switches.
NOTE
You cannot mix R-series and non R-series MSM or I/O modules in the chassis.
Table 41: BlackDiamond 12800 series switch MSM LED behavior

LED Color Indicates
SYS Amber blinking Diagnostic test in progress.
MSTR Off Normal operation for diagnostics.
After the MSM completes the diagnostic test, or the diagnostic test is terminated, the SYS LED is reset.
During normal operation, the SYS LED blinks green.

Using the System Health Checker
LED Behavior—Summit Family of Switches

Table 42 describes the Summit family of switches LED behavior during a diagnostic test.
Table 42: Summit family of switches LED behavior

LED Color Indicates
MGMT Green blinking Normal operation is occurring.
Amber blinking Diagnostic test in progress.
While diagnostic tests are running, the MGMT LED blinks amber. If a diagnostic test fails, the MGMT
LED continues to blink amber. During normal operation, the MGMT LED blinks green.
Displaying Diagnostic Test Results

To display the status of the last diagnostic test run on the switch, use the following command:
show diagnostics {slot [<slot> | A | B]}
NOTE
The slot, A, and B parameters are available only on modular switches.

System health check is a useful tool to monitor the overall health of your system. Depending on your
platform, the software performs a proactive, preventive search for problems by polling and reporting
the health of system components, including I/O and management module processes, power supplies,
power supply controllers, and fans. By isolating faults to a specific module, backplane connection,
control plane, or component, the system health checker notifies you of a possible hardware fault.
This section describes the system health check functionality of the following platforms:
● BlackDiamond 10808 switch
● Summit family of switches
This section also describes the following topics:

● Enabling Backplane Diagnostic Packets on the Switch—Modular Switches Only on page 247
● Configuring Backplane Diagnostic Packets on the Switch—Modular Switches Only on page 248
● Disabling Backplane Diagnostic Packets on the Switch—Modular Switches Only on page 248
● Displaying the System Health Check Setting—All Platforms on page 248

Understanding the System Health Checker—BlackDiamond

10808 and BlackDiamond 12800 Series Switches Only
The BlackDiamond 10808 and the BlackDiamond 12800 series switches support extensive error-checking
and monitoring capabilities. Packet and system memories are protected by an error correction code
(ECC). ECC is capable of correcting all single-bit errors and detecting all other memory errors. The data
path is protected by checksums and parity checks. The system automatically corrects correctable
memory errors and kills packets that encounter checksum and parity errors during processing. Errored
packets are not propagated through the system.
The primary responsibility of the system health checker is to monitor and poll the ASIC error registers.
The system health checker processes, tracks, and reads the memory, parity, and checksum error counts.
The ASICs maintain counts of correctable and uncorrectable memory errors, as well as packets that
encountered checksum and parity errors. In a running system, some of these error counts may show
non-zero values. Occasional increments of these counters does not mean faulty hardware is detected or
that hardware requires replacement. If you see persistent increments of these counters, contact Extreme
Networks Technical Support.
In addition, you can enable the system health checker to check the backplane, CPU, and I/O modules
by periodically sending diagnostic packets and checking the validity of the looped back diagnostic
packets.
In summary, two modes of health checking are available: polling and backplane diagnostic packets.
These methods are briefly described in the following:
● Polling is always enabled on the system and occurs every 60 seconds by default. The system health
checker polls and tracks the ASIC counters that collect correctable and uncorrectable packet memory
errors, checksum errors, and parity errors on a per ASIC basis. By reading and processing the
registers, the system health check detects and associates faults to specific system ASICs.
● Backplane diagnostic packets are disabled by default. If you enable this feature, the system health
checker tests the packet path for a specific I/O module every 6 seconds by default. The MSM sends
and receives diagnostic packets from the I/O module to determine the state and connectivity. (The
other I/O modules with backplane diagnostic packets disabled continue polling every 60 seconds by
default.)
For more information about enabling and configuring backplane diagnostics, see the following
sections:
■ Enabling Backplane Diagnostic Packets on the Switch—Modular Switches Only on page 247
■ Configuring Backplane Diagnostic Packets on the Switch—Modular Switches Only on page 248
System health check errors are reported to the syslog. If you see an error, contact Extreme Networks
Technical Support.
Understanding the System Health Checker—BlackDiamond 8800

Series Switch Only
On the BlackDiamond 8800 series switch, the system health checker tests the backplane, the CPUs on
the MSM modules, the I/O modules, the processes running on the switch, and the power supply
controllers by periodically forwarding packets and checking for the validity of the forwarded packets.

Two modes of health checking are available: polling (also known as control plane health checking) and
backplane diagnostic packets (also known as data plane health checking). These methods are briefly
described in the following:
● Polling is always enabled on the system and occurs every 5 seconds by default. The polling value is
not a user-configured parameter. The system health check polls the control plane health between
MSMs and I/O modules, monitors memory levels on the I/O module, monitors the health of the
I/O module, and checks the health of applications and processes running on the I/O module. If the
system health checker detects an error, the health checker notifies the MSM.
checker tests the data link for a specific I/O module every 5 seconds by default. The MSM sends and
receives diagnostic packets from the I/O module to determine the state and connectivity.
If you disable backplane diagnostics, the system health checker stops sending backplane diagnostic
packets.
For more information about enabling and configuring backplane diagnostics, see the following
sections:
■ Enabling Backplane Diagnostic Packets on the Switch—Modular Switches Only on page 247
■ Configuring Backplane Diagnostic Packets on the Switch—Modular Switches Only on page 248
Technical Support.
NOTE
There are no health checking tests related to the stacking links in a SummitStack.
Understanding the System Health Checker—Summit Family of

Switches Only
On the Summit family of switches, the system health checker polls and reads the switch fabric and CPU
registers. Unlike the modular platforms, only polling is available on the Summit family of switches.
Polling is always enabled on the system and occurs in the background every 10 seconds; the polling
value is not a user-configured parameter.
Technical Support. There are no health checking tests related to the stacking links in a SummitStack.
Enabling Backplane Diagnostic Packets on the Switch—Modular

Switches Only
To enable backplane diagnostic packets, use the following command:
enable sys-health-check slot <slot>
● BlackDiamond 10808 and BlackDiamond 12800 series switches—By default, the system health
checker tests the packet path every 6 seconds for the specified slot.
● BlackDiamond 8800 series switches—By default, the system health checker tests the data link
(BlackDiamond 8800 original modules) or the 10 Gbps links (BlackDiamond e-series and a-series
modules) every 5 seconds for the specified slot.

NOTE
Enabling backplane diagnostic packets increases CPU utilization and competes with network traffic for resources.
Configuring Backplane Diagnostic Packets on the Switch—

To configure the frequency of sending backplane diagnostic packets, use the following command:
configure sys-health-check interval <interval>
NOTE
Extreme Networks does not recommend configuring an interval of less than the default interval. Doing so can cause
excessive CPU utilization.
Disabling Backplane Diagnostic Packets on the Switch—Modular

Switches Only
To disable backplane diagnostic packets, use the following command:
disable sys-health-check slot <slot>
checker discontinues sending backplane diagnostic packets and returns the polling frequency to 60
seconds on the specified slot. Only polling is enabled.
● BlackDiamond 8800 series switch—By default, the system health checker discontinues sending
backplane diagnostic packets to the specified slot. Only polling is enabled.
Displaying the System Health Check Setting—All Platforms

To display the system health check setting, including polling and how ExtremeXOS handles faults on
the switch, use the following command:
show switch
As previously described, polling is always enabled on the switch.
The system health check setting, displayed as SysHealth check, shows the polling setting and how
ExtremeXOS handles faults. The polling setting appears as Enabled, and the fault handling setting
appears in parenthesis next to the polling setting. For more information about the fault handling setting,
see the following sections: “Configuring Hardware Recovery—SummitStack and Summit Family of
Switches Only” on page 251 and “Configuring Module Recovery—Modular Switches Only” on
page 254.
In the following truncated output from a BlackDiamond 8810 switch, the system health check setting
appears as SysHealth check: Enabled (Normal):
SysName: TechPubs Lab

SysName: BD-8810Rack3
SysLocation:
SysContact: [email protected], +1 888 257 3000

System MAC: 00:04:96:1F:A2:60
SysHealth check: Enabled (Normal)

Recovery Mode: None
System Watchdog: Enabled
System Health Check Examples: Backplane Diagnostics—Modular

Switches Only
This section provides examples for using the system health checker on the BlackDiamond 10808 switch,
BlackDiamond 12800 series switches, and BlackDiamond 8800 series switch. For more detailed
information about the system health check commands, see the chapter “Commands for Status
Monitoring and Statistics” in the ExtremeXOS Command Reference Guide.
Examples on the BlackDiamond 10808 and the BlackDiamond 12800 Series Switches
This section describes a series of two examples for:
● Enabling and configuring backplane diagnostics
● Disabling backplane diagnostics
Enabling and Configuring Backplane diagnostics. The following example:
● Enables backplane diagnostic packets on slot 3

● Modifies the polling interval from 60 seconds to 6 seconds
● Configures backplane diagnostic packets to be sent every 7 seconds and polling to occur every 7
seconds
1 Enable backplane diagnostic packets on slot 3 using the following command:
enable sys-health-check slot 3
When you enable backplane diagnostic packets on slot 3, the polling timer changes from its current
default value of 60 seconds to 6 seconds; 6 seconds is the default for sending backplane diagnostic
packets.
2 Configure backplane diagnostic packets to be sent every 7 seconds and update the polling rate to 7
seconds using the following command:
configure sys-health-check interval 7
NOTE
Extreme Networks does not recommend configuring an interval of less than 6 seconds. Doing this can cause
Disabling Backplane Diagnostics. Building upon the previous example, the following example disables
backplane diagnostics on slot 3:
disable sys-health-check slot 3
Backplane diagnostic packets are no longer sent, and the polling interval goes from 7 seconds to 70
seconds.

To return to the "default" settings of sending backplane diagnostic packets every 6 seconds (when
enabled) and polling the system every 60 seconds, specify 6 for the interval using the following
command:
Example on the BlackDiamond 8800 Series Switch

This section describes a series of two examples for:
● Enabling and configuring backplane diagnostics
● Disabling backplane diagnostics
Enabling and Configuring Backplane Diagnostics. The following example:

● Enables backplane diagnostic packets on slot 3
● Configures backplane diagnostic packets to be sent every 7 seconds
1 Enable backplane diagnostic packets on slot 3 using the following command:
enable sys-health-check slot 3
When you enable backplane diagnostic packets on slot 3, the timer runs at the default rate of 5
seconds.
2 Configure backplane diagnostic packets to be sent every 7 seconds using the following command:
NOTE
Extreme Networks does not recommend configuring an interval of less than 5 seconds. Doing this can cause
Disabling Backplane Diagnostics. Building upon the previous example, the following example disables
backplane diagnostics on slot 3:
disable sys-health-check slot 3
Backplane diagnostic packets are no longer sent, but the configured interval for sending backplane
diagnostic packets remains at 7 seconds. The next time you enable backplane diagnostic packets, the
health checker sends the backplane diagnostics packets every 7 seconds.
To return to the "default" setting of 5 seconds, configure the frequency of sending backplane diagnostic
packets to 5 seconds using the following command:
Setting the System Recovery Level

Depending on your switch model, you can configure the switch, MSM, or I/O module to take action if
a fault detection exception occurs. The following sections describe how to set the software and
hardware recovery levels on the switch, MSM, and I/O modules.

NOTE
You configure MSM and I/O module recovery only on the BlackDiamond 10808 switch, the BlackDiamond 12800
series switches, and the BlackDiamond 8800 series switches.
Configuring Software Recovery

You can configure the system to either take no action or to automatically reboot the switch after a
software task exception, using the following command:
configure sys-recovery-level [all | none]

● all—Configures ExtremeXOS to log an error to the syslog and automatically reboot the system after
any software task exception.
● none—Configures the system to take no action if a software task exception occurs. The system does
not reboot, which can cause unexpected switch behavior. On a SummitStack, the sys-recovery-level
setting applies to all active nodes.
NOTE
Use this parameter only under the guidance of Extreme Networks Technical Support personnel.
The default setting is all. Extreme Networks strongly recommends using the default setting.
Displaying the Software Recovery Setting

To display the software recovery setting on the switch, use the following command:
show switch
This command displays general switch information, including the software recovery level. The
following truncated output from a Summit X450 series switch displays the software recovery setting
(displayed as Recovery Mode):

SysLocation:
System MAC: 00:04:96:1F:A4:0E
Recovery Mode: All
NOTE
All platforms display the software recovery setting as Recovery Mode.
Configuring Hardware Recovery—SummitStack and Summit

Family of Switches Only
You can configure the Summit family of switches or SummitStack to take no action, automatically
reboot, or shut down if the switch detects a hardware fault.

To configure how the switch recovers from hardware problems on a stand-alone Summit family switch,
configure sys-recovery-level switch [none | reset | shutdown]
To configure hardware recovery on a particular active node in the SummitStack, use the following
command:
configure sys-recovery-level slot <slot> [none | reset | shutdown]

● none—Configures the switch to maintain its current state regardless of the detected fault. The switch
does not reboot or shutdown. ExtremeXOS logs fault and error messages to the syslog.
● reset—Configures the switch to reboot upon detecting a hardware fault. ExtremeXOS logs fault,
error, system reset, and system reboot messages to the syslog.
● shutdown—Configures the switch to shut down upon detecting a hardware fault. All ports are taken
offline in response to the reported errors; however, the management port remains operational for
debugging purposes only. If the switch shuts down, it remains in this state across additional reboots
or power cycles until you explicitly clear the shutdown state. See “Clearing the Shutdown State” on
page 253 for more information. ExtremeXOS logs fault, error, system reset, system reboot, and
system shutdown messages to the syslog.
The default setting is reset.
Beginning with ExtremeXOS 11.5, you can now configure how ExtremeXOS handles a detected fault
depending on the sys-recovery-level setting. To configure how ExtremeXOS handles faults, use the
configure sys-health-check all level [normal | strict] command. For detailed information
about this command, see the ExtremeXOS Command Reference Guide.
To view the system health check settings on the switch, use the show switch command as described in
“Displaying the System Health Check Setting—All Platforms” on page 248.
Confirmation Messages Displayed

If you configure the hardware recovery setting to either none (ignore) or shut down, the switch prompts
you to confirm this action. The following is a sample shutdown message:
Are you sure you want to shutdown on errors? (y/n)
Enter y to confirm this action and configure the hardware recovery level. Enter n or press [Enter] to
cancel this action.
Messages Displayed at the Startup Screen

If you configure the shutdown feature and a hardware error is detected, the system displays an
explanatory message on the startup screen. The following truncated sample output shows the startup
screen if a stand-alone switch is shut down as a result of the hardware recovery configuration:
All switch ports have been shut down.
Use the "clear sys-recovery-level" command to restore all ports.
! SummitX450-24x.1 #

When an exclamation point (!) appears in front of the command line prompt, it indicates that the entire
stand-alone switch is shut down as a result of your hardware recovery configuration and a switch error.
Displaying the Hardware Recovery Setting

To display the hardware recovery setting, use the following command:
show switch
If you change the hardware recovery setting from the default (reset) to either none (ignore) or
shutdown, the switch expands the Recovery Mode output to include a description of the hardware
recovery mode. If you keep the default behavior or return to reset, the Recovery Mode output lists only
the software recovery setting.
The following truncated output from a Summit X450 series switch displays the software recovery and
hardware recovery settings (displayed as Recovery Mode):

SysLocation:
System MAC: 00:04:96:1F:A5:71
Recovery Mode: All, Ignore
To see the output of "show switch" command for a particular node other than the master, you should
log into that node and run the "show switch" command.
If you configure the hardware recovery setting to none, the output displays “Ignore” to indicate that no
corrective actions will occur on the switch. “Ignore” appears only if you configure the hardware
recovery setting to none.
If you configure the hardware recovery setting to shut down, the output displays “Shutdown” to
indicate that the switch will shut down if fault detection occurs. “Shutdown” appears only if you
configure the hardware recovery setting to shut down.
If you configure the hardware recovery setting to reset, the output only displays the software recovery
mode.
Clearing the Shutdown State

If you configure the switch to shut down upon detecting a hardware fault, and the switch enters the
shutdown state, you must explicitly clear the shutdown state and reboot for the switch to become
functional. To clear the shutdown state, use the following command:
clear sys-recovery-level
On a SummitStack, use the command:
clear sys-recovery-level slot <slot>
The switch prompts you to confirm this action. The following is a sample confirmation message:
Are you sure you want to clear sys-recovery-level? (y/n)
Enter y to confirm this action and clear the shut down state. Enter n or press [Enter] to cancel this
action.

After you clear the shutdown state, use the reboot command to bring the switch and ports back
online. After you use the reboot command, the switch is operational.
Configuring Module Recovery—Modular Switches Only

You can configure the MSMs or I/O modules installed in the BlackDiamond 10808 switch,
BlackDiamond 12800 series switches, or BlackDiamond 8800 series switches to take no action, take ports
offline in response to errors, automatically reset, shutdown, or if dual MSMs are installed failover to the
other MSM if the switch detects a hardware fault. This enhanced level of recovery detects faults in the
ASICs as well as packet buses.
To configure module recovery, use the following command:
configure sys-recovery-level slot <slot_number> [none | reset | shutdown]

● none—Configures the MSM or I/O module to maintain its current state regardless of the detected
fault. The offending MSM or I/O module is not reset. ExtremeXOS logs fault and error messages to
the syslog and notifies you that the errors are ignored. This does not guarantee that the module
remains operational; however, the switch does not reboot the module.
● reset—Configures the offending MSM or I/O module to reset upon fault detection. ExtremeXOS
logs fault, error, system reset, and system reboot messages to the syslog.
● shutdown—Configures the switch to shut down all slots/modules configured for shutdown upon
fault detection. On the modules configured for shutdown, all ports in the slot are taken offline in
response to the reported errors; however, the MSMs remain operational for debugging purposes
only. ExtremeXOS logs fault, error, system reset, system reboot, and system shutdown messages to
the syslog.
The default setting is reset.
Depending on your configuration, the switch resets the offending MSM or I/O module if a hardware
fault detection occurs. An offending MSM is reset any number of times and is not permanently taken
offline. On the BlackDiamond 10808 and BlackDiamond 12800 series switches, an offending I/O module
is reset a maximum of three times. On the BlackDiamond 8800 series switch, an offending I/O module
is reset a maximum of five times. After the maximum number of resets, the I/O module is permanently
taken offline. For more information, see “Module Recovery Actions—BlackDiamond 8800 Series Switch
Only” on page 255.
Beginning with ExtremeXOS 11.5, you can now configure how ExtremeXOS handles a detected fault
based on the configuration of the configure sys-recovery-level slot <slot_number> [none |
reset | shutdown] command.
To configure how ExtremeXOS handles faults, use the configure sys-health-check all level
[normal | strict] command. For detailed information about this command, see the ExtremeXOS
To view the system health check settings on the switch, use the show switch command as described in
“Displaying the System Health Check Setting—All Platforms” on page 248.

Confirmation Messages Displayed

If you configure the hardware recovery setting to either none (ignore) or shutdown, the switch prompts
you to confirm this action. The following is a sample shutdown message:
Are you sure you want to shutdown on errors? (y/n)
Enter y to confirm this action and configure the hardware recovery level. Enter n or press [Enter] to
cancel this action.
Understanding the Shut Down Recovery Mode

Beginning with ExtremeXOS 11.5, you can configure the switch to shut down one or more I/O modules
upon fault detection by specifying the shutdown option. If you configure one or more slots to shut
down and the switch detects a hardware fault, all ports in all of the configured shut down slots are
taken offline in response to the reported errors. (MSMs are available for debugging purposes only.)
The affected I/O module remains in the shutdown state across additional reboots or power cycles until
you explicitly clear the shutdown state. If a module enters the shutdown state, the module actually
reboots and the show slot command displays the state of the slot as Initialized; however, the ports are
shut down and taken offline. For more information about clearing the shutdown state, see “Clearing the
Shutdown State” on page 260.
Messages Displayed at the Startup Screen

If you configure the shutdown feature and a hardware error is detected, the system displays an
explanatory message on the startup screen. The following truncated sample output shows the startup
screen if any of the slots in a modular switch are shut down as a result of the system recovery
configuration:
The I/O modules in the following slots are shut down: 1,3
Use the "clear sys-recovery-level" command to restore I/O modules
! BD-8810.1 #
When an exclamation point (!) appears in front of the command line prompt, it indicates that one or
more slots shut down as a result of your system recovery configuration and a switch error.
Module Recovery Actions—BlackDiamond 8800 Series Switch Only

Table 43 describes the actions module recovery takes based on your module recovery setting. For
example, if you configure a module recovery setting of reset for an I/O module, the module is reset a
maximum of five times before it is taken permanently offline.
From left to right, the columns display the following information:

● Module Recovery Setting—This is the parameter used by the configure sys-recovery-level
slot command to distinguish the module recovery behavior.
● Hardware—This indicates the hardware that you may have installed in your switch.
● Action Taken—This describes the action the hardware takes based on the module recovery setting.
Table 43: Module recovery actions for the BlackDiamond 8800 series switch
Module Recovery Setting Hardware Action Taken
none

Table 43: Module recovery actions for the BlackDiamond 8800 series switch (Continued)
Module Recovery Setting Hardware Action Taken
Single MSM The MSM remains powered on in its current state.
This does not guarantee that the module remains operational;
however, the switch does not reboot the module.
Dual MSM The MSM remains powered on in its current state.
I/O Module The I/O module remains powered on in its current state. The switch
sends error messages to the log and notifies you that the errors are
ignored.
reset
Single MSM Resets the MSM.
Dual MSM Resets the primary MSM and fails over to the backup MSM.
I/O Module Resets the I/O module a maximum of five times. After the fifth
time, the I/O module is permanently taken offline.
shutdown
Single MSM The MSM is available for debugging purposes only (the I/O ports
also go down); however, you must clear the shutdown state using
the clear sys-recovery-level command for the MSM to
become operational.
After you clear the shutdown state, you must reboot the switch.
For more information see, “Clearing the Shutdown State” on
page 260.
Dual MSM The MSMs are available for debugging purposes only (the I/O ports
also go down); however, you must clear the shutdown state using
the clear sys-recovery-level command for the MSM to
become operational.
After you clear the shutdown state, you must reboot the switch.
page 260.
I/O Module Reboots the I/O module. When the module comes up, the ports
remain inactive because you must clear the shutdown state using
the clear sys-recovery-level command for the I/O module
to become operational.
After you clear the shutdown state, you must reset each affected I/O
module or reboot the switch.
page 260.
Module Recovery Actions—BlackDiamond 10808 and BlackDiamond 12800 Series

Switches Only
Table 44 describes the actions module recovery takes based on your module recovery and software
recovery settings. For example, if you configure a module recovery setting of reset and a system
recovery setting of all for an I/O module, the module is reset a maximum of three times before it is
taken permanently offline.

From left to right, the columns display the following information:

● Module Recovery Setting—This is the parameter used by the configure sys-recovery-level
slot command to distinguish the module recovery behavior.
● System Recovery Setting—This is the parameter used by the configure sys-recovery-level
command to distinguish the software recovery behavior.
● Hardware—This lists the hardware that you may have installed in your switch
● Action Taken—This describes the action the hardware takes based on the module recovery and
software recovery settings.
Table 44: Module recovery actions for the BlackDiamond 10808 and the BlackDiamond 12800
series switches
Module Recovery Setting System Recovery Setting Hardware Action Taken
reset all
Dual MSM Resets the primary MSM and fails over to
the backup MSM.
I/O Module Resets the I/O module a maximum of three
times. After the third time, the I/O module
is permanently taken offline.
reset none
Dual MSM Fails over to the backup MSM.
I/O Module Resets the I/O module.
shutdown all
Single MSM The MSM is available for debugging
purposes only; however, you must clear the
shutdown state using the clear sys-
recovery-level command for the MSM
After you clear the shutdown state, you
must reboot the switch.
For more information see, “Clearing the
Dual MSM The MSMs are available for debugging
purposes only; however, you must clear the
shutdown state using the clear sys-
recovery-level command for the MSM
must reboot the switch.

Table 44: Module recovery actions for the BlackDiamond 10808 and the BlackDiamond 12800
series switches (Continued)
Module Recovery Setting System Recovery Setting Hardware Action Taken
I/O Module Reboots the I/O module. When the module
comes up, the ports remain inactive
because you must clear the shutdown state
using the clear sys-recovery-level
command for the I/O module to become
operational.
must reset each affected I/O module or
reboot the switch.
shutdown none
Single MSM The MSM is available for debugging
purposes only.
Dual MSM The MSMs are available for debugging
purposes only.
I/O Module Reboots the I/O module. When the module
comes up, the ports remain inactive
because you must clear the shutdown state
using the clear sys-recovery-level
command for the I/O module to become
operational.
must reset each affected I/O module or
reboot the switch.
Displaying the Module Recovery Setting

To display the module recovery setting, use the following command:
show slot
Beginning with ExtremeXOS 11.5, the show slot output has been modified to include the shutdown
configuration. If you configure the module recovery setting to shut down, the output displays an “E”
flag that indicates any errors detected on the slot disables all ports on the slot. The “E” flag appears
only if you configure the module recovery setting to shut down.
NOTE
If you configure one or more slots for shut down and the switch detects a hardware fault on one of those slots, all of
the configured slots enter the shutdown state and remain in that state until explicitly cleared.
If you configure the module recovery setting to none, the output displays an “e” flag that indicates no
corrective actions will occur for the specified MSM or I/O module. The “e” flag appears only if you
configure the module recovery setting to none.

The following sample output displays the module recovery action. In this example, notice the flags
identified for slot 10:
Slots Type Configured State Ports Flags
-------------------------------------------------------------------------------
Slot-1 G48P G48P Operational 48 MB S
Slot-2 G24X G24X Operational 24 MB S
Slot-3 G48T G48T Operational 48 MB S
Slot-4 Empty 0
Slot-7 Empty 0
Slot-8 G48Te G48Te Operational 48 MB S
Slot-9 G48Ta Operational 48 MB S
Slot-10 G48Pe G48Pe Operational 48 MB S E
MSM-A MSM-G8X Operational 0 S
MSM-B MSM-G8X Operational 0 S
Flags : M - Backplane link to Master MSM is Active

B - Backplane link to Backup MSM is also Active
D - Slot Disabled, S - Slot Secured
I - Insufficient Power (refer to "show power budget")
e - Errors on slot will be ignored (no corrective action initiated)
E - Errors on slot will disable all ports on slot
NOTE
In ExtremeXOS 11.4 and earlier, if you configure the module recovery setting to none, the output displays an “E”
flag that indicates no corrective actions will occur for the specified MSM or I/O module. The “E” flag appears only if
you configure the module recovery setting to none.
Displaying Detailed Module Recovery Information

To display the module recovery setting for a specific port on a module, including the current recovery
mode, use the following command:
show slot <slot>
In addition to the information displayed with show slot, this command displays the module recovery
setting configured on the slot. The following truncated output displays the module recovery setting
(displayed as Recovery Mode) for the specified slot:
Slot-10 information:
State: Operational
Download %: 100
Flags: MB S E
Serial number: 800158-00-01 06014-00022
Hw Module Type: G48Pe
SW Version: 11.5.0.4
SW Build: v1150b4
Configured Type: G48Pe
Ports available: 48
Recovery Mode: Shutdown
Flags : M - Backplane link to Master MSM is Active

B - Backplane link to Backup MSM is also Active

D - Slot Disabled, S - Slot Secured

I - Insufficient Power (refer to "show power budget")
e - Errors on slot will be ignored (no corrective action initiated)
E - Errors on slot will disable all ports on slot
Clearing the Shutdown State

If you configure one or more modules to shut down upon detecting a hardware fault, and the switch
enters the shutdown state, you must explicitly clear the shut down state and reset the affected modules
for the switch to become functional. To clear the shutdown state, use the following command:
clear sys-recovery-level
The switch prompts you to confirm this action. The following is a sample confirmation message:
Are you sure you want to clear sys-recovery-level? (y/n)
Enter y to confirm this action and clear the shutdown state. Enter n or press [Enter] to cancel this action.
After using the clear sys-recovery-level command, you must reset each affected module.
If you configured only a few I/O modules to shutdown, reset each affected I/O module as follows:
1 Disable the slot using the disable slot <slot> command.
2 Re-enable the slot using the enable slot <slot> command.
NOTE
You must complete this procedure for each module that enters the shutdown state.
If you configured all I/O modules or one or more MSMs to shutdown, use the reboot command to
reboot the switch and reset all affected I/O modules.
After you clear the shutdown state and reset the affected module, each port is brought offline and then
back online before the module and the entire system is operational.
Troubleshooting Module Failures

If you experience an I/O module failure, use the following troubleshooting methods when you can
bring the switch offline to solve or learn more about the problem:
● Restarting the I/O module—Use the disable slot <slot> command followed by the enable
slot <slot> command to restart the offending I/O module. By issuing these commands, the I/O
module and its associated fail counter is reset. If the module does not restart, or you continue to
experience I/O module failure, contact Extreme Networks Technical Support.
● Running diagnostics—Use the run diagnostics normal <slot> command to run diagnostics on
the offending I/O module to ensure that you are not experiencing a hardware issue. If the module
continues to enter the failed state, please contact Extreme Networks Technical Support. For more
information about switch diagnostics, see “Performing Switch Diagnostics” on page 235.
If you experience an MSM failure, contact Extreme Networks Technical Support.

Using ELSM
Using ELSM
ExtremeXOS 11.4 introduces support for the Extreme Link Status Monitoring (ELSM) protocol. ELSM is
an Extreme Networks proprietary protocol that monitors network health by detecting CPU and remote
link failures. ELSM is available only on Extreme Networks devices and operates on a point-to-point
basis. You configure ELSM on the ports that connect to other network devices and on both sides of the
peer connection.

● About ELSM on page 261
● ELSM Hello Messages on page 262
● ELSM Port States on page 262
● Link States on page 263
● ELSM Link States on page 263
● ELSM Timers on page 264
● Configuring ELSM on a Switch on page 265
● Displaying ELSM Information on page 268
● Using ELSM with Layer 2 Control Protocols on page 270
● ELSM Configuration Example on page 270
About ELSM
ELSM monitors network health by exchanging various hello messages between two ELSM peers. ELSM
uses an open-ended protocol, which means that an ELSM-enabled port expects to send and receive hello
messages from its peer. The Layer 2 connection between ports determines the peer connection. Peers
can be either directly connected or separated by one or more hubs. If there is a direct connection
between peers, they are considered neighbors.
If ELSM detects a failure, the ELSM-enabled port responds by blocking traffic on that port. For example,
if a peer stops receiving messages from its peer, ELSM brings down that connection by blocking all
incoming and outgoing data traffic on the port and notifying applications that the link is down.
In some situations, a software or hardware fault may prevent the CPU from transmitting or receiving
packets, thereby leading to the sudden failure of the CPU. If the CPU is unable to process or send
packets, ELSM isolates the connections to the faulty switch from the rest of the network. If the switch
fabric sends packets during a CPU failure, the switch may appear healthy when it is not. For example, if
hardware forwarding is active and software forwarding experiences a failure, traffic forwarding may
continue. Such failures can trigger control protocols such as Extreme Standby Router Protocol (ESRP) or
Ethernet Automatic Protection Switching (EAPS) to select different devices to resume forwarding. This
recovery action, combined with the CPU failure, can lead to loops in a Layer 2 network.
Configuring ELSM on Extreme Networks devices running ExtremeXOS is backward compatible with
Extreme Networks devices running ExtremeWare.

ELSM Hello Messages

ELSM uses two types of hello messages to communicate the health of the network to other ELSM ports.
The following describes the hello messages:
● Hello+ — The ELSM-enabled port receives a hello message from its peer and no problem is detected.
● Hello- — The ELSM-enabled port has not received a hello message from its peer.
In addition to the ELSM port states described in the next section, ELSM has hello transmit states. The
hello transmit states display the current state of transmitted ELSM hello messages and can be one of the
following:
● HelloRx(+)—The ELSM-enabled port is up and receives Hello+ messages from its peer. The port
remains in the HelloRx+ state and restarts the HelloRx timer each time it receives a Hello+ message.
If the HelloRx timer expires, the hello transmit state enters HelloRx(-). The HelloRx timer is 6 * hello
timer, which by default is 6 seconds.
● HelloRx(-)—The ELSM-enabled port either transitions from the initial ELSM state or is up but has
not received hello messages because there is a problem with the link or the peer is missing.
For information about displaying ELSM hello messages and hello transmit states, see “Displaying ELSM
Information” on page 268.
ELSM Port States

Each ELSM-enabled port exists in one of the following states:
● Up—Indicates a healthy remote system and this port is receiving Hello+ messages from its peer.
If an ELSM-enabled port enters the Up state, the up timer begins. Each time the port receives a
Hello+ message from its peer, the up timer restarts and the port remains in the Up state. The up
timer is 6* hello timer, which by default is 6 seconds.
● Down—Indicates that the port is down, blocked, or has not received Hello+ messages from its peer.
If an ELSM-enabled port does not receive a hello message from its peer before the up timer expires,
the port transitions to the Down state. When ELSM is down, data packets are neither forwarded nor
transmitted out of that port.
● Down-Wait—Indicates a transitional state.
If the port enters the Down state and later receives a Hello+ message from its peer, the port enters
the Down-Wait state. If the number of Hello+ messages received is greater than or equal to the hold
threshold (by default 2 messages), the port transitions to the Up state. If the number of Hello+
messages received is less than the hold threshold, the port enters the Down state.
● Down-Stuck—Indicates that the port is down and requires user intervention.
If the port repeatedly flaps between the Up and Down states, the port enters the Down-Stuck state.
Depending on your configuration, there are two ways for a port to transition out of this state:
By default, automatic restart is enabled, and the port automatically transitions out of this state. For
more information, see the enable elsm ports <portlist> auto-restart command.
If you disabled automatic restart, and the port enters the Down-Stuck state, you can clear the stuck
state and enter the Down state by using one of the following commands:
■ clear elsm ports <portlist> auto-restart
■ enable elsm ports <portlist> auto-restart

Using ELSM
NOTE
If you reboot the peer switch, its ELSM-enabled peer port may enter the Down-Stuck state. If this occurs, clear
the stuck state using one of the following commands: clear elsm ports <portlist> auto-restart or
enable elsm ports <portlist> auto-restart.
For information about displaying ELSM port states, see “Displaying ELSM Information” on page 268.
Link States
The state of the link between ELSM-enabled (peer) ports is known as the link state. The link state can be
one of the following:
● Ready—Indicates that the port is enabled but there is no physical link
● Active—Indicates that the port is enabled and the physical link is up
To view the state of the link between the peer ports, use the following commands:
● show elsm ports <all | portlist>
● show ports {<port_list>} information {detail}
If you use the show elsm ports <all | portlist> command, the Link State row displays link
state information.
If you use the show ports show ports {<port_list>} information command, the Link State
column displays link state information.
If you use the show ports {<port_list>} information command and specify the detail option, the
ELSM Link State row displays ELSM link state information. For more information, see “ELSM Link
States” on page 263.
For more information about these show commands, see “Displaying ELSM Information” on page 268.
ELSM Link States

The state of the ELSM logical link is known as the ELSM link state. The ELSM link state can be one of
the following:
● ELSM is enabled and the ELSM peer ports are up and communicating
● ELSM is enabled but the ELSM peer ports are not up or communicating
● ELSM is disabled
To view the current state of ELSM on the switch, use the following commands:
● show elsm
● show elsm ports <all | portlist>
● show ports {<port_list>} information {detail}
If you use the show elsm commands, the following terms display the ELSM link state:
● Up—Indicates that ELSM is enabled and the ELSM peer ports are up and communicating; the ELSM
link state is up. In the up state, the ELSM-enabled port sends and receives hello messages from its
peer.

● Down—Indicates that ELSM is enabled, but the ELSM peers are not communicating; the ELSM link
state is down. In the down state, ELSM transitions the peer port on this device to the down state.
ELSM blocks all incoming and outgoing switching traffic and all control traffic except ELSM PDUs.
If ELSM is disabled, the switch does not display any information.
If you use the show ports {<port_list>} information {detail} command, the following columns
display the current state of ELSM on the switch:
● Flags
■ L—Indicates that ELSM is enabled on the switch
■ - —Indicates that ELSM is disabled on the switch
● ELSM
■ up—Indicates that ELSM is enabled and the ELSM peer ports are up and communicating; the
ELSM link state is up. In the up state, the ELSM-enabled port sends and receives hello messages
from its peer.
■ dn—Indicates that ELSM is enabled, but the ELSM peers are not communicating; the ELSM link
state is down. In the down state, ELSM transitions the peer port on this device to the down state.
ELSM blocks all incoming and outgoing switching traffic and all control traffic except ELSM
PDUs.
■ - —Indicates that ELSM is disabled on the switch.
If you specify the optional detail parameter, the following ELSM output is called out in written
explanations versus displayed in a tabular format:
● ELSM Link State (displayed only if ELSM is enabled on the switch)
■ Up—Indicates that ELSM is enabled and the ELSM peer ports are up and communicating; the
ELSM link state is up. In the up state, the ELSM-enabled port sends and receives hello messages
from its peer.
■ Down—Indicates that ELSM is enabled, but the ELSM peers are not communicating; the ELSM
link state is down. In the down state, ELSM transitions the peer port on this device to the down
state. ELSM blocks all incoming and outgoing switching traffic and all control traffic except
ELSM PDUs.
● ELSM
■ Enabled—Indicates that ELSM is enabled on the switch
■ Disabled—Indicates that ELSM is disabled on the switch
For more information about these show commands, see “Displaying ELSM Information” on page 268.
ELSM Timers
To determine whether there is a CPU or link failure, ELSM requires timer expiration between the ELSM
peers. Depending on the health of the network, the port enters different states when the timers expire.
For more information about the ELSM port states, see “ELSM Port States” on page 262.
Table 45 describes the ELSM timers. Only the hello timer is user-configurable; all other timers are
derived from the hello timer. This means that when you modify the hello timer, you also modify the
values for down, up, and HelloRx timers.

Using ELSM
Table 45: ELSM timers

Timer Description
Hello The ELSM hello timer is the only user-configurable timer and specifies the time in seconds between
consecutive hello messages. The default value is 1 second, and the range is 100 milliseconds to
255 seconds.
Down The ELSM down timer specifies the time it takes the ELSM port to cycle through the following
states:
• Down—Indicates that the port is down, blocked, or has not received Hello+ messages from its
peer.
• Down-Wait—Indicates a transitional state.
• Up—Indicates a healthy remote system and this port is receiving Hello+ messages from its peer.
By default, the down timer is (2 + hold threshold) * hello timer, which is 4 seconds. If the hold
threshold is set to 2 and the hello timer is set to 1 second, it takes 4 seconds for the ELSM port
receiving messages to cycle through the states.
After the down timer expires, the port checks the number of Hello+ messages against the hold
threshold. If the number of Hello+ messages received is greater than or equal to the configured hold
threshold, the ELSM receive port moves from the Down-Wait state to the Up state.
If the number of Hello+ messages received is less than the configured hold threshold, the ELSM
receive port moves from the Down-Wait state back to the Down state and begins the process again.
Up The ELSM up timer begins when the ELSM-enabled port enters the UP state. Each time the port
receives a Hello+ message, the timer restarts.
Up timer is the UpTimer threshold * hello timer. It is configurable and the range is 3 to 60
seconds.
By default, the UpTimer threshold is 6. Therefore, the default up timer is 6 seconds (6*1).
HelloRx The ELSM HelloRx timer specifies the time in which a hello message is expected. If the port does
not receive a hello message from its peer, there is the possibility of a CPU or link failure.
By default the HelloRx timer 6 * hello timer, which is 6 seconds.
Configuring ELSM on a Switch

This section describes the commands used to configure ELSM on the switch and contains the following
topics:
● Enabling ELSM on page 266
● Configuring the ELSM Hello Timer on page 266
● Configuring the ELSM Hold Threshold on page 266
● Configuring Automatic Restart on page 267
● Disabling ELSM on page 268
NOTE
Extreme Networks does not recommend enabling the Link Aggregation Control Protocol (LACP) and ELSM on the
same port. For more information about LACP, see Chapter 6, “Configuring Slots and Ports on a Switch.”
NOTE
ELSM and mirroring are mutually exclusive. You can enable either ELSM, or mirroring, but not both.

Enabling ELSM
ELSM works between two connected ports (peers), and each ELSM instance is based on a single port.
The Layer 2 connection between the ports determines the peer. You can have a direct connection
between the peers or hubs that separate peer ports. In the first instance, the peers are also considered
neighbors. In the second instance, the peer is not considered a neighbor.
To enable ELSM on one or more ports, use the following command:
enable elsm ports <portlist>
When you enable ELSM on a port, ELSM immediately blocks the port and it enters the Down state.
When the port detects an ELSM-enabled peer, the peer ports exchange ELSM hello messages. At this
point, the ports enter the transitional Down-Wait state. If the port receives Hello+ messages from its
peer and does not detect a problem, the peers enter the Up state. If a peer detects a problem or there is
no peer port configured, the port enters the Down state.
For more information about the types of ELSM hello messages, see “ELSM Hello Messages” on
page 262. For information about configuring the ELSM hello timer, see the next section.
Configuring the ELSM Hello Timer

The ELSM hello timer is the only user-configurable timer and specifies the time in seconds between
consecutive hello messages. The default value is 1 second, and the range is 1 to 128 seconds. Although
other timers rely on the hold timer for their values, you do not explicitly configure the down, up, or
HelloRx timers. If you modify the hello timer on one port, Extreme Networks recommends that you use
the same hello timer value on its peer port.
A high hello timer value can increase the time it takes for the ELSM-enabled port to enter the Up state.
The down timer is (2 + hold threshold) * hello timer. Assuming the default value of 2 for the hold
threshold, configuring a hello timer of 128 seconds creates a down timer of (2 + 2) 128, or 512 seconds.
In this scenario it would take 512 seconds for the port to transition from the Down to the Up state.
To configure the ELSM hello timer, use the following command:
configure elsm ports <portlist> hellotime <hello_time>
Configuring the ELSM Hold Threshold

The ELSM hold threshold determines the number of Hello+ messages the ELSM peer port must receive
to transition from the Down-Wait state to the Up state. For example, a threshold of 1 means the ELSM
port must receive at least one Hello+ message to transition from the Down-Wait state to the Up state.
The default is two messages, and the range is one to three messages.
After the down timer expires, the port checks the number of Hello+ messages against the hold
threshold. If the number of Hello+ messages received is greater than or equal to the configured hold
threshold, the ELSM receive port moves from the Down-Wait state to the Up state.
If the number of Hello+ messages received is less than the configured hold threshold, the ELSM receive
port moves from the Down-Wait state back to the Down state and begins the process again.
If you modify the hold threshold on one port, Extreme Networks recommends that you use the same
hold threshold value on its peer port.

Using ELSM
You configure the hold threshold on a per-port basis, not on a per-switch basis.
To configure the ELSM hold threshold, use the following command:
configure elsm ports <portlist> hold-threshold <hold_threshold>
Configuring Automatic Restart

You must explicitly configure automatic restart on each ELSM-enabled port; this is not a global
configuration.
By default, ELSM automatic restart is enabled. If an ELSM-enabled port goes down, ELSM bypasses the
Down-Stuck state and automatically transitions the down port to the Down state, regardless of the
number of times the port goes up and down. For information about the port states, see “ELSM Port
States” on page 262.
If you disable ELSM automatic restart, the ELSM-enabled port can transition between the following
states multiple times: Up, Down, and Down-Wait. When the number of state transitions is greater than
or equal to the sticky threshold, the port enters the Down-Stuck state. The ELSM sticky threshold
specifies the number of times a port can transition between the Up and Down states. The sticky
threshold is not user-configurable and has a default value of 1. That means a port can transition only
one time from the Up state to the Down state. If the port attempts a subsequent transition from the Up
state to the Down state, the port enters the Down-Stuck state.
If the port enters the Down-Stuck state, you can clear the stuck state and enter the Down state by using
one of the following commands:
clear elsm ports <portlist> auto-restart

enable elsm ports <portlist> auto-restart
If you use the enable elsm ports <portlist> auto-restart command, automatic restart is always
enabled; you do not have to use the clear elsm ports <portlist> auto-restart command to clear
the stuck state.
Disabling Automatic Restart. To disable automatic restart, use the following command:
disable elsm ports <portlist> auto-restart
Extreme Networks recommends that you use the same automatic restart configuration on each peer
port.
Re-Enabling Automatic Restart. To re-enable automatic restart, use the following command:
Extreme Networks recommends that you use the same automatic restart configuration on each peer
port.
Re-Setting the Sticky Threshold. The following user events clear, re-set the sticky threshold:
● Enabling automatic restart on the port using the following command:

● Clearing the port that is in the Down-Stuck state using the following command:
clear elsm ports <portlist> auto-restart

● Disabling and re-enabling the port using the following commands:

disable ports [<port_list | all]
enable ports [<port_list | all]
● Disabling ELSM on the port using the following command:
disable elsm ports <portlist>
Disabling ELSM
To disable ELSM on one or more ports, use the following command:
disable elsm ports <portlist>
When you disable ELSM on the specified ports, the ports no longer send ELSM hello messages to their
peers and no longer maintain ELSM states.
Displaying ELSM Information

To display summary information for all of the ELSM-enabled ports on the switch, use the following
command:
show elsm
This command displays in a tabular format the operational state of ELSM on the configured ports.
If ports are configured for ELSM (ELSM is enabled), the switch displays the following information:
● Port—The port number of the ELSM-enabled port.
● ELSM State—The current state of ELSM on the port.
For information about the port states, see “ELSM Port States” on page 262.
● Hello time—The current value of the hello timer, which by default is 1 second.
For information about configuring the hello timer, see “Configuring the ELSM Hello Timer” on
page 266.
If no ports are configured for ELSM (ELSM is disabled), the switch does not display any information.
To display detailed information for one or more ELSM-enabled ports on the switch, use the following
command:
show elsm ports <all | portlist>
In addition to the port, ELSM state, and hello timer information, this command displays in a tabular
format the following:
● Link State—The state of the link between ELSM-enabled ports.
For information about the link states, see “Link States” on page 263.
● ELSM Link State—The current state of the ELSM logical link on the switch.
For more information, see “ELSM Link States” on page 263.
● Hello Transmit State—The current state of ELSM hello messages being transmitted.
● Hold Threshold—The number of Hello+ messages required by the ELSM-enabled port to transition
from the Down-Wait state to the Up state within the hold threshold.

Using ELSM
● UpTimer Threshold—The number of hello times that span without receiving Hello+ packets before a
port changes its ELSM state from Up to Down.
● Auto Restart—The current state of ELSM automatic restart on the port.
● Sticky Threshold—The number of times a port can transition between the Up and Down states. The
sticky threshold is not user-configurable and has a default value of 1.
● Sticky Threshold Counter—The number of times the port transitions from the Up state to the Down
state.
● Down Timeout—The actual waiting time (msecs or secs) before a port changes its ELSM state from
Down to Up after receiving the first Hello+ packet. It is equal to [Hello Time * (Hold Threshold+2)].
● Up Timeout—The actual waiting time (msecs or secs) before a port changes its ELSM state from Up
to Down after receiving the last Hello+ packets. It is equal to [Hello Time * UpTimer Threshold].
The remaining output displays counter information. Use the counter information to determine the
health of the ELSM peers and how often ELSM has gone up or down. The counters are cumulative.
● RX Hello+—The number of Hello+ messages received by the port.
● Rx Hello- —The number of Hello- messages received by the port.
● Tx Hello+—The number of Hello+ messages sent by the port.
● Tx Hello- —The number of Hello- messages sent by the port.
● ELSM Up/Down Count—The number of times ELSM has been up or down.
To display summary port information in a tabular format for one or more ELSM-enabled ports, use the
following command:
show ports {<port_list>} information {detail}
This command displays the following ELSM information:

● Flags
■ L—ELSM is enabled on the switch
■ - —ELSM is disabled on the switch
● ELSM
■ up—ELSM is enabled and the ELSM link state is up.
■ dn—ELSM is enabled and the ELSM link state is down.
■ - —ELSM is disabled on the switch.
To display summary port information called out in written explanations versus displayed in a tabular
format, use the following command and specify the optional detail parameter:
show ports {<port_list>} information detail
This command displays the following ELSM information:

● ELSM Link State (displayed only if ELSM is enabled on the switch).
■ Up—ELSM is enabled and the ELSM link state is up.
■ Down—ELSM is enabled and the ELSM link state is down.
● Link State—The state of the link between ELSM-enabled ports.
For information about the link states, see “Link States” on page 263.

● ELSM
■ Enabled—ELSM is enabled on the switch
■ Disabled—ELSM is disabled on the switch
Clearing ELSM Counters

Before clearing the ELSM counters, you should use the show elsm and show elsm ports commands to
view the ELSM information. To clear only the ELSM-related counters gathered by the switch, use the
following command:
clear elsm {ports <portlist>} counters
You can also use the clear counters command, which clears all of the counters on the device,
including those associated with ELSM.
Using ELSM with Layer 2 Control Protocols

You can use ELSM with Layer 2 control protocols such as STP, ESRP, EAPS, and so on to improve the
recovery of Layer 2 loops in the network. ELSM detects remote link failures if the established link is
through a Layer 2 transport cloud and the ELSM endpoints traverse the Layer 2 cloud in a point-to-
point fashion, as shown in Figure 7.
Figure 7: ELSM through a Layer 2 cloud
Switch A Switch B
L2 cloud
EX_169
A sudden failure of the switch CPU may cause the hardware to continue forwarding traffic. For
example, ESRP may select a second device for forwarding traffic thereby creating a Layer 2 loop in the
network. If you configure ELSM and the CPU fails, ELSM closes the connection to the faulty device to
prevent a loop.
ELSM Configuration Example

The following example configures ELSM on two ports connected directly to each other and assumes the
following:
NOTE
In the following sample configurations, any lines marked (Default) represent default settings and do not need to be
explicitly configured.

Viewing Fan Information
Switch A Configuration
● ELSM-enabled port—Port 1
● Hello timer—2 seconds
● Hello threshold—2 hello messages
enable elsm ports 1
configure elsm ports 1 hellotime 2
configure elsm ports 1 hello-threshold 2 (Default)
Switch B Configuration
● ELSM-enabled port—Slot 2, port 1
● Hello timer—2 seconds
● Hello threshold—2 hello messages
enable elsm ports 2:1
configure elsm ports 2:1 hellotime 2
configure elsm ports 2:1 hello-threshold 2 (Default)
After you enable ELSM on the ports, the peers exchange hello messages with each other as displayed in
Figure 8.
Figure 8: Extreme Networks switches with ELSM-enabled ports exchanging hello messages
Hello message
Switch A Switch B
Hello message
EX_170
Viewing Fan Information

You can view detailed information about the fans installed in your switch. Depending on your switch
model, different information may be displayed.
To view detailed information about the health of the fans, use the following command:
show fans
The switch collects and displays the following fan information:

● State—The current state of the fan. Options are:
■ Empty: There is no fan installed.
■ Failed: The fan failed.
■ Operational: The fan is installed and working normally.
● NumFan—The number of fans in the fan tray.
● Fan Name, displayed as Fan-1, Fan-2, and so on (modular switches also include a description of the
location, for example, Upper or Upper-Right)—Specifies the individual state for each fan in a fan
tray and its current speed in revolutions per minute (rpm).

On modular switches, the output also includes the following information:

● PartInfo—Information about the fan tray, including the:
■ Serial number—A collection of numbers and letters, that make up the serial number of the fan.
This is the first series of numbers and letters in the display.
■ Part number—A collection of numbers and letters, that make up the part number of the fan. This
is the second series of numbers and letters in the display.
● Revision—The revision number of the fan.
● Odometer—Specifies the power-on date and how long the fan tray has been operating since it was
first powered-on.
● Temperature—Specifies, in celsius, the current temperature of the fan. (BlackDiamond 10808 switch
only.) For more information see, “Fan Tray Temperature—BlackDiamond 10808 Switch Only” on
page 273.
Viewing the System Temperature

Depending on your switch model, you can view the temperature in Celsius of the I/O modules,
management modules, power controllers, power supplies, and fan trays installed in your switch. In
addition, depending on the software version running on your switch, additional or different
temperature information might be displayed.
To view the system temperature, use the following command:
show temperature
System Temperature Output—Modular Switches and SummitStack

Only
On a modular switch, the output includes the current temperature and operating status of the I/O
modules, management modules, and power controllers. On a SummitStack, the output includes the
current temperature and operating status of all active nodes and their option cards (if any).
The following output shows a sample display of the current temperature and operating status of the
installed modules and power controllers:
Field Replaceable Units Temp (C) Status

------------------------------------------------
Slot-1 : 10G6X 36.37 Normal
Slot-2 : G60X 35.31 Normal
Slot-3 :
Slot-4 :
Slot-5 :
Slot-8 :
MSM-A : MSM-1XL 31.37 Normal
MSM-B : MSM-1XL 29.75 Normal
PSUCTRL-1 :
PSUCTRL-2 : 29.00 Normal
Temp Range: -10.00 (Min), 0.00-50.00 (Normal), 60.00 (Max)

Viewing the System Temperature
The switch monitors the temperature of each component and generates a warning if the temperature
exceeds the normal operating range. If the temperature exceeds the minimum/maximum limits, the
switch shuts down the overheated module.
System Temperature Output—Summit Family of Switches Only

On the Summit family of switches, the output includes the current temperature and operating status of
the switch and the XGM-2xn card.
The following sample output displays the current temperature and operating status of the switch:
Field Replaceable Units Temp (C) Status

----------------------------------------------------
Switch : SummitX450-24x 32.00 Normal
XGM-2xn-1 :
Temp Range: -10.00 (Min), 0.00-60.50 (Normal), 62.90 (Max)
The switch monitors its temperature and generates a warning if the temperature exceeds the normal
operating range. If the temperature exceeds the maximum limit, the show switch output indicates the
switch in an OPERATIONAL (Overheat) mode, and the show temperature output indicates an error
state due to overheat.
Power Supply Temperature—Modular Switches Only

To view the current temperature of the power supplies installed in a BlackDiamond 10808 switch,
BlackDiamond 12800 series switch, or BlackDiamond 8800 series switch, use the following command:
The following is sample output of temperature information:
PowerSupply 1 information:
...
Temperature: 30.1 deg C
...
Fan Tray Temperature—BlackDiamond 10808 Switch Only

To view the current temperature and status of the fan trays installed in the BlackDiamond 10808 switch,
show fans
The following is sample output of the fan tray temperature information on a BlackDiamond 10808
switch:
Right(Rear-facing) FanTray 1 information:

...
Temperature: 34.25 deg C
...

Using the Event Management System/Logging

We use the general term, event, for any type of occurrence on a switch that could generate a log
message or require an action. For example, a link going down, a user logging in, a command entered on
the command line, or the software executing a debugging statement, are all events that might generate a
log message. The system for saving, displaying, and filtering events is called the Event Management
System (EMS). With EMS, you have many options about which events generate log messages, where the
messages are sent, and how they are displayed.
Using EMS you can:

● Send event messages to a number of logging targets (for example, syslog host and NVRAM)
● Filter events per target, by:
■ Component, subcomponent, or specific condition (for example, BGP messages, IGMP.Snooping
messages, or the IP.Forwarding.SlowPathDrop condition)
■ Match expression (for example, any messages containing the string “user5”)
■ Matching parameters (for example, only messages with source IP addresses in the 10.1.2.0/24
subnet)
■ Severity level (for example, only messages of severity critical, error, or warning)
● Change the format of event messages (for example, display the date as “12-May-2005” or “2005-05-
12”)
● Display log messages in real time and filter the messages that are displayed, both on the console and
from Telnet sessions
● Display stored log messages from the memory buffer or NVRAM
● Upload event logs stored in memory buffer or NVRAM to a TFTP server
● Display counts of event occurrences, even those not included in filter
● Display debug information using a consistent configuration method
Beginning with ExtremeXOS 11.2, EMS supports IPv6 as a parameter for filtering events.
Sending Event Messages to Log Targets

You can specify seven types of targets to receive log messages:
● Console display
● Current session (Telnet or console display)
● Memory buffer (can contain 200 to 20,000 messages)
● NVRAM (messages remain after reboot)
● Primary MSM
● Backup MSM
● Syslog host
The first six types of targets exist by default; but before enabling any syslog host, you must add the
host’s information to the switch using the configure syslog command. Extreme Networks EPICenter
can be a syslog target.

By default, the memory buffer and NVRAM targets are already enabled and receive messages. To start
sending messages to the targets, use the following command:
enable log target [console | memory-buffer | nvram | primary-msm |primary-node|
backup-msm | backup-node| session | syslog [all | <ipaddress> | <ipPort>] {vr
<vr_name>} [local0 ... local7]]]
After you enable this feature, the target receives the messages it is configured for. See “Target
Configuration” later in this chapter for information on viewing the current configuration of a target. The
memory buffer can contain only the configured number of messages, so the oldest message is lost when
a new message arrives, when the buffer is full.
To stop sending messages to the target, use the following command:
disable log target [console | memory-buffer | nvram | primary-msm | primary-node |

backup-msm | backup-node | session | syslog [all | <ipaddress> | <ipPort>] {vr
<vr_name>} [local0 ... local7]]]
NOTE
Refer to your UNIX documentation for more information about the syslog host facility.
Dual MSM Systems—Modular Switches Only

A system with dual MSMs keeps the two MSMs synchronized by executing the same commands on
both. However, the full data between the EMS servers is not synchronized. The reason for this design
decision is to make sure that the control channel will not be overloaded when a high number of log
messages are generated.
To capture events generated by one MSM onto the other MSM, two additional targets are shown in the
target commands—one called master-msm and one called backup-msm. The first target is active only on
the non-primary (backup) EMS server and is used to send matching events to the primary EMS server.
The other target is active only on the primary EMS server and is used to send matching events to all
other EMS servers.
If the condition for the backup-msm target is met by a message generated on the primary MSM, the
event is sent to the backup MSM. When the backup MSM receives the event, it will see if any of the
local targets (NVRAM, memory, or console) are matched. If so that event gets processed. The session
and syslog targets are disabled on the backup MSM, as they are handled on the primary. If the
condition for the primary-msm target is met by a message generated on the backup, the event is sent to
the primary MSM.
Note that the backup-msm target is active only on the primary MSM, and the primary-msm target is
active only on the backup MSM.
Filtering Events Sent to Targets

Not all event messages are sent to every enabled target. Each target receives only the messages that it is
configured for.

Target Configuration
To specify the messages to send to an enabled target, you will set a message severity level, a filter
name, and a match expression. These items determine which messages are sent to the target. You can
also configure the format of the messages in the targets. For example, the console display target is
configured to get messages of severity info and greater, the NVRAM target gets messages of severity
warning and greater, and the memory buffer target gets messages of severity debug-data and greater.
All the targets are associated by default with a filter named DefaultFilter, that passes all events at or
above the default severity threshold. All the targets are also associated with a default match expression
that matches any messages (the expression that matches any message is displayed as Match : (none)
from the command line). And finally, each target has a format associated with it.
To display the current log configuration of the targets, use the following command:
show log configuration target {console | memory-buffer | nvram | primary-msm |

primary-node | backup-msm | backup-node | session | syslog {<ipaddress> | <ipPort> |
vr <vr_name>} {[local0 ... local7]}}
To configure a target, you use specific commands for severity, filters, and formats. In addition, you can
configure the source IP address for a syslog target. Configuring the source IP address allows the
management station or syslog server to identify from which switch it received the log messages. To
configure the source IP address for a syslog target, use the following command:
configure log target syslog [all | <ipaddress> | <ipPort>] {vr <vr_name>} {local0 ...
local7} from <source-ip-address>
The following sections describe the commands required for configuring filters, formats, and severity.
Severity
Messages are issued with one of the severity levels specified by the standard Berkeley Software
Distribution (BSD) syslog values (RFC 3164)—critical, error, warning, notice, and info—plus three
severity levels for extended debugging—debug-summary, debug-verbose, and debug-data. Note that
RFC 3164 syslog values emergency and alert are not needed because critical is the most severe
event in the system.
The three severity levels for extended debugging—debug-summary, debug-verbose, and debug-data—
require that debug mode be enabled (which may cause a performance degradation). See “Displaying
Debug Information” on page 286 for more information about debugging.
Table 46: Severity levels assigned by the switch

Level Description
Critical A serious problem has been detected that is compromising the operation of the
system; the system cannot function as expected unless the situation is remedied. The
switch may need to be reset.
Error A problem has been detected that is interfering with the normal operation of the
system; the system is not functioning as expected.
Warning An abnormal condition, not interfering with the normal operation of the system, has
been detected that indicate that the system or the network in general may not be
functioning as expected.
Notice A normal but significant condition has been detected, which signals that the system is
functioning as expected.

Table 46: Severity levels assigned by the switch (Continued)

Level Description
Info (Informational) A normal but potentially interesting condition has been detected, which signals that
the system is functioning as expected; this level simply provides potentially detailed
information or confirmation.
Debug-Summary A condition has been detected that may interest a developer seeking the reason
underlying some system behavior.
Debug-Verbose A condition has been detected that may interest a developer analyzing some system
behavior at a more verbose level than provided by the debug summary information.
Debug-Data A condition has been detected that may interest a developer inspecting the data
underlying some system behavior.
You can use more than one command to configure the severity level of the messages sent to a target.
The most direct way to set the severity level of all the sent messages is to use the following command:
configure log target [console | memory-buffer | nvram | primary-msm | primayr-node |

backup-msm | backup-node | session | syslog [all | <ipaddress> | <ipPort> {vr
<vr_name>} [local0 ... local7]]] {severity <severity> {only}}
When you specify a severity level, messages of that severity level and greater are sent to the target. If
you want only those messages of the specified severity to be sent to the target, use the keyword only.
For example, specifying severity warning will send warning, error, and critical messages to the
target, but specifying severity warning only sends only warning messages.
You can also use the following command to configure severity levels, which associate a filter with a
target:
configure log target [console | memory-buffer | primary-msm | primary-node | backup-

msm | backup-node | nvram | session | syslog [all | <ipaddress> | <ipPort> {vr
<vr_name>} [local0 ... local7]]] filter <filter-name> {severity <severity> {only}}
When you specify a severity level as you associate a filter with a target, you further restrict the
messages reaching that target. The filter may allow only certain categories of messages to pass. Only the
messages that pass the filter and then pass the specified severity level reach the target.
Finally, you can specify the severity levels of messages that reach the target by associating a filter with a
target. The filter can specify exactly which message it will pass. Constructing a filter is described in
“Filtering By Components and Conditions” on page 279.
Components and Conditions

The event conditions detected by ExtremeXOS are organized into components and subcomponents. To
get a listing of the components and subcomponents in your release of ExtremeXOS, use the following
command:
show log components {<event component>} {version}
For example, to get a list of the components and subcomponents in your system, use the following
command:
show log components

The following is partial output from this command:

Severity
Component Title Threshold
------------------- ---------------------------------------------- -------------
...
...
STP Spanning-Tree Protocol (STP) Error
InBPDU STP In BPDU subcomponent Warning
OutBPDU STP Out BPDU subcomponent Warning
System STP System subcomponent Error
...
...
The display above lists the components, subcomponents, and the default severity threshold assigned to
each. In EMS, you use A period (.) is used to separate component, subcomponent, and condition names.
For example, you can refer to the InBPDU subcomponent of the STP component as STP.InBPDU. On the
CLI, you can abbreviate or TAB complete any of these.
A component or subcomponent often has several conditions associated with it. To see the conditions
associated with a component, use the following command:
show log events [<event condition> | [all | <event component>] {severity <severity>
{only}}] {details}
For example, to see the conditions associated with the STP.InBPDU subcomponent, use the following
command:
show log events stp.inbpdu

Comp SubComp Condition Severity Parameters
------- ----------- ----------------------- ------------- ----------
STP InBPDU Drop Error 2 total
STP InBPDU Dump Debug-Data 3 total
STP InBPDU Trace Debug-Verbose 2 total
STP InBPDU Ign Debug-Summary 2 total
STP InBPDU Mismatch Warning 2 total
The display above lists the five conditions contained in the STP.InBPDU component, the severity of the
condition, and the number of parameters in the event message. In this example, the severities of the
events in the STP.InBPDU subcomponent range from error to debug-summary.
When you use the details keyword, you see the message text associated with the conditions. For
example, if you want to see the message text and the parameters for the event condition
STP.InBPDU.Trace, use the following command:
show log events stp.inbpdu.trace details
Comp SubComp Condition Severity Parameters

------- ----------- ----------------------- ------------- ----------
STP InBPDU Trace Debug-Verbose 2 total
0 - string
1 - string (printf)
Port=%0%: %1%

The Comp heading shows the component name, the SubComp heading shows the subcomponent (if any),
the Condition heading shows the event condition, the Severity heading shows the severity assigned
to this condition, the Parameters heading shows the parameters for the condition, and the text string
shows the message that the condition will generate. The parameters in the text string (for example, %0%
and %1% above) will be replaced by the values of these parameters when the condition is encountered
and displayed as the event message.
Filtering By Components and Conditions. You may want to send the messages that come from a specific
component that makes up ExtremeXOS or to send the message generated by a specific condition. For
example, you might want to send only those messages that come from the STP component, or send the
message that occurs when the IP.Forwarding.SlowPathDrop condition occurs. Or you may want to
exclude messages from a particular component or event. To do this, you construct a filter that passes
only the items of interest, and you associate that filter with a target.
The first step is to create the filter using the create log filter command. You can create a filter
from scratch, or copy another filter to use as a starting point. (It may be easiest to copy an existing filter
and modify it.) To create a filter, use the following command:
create log filter <name> {copy <filter name>}
If you create a filter from scratch, that filter initially blocks all events until you add events (either the
events from a component or a specific event condition) to pass. You might create a filter from scratch if
you want to pass a small set of events and to block most events. If you want to exclude a small set of
events, use the default filter that passes events at or above the default severity threshold (unless the
filter has been modified), named DefaultFilter, that you can copy to use as a starting point for your filter.
After you create your filter, you configure filter items that include or exclude events from the filter.
Included events are passed; excluded events are blocked. To configure your filter, use the following
command:
configure log filter <name> [add | delete] {exclude} events [<event-condition> | [all
| <event-component>] {severity <severity> {only}}]
For example, if you create the filter myFilter from scratch, use the following command to include events:
configure log filter myFilter add events stp
All STP component events of at least the default threshold severity passes myFilter (for the STP
component, the default severity threshold is error). You can further modify this filter by specifying
additional conditions.
For example, assume that myFilter is configured as before, and assume that you want to exclude the
STP.CreatPortMsgFail event. To add that condition, use the following command:
configure log filter myFilter add exclude events stp.creatportmsgfail
You can also add events and subcomponents to the filter. For example, assume that myFilter is
configured as before, and you want to include the STP.InBPDU subcomponent. To add that condition,
configure log filter myFilter add events stp.inbpdu
You can continue to modify this filter by adding more filter items. The filters process events by
comparing the event with the most recently configured filter item first. If the event matches this filter
item, the incident is either included or excluded, depending on whether the exclude keyword was

used. if necessary, subsequent filter items on the list are compared. If the list of filter items is exhausted
with no match, the event is excluded and is blocked by the filter.
To view the configuration of a filter, use the following command:

show log configuration filter {<filter name>}
The following is sample output from this command (for the earlier filter):
Log Filter Name: myFilter

I/ Severity
E Comp. Sub-comp. Condition CEWNISVD
- ------- ----------- ----------------------- --------
I STP InBPDU --------
E STP CreatPortMsgFail -E------
I STP --------
Include/Exclude: I - Include, E - Exclude

Component Unreg: * - Component/Subcomponent is not currently registered
Severity Values: C - Critical, E - Error, W - Warning, N - Notice, I - Info
Debug Severity : S - Debug-Summary, V - Debug-Verbose, D - Debug-Data
+ - Debug Severities, but log debug-mode not enabled
If Match parameters present:
Parameter Flags: S - Source, D - Destination, (as applicable)
I - Ingress, E - Egress, B - BGP
Parameter Types: Port - Physical Port list, Slot - Physical Slot #
MAC - MAC address, IP - IP Address/netmask, Mask - Netmask
VID - Virtual LAN ID (tag), VLAN - Virtual LAN name
L4 - Layer-4 Port #, Num - Number, Str - String
Nbr - Neighbor, Rtr - Routerid, EAPS - EAPS Domain
Proc - Process Name
Strict Match : Y - every match parameter entered must be present in the event
N - match parameters need not be present in the event
The show log configuration filter command shows each filter item, in the order that it will be
applied and whether it will be included or excluded. The above output shows the three filter items, one
including events from the STP.InBPDU component, one excluding the event STP.CreatPortMsgFail, and
the next including the remaining events from the STP component. The severity value is shown as “*”,
indicating that the component’s default severity threshold controls which messages are passed. The
Parameter(s) heading is empty for this filter because no match is configured for this filter. Matches are
described in “Matching Expressions” next.
Each time a filter item is added to or deleted from a given filter, the specified events are compared
against the current configuration of the filter to try to logically simplify the configuration. Existing items
will be replaced by logically simpler items if the new item enables rewriting the filter. If the new item is
already included or excluded from the currently configured filter, the new item is not added to the
filter.

Matching Expressions
You can configure the switch so messages reaching the target match a specified match expression. The
message text is compared with the configured match expression to determine whether to pass the
message on. To require that messages match a match expression, use the following command:
configure log target [console | memory-buffer | nvram | primary-msm | primary-node|
backup-msm | backp-node | session | syslog [all | <ipaddress> | <ipPort> {vr
<vr_name>} [local0 ... local7]]] match [any |<match-expression>]
The messages reaching the target will match the match-expression, a simple regular expression. The
formatted text string that makes up the message is compared with the match expression and is passed
to the target if it matches. This command does not affect the filter in place for the target, so the match
expression is compared only with the messages that have already passed the target’s filter. For more
information on controlling the format of the messages, see “Formatting Event Messages” on page 283.
Simple Regular Expressions. A simple regular expression is a string of single characters including the dot
character (.), which are optionally combined with quantifiers and constraints. A dot matches any single
character, while other characters match only themselves (case is significant). Quantifiers include the star
character (*) that matches zero or more occurrences of the immediately preceding token. Constraints
include the caret character (^) that matches at the beginning of a message and the currency character ($)
that matches at the end of a message. Bracket expressions are not supported. There are a number of
sources available on the Internet and in various language references describing the operation of regular
expressions. Table 47 shows some examples of regular expressions.
Table 47: Simple regular expressions

Regular Expression Matches Does Not Match
port port 2:3 poor
import cars por
portable structure pot
..ar baar bar
bazaar
rebar
port.*vlan port 2:3 in vlan test
add ports to vlan
port/vlan
myvlan$ delete myvlan myvlan port 2:3
error in myvlan ports 2:4,3:4 myvlan link down
Matching Parameters
Rather than using a text match, EMS allows you to filter more efficiently based on the parameter values
of the message. In addition to event components and conditions and severity levels, each filter item can
also use parameter values to further limit which messages are passed or blocked. The process of
creating, configuring, and using filters has already been described in “Filtering By Components and
Conditions” on page 279, so this section describes matching parameters with a filter item.
To configure a parameter match filter item, use the following command:
configure log filter <name> [add | delete] {exclude} events [<event-condition> | [all
| <event-component>] {severity <severity> {only}}] [match | strict-match] <type>
<value>

Each event in ExtremeXOS is defined with a message format and zero or more parameter types. The
show log events all command can be used to display event definitions (the event text and
parameter types). Only those parameter types that are applicable given the events and severity specified
are exposed on the CLI. The syntax for the parameter types (represented by <type> in the command
syntax above) is:
[address-family [ipv4-multicast | ipv4-unicast | ipv6-multicast | ipv6-unicast]

| bgp-neighbor <ip address>
| bgp-routerid <ip address>
| eaps <eaps domain name>
| {destination | source} [ipaddress <ip address> | L4-port <L4-port>| mac-address
<mac-address>]
| esrp <esrp domain name>
| {egress | ingress} [slot <slot number> | ports <portlist>]
| ipaddress <ip address>
| L4-port <L4-port>
| mac-address <mac_address>
| netmask <netmask>
| number <number>
| port <portlist>
| process <process name>
| slot <slotid>
| string <match expression>
| vlan <vlan name>
| vlan tag <vlan tag>]
Beginning with ExtremeXOS 11.2, you can specify the ipaddress type as IPv4 or IPv6, depending on
the IP version. The following examples show how to configure IPv4 addresses and IPv6 addresses:
● IPv4 address
To configure an IP address, with a mask of 32 assumed, use the following command:
configure log filter myFilter add events all match ipaddress 12.0.0.1
To configure a range of IP addresses with a mask of 8, use the following command:
configure log filter myFilter add events all match ipaddress 12.0.0.0/8
● IPv6 address
To configure an IPv6 address, with a mask of 128 assumed, use the following command:
configure log filter myFilter add events all match ipaddress 3ffe::1
To configure a range of IPv6 addresses with a mask of 16, use the following command:
configure log filter myFilter add events all match ipaddress 3ffe::/16
● IPv6 scoped address
IPv6 scoped addresses consist of an IPv6 address and a VLAN. The following examples identify a
link local IPv6 address.
To configure a scoped IPv6 address, with a mask of 128 assumed, use the following command:
configure log filter myFilter add events all match ipaddress fe80::1%Default
To configure a range of scoped IPv6 addresses with a mask of 16, use the following command:
configure log filter myFilter add events all match ipaddress fe80::/16%Default
To configure a scoped IPv6 address with any VLAN, use the following command:
configure log filter myFilter add events all match ipaddress fe80::/16%*

To configure any scoped IPv6 address with a specific VLAN, use the following command:
configure log filter myFilter add events all match ipaddress fe80::/0%Default
NOTE
In the previous example, if you specify the VLAN name, it must be a full match; wild cards are not allowed.
The <value> depends on the parameter type specified. As an example, an event may contain a physical
port number, a source MAC address, and a destination MAC address. To allow only those RADIUS
incidents, of severity notice and above, with a specific source MAC address, use the following
command:
configure log filter myFilter add events aaa.radius.requestInit severity notice match
source mac-address 00:01:30:23:C1:00
The string type is used to match a specific string value of an event parameter, such as a user name. A
string can be specified as a simple regular expression.
Match Versus Strict-Match. The match and strict-match keywords control the filter behavior for those
incidents with event definition that does not contain all the parameters specified in a configure log
filter events match command.
This is best explained with an example. Suppose an event in the XYZ component, named XYZ.event5,
contains a physical port number, a source MAC address, but no destination MAC address. If you
configure a filter to match a source MAC address and a destination MAC address, XYZ.event5 will
match the filter when the source MAC address matches regardless of the destination MAC address
because the event contains no destination MAC address. If you specify the strict-match keyword,
then the filter will never match event XYZ.event5 because this event does not contain the destination
MAC address.
In other words, if the match keyword is specified, an incident will pass a filter so long as all parameter
values in the incident match those in the match criteria, but all parameter types in the match criteria
need not be present in the event definition.
Formatting Event Messages

Event messages are made up of a number of items. The individual items can be formatted; however,
EMS does not allow you to vary the order of the items. To format the messages for a particular target,
configure log target [console | memory-buffer | nvram | session | syslog [all |
<ipaddress> | <ipPort>] {vr <vr_name>} {local0 ... local7}]] format [timestamp
[seconds | hundredths | none] | date [dd-mm-yyyy | dd-Mmm-yyyy | mm-dd-yyyy | Mmm-dd |
yyyy-mm-dd | none] | severity | event-name [component | condition | none |
subcomponent] | host-name | priority | process-name | process-slot | source-line
Using the default format for the session target, an example log message might appear as:
06/25/2004 22:49:10.63 <Info:dm.Info> MSM-A: PowerSupply:4 Powered On
If you set the current session format using the following command:
configure log target session format timestamp seconds date mm-dd-yyyy event-name
component

The same example would appear as:
06/25/2004 22:49:10 <dm> PowerSupply:4 Powered On
To provide some detailed information to technical support, set the current session format using the
following command:
configure log target session format timestamp hundredths date mmm-dd event-name
condition process-name source-line
The same example then appears as:
Jun 25 22:49:10.63 <dm.info> devmgr: (dm.c:134) PowerSupply:4 Powered On
Displaying Real-Time Log Messages

You can configure the system to maintain a running real-time display of log messages on the console
display or on a (Telnet) session. To turn on the log display on the console, use the following command:
enable log target console
This setting may be saved to the FLASH configuration and is restored on boot-up (to the console
display session).
To turn on log display for the current session, use the following command:
enable log target session
This setting only affects the current session and is lost when you log off the session.
The messages that are displayed depend on the configuration and format of the target. For information
on message filtering, see “Filtering Events Sent to Targets” on page 275. for information on message
formatting, see “Formatting Event Messages” on page 283.
Displaying Event Logs

The log stored in the memory buffer and the NVRAM can be displayed on the current session (either
the console display or telnet). To display the log, use the following command:
show log {messages [memory-buffer | nvram]} {events {<event-condition> | <event-
component>]} {<severity> {only}} {starting [date <date> time <time> | date <date> |
time <time>]} {ending [date <date> time <time> | date <date> | time <time>]} {match
<regex>} {chronological}
You can use many options to select those log entries of interest. You can select to display only those
messages that conform to the specified:
● Severity
● Starting and ending date and time
● Match expression
The displayed messages can be formatted differently from the format configured for the targets, and
you can choose to display the messages in order of newest to oldest or in chronological order (oldest to
newest).

Uploading Event Logs

The log stored in the memory buffer and the NVRAM can be uploaded to a TFTP server. Use the
following command to upload the log:
upload log <ipaddress> {vr <vr_name>} <filename> {messages [memory-buffer | nvram]
{events {<event-condition> | <event_component>}}} {<severity> {only}} {match <regex>}
{chronological}
You must specify the TFTP host and the filename to use in uploading the log. There are many options
you can use to select the log entries of interest. You can select to upload only those messages that
conform to the specified:
● Severity
● Match expression
The uploaded messages can be formatted differently from the format configured for the targets, and
you can choose to upload the messages in order of newest to oldest or in chronological order (oldest to
newest).
Displaying Counts of Event Occurrences

EMS adds the ability to count the number of occurrences of events. Even when an event is filtered from
all log targets, the event is counted. To display the event counters, use the following command:
show log counters {<event condition> | [all | <event component>]} {include | notified
| occurred} {severity <severity> {only}}}
The system displays two counters. One counter displays the number of times an event has occurred,
and the other displays the number of times that notification for the event was made to the system for
further processing. Both counters reflect totals accumulated since reboot or since the counters were
cleared using the clear log counters or clear counters command.
The show log counters command also displays an included flag (the column titled In in the output).
The included flag is set to Y(es) if one or more targets are receiving notifications of this event without
regard to matching parameters.
The keywords include, notified, and occurred display events only with non-zero counter values for
the corresponding counter.
The output of the command:
show log counters stp.inbpdu severity debug-summary
is similar to the following:

Comp SubComp Condition Severity Occurred In Notified
------- ----------- ----------------------- ------------- -------- -- --------
STP InBPDU Drop Error 0 Y 0
STP InBPDU Ign Debug-Summary 0 N 0
STP InBPDU Mismatch Warning 0 Y 0
Occurred : # of times this event has occurred since last clear or reboot
Flags : (*) Not all applications responded in time with there count values
In(cluded): Set to Y(es) if one or more targets filter includes this event
Notified : # of times this event has occurred when 'Included' was Y(es)

The output of the command:
show log counters stp.inbpdu.drop
is similar to the following:

Comp SubComp Condition Severity Occurred In Notified
------- ----------- ----------------------- ------------- -------- -- --------
STP InBPDU Drop Error 0 Y 0
Occurred : # of times this event has occurred since last clear or reboot
Flags : (*) Not all applications responded in time with there count values
In(cluded): Set to Y(es) if one or more targets filter includes this event
Notified : # of times this event has occurred when 'Included' was Y(es)
Displaying Debug Information

By default, a switch does not generate events of severity Debug-Summary, Debug-Verbose, and Debug-
Data unless the switch is in debug mode. Debug mode causes a performance penalty, so it should only
be enabled for specific cases where it is needed. To place the switch in debug mode, use the following
command:
enable log debug-mode
When the switch is in debug-mode, any filters configured for your targets still affect which messages
are passed on or blocked.
Logging Configuration Changes

ExtremeXOS allows you to record all configuration changes and their sources that are made using the
CLI by way of telnet or the local console. The changes cause events that are logged to the target logs.
Each log entry includes the user account name that performed the change and the source IP address of
the client (if telnet was used). Configuration logging applies only to commands that result in a
configuration change.
To enable configuration logging, use the following command:
enable cli-config-logging
To disable configuration logging, use the following command:
disable cli-config-logging
CLI configuration logging is disabled by default.
Using sFlow
sFlow® is a technology for monitoring traffic in data networks containing switches and routers. It relies
on statistical sampling of packets from high-speed networks, plus periodic gathering of the statistics. A
User Datagram Protocol (UDP) datagram format is defined to send the information to an external entity
for analysis. sFlow consists of a Management Information Base (MIB) and a specification of the packet

Using sFlow
format for forwarding information to a remote agent. Details of sFlow specifications can be found in
RFC 3176, and specifications and more information can be found at the following website:
http://www.sflow.org
The ExtremeXOS implementation is based on sFlow version 5, which is an improvement from the
revision specified in RFC 3176. Additionally, the switch hardware allows you to set the hardware
sampling rate independently for each module on the switch, instead of requiring one global value for
the entire switch. The switch software also allows you to set the individual port sampling rates, so you
can fine-tune the sFlow statistics gathering. Per the RFC, sFlow sampling is done on ingress only.
NOTE
sFlow and mirroring are mutually exclusive on the BlackDiamond 8800 original modules and the Summit X450
series switch, whether or not it is included in a SummitStack. You can enable either sFlow, or mirroring, but not
both. This restriction is not applicable to the BlackDiamond 10808 or the BlackDiamond 12800 series switches.
sFlow and mirroring are not mutually exclusive on the BlackDiamond 8800 a-series and e-series modules and the
Summit X450a, X450e, and X250e series switches, whether or not they are included in a SummitStack. You can
enable sFlow and mirroring at the same time.
If you have a combination of BlackDiamond 8800 original, a-series, and e-series modules installed with sFlow
configured globally on the switch, you can enable mirroring on original I/O module ports provided you do not enable
sFlow on original I/O module ports. Similarly, If you have a combination of Summit X450 series switches with
Summit X450a, X450e, or X250e switches in a SummitStack with sFlow configured globally on the stack, you can
enable mirroring on Summit X450 series switch ports; provided that you do not enable sFlow on Summit X450
series switch ports.
Beginning with ExtremeXOS 11.5, you can enable sFlow and mirroring at the same time on the
following platforms:
● BlackDiamond 8800 e-series modules
● BlackDiamond 8800 a-series modules
However, you should be aware of a few limitations in the current release. The current release supports:
● Generic port statistics reported to the sFlow collector
● Non-extended data
● Only those packets that do not match an ACL rule are considered for sampling
NOTE
This limitation applies only to the BlackDiamond 12800 series switches and the BlackDiamond 10808 switch.
● Only port-based sampling

● No MIB support

● Licensing on page 288
● Sampling Mechanisms on page 288

● Configuring sFlow on page 289

● Additional sFlow Configuration Options on page 291
● sFlow Configuration Example on page 293
● Displaying sFlow Information on page 293
Licensing
You need an Advanced Edge, a Core, or an Advanced Core license to configure and use the sFlow
features described in this section.
sFlow functionality is not available with an Edge license.
The BlackDiamond 10808 switch with an MSM-1 module or an MSM1-XL module ships with a Core or
Advanced Core license, respectively.
The BlackDiamond 12804 switch with an MSM-5 module or an MSM-5R module ships with a Core
license. The Black Diamond 12802 switch ships with an Advanced Edge license.
The BlackDiamond 8800 series switch with an MSM-G8X module or an MSM-48 module, Summit X450
series switch and Summit X450a series switch ship with an Advanced Edge license.
The Summit X450e and X250e series switches ship with an Edge license, which does not support sFlow.
You can, however, upgrade to an Advanced Edge license that supports sFlow.
NOTE
If you have a Summit X450e series switch, you must upgrade to an Advanced Edge license to configure and use the
sFlow features described in this section.
For more information about software licensing, including how to obtain and upgrade your license, see
Chapter 2, “ExtremeXOS Overview.”
Sampling Mechanisms
In ExtremeXOS 11.4 and earlier, statistical sampling was done in software on the following platforms:
● BlackDiamond 8800 original modules
● Summit X450 series switch
Statistical sampling was done in hardware on the following platforms:

Beginning with ExtremeXOS 11.5, the following platforms additionally support hardware-based
sampling at a programmed interval:

Using sFlow
● Summit X250e series switches (beginning with ExtremeXOS 12.0)
With hardware-based sampling, the data path for a packet that traverses the switch does not require
processing by the CPU. Fast path packets are handled entirely by ASICs and are forwarded at wire
speed rate.
With software-based sampling, the data path for the packets is still fast path; however, the switch copies
all packets to the CPU for sampling (instead of only those that have been marked for sampling).
Software sampling requires intensive CPU processing.
Configuring sFlow
ExtremeXOS allows you to collect sFlow statistics on a per port basis. An agent, residing locally on the
switch, sends data to a collector that resides on another machine. You configure the local agent, the
address of the remote collector, and the ports of interest for sFlow statistics gathering. You can also
modify default values for how frequently on average a sample is taken and the maximum number of
samples allowed before throttling the sample gathering.
To configure sFlow on a switch, you must do the following tasks:

● Configure the local agent
● Configure the addresses of the remote collectors
● Enable sFlow globally on the switch
● Enable sFlow on the desired ports
Optionally, you may also change the default values of the following items:
● How often the statistics are collected
● How frequently a sample is taken, globally or per port
● How many samples per second can be sent to the CPU
NOTE
sFlow uses QP2 to sample traffic on the BlackDiamond 8800 original modules and the Summit X450 series switch
whether or not included in a SummitStack. Any traffic grouping on those platforms using QP2 may encounter
unexpected results when sFlow is enabled. For more information about QoS, see Chapter 14, “Quality of Service.”
Configuring the Local Agent

The local agent is responsible for collecting the data from the samplers and sending that data to the
remote collector as a series of UDP datagrams. The agent address is stored in the payload of the sFlow
data, and is used by the sFlow collector to identify each agent uniquely. By default, the agent uses the
management port IP address as its IP address. To change the agent IP address, use the following
command:
configure sflow agent {ipaddress} <ip-address>
To unconfigure the agent, use the following command:
unconfigure sflow agent

Configuring the Remote Collector Address

You can specify up to four remote collectors to send the sFlow data to. Typically, you would configure
the IP address of each collector. You may also specify a UDP port number different from the default
value of 6343, and/or a virtual router different from the default of VR-Mgmt. When you configure a
collector, the system creates a database entry for that collector that remains until the collector is
unconfigured. All the configured collectors are displayed in the show sflow {configuration}
command. To configure the remote collector, use the following command:
configure sflow collector {ipaddress} <ip-address> {port <udp-port-number>} {vr
<vrname>}
To unconfigure the remote collector and remove it from the database, use the following command:
unconfigure sflow collector {ipaddress} <ip-address> {port <udp-port-number>} {vr

<vrname>}
Enabling sFlow Globally on the Switch

Before the switch starts sampling packets for sFlow, you must enable sFlow globally on the switch. To
enable sFlow globally, use the following command:
enable sflow
If you enable sFlow globally on the Summit X450 series switch or the BlackDiamond 8800 original
modules, sFlow uses QP2 to sample traffic. When enabling sFlow, the switch displays a message similar
to the following:
SFLOW will use Qos Profile QP2.
To disable sFlow globally, use the following command:
disable sflow
When you disable sFlow globally, the individual ports are also put into the disabled state. If you later
enable the global sFlow state, individual ports return to their previous state.
Enabling sFlow on the Desired Ports

To enable sFlow on specific ports, use the following command:
enable sflow ports <port_list>
You may enable and disable sFlow on ports irrespective of the global state of sFlow, but samples are
not taken until both the port state and the global state are enabled.
On the BlackDiamond 8800 series switch with a combination of a-series, e-series, and original modules
installed, or on a SummitStack in which a combination of Summit X450 switches with either Summit
X450a, X450e, or X250e series switches is active, please note the following sFlow restrictions:
● If you enable sFlow globally on the switch and on original I/O module ports or Summit X450 series
ports and you try to enable mirroring on original, a-series, or e-series I/O module ports, or on any
ports on a Summit family switch in the SummitStack, the switch displays a message similar to the
following:
Error: Mirroring is not compatible with SFlow. Mirroring not enabled!
Configuration failed on backup MSM, command execution aborted!

Using sFlow
If you see the above message, you must disable sFlow on all of the original I/O module ports.
● If you enable sFlow globally on the switch and on a-series or e-series I/O module ports or Summit
X450a, X450e, or X250e series switch ports, and you try to enable mirroring on original I/O module
ports or on Summit X450 series switch ports in the SummitStack, the switch does not display an
error. You can successfully enable mirroring on original I/O module ports or Summit X450 series
switch ports in a SummitStack, provided you do not enable sFlow on original I/O module ports or
Summit X450 series switch ports in a SummitStack.
● If you enable sFlow on original I/O module ports or Summit X450 series switch ports in a
SummitStack, sFlow uses QP2 to sample traffic. When enabling sFlow on those ports, the switch
displays a message similar to the following:
SFLOW will use Qos Profile QP2.
For more information about the BlackDiamond 8800 a-series, e-series modules, and original modules, or
the Summit family of switches, see Chapter 2, “ExtremeXOS Overview,” and the Extreme Networks
Consolidated ExtremeXOS Hardware Guide. For more information about mirroring, see Chapter
6, “Configuring Slots and Ports on a Switch.” For more information about QOS, see Chapter
14, “Quality of Service.”
To disable sFlow on ports, use the following command:
disable sflow ports <portlist>
Additional sFlow Configuration Options

You can configure three global options to different values from the defaults. These options affect how
frequently the sFlow data is sent to the remote collector, how frequently packets are sampled, and the
maximum number of sFlow samples that could be processed in the CPU per second.
You can also configure how frequently packets are sampled per port.
Polling Interval
Each port counter is periodically polled to gather the statistics to send to the collector. If there is more
than one counter to be polled, the polling is distributed in such a way that each counter is visited once
during each polling interval, and the data flows are spaced in time. For example, assume that the
polling interval is 20 seconds and there are 40 counters to poll. Two ports will be polled each second,
until all 40 are polled. To configure the polling interval, use the following command:
configure sflow poll-interval <seconds>
Global Sampling Rate

The global sampling rate is the rate that newly enabled sFlow ports will have their sample rate set to.
Changing this rate does not affect currently enabled sFlow ports. The default sample rate is 8192, so by
default sFlow samples one packet out of every 8192 received. To configure the switch to use a different
sampling rate, use the following command:
configure sflow sample-rate <number>
For example, if you set the sample rate number to 16384, the switch samples one out of every 16384
packets received. Higher numbers mean fewer samples and longer times between samples. If you set
the number too low, the number of samples can be very large, which increases the load on the switch.

Do not configure the sample rate to a number lower than the default unless you are sure that the traffic
rate on the source is low.
Summit X450a and X450e Series Switches and BlackDiamond 8800 a-Series and e-Series Modules Only. The
minimum rate that these platforms sample is 1 out of every 256 packets. If you configure a rate to be
less than 256, the switch automatically rounds up the sample rate to 256.
Per Port Sampling Rate

The per port sampling rate overrides the system-wide value set in the configure sflow sample-rate
command. The rate is rounded off to the next power of two, so if 400 is specified, the sample rate is
configured as 512. The valid range is 1 to 536870912. To set the sampling rate on individual ports, use
configure sflow ports <portlist> sample-rate <number>
BlackDiamond 10808 Switch Only. At the hardware level, all ports on the same slot are sampled at the
same rate, so if one port is configured to sample less frequently than another on the same slot, the extra
samples are discarded. This is indicated in the output of the show sflow {configuration} command
as the sub-sampling factor. For example, if one port is configured to sample one packet per every 8192
packets, and the second port on the same slot is configured to sample one packet per every 16384
packets, the second port shows a sub-sampling factor of two.
Summit Family of Switches and BlackDiamond 8800 a-Series and e-Series Modules Only. All ports on the
switch or the same I/O module are sampled individually.
Maximum CPU Sample Limit

A high number of samples can cause a heavy load on the switch CPU. To limit the load, there is a CPU
throttling mechanism to protect the switch.
On a modular switch, whenever the limit is reached, the sample rate value is doubled on the slot from
which the maximum number of samples are received. For ports on that slot that are sampled less
frequently, the sampling rate is not changed; the sub-sampling factor is adjusted downward.
On a stand-alone switch, whenever the limit is reached, the sample rate value is doubled on the ports
from which the maximum number of samples are received. For ports that are sampled less frequently,
the sampling rate is not changed; the sub-sampling factor is adjusted downward.
To configure the maximum CPU sample limit, use the following command:
configure sflow max-cpu-sample-limit <rate>
Unconfiguring sFlow
To reset the any configured values for sFlow to their default values and remove from sFlow any
configured collectors and ports, use the following command:
unconfigure sflow

Using RMON
sFlow Configuration Example

In a service provider environment, you can configure sFlow to sample packets at the edge of the
network to determine the hourly usage for each IP address in the data center. You can capture Web
traffic, FTP traffic, mail traffic, and all bits of data that travel across the service provider’s edge routers
to their customer’s (end user’s) servers.
The example in this section assumes that you already have an sFlow data collector installed somewhere
in your network. In many environments, the sFlow data collector is on a network PC.
The following sFlow configuration example for a service provider environment:

● Configures the IP address of the sFlow data collector.
NOTE
In many environments, the sFlow data collector is not directly connected to the switch. Make sure to specify the
VR used to forward traffic between the sFlow collector and the switch. In most cases the VR is vr-mgmt.
● Configures the sampling rate on an edge port.

● Enables sFlow on the edge port.
● Enables sFlow globally on the switch.
configure sflow collector 55.55.55.69 vr vr-mgmt
configure sflow ports 4:12 sample-rate 1024
enable sflow ports 4:12
enable sflow
Displaying sFlow Information

To display the current configuration of sFlow, use the following command:
show sflow {configuration}
To display the sFlow statistics, use the following command:
show sflow statistics
Using RMON
Using the Remote Monitoring (RMON) capabilities of the switch allows network administrators to
improve system efficiency and reduce the load on the network.
The following sections explain more about the RMON concept and the RMON features supported by
the switch.
NOTE
You can only use the RMON features of the system if you have an RMON management application and have enabled
RMON on the switch.

About RMON
RMON is the common abbreviation for the Remote Monitoring Management Information Base (MIB)
system defined by the Internet Engineering Task Force (IETF) documents RFC 1757 and RFC 2021,
which allows you to monitor LANs remotely.
A typical RMON setup consists of the following two components:

● RMON agent
● Management workstation
RMON Agent
An RMON agent is an intelligent software agent that continually monitors port statistics and system
variables. The agent transfers the information to a management workstation on request, or when a
predefined threshold is crossed.
Information collected by RMON includes Ethernet port statistics and history and the software version
and hardware revision of the device. RMON generates alarms when threshold levels are met and then
logs those events to the log. RMON can also send traps to the destination address configured by the
management workstation. You can also use RMON to trigger a system reboot.
Management Workstation
A management workstation communicates with the RMON agent and collects the statistics from it. The
workstation does not have to be on the same network as the RMON agent and can manage the agent by
in-band or out-of-band connections.
If you enable RMON on the switch, you can use a management workstation to review port statistics
and port history, no configuration of the management workstation is necessary. However, you must use
a management workstation to configure the alarm and event entries.
Supported RMON Groups of the Switch

The IETF defines nine groups of Ethernet RMON statistics. The switch supports the following four of
these groups, as defined in RFC 1757:
● Statistics
● History
● Alarms
● Events
The switch also supports the following parameters for configuring the RMON agent and the trap
destination table, as defined in RFC 2021:
● probeCapabilities
● probeSoftwareRev
● probeHardwareRev
● probeDateTime
● probeResetControl
● trapDestTable

Using RMON
The following sections describe the supported groups, the RMON probe configuration parameters, and
the trap destination parameter in greater detail.
Statistics
The RMON Ethernet Statistics group provides traffic and error statistics showing packets, bytes,
broadcasts, multicasts, and errors on an Ethernet port.
Information from the Statistics group is used to detect changes in traffic and error patterns in critical
areas of the network.
History
The History group provides historical views of network performance by taking periodic samples of the
counters supplied by the Statistics group. The group features user-defined sample intervals and bucket
counters for complete customization of trend analysis.
The group is useful for analysis of traffic patterns and trends on an Ethernet port, and to establish
baseline information indicating normal operating parameters.
Alarms
The Alarms group provides a versatile, general mechanism for setting threshold and sampling intervals
to generate events on any RMON variable. Both rising and falling thresholds are supported, and
thresholds can be on the absolute value of a variable or its delta value.
Note, creating an entry in the alarmTable does not validate the alarmVariable and does not generate a
badValue error message.
Alarms inform you of a network performance problem and can trigger automated action responses
through the Events group.
Events
The Events group creates entries in an event log and/or sends SNMP traps to the management
workstation. An event is triggered by an RMON alarm. The action taken can be configured to ignore it,
to log the event, to send an SNMP trap to the receivers listed in the trap receiver table, or to both log
and send a trap. The RMON traps are defined in RFC 1757 for rising and falling thresholds.
Effective use of the Events group saves you time. Rather than having to watch real-time graphs for
important occurrences, you can depend on the Event group for notification. Through the SNMP traps,
events can trigger other actions, which provides a mechanism for an automated response to certain
occurrences.
RMON Probe Configuration Parameters

The RMON probe configuration parameters supported in ExtremeXOS are a subset of the probe
configuration group as defined in RFC 2021. The probe configuration group controls and defines the
operation of the RMON agent.

You can configure the following objects:

● probeCapabilities—If you configure the probeCapabilities object, you can view the RMON MIB
groups supported on at least one interface by the probe.
● probeSoftwareRev—If you configure the probeSoftwareRev object, you can view the current
software version of the monitored device.
● probeHardwareRev—If you configure the probeHardwareRev object, you can view the current
hardware version of the monitored device.
● probeDateTime—If you configure the probeDateTime object, you can view the current date and time
of the probe. For example, Friday December 31, 2004 at 1:30:15 PM EST is displayed as: 2004-12-
31,13:30:15.0
If the probe is aware of time zones, the display also includes the Greenwich Mean Time (GMT)
offset. For example, Friday, December 31, 2004, 1:30:15 PM EST with the offset known is displayed
as: 2004-12-31,13:30:15.0, -4.0
If time information is unavailable or unknown, the time is not displayed.
● probeResetControl—If you configure the probeResetControl object, you can restart a managed device
that is not running normally. Depending on your configuration, you can do one of the following:
■ Warm boot—A warm boot restarts the device using the current configuration saved in non-
volatile memory.
■ Cold boot—A cold boot causes the device to reset the configuration parameters stored in non-
volatile memory to the factory defaults and then restarts the device using the restored factory
default configuration.
trapDestTable
The trapDestTable contains information about the configured trap receivers on the switch and stores
this information in non-volatile memory. To configure one or more trap receivers, see “Using the Simple
Network Management Protocol,” in Chapter 4, “Managing the Switch.”
Configuring RMON
RMON requires one probe per LAN segment, and stand-alone RMON probes traditionally have been
expensive. Therefore, the approach taken by Extreme Networks has been to build an inexpensive
RMON probe into the agent of each system. This allows RMON to be widely deployed around the
network without costing more than traditional network management. The switch accurately maintains
RMON statistics at the maximum line rate of all of its ports.
To enable or disable the collection of RMON statistics on the switch, use one of the following
commands:
enable rmon
disable rmon
By enabling RMON, the switch begins the processes necessary for collecting switch statistics. By default,
RMON is disabled. However, even in the disabled state, the switch collects etherStats and you can
configure alarms and events.
RMON saves the history, alarm, and event configurations to the configuration file. Runtime data is not
stored in the configuration file and is subsequently lost after a system restart.

Using RMON
Event Actions
The actions that you can define for each alarm are shown in Table 48.
Table 48: Event actions
Action High Threshold
no action
log Sends a log message.
log-and-trap Sends a both a log message and a trap to all trap
receivers.
snmp-trap Sends a trap to all trap receivers.
To be notified of events using SNMP traps, you must configure one or more trap receivers, as described
in the section, “Using the Simple Network Management Protocol,” in Chapter 4, “Managing the
Switch.”
Displaying RMON Information

To view the status of RMON polling on the switch (the enable/disable state for RMON polling), use the
following command:
show management
To view the RMON memory usage statistics for a specific RMON feature (for example, statistics, events,
logs, history, or alarms) or for all features, use the following command:
show rmon memory {detail | <memoryType>}


12 Virtual LANs

● Overview of Virtual LANs on page 299
● Types of VLANs on page 300
● VLAN Names on page 307
● Configuring VLANs on the Switch on page 308
● Displaying VLAN Settings on page 311
Setting up Virtual Local Area Networks (VLANs) on the switch eases many time-consuming tasks of
network administration while increasing efficiency in network operations.
Overview of Virtual LANs
NOTE
Beginning with ExtremeXOS 11.2, the software supports using IPv6 addresses, in addition to IPv4 addresses. You
can configure the VLAN with an IPv4 address, IPv6 address, or both. See Chapter 7, “IPv6 Unicast Routing,” for
complete information on using IPv6 addresses.
The term VLAN is used to refer to a collection of devices that communicate as if they were on the same
physical LAN. Any set of ports (including all ports on the switch) is considered a VLAN. LAN
segments are not restricted by the hardware that physically connects them. The segments are defined by
flexible user groups that you create with the command line interface (CLI).
Benefits
NOTE
The system switches traffic within each VLAN using the Ethernet MAC address. The system routes traffic between
two VLANs using the IP addresses.
Implementing VLANs on your networks has the following advantages:

● VLANs help to control traffic—With traditional networks, broadcast traffic that is directed to all
network devices, regardless of whether they require it, causes congestion. VLANs increase the
efficiency of your network because each VLAN can be set up to contain only those devices that must
communicate with each other.
● VLANs provide extra security—Devices within each VLAN can communicate only with member
devices in the same VLAN. If a device in VLAN Marketing must communicate with devices in VLAN
Sales, the traffic must cross a routing device.

Virtual LANs
● VLANs ease the change and movement of devices—With traditional networks, network
administrators spend much of their time dealing with moves and changes. If users move to a
different subnetwork, the addresses of each endstation must be updated manually.
Virtual Routers and VLANs—BlackDiamond 10808 and 12800

Series Switches Only
NOTE
You create virtual routers only on the Black Diamond 10808 and 12800 series switches; the BlackDiamond 8800
series switches, SummitStack, and the Summit family of switches do not support user-created virtual routers.
ExtremeXOS supports virtual routers. Beginning with XOS version 11.2, each port can belong to
multiple virtual routers; that is port can belong to more than one virtual router. Ports can belong to
different VLANs that are in different virtual routers.
If you do not specify a virtual router when you create a VLAN, the system creates that VLAN in the
default virtual router (VR-Default). The management VLAN is always in the management virtual router
(VR-Mgmt).
After you create virtual routers, ExtremeXOS software allows you to designate one of these virtual
routers as the domain in which all your subsequent configuration commands, including VLAN
commands, are applied. After you create virtual routers, ensure that you are creating each VLAN in the
desired virtual router domain. Also, ensure that you are in the correct virtual router domain before you
begin modifying each VLAN.
For information on configuring and using virtual routers, see Chapter 10, “Virtual Routers.”
Types of VLANs
NOTE
Beginning with ExtremeXOS version 11.3, you may have netlogin dynamic VLANs and, on the Summit family of
switches and BlackDiamond 8800 series switches only, netlogin MAC-based VLANs. See Chapter 15, “Network
Login,” for complete information on netlogin.
VLANs can be created according to the following criteria:

● Physical port
● IEEE 802.1Q tag
● Ethernet, LLC SAP, or LLC/SNAP Ethernet protocol type
● A combination of these criteria

Types of VLANs
Port-Based VLANs
In a port-based VLAN, a VLAN name is given to a group of one or more ports on the switch.
At boot-up, all ports are members of the port-based VLAN default. Before you can add any port to
another port-based VLAN, you must remove it from the default VLAN, unless the new VLAN uses a
protocol other than the default protocol any. A port can be a member of only one port-based VLAN.
On the Extreme Networks switch in Figure 9, ports 9 through 14 are part of VLAN Marketing; ports 25
through 29 are part of VLAN Sales; and ports 21 through 24 and 30 through 32 are in VLAN Finance.
Figure 9: Example of a port-based VLAN on an Extreme Networks switch
Marketing Finance
Sales
EX_060
For the members of different IP VLANs to communicate, the traffic must be routed by the switch, even
if the VLANs are physically part of the same I/O module. This means that each VLAN must be
configured as a router interface with a unique IP address.
NOTE
support untagged frames. If your configuration has untagged frames, but the wrong 10 Gbps module, the system
fails that slot; to use the earlier revision of the 10 Gbps module, you must remove the untagged ports from the
VLAN and reset the module. To display the serial number of the module, use the show slot <slot_number>
command.
Spanning Switches with Port-Based VLANs

To create a port-based VLAN that spans two switches, you must do two things:
1 Assign the port on each switch to the VLAN.
2 Cable the two switches together using one port on each switch per VLAN.

Virtual LANs
Figure 10 illustrates a single VLAN that spans a BlackDiamond switch and another Extreme Networks
switch. All ports on the system 1 switch belong to VLAN Sales. Ports 1 through 29 on the system 2
switch also belong to VLAN Sales. The two switches are connected using slot 8, port 4 on system 1 (the
BlackDiamond switch), and port 29 on system 2 (the other switch).
Figure 10: Single port-based VLAN spanning two switches

Sales
System 1
System 2
EX_061
To create multiple VLANs that span two switches in a port-based VLAN, a port on system 1 must be
cabled to a port on system 2 for each VLAN you want to have span across the switches. At least one
port on each switch must be a member of the corresponding VLANs, as well.
Figure 11 illustrates two VLANs spanning two switches. On system 2, ports 25 through 29 are part of
VLAN Accounting; ports 21 through 24 and ports 30 through 32 are part of VLAN Engineering. On
system 1, all port on slot 1 are part of VLAN Accounting; all ports on slot 8 are part of VLAN
Engineering.
Figure 11: Two port-based VLANs spanning two switches

System 1
Accounting Engineering
System 2
EX_063
VLAN Accounting spans system 1 and system 2 by way of a connection between system 2, port 29 and
system 1, slot 1, port 6. VLAN Engineering spans system 1 and system 2 by way of a connection between
system 2, port 32, and system 1, slot 8, port 6.
Using this configuration, you can create multiple port-based VLANs that span multiple switches, in a
daisy-chained fashion. Each switch must have a dedicated port for each VLAN. Each dedicated port
must be connected to a port that is a member of its VLAN on the next switch.

Types of VLANs
Tagged VLANs
Tagging is a process that inserts a marker (called a tag) into the Ethernet frame. The tag contains the
identification number of a specific VLAN, called the VLANid (valid numbers are 1 to 4094).
NOTE
The use of 802.1Q tagged packets may lead to the appearance of packets slightly bigger than the current IEEE
802.3/Ethernet maximum of 1,518 bytes. This may affect packet error counters in other devices and may also lead
to connectivity problems if non-802.1Q bridges or routers are placed in the path.
Uses of Tagged VLANs

Tagging is most commonly used to create VLANs that span switches. The switch-to-switch connections
are typically called trunks. Using tags, multiple VLANs can span multiple switches using one or more
trunks. In a port-based VLAN, each VLAN requires its own pair of trunk ports, as shown in Figure 11.
Using tags, multiple VLANs can span two switches with a single trunk.
Another benefit of tagged VLANs is the ability to have a port be a member of multiple VLANs. This is
particularly useful if you have a device (such as a server) that must belong to multiple VLANs. The
device must have a Network Interface Card (NIC) that supports IEEE 802.1Q tagging.
A single port can be a member of only one port-based VLAN. All additional VLAN membership for the
port must be accompanied by tags.
Assigning a VLAN Tag

Each VLAN may be assigned an 802.1Q VLAN tag. As ports are added to a VLAN with an 802.1Q tag
defined, you decide whether each port will use tagging for that VLAN. The default mode of the switch
is to have all ports assigned to the VLAN named default with an 802.1Q VLAN tag (VLANid) of 1
assigned.
Not all ports in the VLAN must be tagged. As traffic from a port is forwarded out of the switch, the
switch determines (in real time) if each destination port should use tagged or untagged packet formats
for that VLAN. The switch adds and strips tags, as required, by the port configuration for that VLAN.
NOTE
Packets arriving tagged with a VLANid that is not configured on a port will be discarded.

Virtual LANs
Figure 12 illustrates the physical view of a network that uses tagged and untagged traffic.
Figure 12: Physical diagram of tagged and untagged traffic

System 1
M S
802.1Q
Tagged server
M = Marketing M
M
S = Sales S
= Tagged port S
Marketing & Sales S S
System 2 EX_064
Figure 13 is a logical diagram of the same network.
Figure 13: Logical diagram of tagged and untagged traffic

Sales
Marketing System 1
System 1
System 1 Ports 5-8, 13-16 & 32
Port 25 *
Ports 1-4 & 9-12
Port 29 *
System 2 System 2
System 2
Slot 1, Port 2 Slot 1, Port 3
Slot 1, Port 1 *
Slot 2, Ports 1-8 & 17-24 Slot 1, Port 4
Slot 2, Ports 9-16 & 25-32
*Tagged Ports EW_025
In Figure 12 and Figure 13:

● The trunk port on each switch carries traffic for both VLAN Marketing and VLAN Sales.
● The trunk port on each switch is tagged.
● The server connected to port 25 on system 1 has a NIC that supports 802.1Q tagging.
● The server connected to port 25 on system 1 is a member of both VLAN Marketing and VLAN Sales.
● All other stations use untagged traffic.
As data passes out of the switch, the switch determines if the destination port requires the frames to be
tagged or untagged. All traffic coming from and going to the server is tagged. Traffic coming from and
going to the trunk ports is tagged. The traffic that comes from and goes to the other stations on this
network is not tagged.
Mixing Port-Based and Tagged VLANs

You can configure the switch using a combination of port-based and tagged VLANs. A given port can
be a member of multiple VLANs, with the stipulation that only one of its VLANs uses untagged traffic.
In other words, a port can simultaneously be a member of one port-based VLAN and multiple tag-
based VLANs.

Types of VLANs
NOTE
For the purposes of VLAN classification, packets arriving on a port with an 802.1Q tag containing a VLANid of 0 are
treated as untagged.
Protocol-Based VLANs
Protocol-based VLANs enable you to define a packet filter that the switch uses as the matching criteria
to determine if a particular packet belongs to a particular VLAN.
Protocol-based VLANs are most often used in situations where network segments contain hosts running
multiple protocols. For example, in Figure 14, the hosts are running both the IP and NetBIOS protocols.
The IP traffic has been divided into two IP subnets, 192.207.35.0 and 192.207.36.0. The subnets are
internally routed by the switch. The subnets are assigned different VLAN names, Finance and Personnel,
respectively. The remainder of the traffic belongs to the VLAN named MyCompany. All ports are
members of the VLAN MyCompany.
Figure 14: Protocol-based VLANs
1 2 3 4 A B 5 6 7 8
192.207.35.1 192.207.36.1
My Company
192.207.35.0 192.207.36.0
Finance Personnel
1 2 3 4
= IP traffic
= All other traffic
EX_065
Predefined Protocol Filters

The following protocol filters are predefined on the switch:
● IP (IPv4)
● IPv6 (11.2 IPv6)
● MPLS
● IPX
● NetBIOS
● DECNet

Virtual LANs
● IPX_8022
● IPX_SNAP
● AppleTalk
Protocol filters on the BlackDiamond 8800 series switches, SummitStack, and the Summit series switch only.
These devices do not forward packets with a protocol-based VLAN set to AppleTalk. To ensure that
AppleTalk packets are forwarded on the device, create a protocol-based VLAN set to “any” and define
other protocol-based VLANs for other traffic, such as IP traffic. The AppleTalk packets will pass on the
“any” VLAN, and the other protocols will pass traffic on their specific protocol-based VLANs.
NOTE
Beginning with ExtremeXOS version 11.5, you can forward AppleTalk-based VLAN packets using the BlackDiamond
8800 a-series and e-series modules and the Summit X450a and X450e series switches.
Defining Protocol Filters

If necessary, you can define a customized protocol filter based on EtherType, Logical Link Control
(LLC), and/or Subnetwork Access Protocol (SNAP). Up to six protocols may be part of a protocol filter.
To define a protocol filter:
1 Create a protocol using the following command:
create protocol <name>
For example:
create protocol fred
The protocol name can have a maximum of 32 characters.
2 Configure the protocol using the following command:
configure protocol <name> add [etype | llc | snap] <hex> {[etype | llc | snap]
<hex>}
Supported protocol types include:
■ etype—EtherType
The values for etype are four-digit hexadecimal numbers taken from a list maintained by the
IEEE. This list can be found at the following URL:
http://standards.ieee.org/regauth/ethertype/index.html
■ llc—LLC Service Advertising Protocol (SAP)
The values for llc are four-digit hexadecimal numbers that are created by concatenating a two-
digit LLC Destination SAP (DSAP) and a two-digit LLC Source SAP (SSAP).
■ snap—Ethertype inside an IEEE SNAP packet encapsulation
The values for snap are the same as the values for etype, described previously.
For example:
configure protocol fred add llc feff
configure protocol fred add snap 9999
A maximum of 15 protocol filters, each containing a maximum of 6 protocols, can be defined. No more
than 7 protocols can be active and configured for use.

VLAN Names
NOTE
For more information on SNAP for Ethernet protocol types, see TR 11802-5:1997 (ISO/IEC) [ANSI/IEEE std.
802.1H, 1997 Edition].
Deleting a Protocol Filter

If a protocol filter is deleted from a VLAN, the VLAN is assigned a protocol filter of any. You can
continue to configure the VLAN. However, no traffic is forwarded to the VLAN until a protocol is
assigned to it.
Precedence of Tagged Packets Over Protocol Filters

If a VLAN is configured to accept tagged packets on a particular port, incoming packets that match the
tag configuration take precedence over any protocol filters associated with the VLAN.
Default VLAN
The switch ships with one default VLAN that has the following properties:
● The VLAN name is default.
● It contains all the ports on a new or initialized switch.
● The default VLAN is untagged on all ports. It has an internal VLANid of 1; this value is user-
configurable.
NOTE
You can choose to boot the BlackDiamond 12800 series switches without the ports being added to the default VLAN
and with the ports disabled. This information is saved in NVRAM and is not saved in the “.cfg” configuration files.
Use the commands configure switch ports initial-mode disabled to disable and configure
switch ports initial-mode enabled to enable, and show switch to display the configuration.
VLAN Names
Each VLAN is given a name that can be up to 32 characters. VLAN names use standard alphanumeric
characters.
VLAN names must begin with an alphabetical letter. The names can be no longer than 32 characters
and must begin with an alphabetic character. The remainder of the name can be alphanumeric or
contain underscore (_) characters. VLAN names cannot be keywords.
NOTE
If you use the same name across categories (for example, STPD and EAPS names), Extreme Networks recommends

Virtual LANs
VLAN names can be specified using the tab key for command completion.
VLAN names are locally significant. That is, VLAN names used on one switch are only meaningful to
that switch. If another switch is connected to it, the VLAN names have no significance to the other
switch.
NOTE
Extreme Networks recommends that you use VLAN names consistently across your entire network.
You must use mutually exclusive names for:

● VLANs
● vMANs
● IPv6 tunnels
● SVLANs
● BVLANs
Renaming a VLAN
To rename an existing VLAN, use the following command:
configure vlan <vlan_name> name <name>
The following rules apply to renaming VLANs:

● You cannot change the name of the default VLAN.
● You cannot create a new VLAN named default.
Configuring VLANs on the Switch
NOTE
support untagged frames. To display the serial number of the module, use the show slot <slot_number>
command.
This section describes how to create, name, and enable or disable a VLAN. This part covers the
following areas:
● Creating and Configuring VLANs on page 309
● Enabling and Disabling VLANs on page 309
● VLAN Configuration Examples on page 310
Beginning with ExtremeXOS version 11.4, the system returns the following message if the ports you are
adding are already EAPS primary or EAPS secondary ports:
WARNING: Make sure Vlan1 is protected by EAPS, Adding EAPS ring ports to a VLAN
could cause a loop in the network.
Do you really want to add these ports? (y/n)

Configuring VLANs on the Switch
Creating and Configuring VLANs
NOTE
If you plan to use this VLAN as a control VLAN for an EAPS domain, do NOT assign an IP address to the VLAN.
This section describes the commands associated with setting up VLANs on the switch. To configure a
VLAN:
1 Create and name the VLAN.
NOTE
Each IP address and mask assigned to a VLAN must represent a unique IP subnet. You cannot configure the
same IP subnet on different VLANs on the same virtual router.
2 Assign an IP address and mask (if applicable) to the VLAN, if needed.
NOTE
Beginning with ExtremeXOS 11.2, the software supports using IPv6 addresses, in addition to IPv4 addresses.
You can configure the VLAN with an IPv4 address, IPv6 address, or both. See Chapter 7, “IPv6 Unicast
Routing,” for complete information on using IPv6 addresses.
3 Assign a VLANid, if any ports in this VLAN will use a tag.

4 Assign one or more ports to the VLAN.
As you add each port to the VLAN, decide if the port will use an 802.1Q tag.
5 For management VLAN on the switch, configure the default iproute for virtual router VR-Mgmt.
NOTE
See Chapter 6, “IPv4 Unicast Routing,” for information on adding secondary IP addresses to VLANs.
Enabling and Disabling VLANs

Beginning with ExtremeXOS 11.4, you can enable or disable individual VLANs. The default setting is
that all VLANs are enabled.
To disable a VLAN, issue the following CLI command:
disable vlan <vlan-name>
After you have disabled a VLAN and want to re-enable that VLAN, use the following CLI command:
enable vlan <vlan-name>

Virtual LANs
Guidelines for Disabling a VLAN

This command allows you to administratively enable or disable specified VLANs. The following
guidelines apply to working with disabling VLANs:
● Disabling a VLAN stops all traffic on all ports associated with the specified VLAN.
● You cannot disable any VLAN that is running any Layer 2 protocol traffic at all.
When you attempt to disable a VLAN running Layer 2 protocol traffic (for example, the VLAN
accounting), the system returns a message similar to the following:
VLAN accounting cannot be disabled because it is actively use by an L2 Protocol
● You can disable the default VLAN; ensure that this is necessary before disabling the default VLAN.
● You cannot disable the management VLAN.
● Although you can remove ports from a disabled VLAN, you cannot add ports to a disabled VLAN or
bind Layer 2 protocols to that VLAN.
When you attempt to add ports or bind L2 protocols to a disabled VLAN (for example, the VLAN
accounting), the system returns a message similar to the following:
VLAN accounting is disabled. Enable VLAN before adding ports.
VLAN Configuration Examples
NOTE
To add an untagged port to a VLAN you create, you must first delete that port from the default vlan. if you
attempt to add an untagged port to a VLAN before deleting it from the default VLAN, you see the following error
message:
Error: Protocol conflict when adding untagged port 1:2. Either add this port as
tagged or assign another protocol to this VLAN.
The following modular switch example creates a port-based VLAN:

● Named accounting
● IPv4 address 132.15.121.1
● Slot 2, ports 1, 2, 3, and 6, and slot 4, ports 1 and 2
create vlan accounting
configure accounting ipaddress 132.15.121.1
configure default delete port 2:1-2:3,2:6,4:1,4:2
configure accounting add port 2:1-2:3,2:6,4:1,4:2
NOTE
Because VLAN names are unique, you do not need to enter the keyword vlan after you have created the unique
VLAN name. You can use the VLAN name alone (unless you are also using this name for another category such as
STPD or EAPS, in which case Extreme Networks recommends including the keyword vlan).
The following stand-alone switch example creates a port-based VLAN with an IPv6 address:
● Named development
● IPv6 address 2001:DB8::8:800:200C:417A

Displaying VLAN Settings
● Ports 1, 2, and 3
create vlan development
configure development ipaddress 2001:0DB8::8:800:200C:417A/64
configure default delete port 1-3
configure development add port 1-3
The following modular switch example creates a protocol-based VLAN named ipsales. Slot 5, ports 6
through 8, and slot 6, ports 1, 3, and 4-6 are assigned to the VLAN. In this example, you can add
untagged ports to a new VLAN without first deleting them from the default VLAN, because the new
VLAN uses a protocol other than the default protocol.
create vlan ipsales

configure ipsales protocol ip
configure ipsales add port 5:6-5:8,6:1,6:3-6:6
The following modular switch example defines a protocol filter, myprotocol and applies it to the VLAN
named myvlan. This is an example only, and has no real-world application.
create protocol myprotocol

configure protocol myprotocol add etype 0xf0f0
configure protocol myprotocol add etype 0xffff
create vlan myvlan
configure myvlan protocol myprotocol
To disable the protocol-based VLAN (or any VLAN) in the above example, use the following command:
disable vlan myprotocol
To re-enable the VLAN, use the following command:
enable vlan myprotocol
Displaying VLAN Settings
NOTE
The BlackDiamond 8800 series switches, SummitStack, and the Summit family of switches do not support user-
created virtual routers.
To display VLAN settings, use the following command:
show vlan {detail {ipv4 | ipv6} | <vlan_name> {ipv4 | ipv6} | virtual-router <vr-
router> | <vlan_name> stpd | security}
The show command displays information about each VLAN, which includes:
● Name
● VLANid
● Enabled or disabled
● How the VLAN was created
● Primary IPv4 address

Virtual LANs
● Secondary IP address (if configured)

● IPv6 addresses (if configured)
● Virtual router that VLAN belongs with
● IPX address (if configured).
● STPD information
● Protocol information
● QoS profile information
● Rate shaping information
● NetLogin information
● Ports assigned
● Tagged/untagged status for each port
● How the ports were added to the VLAN
● Number of VLANs configured on the switch
● EAPs information
● ESRP information
● IP forwarding information
● Multicasting information
● Routing protocol information
Use the detail option to display the detailed format.
NOTE
To display IPv6 information, you must use either the show vlan detail command or show vlan command
with the name of the specified VLAN.
You can display additional useful information on VLANs configured with IPv6 addresses by issuing the
show ipconfig ipv6 vlan <vlan_name>.
Displaying Protocol Information

To display protocol information, use the following command:
show protocol {<name>}
This show command displays protocol information, which includes:

● Protocol name
● Type
● Value

13 vMAN and Tunneling
This section describes how to configure the virtual metropolitan area network (vMAN) feature,
discussed in the IEEE 802.1ad specification, and contains the following topics:
● vMAN Features on page 318
● vMAN Configuration on page 326
● vMAN Examples on page 332
NOTE
Beginning with ExtremeXOS version 11.4, you can configure many aspects of vMANs using ACLs on the
BlackDiamond 10808 and 12800 series switches. See Chapter 12, “Access Lists (ACLs),” for complete information
on ACLs.
Overview
This section provides vMAN overview information and describes the following topics in greater detail:
● vMAN Overview on page 313
● BlackDiamond 12800 Series Switch vMAN Enhancements on page 315
● vMANs on the BlackDiamond 8800 Series Switches, SummitStack, and the Summit Family of
● vMANs on the BlackDiamond 10808 and BlackDiamond 12800 Series Switches Only on page 317
vMAN Overview
The 802.1ad IEEE standard which is an amendment to 802.1q, provides metro Ethernet service
providers a framework to carry VLAN traffic from multiple customers across their common network.
Extreme Network's implementation of this standard is referred to as the vMAN feature, which allows
tagging Ethernet frames to allow multiple VLANs, or separate broadcast domains to coexist on a single
physical wire.
The traditional VLAN implementation allows for one tag in the Ethernet frame. With the vMAN
feature, the tagging has been expanded to allow one more additional tag such that VLANs can be
encapsulated within vMANs. In a typical vMAN scenario with two tags, the inner tag is referred to as
the customer tag and the outer tag us referred to as the service tag. This tagging technique is sometimes
referred to as VLAN tunneling, VLAN stacking, and Q-in-Q.
Beginning with ExtremeXOS version 11.4, vMAN provided a framework for performing vMAN
translations on the Extreme switches. With ExtremeXOS version 12.0, enhancements to this feature are
available on the Extreme BlackDiamond 12800 series switches.

vMAN and Tunneling
The vMAN functionality was built on the existing ExtremeXOS policy manager by creating a new type
of ACL entry for controlling vMAN frames. vMAN ACLs define a set of match conditions and
modifiers that can be applied to vMAN traffic. These conditions allow specific traffic flows to be
identified, and the modifiers allow a translation to be performed on the traffic. This forms a mapping
between the original vMAN frame and the translated frame.
To use the vMAN feature, you configure an encapsulation for all the traffic on the specified vMAN. The
encapsulation allows the vMAN traffic to be switched over an Layer 2 infrastructure. To encapsulate the
packet, the system adds a vMAN header that forms an outer VLAN header to the Ethernet frame. The
traffic is switched through the infrastructure based on the vMAN header. The egress port of the entire
vMAN removes the vMAN header, and the frame proceeds through the rest of the network with the
original VLAN header.
When vMAN is enabled on a network port, that port adds the vMAN tag to all the outgoing frames
and incoming frames are checked to make sure their outer tag matches the vMAN Ethertype of the
network port. This provides customers the ability to tunnel VLAN traffic across a service provider's
network. The vMAN service tag (STAG) and the vLAN customer tag (CTAG) in the L2 header of the
packet is shown in Figure 15.
Figure 15: vMAN Overview
vMAN ports are assigned in the same manner as they are assigned in VLAN (see Chapter 6, “Virtual
LANs” for further details). On untagged or access ports, packets can move in both the ingress and
egress directions without the vMAN STAG, whereas the opposite is true for tagged or network ports.
Figure 16 shows the packet format for the untagged (access) and tagged (network) ports.

Overview
Figure 16: Expected Format for vMAN Tagged Ports
If your vMAN transits a third-party device (in other words, a device other than an Extreme Networks
device), you must configure the EtherType as the Ethernet type that the third-party device uses.
The system adds a 4-byte vMAN header on all packets, both originally tagged and untagged packets
arriving at the vMAN port. Beginning with ExtremeXOS 11.3, the system supports all vMAN
EtherTypes, including the standard VLAN Ethernet type of 0x8100.
The vMAN tunnel begins at the ingress, or customer access, port and terminates at the egress, or trunk,
port. Traffic flows from the egress trunk port onto the network thereafter without the vMAN tag.
Ensure that all the switch-to-switch ports in the vMAN tunnel are configured as tagged ports.
Configure the vMAN ingress, or customer access, port as an untagged port (although this port does
accept tagged packets). You must configure the vMAN tunnel egress, or trunk, port as an untagged port
so that the vMAN header is stripped from the frame.
NOTE
You must configure the vMAN tunnel egress, or trunk, port as untagged so that the vMAN header is stripped from
the frame.
In ExtremeXOS software version 12.0, vMAN support includes the following features on the
BlackDiamond 12800 series switches:
● Inter vMAN forwarding using vMAN Ingress ACL

● LAG Filtering using vMAN Egress ACL
● STAG Ethertype Translation
BlackDiamond 12800 Series Switch vMAN Enhancements

Beginning with ExtremeXOS 12.0, an s-EtherType translation feature was added to allow a second
EtherType of 0x8100 to be used for vMANs. Now a port can choose to use the primary or secondary
EtherType for the s-tag header, forcing the ingress packets to adhere to the EtherType of the port when
it belongs to a vMAN. Similarly, egress packets must have the s-tag headers matching the EtherType of
the port in the default case when no ACL changing is applied to the port.
Each tunnel port that accesses the user can support (or belong to) only one vMAN tunnel; the remaining
ports throughout the vMAN tunnel can support many vMANs.

vMAN and Tunneling
Beginning with ExtremeXOS 12.0, a customer domain feature was added for the BlackDiamond 12800
series switch platform to provide you with the flexibility to enable forwarding be based on a second
domain. This enables multiple vMANs to learn on a specific domain. In the case where you have
multiple vMANs and want all their learning to be in the same domain, you can use this feature to
combined them to a single domain for forwarding purposes.
Beginning with ExtremeXOS 12.0, the link aggregation (LAG) filtering feature enables you to pick a
specific port in a trunk group, and to specify a backup port in the event that the first preferred port
goes down. If a specific backup port is not desired, you can use the default link aggregation hash for
backup, or specify no backup port at all.
vMANs on the BlackDiamond 8800 Series Switches,

SummitStack, and the Summit Family of Switches Only
You can enable or disable jumbo frames before configuring the vMANs on the BlackDiamond 8800
series switches, SummitStack, and the Summit family of switches. On these switches, you can enable or
disable jumbo frames on individual ports. See Chapter 6, “Configuring Slots and Ports on a Switch,” for
more information on configuring jumbo frames on these devices.
Beginning with ExtremeXOS 11.6, you can configure vMANs and untagged VLANs on the same physical
port regardless of the Ethernet type for the vMAN on any BlackDiamond 8800 switches, SummitStack,
and the Summit family of switches.
Beginning with ExtremeXOS 11.5, you can configure vMANs and VLANs on the same module or on the
same stand-alone switch regardless of the vMAN Ethernet type. However, vMANs and VLANS cannot
be configured on the same physical port, unless you change the vMAN Ethernet type to 0x8100. This
capability is supported on the following platforms, which were introduced in ExtremeXOS version 11.5:
● Summit X450e series switch
Beginning with ExtremeXOS 11.4, you can configure vMANs and VLANs on the same module,
SummitStack, or on the same Summit family switch, whether or not included in a SummitStack, (and
the same port) after you configure the vMAN Ethernet type as 0x8100. You must specifically configure
this vMAN Ethernet type; the default vMAN Ethernet type is 0x88a8. If you attempt to configure
vMANs and VLANs on the same module or stand-alone box without changing the Ethernet type to
0x8100, the system returns an error. Depending on which (VLAN/vMAN) you are trying to add, the
system returns the following error message:
Error: Port X;X cannot be added to (Vlan/vMan) XXX because ports on slot X are (vMan/
Vlan) members and the vMan ethertype is not 8100.
If you already configured VLANs and vMANs on the same module or stand-alone switch using
ExtremeXOS 11.4, you cannot change the vMAN Ethernet type from 0X8100 without first removing
either the VLAN or vMAN configuration.
Before ExtremeXOS version 11.4, you cannot configure both VLANs and vMANs on the same slot on
the BlackDiamond 8800 family of switches; all ports on each slot must belong exclusively to either
VLANs or vMANs. If you do configure both VLANs and vMANs on the same slot, the system returns
an error message. The vMAN can span multiple modules, but you cannot configure VLANs and
vMANs on the same module. Also, on a Summit X450 series switch, you cannot configure both VLANs

Overview
and vMANs on the same switch; all ports on each discrete Summit X450 series switch must belong
exclusively to either VLANs or vMANs.
vMAN multicasting with IP addresses

Beginning with ExtremeXOS software version 11.4, you can use a common uplink to carry both VLAN
and vMAN traffic and provide multicast services from a vMAN through a separate VLAN. The vMAN
traffic cannot be double-tagged if you want to use IP multicast routing between vMANs and VLANs.
You can also perform IGMP snooping on packets that arrive untagged on vMAN ports for multicast
registration.
To enable this configuration:

1 Enable jumbo frames on the switch.
2 If you are using ExtremeXOS version 11.4 and you want to configure VLANs and vMANs on the
same module or stand-alone switch, change the vMAN Ethernet type to 0x8100.
Beginning with ExtemeWare XOS 11.5, step 2 is not necessary on the BlackDiamond a-series and e-
series modules and the Summit X450a and X450e series switches whether or not they are included in
a SummitStack.
3 Assign an IP address to both the VLAN and vMAN.
4 Enable IP forwarding on both the VLAN and vMAN.
5 Enable IP multicast forwarding on both the VLAN and vMAN.
6 Enable and configure multicasting on both the VLAN and vMAN.
NOTE
See Chapter 13, “Multicast Routing and Switching,” for information on configuring and using multicasting.
After you enable multicast forwarding on the vMAN, you can configure and use IGMP snooping.
vMANs on the BlackDiamond 10808 and

BlackDiamond 12800 Series Switches Only
NOTE
Beginning with ExtremeXOS version 11.4, you can use ACLs to create and configure vMANs on the BlackDiamond
10808 and 12800 series switches. See Chapter 12, “Access Lists (ACLs),” for more information on using ACLs for
vMANs.
The system automatically enables the specified ports for jumbo frames when you add ports to the
vMANs.
A physical port configured in a vMAN as an untagged port cannot also belong to a VLAN; however, a
port configured in a vMAN as a tagged port can simultaneously belong to a VLAN.
On the BlackDiamond 10808 and 12800 series switches, all ports added to a specified vMAN must be in
the same virtual router. For more information on displaying, configuring, and using virtual routers, see
Chapter 10, “Virtual Routers.”

vMAN and Tunneling
vMAN multicasting with IP addresses

Beginning with ExtremeXOS software version 11.3, you can assign an IP address to a specified vMAN
to enable multicasting. To enable multicasting on the specified vMAN after you have assigned an IP
address:
1 Enable IP forwarding.
2 Enable IP multicast forwarding.
3 Enable and configure multicasting.
NOTE
See Chapter 13, “Multicast Routing and Switching,” for information on configuring and using multicasting.
After you enable multicast forwarding on the vMAN, you can configure and use IGMP snooping.
Licensing
You need an Advanced Edge, a Core, or an Advanced Core license to configure and use the vMAN
features described in this section.
vMAN functionality is not available with an Edge license.
The BlackDiamond 12800 series switches with an MSM-5 module or an MSM-5R module ships with a
Core license.
The BlackDiamond 8800 series switches, SummitStack, Summit series switches, and Summit X450a
series switches ship with an Advanced Edge license.
The Summit X450e and X250e series switches ship with an Edge license, which does not support vMAN
functionality. You can, however, upgrade to an Advanced Edge license that supports vMANs.
vMAN Features
This sections describes the vMAN features, and covers the following topics in greater detail:
● Inter vMAN Forwarding Using vMAN Ingress ACL on page 319
● vMAN Flood Groups on page 319
● vMAN Learning Domain on page 320
● STAG Ethertype Translation on page 321
● LAG Filtering using vMAN Egress ACL on page 321
● MAC-in-MAC Tunneling—BlackDiamond 10808 and 12800 Series Switches Only on page 322
● QoS Queue on Egress Port with vMAN packets on page 325

vMAN Features
● Egress Queue on the BlackDiamond 10808 and 12800 Series Switches Only on page 325
● Egress Queue on the BlackDiamond 8800 Series Switches, SummitStack, and the Summit Family of
Inter vMAN Forwarding Using vMAN Ingress ACL

On the BlackDiamond 12800 series switches with ExtremeXOS version 12.0 and later, you can forward
traffic across vMANs using two new features: vMAN flood groups and vMAN learning domain.
vMAN Flood Groups

Flood groups are a feature that allow you to flood across multiple vMANs. They are used to determine
the route for incoming packets. There are two kinds of flood groups that you can configure: extended
flood groups and virtual flood groups. You can modify these flood groups to suit your requirements and
apply them to a vMAN port. This enables packets from the vMAN to be flooded based on these flood
groups, allowing a single packet to be distributed to multiple ports as determined from the flood group.
Extended flood groups allow any combination of configured vMAN SVIDs and ports. Figure 17 shows
an example of a packet coming in with service vLAN ID (SVID) 10 and getting flooded across all the
vMANs with the outgoing SVIDs belonging to the respective vMANs.
Figure 17: Extended Flood Groups
Virtual flood groups allow you to use any SVID and port combination. The SVIDs are virtual in that
they are not configured on any vMAN on the switch. This allows you to force an SVID on an outgoing
packet. Figure 18 shows how a virtual flood group allows you to configure an outgoing SVID on a port.
NOTE
In Figure 18 there is no vMAN associated with the virtual SVIDs.

vMAN and Tunneling
Figure 18: Virtual Flood Groups
vMAN Learning Domain

To achieve the concept of forwarding a packet across multiple domains, the learning domain vMAN was
introduced. If you want to forward packets across vMANs, you need to have a learning domain vMAN,
which includes all the ports of the individual vMANs added to it as tagged ports. This allows the
switch to use this learning domain in the forwarding context so packets can be forwarded across the
interested vMANs.
Figure 19 shows the learning domain Customer-1 allowing switching between vMANs VMAN1,
VMAN2 and VMAN3.
Figure 19: vMAN Learning Domain
A vMAN ingress ACL containing the specification for the flood groups and the learning domain is
applied to the ingress vMAN port to achieve the inter-vMAN forwarding ability.

vMAN Features
LAG Filtering using vMAN Egress ACL

On the BlackDiamond 12800 series switches with ExtremeXOS version 12.0 and later, the load sharing
feature has been extended to allow LAG filtering using vMAN egress ACLs.
Customers normally do not have control over which port in their load sharing ports traffic needs to be
sent. This feature allows you to select a port for a specific traffic stream. This is achieved by using a
vMAN Egress ACL with the LAG port specifications in it.
LAG filtering allows you to pick a specific port in the trunk group as your primary port and also
specify another port as a backup port in case of failure of the primary port. You also have the option to
have no backup port, use the entire LAG group, or use its internal logic to find the backup port. One
important thing to note is that the trunk group will need to use address based hash algorithm which
can specified by the user when the load sharing port is created. The port selection behavior holds true
for flooded traffic even when the flooded ports are trunk or link aggregation groups. You can constrain
the list of possible ports in the trunk group for each service selection.
A typical LAG egress policy looks as follows. With this rule any outgoing packet with an STAG
containing an SVID of 30 and going out of the load sharing group of ports 1:6 and 1:7, will use link 1:6
and in the case of 1:6 being down will use 1:7.
entry policy_egress {
if match all {
svid 30;
} then {
link 1:6;
back-up-link 1:7;
count counter-1;
}
}
An important condition to observe is that in the if clause only SVID can be used for the vMAN LAG
feature.
STAG Ethertype Translation

The Extreme switches support the selection of the vMAN Ethertype value by the user. This Ethertype is
used in the service TAG (STAG) to identify the vMAN frame to the switch. Figure 15 shows the STAG
in the frame. With ExtremeXOS version 12.0 and later on the BlackDiamond 12800 series switches,
support has been added for selection of an additional secondary Ethertype. This is set to a value of
0x8100. With this enhancement, you can configure a port to be a primary or secondary vMAN port.
This allows the switch to tunnel packets across two different flavored networks at any given time. A
typical scenario would be when the switch needs to talk to two different implementations of 802.1ad.
The default vMAN Ethertype of the switch is set to 0x88A8. This can be changed to any desired value.
The secondary Ethertype on switches that support this feature can be set to a fixed value of 0x8100. On
the access or untagged ports, the switch internally adds the 4 byte STAG header to each frame and
switches to the desired port. On a network or tagged port, the switch accepts a frame matching the
vMAN Ethertype of the port. Packets leaving through the access or untagged ports have their 4 byte
STAG header stripped off while packets leaving the network ports have the STAG header with the
Ethertype in it matching the exiting port's Ethertype.

vMAN and Tunneling
MAC-in-MAC Tunneling—BlackDiamond 10808 and 12800 Series

Switches Only
NOTE
After you configure a port as part of a BVLAN or SVLAN, you cannot apply any other ACLs to that port.
Beginning with ExtremeXOS version 11.4, the software anticipates the emerging IEEE 802.1ah Backbone
Bridge standard, also known as MAC-in-MAC tunneling. This standard allows Internet Service
Providers (ISPs) to use Ethernet to create a separate backbone over which the subscriber’s frames are
transported (see Figure 20).
NOTE
You must use IEEE 802.1ad vMAN tunneling (also known as Q-in-Q tunneling) in tandem with this feature. The
802.1ah-configured switch can receive and transmit only 802.1ad frames.
Figure 20: Hierarchical 802.1ad/802.1ah services
IEEE 802.1ad IEEE 802.1ah IEEE 802.1ad

User networks User networks
ISP Core backbone ISP
ISP
Bridge Bridge Bridge Bridge

Laptop
C-DA C-SA C-TAG DATA Servers

Customer frame
C-DA C-SA S-TAG C-TAG DATA

IEEE 802.1ad (Q-in-Q) frame
B-DA B-SA B-TAG I-TAG C-DA C-SA C-TAG DATA

IEEE 802.1ah (MAC-in-MAC) frame
C-DA C-SA S-TAG C-TAG DATA

IEEE 802.1ad (Q-in-Q) frame
C-DA C-SA C-TAG DATA

Customer frame
EX_113A
Each provider has a separate ID (the service ID), or ISID. This ISID maps to the I-tag on the 802.1ah
frame; thus allowing many more S-tags than the vMAN-limited 4094. These tools permit true
hierarchical scaling and full isolation of provider infrastructure from subscriber broadcast domains.
This feature works with the service provider having already created a service ID tag, using vMAN
tagging, to isolate the traffic for a particular customer; this tag is referred to as the S-tag (the original
customer tag is referred to as the C-tag). Then the 802.1ah prepends L2 information onto the existing
vMAN frame, which creates a MAC-in-MAC tunnel or provider backbone, for the ISP to send all its
traffic through the network (see Figure 21).
The 802.1ah frame maps and replaces the original service provider tag (S-tag) on the 802.1ad frame to an
I-tag on the 802.1ah frame. This mapping allows many more S-tags than the 4094 limited by vMANs; it

vMAN Features
allows duplicate S-tags on different ingress ports, each of which is mapped to a different I-tag on the
802.1ah frame. The Ethernet type of the 802.1ah frame is 0x88B5 by default; you can change this by
configuring with the CLI.
Figure 21: IEEE 802.1ah MAC-in-MAC encapsulation
IEEE 802.1ad frame:

C-DA C-SA S-TAG C-TAG DATA FCS
B-DA B-SA B-TAG I-TAG 802.1ad frame (with or without FCS) FCS
IEEE 802.1ah frame:
(S-TAG translates to I-TAG)

EX_114A
As shown in Figure 21, the material prepended consists of the following:

● I-tag—used to identify the service provider in the scope of MAC-in-MAC (the BVLAN is the tunnel
inside of which the system uses the I-tag to identify the service provider using the I-tag to S-tag
mapping and replacement)
● B-tag—is the tag for the backbone, MAC-in-MAC vMAN
● B-SA—is the MAC source address for the MAC-in-MAC bridge
● B-DA—is the MAC destination address for the MAC-in-MAC bridge
In the Extreme Networks’ implementation, the original VLAN tag is the customer’s VLAN and remains
unchanged from end-to-end.
Figure 22 shows more detail on the information in the B-TAG and I-TAG areas of the 802.1ah frame.
Figure 22: B-TAG and I-TAG details
B-DA B-SA B-TAG I-TAG 802.1ad frame FCS

88A8 0 10 88B5 4 1000
S-TAG value ISID value

(mapped S-TAG to I-TAG value)
EW_115B
You configure an SVLAN, which carries the traffic for the specific service provider and is identified by
the S-tag on the 802.1ad, or vMAN, packet. Using 802.1ah, the system maps the S-tag (on the original
vMAN packet) to the I-tag on the 802.1ah frame using the ISID, if configured, or just by copying the
original S-tag value into the I-tag. You must configure an ISID if you have duplicate S-tag values on the
same BVLAN ingressing port or if you need more than 4094 S-tag values for one BVLAN.
Then, you configure a backbone VLAN, or BVLAN, which the ISP uses as a separate backbone for all
the service traffic. Finally, you associate the specified SVLAN with the ISP’s BVLAN. (You can
configure an SVLAN and a BVLAN on the same physical port as long as there is no membership
relationship between the specific SVLAN and BVLAN. For example, if SVLAN s1 is in BVLAN b1, they
cannot share the same physical port.)

vMAN and Tunneling
The MAC-binding tables associate the ID and MAC address of each BVLAN with the ID and MAC
address of each customer host behind that BVLAN. Because of the mapping of the I-tag in the 802.1ah
packet to the S-tag in the 802.1ad packet, the BVLAN can carry many SVLANs (which carry the original
customer frames).
At the end-point of the MAC-in-MAC tunnel, the system strips off the 802.1ah material, which leaves a
vMAN frame containing an original customer packet and the service provider’s S-tag. At the end-point
of the vMAN tunnel, the system strips off the 802.1ad material, which delivers the original customer
frame to the designated destination.
The MAC-in-MAC feature prepends the ISP’s source and destination MAC address, which is the ISP’s
backbone bridge’s MAC addresses, an ISP B-VLAN tag (with three priority bits for QoS), and an ISP’s
service ID to the subscriber’s frames, which is mapped to the S-tag on the vMAN packet (see Figure 21).
Note that the I-tag that is prepended in the MAC-in-MAC frame is a direct mapping of the S-tag in the
original vMAN frame.
NOTE
There is no interaction between the STPs of the ISP and the subscriber. The subscriber’s BPDUs are tunneled
through the ISP backbone.
The B-VLAN tag identifies the ISP’s backbone MAC-in-MAC tunnel, over which the subscriber’s frames
are transported, and the service ID identifies the service provider in the ISP’s network. The ISP switches
traffic based on its backbone bridge MAC addresses. The service provider’s frames are tunneled by the
MAC-in-MAC feature; the ISP and service provider networks are isolated.
You can also view MAC-in-MAC tunneling as a Layer 2 multipoint tunnel. The backbone interface is
the tunnel entry point and multiple subscribers’ traffic can be tunneled through the MAC-in-MAC
tunnel.
The default Ethernet type for MAC-in-MAC, or 802.1ah, frames is 0x88B5 (the default Ethernet type for
vMAN, or 802.1ad, frames is 0x88A8). If you choose to configure the MAC-in-MAC Ethernet type to
another value than the default 0x88B5, ensure that the 802.1ah Ethernet type is different than the
802.1ad Ethernet type of your switch. You can change the 802.1ah Ethernet type from the default
0x88A0 value by issuing the configure bvlan ethertype <value> command.
NOTE
Ensure that the Ethernet types are different for 802.1ah and 802.1ad on your switch; the default values are
different on the Extreme Networks devices.
You can use any physical topology on the core backbone network, or BVLAN. Although you can assign
IP addresses to backbone interfaces to test connectivity, do not enable IP forwarding. The BVLAN must
be tagged, and only tagged ports can be added to the BVLAN.
Each SVLAN carries the traffic for a different service provider, or service instance. Within the SVLAN,
or vMAN, the customer VLANs remain untouched. The SVLAN is a pure Layer 2 vMAN, which is
transparent to users. Do not install Layer 2 control protocols on these interfaces because all BPDUs are
tunneled through the 802.1ah, or MAC-in-MAC, tunnel. Do not assign any IP addresses to these
SVLAN interfaces and ensure that IP forwarding is disabled.
When you configure the MAC-in-MAC tunnels using the CLI, the system automatically installs
dynamic ACLs. Although these ACLs are not user-configurable, you will see these ACLS displayed

vMAN Features
when you use the show access-list dynamic rules command. (See Chapter 12, “Access Lists
(ACLs),” for information on ACLs.)
In case of a failover from MSM A to MSM B, the network traffic for this feature is not interrupted at all.
The system has hitless failover—network traffic not interrupted in case of a failover.
Guidelines for Using MAC-in-MAC Tunnels

Keep in mind the following guidelines when you are configuring MAC-in-MAC tunnels:
● The ExtremeXOS software supports only the ‘B’ format encapsulation without FCS.
● FMT is set only to 2.
● The ExtremeXOS software does not support Drop Eligible Indicator (DEI) in the B-tag or the
I-tag.
● You can use the same SVLAN tags (duplicate SVLAN tags) on one BVLAN as long as they use
different ingressing BVLAN ports.
● You cannot configure Connectivity Fault Management (CFM), based on IEEE standard 802.1ag, on
ports in the SVLAN or BVLAN.
● You cannot have overlapping ports when you add an SVLAN to a BVLAN. You cannot have any of
the same ports on a specific SVLAN that you add to a specific BVLAN.
● The ACLs on a port are evaluated in the following order:
■ ACLs created using the CLI
■ DOS Protect-installed ACLs
■ Sentriant-installed ACLs
■ CFM-installed ACLs
■ MAC-in-MAC-installed ACLs
■ ACLs applied with a policy file (see Chapter 12, “Access Lists (ACLs),” for precedence among
these ACLs)
NOTE
For information on policy files and ACLs, see Chapter 11, “Policy Manager,” and Chapter 12, “Access Lists (ACLs),”
respectively.
QoS Queue on Egress Port with vMAN packets

The system can maintain any configured 802.1p priority contained on the original header of the packet.
See Chapter 14, “Quality of Service,” for more information on configuring the 802.1p priority.
Egress Queue on the BlackDiamond 10808 and

12800 Series Switches Only
On vMAN packets, the BlackDiamond 10808 and 12800 series switches examines the packet’s inner
802.1p tag and then directs the packet to the appropriate egress queue on the egress port. This is not
configurable.

vMAN and Tunneling
Egress Queue on the BlackDiamond 8800 Series Switches,

On the BlackDiamond 8800 series switches, SummitStack, and the Summit family of switches with
vMAN packets, you can configure the switch to examine the packet’s inner 802.1p tag or diffserv code
point and use that value to direct the packet to the appropriate queue on the egress port. To configure
this feature, use the following command:
enable dot1p examination inner-tag port [all | <port_list>]
To return to the default value of ignoring the 802.1p value on the inner tag, use the following
command:
disable dot1p examination inner-tag ports [all | <port_list>]
NOTE
See Chapter 14, “Quality of Service,” for information on configuring and displaying the current 802.1p and DiffServ
configuration for the inner, or original header, 802.1p value.
By default, the system on the BlackDiamond 8800 series switches, SummitStack, and the Summit family
of switches use the 802.1p value on the outer vMAN header to direct the packet to the queue on the
egress port.
vMAN Configuration
This section discusses vMAN configuration, and describes the following topics in greater detail:
● Guidelines for Configuring vMANs on page 326
● Configuring vMANs—BlackDiamond 8800 Series Switches, SummitStack, and the Summit Family of
● Configuring vMANs—BlackDiamond 10808 on page 329
● Configuring vMAN—BlackDiamond 12800 Series Switches on page 328
● Displaying vMAN Configurations on page 330
● Configuring MAC-in-MAC Tunnels on page 330
Guidelines for Configuring vMANs

The following are some guidelines for configuring vMANs:
● Each tunnel port that accesses the user, or customer port, can support (or belong to) only one vMAN
tunnel; the remaining ports throughout the vMAN tunnel can support many vMANs.
● Duplicate customer’s MAC address ingressing from multiple vMAN ports may disrupt the port
learning association process in the switch.
● vMAN ports can belong to load-sharing groups.

vMAN Configuration
On the BlackDiamond 10808 and the 12800 series switches, if any port in the load-sharing group is
enabled for vMAN, all ports in the group are automatically enabled to handle jumbo size frames.
Also, vMAN is automatically enabled on all ports of the untagged load-sharing group.
● You must use mutually exclusive names for:
■ VLANs
■ vMANs
■ IPv6 tunnels
■ SVLANs
■ BVLANs
● You must configure primary and secondary EtherTypes on the system (on BlackDiamond 12800
series switches only)
● You must configure the port to belong to either the primary (by default) or secondary EtherType (on
BlackDiamond 12800 series switches only)
NOTE
A restriction on the secondary EtherType is that it can only be 0x8100.
Configuring vMANs—BlackDiamond 8800 Series Switches,

NOTE
Beginning with ExtremeXOS version 11.5, you can configure vMANs and VLANs on the same module or switch on
the Summit X450a and X450e series switch and BlackDiamond 8800 a-series and e-series modules. (Beginning
with ExtremeXOS version 11.4 if you change the vMAN Ethernet type to 0x8100, you can configure VLANs and
vMANs on the same module, stand-alone switch, or port.)
Before Extreme Ware XOS version 11.4 on the BlackDiamond 8800 series switch, you could not
configure both VLANs and vMANs on the same slot on the BlackDiamond 8800 series switch; all ports
on each slot had to belong exclusively to either VLANs or vMANs. If you did configure both VLANs
and vMANs on the same slot, the system returned an error message. The vMAN was able to span
multiple modules, but the same module could not have both VLANs and vMANs. Similarly, each
discrete Summit series switch had to have all ports belonging to either VLANs or vMANs; you could
not have vMAN and VLAN ports on the same switch.
Because the BlackDiamond 8800 series switches, SummitStack, and the Summit family of switches
enable jumbo frames switch-wide, you enable jumbo frames before configuring vMANs on that system.
To configure a vMAN:
1 Enable jumbo frames on the switch.
2 If you are using ExtremeXOS version 11.4 and you want to configure VLANs and vMANs on the
same module or stand-alone switch, change the vMAN Ethernet type to 0x8100.
Beginning with ExtemeWare XOS 11.5, step 2 is not necessary on the BlackDiamond a-series and e-
series modules and the Summit X450a and X450e series switch, whether or not included in a
SummitStack. However, if you want to configure a VLAN and vMAN on the same port, you must
change the vMAN Ethernet type to 0x8100.

vMAN and Tunneling
3 Create the tunnel by creating the vMAN.

4 Assign a tag value to the vMAN.
5 Add the ports in the tunnel to the vMAN.
6 Configure vMAN member ports as tagged on switch-to-switch ports and untagged on the ingress
and egress ports of the tunnel.
NOTE
You must configure the vMAN tunnel egress, or trunk, port as untagged so that the vMAN header is stripped
from the frame.
7 Configure the switch to use the 802.1p value on the inner tag to assign the packet to the appropriate
egress queue on the egress port, if desired.
8 If you want to configure multicasting on a specified vMAN:
a Ensure that you changed the vMAN Ethernet type to 0x8100.
b Assign different IP addresses to the VLAN and the vMAN.
c Enable IP forwarding on both the VLAN and the vMAN.
d Enable IP multicasting forwarding on both the VLAN and the vMAN.
e Configure IGMP on both the VLAN and the vMAN.
Configuring vMAN—BlackDiamond 12800 Series Switches

In ExtremeXOS software version 12.0, vMAN support includes the following features on the
BlackDiamond 12800 series switches:
● Inter vMAN Forwarding Using vMAN Ingress ACL on page 328

● LAG Filtering Using vMAN Egress ACL on page 329
● STAG Ethertype Translation on page 329
NOTE
You can configure BlackDiamond 12800 series switches in ExtremeXOS version 12.0 and greater only.
Inter vMAN Forwarding Using vMAN Ingress ACL

To configure inter vMAN forwarding using vMAN ingress ACL:
1 Create vMANs and a learning domain vMAN.
2 Tag the vMANs.
3 Add ports to the vMANs and add all the ports of all the vMANs as tagged to the learning domain
vMAN.
4 Create extended and virtual flood group as needed.
5 Add vMANs or ports to the flood groups.
6 Define an ingress ACL using the flood group and learning domain.
7 Apply the ACL on ingress ports of the vMAN.

vMAN Configuration
LAG Filtering Using vMAN Egress ACL

To configure LAG filtering using vMAN egress ACL:
1 Create vMANs and a learning domain vMAN.
2 Tag the vMANs.
3 Create trunk and load sharing groups using an address-based hash algorithm.
4 Add ports to the vMANs and add all the ports of all the vMANs as tagged to the learning domain
vMAN.
5 Create extended and virtual flood group as needed.
6 Add vMANs or ports to the flood groups if necessary.
7 Create an egress ACL with LAG filtering ports as primary and backup.
8 Apply the ACL to the egress trunk ports.
STAG Ethertype Translation

To configure STAG Ethertype translation:
1 Configure the primary secondary Ethertype values globally on the switch.
2 Select the ports belonging to the secondary Ethertype. (By default, all ports are primary Ethertype
ports.)
3 Create the vMANs
4 Tag the vMANs.
5 Add the ports to the vMANs.
Configuring vMANs—BlackDiamond 10808

The BlackDiamond 10808 and 12800 series switches automatically enables jumbo frames on the ports
that you add to a vMAN.
To configure a vMAN:
1 Create the tunnel by creating the vMAN.
2 Assign a tag value to the vMAN.
3 Add the ports in the tunnel to the vMAN.
4 Configure vMAN member ports as tagged on switch-to-switch ports and untagged on the ingress
and egress ports of the tunnel.
NOTE
You must configure the vMAN tunnel egress, or trunk, port as untagged so that the vMAN header is stripped
from the frame.
5 If you want to configure multicasting on a specified vMAN:

a Assign an IP address to the vMAN.
b Enable IP forwarding.
c Enable IP multicasting forwarding.
d Configure IGMP.

vMAN and Tunneling
Displaying vMAN Configurations

You can display the vMAN configuration and associated EAPS domains by issuing the show vman
command. You can also display vMAN information, as well as all the VLANs, by issuing the show
port information detail display.
NOTE
The display for the show vman command is different depending on the platform and configuration you are using.
See the ExtremeXOS Command Reference Guide for complete information on the show vman command.
To display information on all vMANs, use the following command:
show vman
To display information on a specific vMAN, use the following command:
show vman <vlan_name>

To display the EtherType, use the following command:
show vman <vman_name> ethertype
NOTE
You use this command to display the Ethernet type for MAC-in-MAC tunneling (802.1ah) also.
Configuring MAC-in-MAC Tunnels

To configure MAC-in-MAC tunnels, accomplish the following steps:
● Create the BVLAN.
● Configure the BVLAN.
● Create the SVLAN.
● Configure the SVLAN.
● Configure SVLAN ISID, if desired.
● Add the SVLAN to the BVLAN.
If you choose to change the Ethernet type for the 802.1ah, or MAC-in-MAC frames, Extreme Networks
recommends changing the value before any other 802.1ah configuration. Changing the 802.1ah Ethernet
type may cause the system to flush both the FDB and the MAC-binding tables. To change the 8021.ah
Ethernet type from the default value of 0X88B5, use the following command:
configure bvlan ethertype <value>
To create the BVLAN, use the following command:
create bvlan <bvlan-name>
To configure the BVLAN by adding a tag and adding ports, using the following commands:
configure bvlan <blvan-name> tag <tag>

configure bvlan <blvan-name> add ports <port_list> tagged

vMAN Configuration
NOTE
You must tag the ports you add to the BVLAN.
To delete the BVLAN or any ports assigned to it, use the following commands:
delete bvlan <bvlan-name>

configure bvlan <blvan-name> delete ports [all | <port_list>]
To create the SVLAN, use the following command:
create svlan <svlan-name>
You configure the SVLAN by assigning it a tag and adding ports, which can be either tagged or
untagged. To do this, use the following commands:
configure svlan <svlan-name> tag <tag>

configure svlan <svlan-name> add ports <port_list> {tagged | untagged}
You can assign a specific ISID to each SVLAN in the BVLAN; if you do not assign an ISID, the system
uses the value in the S-tag as the value in the I-tag. Each ISID in the BVLAN must be unique. To
configure the ISID for the SVLAN, use the following command:
configure svlan <svlan-name> isid <isid>
NOTE
You must configure an ISID if you have duplicate S-tag values on the same BVLAN ingressing port or if you need
more than 4094 S-tag values for one BVLAN.
To delete the SVLAN or to remove ports from the SVLAN, use the following commands:
delete svlan <slvan-name>

configure svlan <svlan-name> delete ports [all | <port_list>]
To add the SVLAN to the BVLAN, use the following command:
configure bvlan <bvlan-name> add svlan <svlan-name>
To delete an SVLAN from the BVLAN, use the following command:
configure bvlan <bvlan-name> delete svlan [<svlan-name> | all]
To clear the database that binds the BVLAN MAC address and the MAC address for all the customer
host MAC addresses behind the BVLAN, use the following command:
clear mac-binding {bvlan <bvlan-name> {<bridge-mac>} | svlan <svlan-name> <customer-

mac>}

vMAN and Tunneling
Displaying MAC-in-MAC Tunneling Information

You can display the internal database that binds the BVLAN MAC address and the MAC address for all
the customer host MAC addresses behind the BVLAN, as well as details regarding the MAC-in-MAC
configuration on the switch. This section discusses the following topics:
● Displaying MAC-in-MAC Tunneling Configuration on page 332
● Displaying MAC-in-MAC Tunneling (BVLAN) Ethernet Type on page 332
● Displaying the MAC Binding Database on page 332
Displaying MAC-in-MAC Tunneling Configuration

To display information on BVLANs and SVLANs with MAC-in-MAC tunnels, use the following
commands:
show bvlan {<bvlan_name> {ipv4 | ipv6}} {detail} and show svlan {<svlan_name> {ipv4 |
ipv6}} {detail}
The display from the show bvlan detail command shows all the information shown in the show
bvlan <bvlan_name> command, but displays information for all configured BVLANs. The display
from the show svlan detail command shows all the information shown in the show svlan
<svlan_name> command, but displays information for all configured SVLANs.
Displaying MAC-in-MAC Tunneling (BVLAN) Ethernet Type
NOTE
To display the EtherType for IEEE 802.1ah MAC-in-MAC tunneling, use the command show vman ethertype.
The following is an example of the display from the show vman etherType command when you
configure MAC-in-MAC tunnels (802.1ah):
BlackDiamond 12804.41 # show vman etherType
vman EtherType : 0x88a8
bvlan EtherType: 0x88b5
Displaying the MAC Binding Database

To display the internal database on MAC binding, use the following command:
show mac-binding {bvlan <bvlan-name> {<bridge-mac> {svlan <svlan-name> <customer-

mac>}} | svlan <svlan-name>}
vMAN Examples
This section contains some vMAN examples, and describes the following topics in greater detail:
● MAC-in-MAC Tunneling Example on page 333
● BlackDiamond 10808 Switch Example on page 335
● BlackDiamond 8810 Switch Example on page 336

vMAN Examples
● Inter vMAN Forwarding Using vMAN Ingress ACL Example on page 337
● LAG Filtering using vMAN Egress ACL Example on page 339
● STAG Ethertype Translation Example on page 340
MAC-in-MAC Tunneling Example

The example shown in Figure 23 illustrates MAC-in-MAC tunneling. In this example, two ISPs, each
with several customers, are using one MAC-in-MAC tunnel from the backbone provider. S1, S2, and S3
represent the three ISPs (again, each has it own customers running through that SVLAN), and all are
using the one MAC-in-MAC tunnel, B1, through the core backbone.
Figure 23: MAC-in-MAC tunneling configuration example
S1 tag 10 S1 tag 10
S2 tag 10 S2 tag 10
S3 tag 30 S3 tag 30
BlackDiamond BlackDiamond
Edge 2 Edge 1
Backbone
VMAN1 core switch
Port 5:5 tag Port 5:4 tag
B1 tag 100 B1 tag 100
EX_112A
NOTE
You must enable jumbo frames on all ports before configuring MAC-in-MAC tunneling.
To configure MAC-in-MAC tunneling for this example, use the following CLI commands:
● On the BlackDiamond Edge switch 1, use the following commands:
create svlan s1
create svlan s2
create svlan s3
configure svlan s1 tag 10
configure svlan s1 add port 1:1 untagged
configure svlan s2 add port 1:2 tagged
configure svlan s1 isid 1000

vMAN and Tunneling
create bvlan b1
configure bvlan b1 tag 100
configure bvlan b1 add port 2:2 tagged
configure bvlan b1 add svlan s1
● On the BlackDiamond Edge switch 2, use the following commands:
create svlan s1
create svlan s2
create svlan s3
configure svlan s1 add port 4:1 untagged
create bvlan b1
configure bvlan b1 tag 100
configure bvlan b1 add port 3:1 tagged
● On the Summit switch (core MAC-in-MAC tunnel backbone), use the following commands:
create vman vman1
configure vman vman1 tag 100
configure vman vman1 add port 5:4-5 tagged
After you have accomplished these CLI commands, the switches will have configurations detailed in the
following paragraphs.
The BlackDiamond 12800 series switches (backbone edge 1) has the following configuration:
● Three SVLANs, or service VLANs, as follows:
■ S1 VLAN tag 10 and port 1:1 untagged added to that SVLAN
■ S2 VLAN tag 10 and port 1:2 tagged added to that SVLAN
■ S3 VLAN tag 30 and port 1:3 tagged added to that SVLAN
● A BVLAN, or backbone VLAN, named B1 tag100 and port 2:2 tagged added to that BVLAN
The BlackDiamond 10808 switch (backbone edge 2) has the following configuration:
● Three SVLANs, or service VLANs, as follows:
■ SVLAN s1 tag 10 and port 4:1 untagged added to that SVLAN
■ SVLAN s2 tag 10 and port 4:2 tagged added to that SVLAN
■ SVLAN s3 tag 30 and port 4:3 tagged added to that SVLAN
● A BVLAN, or backbone VLAN, named B1 tag100 and port 3:1tagged added to that BVLAN
The core backbone switch (core MAC-in-MAC tunnel backbone) has the following configuration:
● Port 5:4 is connected to the BlackDiamond 12800 series switch BVLAN b1 port 2:2.
● Port 5:5 is connected to the BlackDiamond 10808 switch BVLAN b1 port 3:2.

vMAN Examples
BlackDiamond 10808 Switch Example

The following example shows the steps to configure vMAN 1 on the Black Diamond 10808 switch
shown in Figure 24. Beginning with ExtremeXOS software version, enable multicasting on the vMAN
on the BlackDiamond 10808 switch.
Figure 24: Sample vMAN configuration on BlackDiamond 10808 switch
Engineering & BlackDiamond 10808 BlackDiamond 6808

Science Building EX 101
The vMAN is from the building to port 1, slot 1 on the BlackDiamond 10808 switch and from port 1,
slot 6 on the BlackDiamond 10808 switch to the BlackDiamond 6808 switch:
create vman vman_tunnel_1

configure vman vman_tunnel_1 tag 100
configure vman vman_tunnel_1 add port 1:1 untagged
configure vman vman_tunnel_1 add port 6:1 tagged
configure vlan vman_tunnel ipaddress 10.120.120.120.1/24
enable ipforwarding vman_tunnel_1
enable ipmcforwarding vman_tunnel_1

vMAN and Tunneling
BlackDiamond 8810 Switch Example

The following example shows the steps to configure vMAN 1 on the BlackDiamond 8810 switch shown
in Figure 25.
Figure 25: Sample vMAN configuration on BlackDiamond 8810 switch
Engineering & BlackDiamond 8810 BlackDiamond 6808

Science Building
XOS001
The vMAN is from the building to port 1, slot 3 on the BlackDiamond 8810 switch and from port 2, slot
3 on the BlackDiamond 8810 switch to the BlackDiamond 6808 switch:
enable jumbo frames

create vman vman_tunnel_1
configure vman vman_tunnel_1 tag 100
configure vman vman_tunnel_1 add port 3:1 untagged
configure vman vman_tunnel_1 add port 3:2 tagged
enable dot1p examination inner-tag port 3:2
The following example configuration demonstrates configuring IP multicast routing between vMANs
and VLANs (when vMAN traffic is not double-tagged) on the BlackDiamond 8800 series switch and the
Summit family of switches. Using this configuration you can use a common uplink to carry both VLAN
and vMAN traffic and to provide multicast services from a vMAN through a separate VLAN (notice
that port 1:1 is in both a VLAN and a vMAN):

configure vman ethertype 0x8100
create vlan mc_vlan
configure vlan mc_vlan tag 77
create vman vman1
configure vlan vman1 ipaddress 10.0.0.1/24
configure vlan mc_vlan ipaddress 11.0.0.1/24
enable ipforwarding vman1
enable ipforwarding mc_vlan
enable ipmcforwarding vman1
enable ipmcforwarding mc_vlan
configure vlan mc_vlan add port 1:1 tag
configure vman vman1 add port 1:1 tag
configure vman vman1 add port 2:1, 2:2, 2:3
NOTE
IGMP reports can be received untagged on ports 2:1, 2:2, and 2:3. Tagged IP multicast data is received on mc_vlan
port 1:1 and is routed using IP multicasting to vman1 ports that subscribe to the IGMP group.

vMAN Examples
NOTE
IGMP snooping (Layer 2 IP multicasting forwarding) does not work on the vMAN ports because there is no double-
tagged IP multicast cache look capability from port 1:1.
Inter vMAN Forwarding Using vMAN Ingress ACL Example

The following example shows the steps to configure inter vMAN forwarding using vMAN ingress ACL
as shown in Figure 26.
Figure 26: Sample inter vMAN forwarding using vMAN ingress ACL
create vman vman10

# configure vman vman10 tag 10
# configure vman vman10 add port 1:1, 1:2 tagged
#
# create vman vman300
#
# configure vman vman4000 add port 1:1, 1:2, 2:1, 2:2 tag
# create flood-group extended extend-1

# configure flood-group ext extend-1 add vman10 port all
#
entry vman10 {
if {
svid 10;
} then {
learning-domain vman4000;
flood-group extend-1 no-echo;
} }
# configure access-list vman10 port 1:1, 1:2 ingress

#
# configure access-list vman300 port 2:1, 2:2 ingress

vMAN and Tunneling
An alternate method for this configuration is shown in Figure 27.
Figure 27: Sample inter vMAN forwarding using vMAN ingress ACL—Alternative
create vman vman10

#
#
# configure vman vman4000 add port 1:1, 1:2, 2:1, 2:2 tagged
#
# create flood-group virtual virt-1
# configure flood-group vir virt-1 add stag 10 port 1:1
#
# configure access-list svid_10 port 1:1, 1:2 ingress
# configure access-list svid_300 port 2:1, 2:2 ingress

vMAN Examples
LAG Filtering using vMAN Egress ACL Example

The following example shows the steps to configure LAG filtering using vMAN egress ACL as shown
in Figure 28.
Figure 28: LAG filtering using vMAN egress ACL example

#
# enable sharing 1:1 grouping 1:1, 1:2, 1:3 algorithm port-based
# enable sharing 2:1 grouping 2:1, 2:2 algorithm port-based
#
# configure vman vman10 add port 1:1, 2:1, 5:1 tagged
#
# configure access-list lag_11 port 1:1 egress
# configure access-list lag_21 port 2:1 egress

vMAN and Tunneling
STAG Ethertype Translation Example

The following example shows the steps to configure STAG Ethertype translation as shown in Figure .
Figure 29: STAG Ethertype translation example
# configure vman ethertype 0x9100 primary

# configure vman ethertype 0x8100 secondary
#
# configure port 2:2 ethertype secondary
#
#
# configure vman vman300 add port 1:1, 2:1, 2:2 tagged
# configure vman vman300 add port 1:2 untagged
#
# configure access-list port_22_i port 2:2 ingress
# configure access-list port_22_o port 2:2 egress

vMAN Examples
An alternate method for this configuration is shown in Figure 30.
Figure 30: STAG Ethertype translation example—alternative

#
#
# create flood-group extended extend-2

vMAN and Tunneling

14 Web-Based Device Management
This section describes the web-based device management tool ScreenPlay, and contains the following
topics:
● “ScreenPlay Overview” on page 343
● “Setting Up ScreenPlay” on page 343
● “ScreenPlay Dashboard” on page 346
● “Configuration” on page 348
● “Statistics and Monitoring” on page 351
● “Administration” on page 354
ScreenPlay Overview
ScreenPlay is a web-based device management tool for all ExtremeXOS-based devices running software
version 12.0 or higher. ScreenPlay is launched as a web page on the device. The ScreenPlay client,
loaded onto the web browser, uses SOAP over HTTP to communicate with the device using XML APIs.
ScreenPlay provides a graphical user interface for the more commonly used CLI commands, with focus
on the statistics and monitoring commands.
Setting Up ScreenPlay
This section describes the setup process you need to go through before you can use ScreenPlay with the
switch, and includes the following topics:
● “HTTP and HTTPS Setup” on page 343
● “Client Setup” on page 344
● “Launching ScreenPlay” on page 345
HTTP and HTTPS Setup

Before you can launch ScreenPlay, you must enable web services. You can use either HTTP or HTTPS to
access ScreenPlay.
NOTE
You must assign an IP address to a VLAN for management access to the switch. For more information, see the
chapter “Configuring Management Access.”

Web-Based Device Management
Web Services Using HTTP

To enable HTTP web services, enter the following command:
SummitX450a-48t.1 # enable web http
The switch is now ready for web service calls using HTTP at the URL "http://<switch_ip>".
Web Services Using HTTPS

To enable HTTPS web service, you must first check to see if the SSL module is installed. Enter the
following command:
BD-8806.1 # show ssl
If the following displays, the SSL module is not installed:
SSL Module: Not Installed

BD-8806.2 #
The SSL module must be installed to enable HTTPS web service. (Refer to the section “Secure Socket
Layer” in Chapter 16, “Security.” See also the section “Guidelines for Activating SSL” in Appendix
A, “Software Upgrade and Boot Options.”)
After the SSL module is installed, create a certificate by entering the following command:
BD-8806.2 # configure ssl certificate privkeylen 1024 country us organization extreme

common-name name1
The HTTPS server is not enabled by default; you must enable it manually. To enable HTTPS web
service, enter the following command:
SummitX450a-48t.1 # enable web https

The switch is now ready for web service calls using HTTPS at the URL "https://<switch_ip>".
Client Setup
You can launch the interface by navigating any standard web browser with the Adobe Flash® Player 9
plug-in installed, such as Mozilla Firefox® (version 1.0 or greater) or Internet Explorer® (version 6.0 or
greater), to the ExtremeXOS-based device.
NOTE
ScreenPlay supports up to six concurrent sessions.

Setting Up ScreenPlay
Launching ScreenPlay
To launch ScreenPlay, enter the URL of the switch in the address window of your browser. When
ScreenPlay launches, the authentication screen displays, as shown in Figure 31.
Figure 31: Login Window
The ScreenPlay login window provides you with the switch IP address. You must enter a user name
and password for access. The default user name is “admin” and the there is no default password (so
enter nothing in that field).

ScreenPlay Dashboard
Next, the ScreenPlay dashboard (Figure 32) appears. The dashboard is the home screen, or opening
screen, of ScreenPlay. This screen provides you with a one-glance-snapshot of switch status, inventory,
and management details.
From the device dashboard, you can navigate to any other portion of the interface that you desire. The
functions available in ScreenPlay are divided up into three major categories:
● Configuration, which covers configuration of ports, VLANs, stacking, and SNMP

● Statistics and Monitoring, which provides you with the capability to generate event logs, monitor
and generate statistics on ports, and perform QoS monitoring
● Administration, which allows you to perform administrative tasks on user accounts and user
sessions
Figure 32: Dashboard

ScreenPlay Dashboard
The main dashboard is divided into four main information panes, and allows you to navigate to all the
other ScreenPlay views, as shown in Figure 33.
Figure 33: Dashboard Details
The dashboard details are listed in Table 49.
Table 49: Dashboard Details

Name Description and Function
1 Menu Bar Allows you to navigate to all other ScreenPlay views. This menu remains
consistent across all views.
2 Main Display Pane Displays the main information for that ScreenPlay feature, as well as any
(main work area) subordinate available functions, usually available through tabs. For example,
when “Statistics & Monitoring > Ports” is selected in the Menu Bar, tabs appear
in the Main Display Pane to allow you display a Utilization Chart, Statistics
Table, or Bandwidth Chart view.
3 Switch Summary Contains summary information about the switch, such as the switch name, the
Pane system type, the number of active sessions, and boot version. This pane remains
consistent across all views.
4 Hardware Displays hardware information about the switch, such as the number of slots in
Information Pane the switch, the number of fan trays, the number of power modules and whether
(health inventory) those modules are active. This pane remains consistent across all views. You can
click items in this pane to get more detailed information about that item. For
example, clicking a slot provides you with such slot information as its state,
serial number, and temperature.

Configuration
You can use the ScreenPlay tool to perform device-level configuration tasks. This section provides an
overview of the four configuration panes available through ScreenPlay:
● Ports Configuration on page 348
● VLAN Configuration on page 349
● Stacking Configuration on page 349
● SNMP Configuration on page 350
Ports Configuration
This feature of the ScreenPlay tool allows you to view and modify some of the basic configurations of
the ports on the device. The Port Configuration screens don't allow you to configure all port attributes.
The features available through the ports configuration screen are:
● Port listing
● Port details
● Enabling and disabling ports
● Basic port set operations
Figure 34 shows the ports configuration screen of the ScreenPlay tool.
Figure 34: Ports Configuration Screen

Configuration
VLAN Configuration
This feature of the ScreenPlay tool allows you to create, modify, and delete VLANs, as well as add ports
to VLANs and configure them. The features available through the VLAN configuration screen are:
● VLAN listing
● VLAN details
● Enabling and disabling VLANs
● Basic VLAN configuration
● Port membership in VLANs
Figure 35 shows the VLAN configuration screen of the ScreenPlay tool.
Figure 35: VLAN Configuration Screen
Stacking Configuration
This feature of the ScreenPlay tool allows you to manage view and monitor ExtremeXOS stackable
switches in a stack configuration. This feature is limited to viewing the node configuration, and
monitoring the stack topology and interconnectivity among the stack members. The configuration
capabilities of actually setting up or maintaining a stack are outside the scope of this feature.

The features available through the stacking configuration screen are:
● Stack topology
● Stack nodes and connectivity
● Stack node details
NOTE
This feature will only be available on switches that support stacking.
Figure 36 shows the port statistics screen of the ScreenPlay tool.
Figure 36: Stacking Configuration Screen
SNMP Configuration
This feature of the ScreenPlay tool allows you to view the SNMP configuration on the switch. The
information provided is useful to view the settings that are used by an SNMP client communicating
with the switch. The tool allows you to view and manipulate the following SNMP features:
● SNMP settings
● SNMP v1 and v2c communities
● SNMP v3 users

Statistics and Monitoring
● SNMP trap receivers

● SNMP statistics
NOTE
There are no capabilities to make changes to the SNMP configuration in this release.
Figure 37 shows the SNMP configuration screen of the ScreenPlay tool.
Figure 37: SNMP Configuration Screen

This section provides an overview of the three statistics and monitoring panes available through
ScreenPlay:
● Event Log on page 352

● Ports on page 352
● QoS on page 353

Event Log
This feature of the ScreenPlay tool provides you with a tabular event log of the activity on the switch.
Figure 38 shows the Event Log screen of the ScreenPlay tool.
Figure 38: Event Log Screen
Ports
This feature of the ScreenPlay tool allows you to view the live statistics of the ports as various kinds of
charts and tables. The tool allows you to view the following port statistic features:
● Port Utilization
● Port Counters
● Port Errors
● Graphical plot

Figure 39 shows the port statistics screen of the ScreenPlay tool.
Figure 39: Port Statistics Screen
QoS
This feature of the ScreenPlay tool allows you to monitor QoS information on a port (both bytes and
packets). Figure 38 shows the QoS monitoring screen of the ScreenPlay tool.

Figure 40: QoS Monitoring Screen
Administration
This section provides an overview of the two administration panes available through ScreenPlay:
● User Accounts on page 354

● User Sessions on page 355
User Accounts
This feature of the ScreenPlay tool allows you to manage user accounts on a switch. This includes the
ability to add, modify and delete user accounts local to the switch and configuring remote RADIUS
AAA, or TACACS servers.
The user accounts screen includes the following views:
● List of User Accounts

● Add and modify an account
● Global password policy configuration form
● Navigation to currently logged-in sessions

Administration
Figure 41 shows the user accounts screen of the ScreenPlay tool.
Figure 41: User Accounts Screen
User Sessions
This ScreenPlay feature allows you to manage Telnet and SSH sessions on a switch. This includes the
ability to view current and historical sessions and allow administrators to kill rogue sessions. The
features available through the sessions management screen are:
● Monitor and manipulate active CLI and XML API sessions
● Clear selected session
● View session history
Figure 42 shows the session management screen of the ScreenPlay tool.

Figure 42: User Sessions Screen

15 Forwarding Database

● Overview of the FDB on page 357
● Differing FDB Table Sizes—BlackDiamond 8800 Series Switches, SummitStack, and the Summit
Family of Switches Only on page 359
● FDB Configuration Examples on page 360
● Configuring the FDB Aging Time on page 360
● Displaying FDB Entries on page 361
● MAC-Based Security on page 362
● Multicast FDB with Multiport Entry—BlackDiamond 8800 Series Switches, SummitStack, and the
Summit Family of Switches Only on page 365
Overview of the FDB
NOTE
See the ExtremeXOS Command Reference Guide for details of the commands related to the FDB.
The switch maintains a database of all MAC addresses received on all of its ports. It uses the
information in this database to decide whether a frame should be forwarded or filtered.

● FDB Contents on page 357
● How FDB Entries Get Added on page 358
● FDB Entry Types on page 358
FDB Contents
Each Forwarding Database (FDB) entry consists of:
● The MAC address of the device
● An identifier for the port and VLAN on which it was received
● The age of the entry
● Flags
Frames destined for MAC addresses that are not in the FDB are flooded to all members of the VLAN.

Forwarding Database
How FDB Entries Get Added

Entries are added into the FDB in the following ways:
● The switch can learn entries by examining packets it receives. The system updates its FDB with the
source MAC address from a packet, the VLAN, and the port identifier on which the source packet is
received.
The ability to learn MAC addresses can be enabled or disabled on a port-by-port basis.
You can also limit the number of addresses that can be learned, or you can “lock down” the current
entries and prevent additional MAC address learning.
NOTE
For more information on port control for learning MAC address, refer to Chapter 5, “Commands for Configuring Slots
and Ports on a Switch.”
● You can enter and update entries using the command line interface (CLI).
● Certain static entries are added by the system upon switch boot-up.
FDB Entry Types

FDB entries may be dynamic or static, and the entries may be permanent or non-permanent. The
following describes the types of entries that can exist in the FDB:
● Dynamic entries—A dynamic entry is learned by the switch by examining packets to determine the
source MAC address, VLAN, and port information. The switch then creates or updates an FDB entry
for that MAC address. Initially, all entries in the database are dynamic, except for certain entries
created by the switch at boot-up.
Entries in the database are removed (aged-out) if, after a period of time (aging time), the device has
not transmitted. This prevents the database from becoming full with obsolete entries by ensuring
that when a device is removed from the network, its entry is deleted from the database. Dynamic
entries are flushed and relearned (updated) when any of the following take place:
■ A VLAN is deleted.
■ A VLAN identifier (VLANid) is changed.
■ A port mode is changed (tagged/untagged).
■ A port is deleted from a VLAN.
■ A port is disabled.
■ A port enters blocking state.
■ A port goes down (link down).
A non-permanent dynamic entry is initially created when the switch identifies a new source MAC
address that does not yet have an entry in the FDB. The entry may then be updated as the switch
continues to encounter the address in the packets it examines. These entries are identified by the “d”
flag in show fdb output.
Dynamic entries age—that is, a dynamic entry is removed from the FDB (aged-out) if the device
does not transmit for a specified period of time (the aging time). This aging process prevents the
FDB from becoming full with obsolete entries by ensuring that when a device is removed from the
network, its entry is deleted from the database. The aging time is configurable. For more information
about setting the aging time, see “Configuring the FDB Aging Time” on page 360.

Differing FDB Table Sizes—BlackDiamond 8800 Series Switches, SummitStack, and the Summit Family of Switches Only
● Static entries—A static entry does not age and does not get updated through the learning process. A
static entry is maintained exactly as it was created. Conditions that cause dynamic entries to be
updated, such as VLAN or port configuration changes, do not affect static entries.
A locked static entry is an entry that was originally learned dynamically, but has been made static
(locked) using the MAC address lock-down feature. It is identified by the “s,” “p,” and “l” flags in
show fdb output and can be deleted using the delete fdbentry command. See “MAC Address
Lockdown” on page 528 for more information about MAC address lock-down.
If the FDB entry aging time is set to 0, all entries in the database are considered static, non-aging
entries. This means that the entries do not age, but they are still deleted if the switch is reset.
NOTE
On the BlackDiamond 8800 series switches, SummitStack, and the Summit family of switches, if the same MAC
address is detected on another virtual port that is not defined in the static FDB entry for the MAC address, that
address is handled as a blackhole entry.
● Permanent entries—Permanent entries are retained in the database if the switch is reset or a power
off/on cycle occurs. Permanent entries must be created by the system administrator through the CLI.
Permanent entries are static, meaning they do not age or get updated.
Differing FDB Table Sizes—BlackDiamond 8800 Series

Switches, SummitStack, and the Summit Family of
Switches Only
The modules included in the BlackDiamond 8800 series switch have FDB tables of two different sizes, as
follows:
● 8K FDB table
■ G48Te module
■ G48Pe module
■ Summit X250e switch
● 16K FDB table
■ G48Ta module
■ G48Xa module
■ 10G4Xa module
■ All original BlackDiamond 8800 modules (G48T, G48P, G24X, and 10G4X)
If you are running with mixture of modules from each group in your chassis and the entry cannot be
added to the G48Te or G48p module but can be added to the other modules, the system displays an
EMS warning-level message similar to the following:
HAL.FDB.Warning> MSM-A: FDB for vlanID1 mac 00:00:03:05:15:04 was not added ot slot 3
- table full.

Forwarding Database
The switches in the Summit family of switches have FDB tables of two different sizes, as follows:
● 8K FDB table
■ Summit X450e series switches
■ Summit X250e switch
● 16K FDB table
■ Summit X450 series switch
■ Summit X450a series switch
If you are running with a mixture of switches from each group in your SummitStack and the entry
cannot be added to the SummitX450e switch but can be added to the other switches, the system
displays an EMS warning-level message similar to the following:
HAL.FDB.Warning> Slot-1: FDB for vlanID1 mac 00:00:03:05:15:04 was not added ot slot 3
- table full.
FDB Configuration Examples

● Adding a Permanent Static Entry on page 360
● Configuring the FDB Aging Time on page 360
● Clearing FDB Entries on page 361
Adding a Permanent Static Entry

The following example adds a permanent static entry to the FDB:
create fdbentry 00:E0:2B:12:34:56 vlan marketing port 3:4
The permanent entry has the following characteristics:

● MAC address is 00:E0:2B:12:34:56.
● VLAN name is marketing.
● Slot number for this device is 3 (only on modular switches).
● Port number for this device is 4.
NOTE
On the BlackDiamond 8800 series switches, SummitStack, and the Summit family of switches, if the MAC address
00:E0:2B:12:34:56 is encountered on any port/VLAN other than VLAN marketing, port 3:4, that address will be
handled as a blackhole entry, and packets from that source will be dropped.
Configuring the FDB Aging Time

You configure the aging time for dynamic FDB entries using the following command:
configure fdb agingtime <seconds>

Displaying FDB Entries
If the aging time is set to 0, all aging entries in the database are defined as static, nonaging entries. This
means the entries will not age out, but non-permanent static entries can be deleted if the switch is reset.
Supported aging is between 15 and 1,000,000 seconds. The default is 5 minutes (300 seconds).
Clearing FDB Entries

You clear dynamic and permanent entries using different CLI commands.
You clear dynamic FDB entries by targeting:

● Specified MAC addresses
● Specified ports
● Specified VLANs,
● All blackhole entries
To clear dynamic entries from the FDB, use the following command:
clear fdb {<mac_addr> | ports <port_list> | vlan <vlan_name> | blackhole}
You clear permanent FDB entries by targeting:

● All permanent entries
● Specified MAC addresses
● Specified VLANs
● All blackhole entries
To clear permanent entries from the FDB, use the following command:
delete fdbentry [all | <mac_address> [vlan <vlan name>]
Displaying FDB Entries

To display FDB entries, use the following command:
show fdb {<mac_addr> {netlogin [all | mac-based-vlans]}| permanent {netlogin [all |
mac-based-vlans]} | ports <port_list> {netlogin [all | mac-based-vlans]}| vlan
<vlan_name> {netlogin [all | mac-based-vlans]} | stats | netlogin {all | mac-based-
vlans]}}

● mac_address—Displays the entry for a particular MAC address.
● netlogin—Displays all FDBs created as a result of the netlogin process. You can display all of these
entries or just the netlogin MAC-based VLAN FDB entries.
NOTE
The MAC-based VLAN netlogin parameter applies only for the Summit family of switches and the BlackDiamond
8800 series switch. See Chapter 15, “Network Login,” for more information on netlogin.
● permanent—Displays all permanent entries, including the ingress and egress QoS profiles.
● ports <portlist>—Displays the entries for a set of ports or slots and ports.

Forwarding Database
● vlan <vlan name>—Displays the entries for a VLAN.

● stats—Displays the number of static, permanent, dynamic, and dropped entries; as well as the
aging time.
With no options, the command displays all FDB entries. (The age parameter does not show on the
display for the backup MSM on modular switches; it does show on the display for the primary MSM.)
MAC-Based Security
MAC-based security allows you to control the way the FDB is learned and populated. By managing
entries in the FDB, you can block and control packet flows on a per-address basis.
MAC-based security allows you to limit the number of dynamically-learned MAC addresses allowed
per virtual port. You can also “lock” the FDB entries for a virtual port, so that the current entries will
not change, and no additional addresses can be learned on the port.
You can also prioritize or stop packet flows based on the source MAC address of the ingress VLAN or
the destination MAC address of the egress VLAN.
NOTE
For detailed information about MAC-based security, see Chapter 16, “Security.”

● Disabling MAC Address Learning on page 362
● Disabling Egress Flooding on page 363
● Displaying Learning and Flooding Settings on page 365
Disabling MAC Address Learning

By default, MAC address learning is enabled on all ports. To disable learning on specified ports, use the
following command:
disable learning port [<port_list> | all]
If MAC address learning is disabled, only broadcast traffic, EDP traffic, and packets destined to a
permanent MAC address matching that port number, are forwarded. Use this command in a secure
environment where access is granted via permanent FDBs per port. Disabling learning on a port causes
the MAC addresses to flood (unless you disable egress flooding) because those addresses will not be
present in the FDB during a destination lookup.
Disabling MAC Address Learning on the BlackDiamond 8800 Series Switches,

SummitStack, and the Summit Family of Switches Only.
When learning is disabled, packets with unknown source MAC addresses are dropped.

MAC-Based Security
Disabling Egress Flooding

With ExtremeXOS software version 11.2, you can enable or disable egress flooding. Under default
conditions, when the system does not find a match in the FDB for a unicast/multicast/broadcast MAC
address in a packet received in a given port, the system forwards that frame to every port in the VLAN
(known as Layer 2 flooding).
However, you can enhance security and privacy as well as improving network performance by
disabling Layer 2 egress flooding on some packets. This is particularly useful when you are working on
an edge device in the network. Limiting flooded egress packets to selected interfaces is also known as
upstream forwarding.
NOTE
Disabling egress flooding can affect many protocols, such as IP and ARP among others.
Figure 43 illustrates a case where you want to disable Layer 2 egress flooding on specified ports to
enhance security and network performance.
Figure 43: Upstream forwarding or disabling egress flooding example
ISP FW/
Security Proxy
Uplink
port 3
Access Link Access Link
port 1 EXOS Switch port 2
Access VLAN
Client 1 Client 2
XOS004A
In this example, the three ports are in an ISP-access VLAN. Ports 1 and 2 are connected to clients 1 and
2, respectively, and port 3 is an uplink to the ISP network. Because clients 1 and 2 are in the same
VLAN, client 1 could possible learn about the other client’s traffic by sniffing client 2’s broadcast traffic;
client 1 could then possibly launch an attack on client 2.
However, when you disable all egress flooding on ports 1 and 2, this sort of attack is impossible, for the
following reasons:
● Broadcast and multicast traffic from the clients is forwarded only to the uplink port.
● Any packet with unlearned destination MAC addresses is forwarded only to the uplink port.
● One client cannot learn any information from the other client. Because egress flooding is disabled on
the access ports, the only packets forwarded to each access port are those packets that are
specifically targeted for one of the ports. There is no traffic leakage.
In this way, the communication between client 1 and client 2 is controlled. If client 1 needs to
communicate with client 2 and has that IP address, client 1 sends out an ARP request to resolve the IP
address for client 2.

Forwarding Database
Guidelines for Enabling or Disabling Egress Flooding

The following guidelines apply to enabling and disabling egress flooding:
● Egress flooding can be disabled on ports that are in a load-sharing group. If that is the situation, the
ports in the group take on the egress flooding state of the master port; each member port of the load-
sharing group has the same state as the master port.
● FDB learning is independent of egress flooding; either can be enabled or disabled independently.
● Disabling unicast (or all) egress flooding to a port also stops packets with unknown MAC addresses
to be flooded to that port.
● Disabling broadcast (or all) egress flooding to a port also stops broadcast packets to be flooded to
that port.
Disabling Egress Flooding on the BlackDiamond 8800 Series Switches, SummitStack,

and the Summit Family of Switches Only
You can enable or disable egress flooding for unicast, multicast, or broadcast MAC addresses, as well as
for all packets on the ports of the BlackDiamond 8800 series switches, SummitStack, and the Summit
family of switches.
Disabling multicasting egress flooding does not affect those packets within an IGMP membership group
at all; those packets are still forwarded out. If IGMP snooping is disabled, multicast packets with static
FDB entries are forwarded according to the FDB entry.
To enable egress flooding for these switches, use the following command:
enable flooding [all_cast | broadcast | multicast | unicast] port [<port_list> | all]
To disable flooding for these switches, use the following command:
disable flooding [all_cast | broadcast | multicast | unicast] port [<port_list> | all]
Disabling Egress Flooding on the BlackDiamond 10808 and

You must enable or disable egress flooding on all packets on the specified port or ports. You cannot
specify broadcast, unicast, or multicast packets; the egress flooding command applies to all packets.
Disabling multicasting egress flooding does not affect those packets within an IGMP membership group
at all; those packets are still forwarded out. If IGMP snooping is disabled, multicast packets are not
flooded.
To enable egress flooding on the BlackDiamond 10808 0r 12800 series switches, use the following
command:
enable flooding all_cast port [<port_list> | all]
To disable egress flooding on the BlackDiamond 10808 or 12800 series switches, use this command:
disable flooding all_cast port [<port_list> | all]

Multicast FDB with Multiport Entry—BlackDiamond 8800 Series Switches, SummitStack, and the Summit Family of Switches
NOTE
When you disable egress flooding on the BlackDiamond 10808 or 12800 series switches, you also turn off
broadcasting.
Displaying Learning and Flooding Settings

To display the status of MAC learning and egress flooding, use the following command:
Multicast FDB with Multiport Entry—BlackDiamond

8800 Series Switches, SummitStack, and the Summit
Family of Switches Only
Beginning with ExtremeXOS version 11.3 on the BlackDiamond 8800 series switches, SummitStack, and
the Summit family of switches, you can create FDB entries to multicast MAC addresses (that is,
01:00:00:00:00:01) and list one or more ports. Use the create fdbentry vlan ports command to enter
the multicast FDB address. After traffic with a multicast MAC destination address enters the switch,
that traffic is multicast to all ports on the list.
However, if the MAC address is in the IP multicast range (for example, 01:00:5e:XX:XX:XX), IGMP
snooping rules take precedence over the multicast static FDB entry. Of course, if you disable IGMP
snooping on all VLANs, the static FDB entry forwards traffic.

Forwarding Database

16 Virtual Routers

● Virtual Routers Overview on page 367
● Using Virtual Routers—BlackDiamond 10808 and BlackDiamond 12000 series Switches Only on
page 370
■ Creating Virtual Routers on page 370
■ Adding Ports to a Single Virtual Router on page 370
■ Adding Ports to Multiple Virtual Routers on page 371
■ Adding Routing Protocols to a Virtual Router on page 371
● Virtual Router Configuration Example on page 373
Virtual Routers Overview

ExtremeXOS supports virtual routers. This capability allows a single physical switch to be split into
multiple virtual routers. This feature separates the traffic forwarded by a virtual router from the traffic
on a different virtual router.
Each virtual router maintains a separate logical forwarding table, which allows the virtual routers to
have overlapping IP addressing. Because each virtual router maintains its own separate routing
information, packets arriving on one virtual router will never be switched to another.
However, virtual routers should not be connected together through a Layer 2 domain. Since there is a
single MAC address per switch in ExtremeXOS, this same MAC address is used for all virtual routers. If
two virtual routers on same switch are connected through a Layer 2 domain, the intermediate Layer 2
switches will learn same MAC address of the switch on different ports, and may send traffic into the
wrong virtual router.
Ports on the switch can either be used exclusively by one virtual router, or can be shared among two or
more virtual routers. One reason to configure a port for the exclusive use of a single virtual router is to
be sure that only packets from that virtual router egress from that port. One reason to configure a port
to be shared by multiple virtual routers is to pass traffic from multiple virtual routers across a shared
link.
With multiple virtual routers contained on a single physical switch, some commands in ExtremeXOS
now require you to specify to which virtual router the command applies. For example, when you use
the ping command, you must specify from which virtual router the ping packets are generated. Many
commands that deal with switch management use the management virtual router by default. See the
ExtremeXOS Command Reference Guide for information on the defaults for individual commands.
NOTE
The term virtual router is also used with the Virtual Router Redundancy Protocol (VRRP). VRRP uses the term to
refer to a single virtual router that spans more than one physical router, which allows multiple switches to provide
redundant routing services to users. For more information about VRRP, see Chapter 20, “VRRP Commands.”

Virtual Routers
Types of Virtual Routers

There are two types of virtual routers in an ExtremeXOS system:
● System virtual routers
These are the special virtual routers created by ExtremeXOS during system boot up, and they cannot
be deleted or renamed. There are three system virtual routers in the ExtremeXOS system.
● User virtual routers
These are the virtual routers created and named by users.
NOTE
User virtual routers are supported only on the BlackDiamond 10808 and BlackDiamond 12000 series switches.
System Virtual Routers

The system virtual routers are the three virtual routers created at boot-up time. These system virtual
routers cannot be deleted or renamed. They are named VR-Mgmt, VR-Control, and VR-Default. The
following describes each system virtual router:
● VR-Mgmt
VR-Mgmt enables remote management stations to access the switch through Telnet, SSH, and SNMP
sessions; and it owns the management port. No other ports can be added to this VR-Mgmt, and the
management port cannot be removed from it.
The Mgmt VLAN is created in the VR-Mgmt during the ExtremeXOS system boot-up. No other
VLAN can be created in this virtual router, and the Mgmt VLAN cannot be deleted from it.
No routing protocol is running or can be added to this virtual router.
This virtual router is called VR-0 in ExtremeXOS releases before 11.0.
● VR-Control
VR-Control is used for internal communications between all the modules and subsystems in the
switch. It has no external visible ports, and you cannot assign any port to it.
This virtual router, VR-Control, has no VLAN interface, and no VLAN can be created for it.
No routing protocol is running or can be added to this virtual router.
● VR-Default
VR-Default is the default virtual router created by the ExtremeXOS system. All data ports in the
switch are assigned to this virtual router by default. Any data port can be added to and deleted from
this virtual router.
Users can create and delete VLANs in this virtual router. The Default VLAN is created in this virtual
router during the ExtremeXOS system boot-up. The Default VLAN cannot be deleted from this
virtual router.
One instance of each routing protocol is spawned for this virtual router during the ExtremeXOS
system boot-up, and these routing instances cannot be deleted.

Virtual Routers Overview
User Virtual Routers—BlackDiamond 10808 and BlackDiamond 12000 series

Switches Only
User virtual routers are the virtual routers created by users in addition to the system virtual routers.
The ability to create user virtual routers was first introduced in ExtremeXOS 11.0.
When a new user virtual router is created, by default, no ports are assigned, no VLAN interface is
created, and no support for any routing protocols is added.
Virtual Router Configuration Domain—BlackDiamond 10808 and

BlackDiamond 12000 series Switches Only
When you create virtual routers, you must configure each virtual router separately, configuring routing
protocols and VLANs for each one. To simplify the configuration process, the concept of a virtual router
configuration domain was introduced in ExtremeXOS 11.0. Under a virtual router configuration
domain, any virtual router commands are applied only to that virtual router. The virtual router
commands consist of all the BGP, OSPF, PIM and RIP commands, and the commands listed in Table 50.
Table 50: Virtual router commands

[enable | disable] ipforwarding
clear iparp *
clear counters iparp *
configure iparp *
configure iparp [add | delete] *
[enable | disable] iparp *
show iparp *
configure iproute [add | delete] *
show iproute *
show ipstats *
rtlookup
create [vlan | vman] <vlan-name>
[enable | disable] igmp
[enable | disable] igmp snooping *
[enable | disable] ipmcforwarding
show igmp
show igmp snooping
show igmp group
show igmp snooping cache
* means that other commands are available with these listed.
The virtual router configuration domain simplifies configuration because you do not have to specify the
virtual router for each individual protocol configuration command. The current configuration domain is
indicated in the command line interface (CLI) prompt by the name of the User virtual router, or no
name if in the VR-Default domain.

Virtual Routers
Using Virtual Routers—BlackDiamond 10808 and

BlackDiamond 12000 series Switches Only
To use the user virtual router functionality in ExtremeXOS, you will need to do the following things:
● Create the virtual router.
● Configure ports to a single virtual router, or to multiple virtual routers.
● Add any required routing protocols to the virtual router.
● Configure the routing protocols and VLANs.
The following sections describe how to do these tasks.
Creating Virtual Routers

To create a user virtual router, use the following command:
create virtual-router
A virtual router name cannot be the same as a VLAN name. You cannot name a user virtual router with
the names VR-Mgmt, VR-Control, or VR-Default because these are the existing default system virtual
routers. For backward compatibility, user virtual routers also cannot be named VR-0, VR-1 or VR-2,
because these three names are the names for the system virtual routers in ExtremeXOS releases before
11.0.
To delete a user virtual router, use the following command:
delete virtual-router
Before you delete a virtual router, you must delete all VLANs created in that virtual router. All of the
ports assigned to this virtual router will be deleted and made available to assign to other virtual
routers. Any routing protocol that is running on the virtual router will be shut down and deleted
gracefully.
Configuring Ports to a Single or to Multiple Virtual Router(s)

By default, all the user data ports belong to the system default virtual router, VR-Default, and belong to
the default VLAN, Default. All these ports are used exclusively by the virtual router, VR-Default. To
configure a port for exclusive use by another virtual router, or for use by multiple virtual routers, it
must first be deleted from VR-Default. You must delete the port from any VLAN it belongs to before
deleting it from a virtual router.
To delete a port from a virtual router, use the following command:
configure vr <vr-name> delete ports <portlist>
Adding Ports to a Single Virtual Router

When you add a port to a virtual router, that port can only be used by that virtual router.

Using Virtual Routers—BlackDiamond 10808 and BlackDiamond 12000 series Switches Only
To add a port to a single virtual router, use the following command:
configure vr add ports
The following is an example of removing all the ports on slot 3 from the default VLAN in the default
virtual router and adding them for the exclusive use of the virtual router helix:
configure vlan default delete ports 3:*

configure vr vr-default delete ports 3:*
configure vr helix add ports 3:*
Adding Ports to Multiple Virtual Routers

A port is available to be configured for multiple virtual routers after it has been deleted from the
default virtual router, VR-Default. To use a port in multiple virtual routers, do not add the port to a
virtual router, as you would in the previous section, “Adding Ports to a Single Virtual Router”, just add
the port to a VLAN in the desired virtual router.
NOTE
See Chapter 14, “Quality of Service,” for details about how multiple virtual routers per port can affect DiffServ and
code replacement.
In the following example we add port 3:5 to the virtual routers VR-green and VR-blue. We’ve already
created the tagged VLAN bldg_200 in VR-green, and the tagged VLAN bldg_300 in VR-blue.
configure vlan default delete ports 3:5

configure vr vr-default delete ports 3:5
configure vlan bldg_200 add ports 3:5 tagged
configure vlan bldg_300 add ports 3:5 tagged
Adding Routing Protocols to a Virtual Router

Unlike the default system virtual router, VR-Default, there are no resources allocated for routing
protocols when a user virtual router is created. You must add the routing protocols needed for your
virtual router before you attempt to configure them. When you add a protocol to a user virtual router, a
process is started to support the protocol.
Adding a protocol to a virtual router does not enable that protocol. You must then specifically enable
and configure any protocol that you add.
To add a protocol to a virtual router, use the following command:
configure vr add protocol
To remove a protocol from a virtual router, use the following command:
configure vr delete protocol

Virtual Routers
Displaying Ports and Protocols

You display the ports, protocols, and the name of the protocol processes for a virtual router by using
show virtual-router
Configuring the Routing Protocols and VLANs

After the virtual router is created, the ports are added, and support for any needed routing protocols is
added, you can configure the virtual router. To simplify configuring the user virtual routers, the
concept of a virtual router configuration domain was added (instead of adding a virtual router keyword
to every command in every routing protocol). Virtual router commands are applied to the current
configuration domain. The virtual router commands consist of all the BGP, OSPF, PIM and RIP
commands, as well as the create vlan and delete vlan commands. Other commands apply to the
switch as a whole.
To enter a virtual router configuration domain, use the following command:
virtual-router
For example, to enter the configuration domain for the virtual router helix, your CLI session would look
similar to this:
* BD10K.13 # virtual-router helix

* (vr helix) BD10K.14 #
The CLI prompt displays the virtual router configuration domain.
Use the virtual-router command with no virtual router name, or use the name VR-Default to return
to the default configuration domain.
Now you can create VLANs, using the following command:
create vlan <vlan_name> {vr <vr-name>}
If you do not specify a virtual router in the create vlan command, the VLAN is created in the virtual
router of the current configuration domain.
NOTE
All VLAN names and VLAN IDs on a switch must be unique, regardless of the virtual router they are created in. You
cannot have two VLANs with the same name, even if they are in different virtual routers.
To display the VLANs in a specific virtual router, use the following command:
show vlan virtual router <vr-name>
which is a specific form of this command:
You can also configure routing protocols, by using the standard ExtremeXOS commands. The routing
configurations of the different virtual routers are independent of each other.

Virtual Router Configuration Example
Virtual Router Configuration Example

In the following example:
● The user virtual router helix is created.
● Ports are removed from the VLAN Default and the virtual router VR-Default.
● Ports are added to the virtual router helix.
● OSPF is added to the virtual router helix.
● The configuration domain is set to helix, so that subsequent virtual router commands affect the
virtual router helix.
● The VLAN helix-accounting is created.
● Ports that belong to the virtual router helix are added to the VLAN helix-accounting.
The CLI prompt is shown in this example to show how the virtual router configuration domain is
displayed. At the end of the example, the virtual router is ready to be configured for OSPF, using
ExtremeXOS commands.
* BD10K.1 # create virtual-router helix

* BD10K.2 # configure vlan default delete ports 3:*
* BD10K.3 # configure vr vr-default delete ports 3:*
* BD10K.4 # configure vr helix add ports 3:*
* BD10K.5 # configure vr helix add protocol ospf
* BD10K.6 # virtual-router helix
* (vr helix) BD10K.7 # create vlan helix-accounting
* (vr helix) BD10K.8 # configure helix-accounting add ports 3:1
* (vr helix) BD10K.9 #

Virtual Routers

17 Policy Manager

● Policy Manager on page 375
● Creating and Editing Policies on page 375
● Checking Policies on page 376
● Refreshing Policies on page 377
● Applying Policies on page 377
Policy Manager
One of the processes that make up the ExtremeXOS system is the policy manager. The policy manager
is responsible for maintaining a set of policy statements in a policy database and communicating these
policy statements to the applications that request them.
Policies are used by the routing protocol applications to control the advertisement, reception, and use of
routing information by the switch. Using policies, a set of routes can be selectively permitted (or
denied) based on their attributes, for advertisements in the routing domain. The routing protocol
application can also modify the attributes of the routing information, based on the policy statements.
Policies are also used by the access control list (ACL) application to perform packet filtering and
forwarding decisions on packets. The ACL application will program these policies into the packet
filtering hardware on the switch. Packets can be dropped, forwarded, moved to a different QoS profile,
or counted, based on the policy statements provided by the policy manager.
Creating and Editing Policies

A policy is created by writing a text file that contains a series of rule entries describing match
conditions and actions to take. Before release 11.0, all policies were created by writing a text file on a
separate machine and then downloading it to the switch. Once on the switch, the file was then loaded
into a policy database to be used by applications on the switch. With release 11.0, policy text files can
also be created and edited directly on the switch.
NOTE
Although ExtremeXOS does not prohibit mixing ACL and routing type entries in a policy file, it is strongly
recommended that you do not mix the entries, and you use separate policy files for ACL and routing policies.
When you create a policy file, name the file with the policy name that you will use when applying the
policy, and use “.pol” as the filename extension. For example, the policy name “boundary” refers to the
text file “boundary.pol”.

Policy Manager
Using the Edit Command

A VI-like editor is available on the switch to edit policies. To edit a policy file on the switch by
launching the editor, use the following command:
edit policy
There are many commands available with the editor. For information about the editor commands, use
any tutorial or documentation about VI. The following is only a short introduction to the editor.
Edit operates in one of two modes; command and input. When a file first opens, you are in the
command mode. To write in the file, use the keyboard arrow keys to position your cursor within the
file, then press one of the following keys to enter input mode:
● i - To insert text ahead of the initial cursor position
● a- To append text after the initial cursor position
To escape the input mode and return to the command mode, press the Escape key.
Several commands can be used from the command mode. The following commands are the most
commonly used:
● dd - To delete the current line
● yy - To copy the current line
● p - To paste the line copied
● :w - To write (save) the file
● :q - To quit the file if no changes were made
● :q! - To forcefully quit the file without saving changes
● :wq - To write and quit the file
Using a Separate Machine

You can also edit policies on a separate machine. Any common text editor can be used to create a policy
file. The file is then transferred to the switch using TFTP and then applied.
To transfer policy files to the switch, use the following command:

tftp [<host-name> | <ip-address>] {-v <vr_name>} [-g | -p] [{-l [internal-memory
<local-file-internal> | memorycard <local-file-memcard> | <local-file>} {-r <remote-
file>} | {-r <remote-file>} {-l [internal-memory <local-file-internal> | memorycard
<local-file-memcard> | <local-file>]}]
Checking Policies
A policy file can be checked to see if it is syntactically correct. To check the policy syntax, use the
following command:
check policy
This command can only determine if the syntax of the policy file is correct and can be loaded into the
policy manager database. Since a policy can be used by multiple applications, a particular application
may have additional constraints on allowable policies.

Applying Policies
Refreshing Policies
When a policy file is changed (such as adding, deleting an entry, adding/deleting/modifying a
statement), the information in the policy database does not change until the policy is refreshed. The
user must refresh the policy so that the latest copy of policy is used.
When the policy is refreshed, the new policy file is read, processed, and stored in the server database.
Any clients that use the policy are updated. To refresh the policy, use the following command:
refresh policy
For ACL policies only, during the time that an ACL policy is refreshed, packets on the interface are
blackholed, by default. This is to protect the switch during the short time that the policy is being
applied to the hardware. It is conceivable that an unwanted packet could be forwarded by the switch as
the new ACL is being setup in the hardware. You can disable this behavior. To control the behavior of
the switch during an ACL refresh, use the following commands:
enable access-list refresh blackhole

disable access-list refresh blackhole
In releases previous to ExtremeXOS 11.4, when ACLs were refreshed, all the ACL entries were
removed, and new ACL entries were created to implement the newly applied policy. Beginning in
release 11.4, the policy manager uses Smart Refresh to update the ACLs. When a change is detected,
only the ACL changes needed to modify the ACLs are sent to the hardware, and the unchanged entries
remain. This behavior avoids having to blackhole packets because the ACLs have been momentarily
cleared. Smart Refresh works well up for up to 200 changes. If the number of changes exceeds 200, you
will see this message: Policy file has more than 200 new rules. Smart refresh can not be
carried out. Following this message, you will see a prompt based on the current blackhole
configuration. If blackhole is disabled you will see the following prompt:
Note, the current setting for Access-list Refresh Blackhole is Disabled. WARNING: If
a full refresh is performed, it is possible packets that should be denied may be
forwarded through the switch during the time the access list is being installed.
Would you like to perform a full refresh?
If blackhole is enabled, you will see the following prompt:
Note, the current setting for Access-list Refresh Blackhole is Enabled.
Would you like to perform a full refresh?
To take advantage of Smart Refresh, disable access-list refresh blackholing.
Applying Policies
ACL policies and routing policies are applied using different commands.

Policy Manager
Applying ACL Policies

A policy intended to be used as an ACL is applied to an interface, and the CLI command option is
named <aclname>. Supply the policy name in place of the <aclname> option. To apply an ACL policy,
configure access-list <aclname> [any | ports <portlist> | vlan <vlanname>] {ingress |

egress}
If you use the any keyword, the ACL is applied to all the interfaces and is referred to as the wildcard
ACL. This ACL is evaluated for any ports without specific ACLs, and it is also applied to any packets
that do not match the specific ACLs applied to the interfaces.
If an ACL is already configured on an interface, the command will be rejected and an error message
displayed.
To remove an ACL from an interface, use the following command:
unconfigure access-list {any | ports <portlist> | vlan <vlanname>} {ingress | egress}
To display which interfaces have ACLs configured, and which ACL is on which interface, use the
following command:
show access-list {any | ports <portlist> | vlan <vlanname>} {ingress | egress}
Applying Routing Policies

To apply a routing policy, use the command appropriate to the client. Different protocols support
different ways to apply policies, but there are some generalities. Policies applied with commands that
use the keyword import-policy control the routes installed into the switch routing table by the
protocol. The following are examples for the BGP and RIP protocols:
configure bgp import-policy [<policy-name> | none]

configure rip import-policy [<policy-name> | none]
Commands that use the keyword route-policy control the routes advertised or received by the
protocol. For BGP and RIP, here are some examples:
configure bgp neighbor [<remoteaddr> | all] {address-family [ipv4-unicast | ipv4-

multicast]} route-policy [in | out] [none | <policy>]
configure bgp peer-group <peer-group-name> route-policy [in | out] [none | <policy>]
configure rip vlan [<vlan-name> | all] route-policy [in | out] [<policy-name> | none]
Other examples of commands that use route policies include:
configure ospf area <area-identifier> external-filter [<policy-map> |none]

configure ospf area <area-identifier> interarea-filter [<policy-map> | none]
configure rip vlan [<vlan-name> | all] trusted-gateway [<policy-name> | none]
To remove a routing policy, use the none option in the command.

18 Access Lists (ACLs)

● ACLs on page 380
● ACL Rule Syntax on page 381
■ Matching All Egress Packets on page 382
■ Comments and Descriptions in ACL Policy Files on page 382
■ Types of Rule Entries on page 383
■ Match Conditions on page 384
■ Actions on page 384
■ Action Modifiers on page 384
■ ACL Rule Syntax Details on page 386
■ IPv6 ACL Address Masks—BlackDiamond 10808 and BlackDiamond 12800 Series Switches Only
on page 391
● vMAN ACLs—BlackDiamond 10808 and BlackDiamond 12800 Series Switches Only on page 392
■ vMAN ACL Actions on page 392
■ vMAN ACL Action Modifiers on page 393
■ vMAN ACL Examples—BlackDiamond 10808 and BlackDiamond 12800 Series Switches Only on
page 394
● Layer-2 Protocol Tunneling ACLs—BlackDiamond 8800 a-Series modules and e-Series and Summit
X250e, X450a, and X450e Switches Only on page 394
● Dynamic ACLs on page 395
■ Creating the Dynamic ACL Rule on page 395
■ Configuring the ACL Rule on the Interface on page 396
● Configuring ACL Priority on page 397
● ACL Evaluation Precedence—BlackDiamond 10808 and BlackDiamond 12800 Series Switches Only
on page 400
■ Rule Evaluation on page 400
■ Precedence of Dynamic ACLs on page 400
■ Precedence of L2/L3/L4 ACL Entries on BlackDiamond10k and 12K Switches on page 400
■ Precedence Among Interface Types on page 401
● ACL Evaluation Precedence—BlackDiamond 8800 Series switches, SummitStack, and the Summit
Family Switches Only on page 401
■ Rule Evaluation on page 401
■ Precedence of Dynamic ACLs on page 402
■ Precedence of L2/L3/L4 ACL Entries on page 402
■ Precedence Among Interface Types on page 402
■ Redundant Rules on page 402

Access Lists (ACLs)
● Applying ACL Policy Files on page 403

■ Displaying and Clearing ACL Counters on page 403
■ Example ACL Rule Entries on page 403
● ACL Mechanisms on page 405
■ ACL Masks and Rules—BlackDiamond 8800 Series Switches and the Summit X450 Series
■ ACL Slices and Rules—BlackDiamond 8800 a-series and e-series Modules and Summit X450a,
X450e, and X250e Series Switches Only on page 411
● Policy Based Routing on page 421
■ Layer 3 Policy Based Redirect on page 421
■ Layer 2 Policy Based Redirect on page 422
■ Configuring Policy Based Routing on page 424
● ACL Troubleshooting—BlackDiamond 8800, SummitStack, and Summit Family of Switches on
page 425
ACLs
ACLs are used to perform packet filtering and forwarding decisions on traffic traversing the switch.
Each packet arriving on an ingress port and/or VLAN is compared to the access list applied to that
interface and is either permitted or denied. On the BlackDiamond 10808 and the BlackDiamond 12800
series switches, packets egressing an interface can also be filtered. However, only a subset of the
filtering conditions available for ingress filtering are available for egress filtering.
In addition to forwarding or dropping packets that match an ACL, the switch can also perform
additional operations like incrementing counters, logging packet headers, mirroring traffic to a monitor
port, sending the packet to a QoS profile, and, for the BlackDiamond 8800 series switches, SummitStack,
and the Summit series switches only, meter the packets matching the ACL to control bandwidth. Using
ACLs has no impact on switch performance (with the minor exception of the mirror-cpu action
modifier).
ACLs are typically applied to traffic that crosses Layer 3 router boundaries, but it is possible to use
access lists within a Layer 2 virtual LAN (VLAN).
ACLs in ExtremeXOS apply to all traffic. This is somewhat different from the behavior in ExtremeWare.
For example, if you deny all the traffic to a port, no traffic, including control packets, such as OSPF or
RIP, will reach the switch and the adjacency will be dropped. You must explicitly allow those types of
packets (if desired). In ExtremeWare, an ACL that denied “all” traffic would allow control packets
(those bound for the CPU) to reach the switch.
ACLs are created in two different ways. One method is to create an ACL policy file and apply that ACL
policy file to a list of ports, a VLAN, or to all interfaces.
NOTE
ACLs applied to a VLAN are actually applied to all ports on the switch, without regard to VLAN membership. The
result is that resources are consumed per port on BlackDiamond and Summit switches with original modules, and
per chip on the BlackDiamond and Summit a- and e- series switches. If a single platform contains both types of
modules, the limitations of the original modules will determine the policy support available, without respect to VLAN
membership.

ACL Rule Syntax
An ACL policy file is a text file that contains one or more ACL rule entries. This first method creates
ACLs that are persistent across switch reboots, can contain a large number of rule entries, and are all
applied at the same time. See “ACL Rule Syntax” on page 381 for information about creating ACL rule
entries. For information about creating policy files, see Chapter 11, “Policy Manager.”
Policy files are also used to define routing policies. Routing policies are used to control the
advertisement or recognition of routes communicated by routing protocols. ACL policy files and
routing policy files are both handled by the policy manager, and the syntax for both types of files is
checked by the policy manager.
NOTE
The second method to create an ACL is to use the CLI to specify a single rule, called a dynamic ACL.
By default, dynamic ACLs persist across reboots; however, you can configure non-persistent dynamic
ACLS that disappear when the switch reboots. Dynamic ACLs consist of only a single rule. Multiple
dynamic ACLs can be applied to an interface. See “Layer-2 Protocol Tunneling ACLs—BlackDiamond
8800 a-Series modules and e-Series and Summit X250e, X450a, and X450e Switches Only” on page 394
for information about creating dynamic ACLs. The precedence of ACLs can be configured by defining
zones and configuring the priority of both the zones and the ACLs within a zone. See “Configuring
ACL Priority” on page 397 for more information.
ACL Rule Syntax

An ACL rule entry consists of:
● A rule entry name, unique within the same ACL policy file or among Dynamic ACLs.
● Zero or more match conditions.
● Zero or one action (permit or deny). If no action is specified, the packet is permitted by default.
● Zero or more action modifiers.
Each rule entry uses the following syntax:
entry <ACLrulename>{
if {
<match-conditions>;
} then {
<action>;
<action-modifiers>;
}
}
The following is an example of a rule entry:
entry udpacl {
if {
source-address 10.203.134.0/24;
destination-address 140.158.18.16/32;
protocol udp;
source-port 190;

Access Lists (ACLs)
destination-port 1200 - 1250;

} then {
permit;
}
}
An ACL rule is evaluated as follows:

● If the packet matches all the match conditions, the action and any action modifiers in the then
statement are taken.
● For ingress ACLs, if a rule entry does not contain any match condition, the packet is considered to
match and the action and any action modifiers in the rule entry’s then statement are taken. For
egress ACLs, if a rule entry does not contain any match condition, no packets will match. See
“Matching All Egress Packets” for more information.
● If the packet matches all the match conditions, and if there is no action specified in the then
statement, the action permit is taken by default.
● If the packet does not match all the match conditions, the action in the then statement is ignored.
Matching All Egress Packets

Unlike ingress ACLs, for egress ACLs, you must specify either a source or destination address, instead
of writing a rule with no match conditions.
For example, an ingress ACL deny all rule could be:
entry DenyAllIngress{
if {
} then {
deny;
}
}
The previous rule would not work as an egress ACL. The following is an example of an egress ACL
deny all rule:
entry DenyAllEgress{
if {
} then {
deny;
}
}
Comments and Descriptions in ACL Policy Files

In ACL policy files, there are two types of textual additions that have no effect on the ACL actions:
comments and descriptions. A comment is ignored by the policy manager and resides only in the policy
file. Comments are not saved in the switch configuration and are not displayed by the "show policy"
command. A description is saved in the policy manager and is displayed when the ACL is displayed.

ACL Rule Syntax
You can display the ACL using the following two commands:
show policy {<policy-name> | detail}

For example, the following policy, saved in the file denyping.pol, contains both a comment and a
description:
# this line is a comment

@description "This line is a description for the denyping.pol"
entry ping_deny_echo-request {
if {
protocol icmp;
icmp-type echo-request;
} then {
deny;
count pingcount_deny;
}
}
Note that the description begins with the tag @description and is a text string enclosed in quotes.
You can apply the policy to port 1, using the following command:
configure access-list denyping port 1
and display the policy using the following command:
show policy denyping
The output of this command is similar to the following:
Policies at Policy Server:

Policy: denyping
@description This line is a description for the denyping.pol
entry ping_deny_echo-request {
if match all {
protocol icmp ;
icmp-type echo-request ;
}
then {
deny ;
count pingcount_deny ;
}
}
Number of clients bound to policy: 1
Client: acl bound once
Types of Rule Entries

In ExtremeXOS, each rule can be one of following types:
● L2 rule—A rule containing only Layer 2 (L2) matching conditions, such as Ethernet MAC address
and Ethernet type
● L3 rule—A rule containing only Layer 3 (L3) matching conditions, such as source or destination IP
address and protocol

Access Lists (ACLs)
● L4 rule—A rule containing both Layer 3 (L3) and Layer 4 (L4) matching conditions, such as TCP/
UDP port number
NOTE
BlackDiamond 10808 and BlackDiamond 12800 series switches only—L2 matching conditions cannot be mixed
with L3/L4 matching conditions in a rule, otherwise, syntax checking will fail.
Match Conditions
You can specify multiple, single, or zero match conditions. If no match condition is specified, all packets
match the rule entry. Among the match conditions commonly used are:
● ethernet-source-address <mac-address>—Ethernet source address
● ethernet-destination-address <mac-address> <mask>—Ethernet destination address and mask
● source-address <prefix>—IP source address and mask
● destination-address <prefix>—IP destination address and mask
● source-port [<port> | <range]—TCP or UDP source port range
● destination-port [<port> | <range>]—TCP or UDP destination port range
Table 51 describes all the possible match conditions.
Actions
The actions are:
● permit—The packet is forwarded.
● deny—The packet is dropped.
The default action is permit, so if no action is specified in a rule entry, the packet is forwarded.
Action Modifiers
Additional actions can also be specified, independent of whether the packet is dropped or forwarded.
These additional actions are called action modifiers. Not all action modifiers are available on all
switches, and not all are available for both ingress and egress ACLs. The action modifiers are:
● count <countername>—Increments the counter named in the action modifier (ingress only)
● log—Logs the packet header
● log-raw—Logs the packet header in hex format
● meter <metername>—Takes action depending on the traffic rate (BlackDiamond 8800 series
switches, SummitStack, and the Summit series switches only)
● mirror—Sends a copy of the packet to the monitor (mirror) port (ingress only)
● mirror-cpu—Mirrors a copy of the packet to the CPU in order to log it
● qosprofile <qosprofilename>—Forwards the packet to the specified QoS profile (ingress only)
● traffic-queue <traffic-queue>—Places the traffic on the specified traffic-queue (BlackDiamond
12800 series only)

ACL Rule Syntax
● redirect <ipv4 addr>—Forwards the packet to the specified IPv4 address (BlackDiamond 10808,
BlackDiamond 12800 series switches, BlackDiamond 8800 a-Series and e-Series modules, and Summit
X450a, X450e, and X250e series switches only)
● replace-dscp—Replace the packet’s DSCP field with the value from the associated QoS profile
● replace-dot1p—Replace the packet’s 802.1p field with the value from the associated QoS profile
● replace-ethernet-destination-address <mac-address>—Replace the packet's destination MAC
address; this is only applicable to layer-2 forwarded traffic (BlackDiamond 8800 a- and e- series
switches, SummitStack, and the Summit a- and e- series switches only)
● redirect-port <port>—Override the forwarding decision and change the egress port used.
Counting Packets
When the ACL entry match conditions are met, the specified counter is incremented. The counter value
can be displayed by the command:
show access-list counter {<countername>} {any | ports <portlist> | vlan <vlanname>}
{ingress | egress}
Logging Packets
Packets are logged only when they go to the CPU, so packets in the fastpath are not automatically
logged. You must use both the mirror-cpu action modifier and the log or log-raw action modifier if
you want to log both slowpath and fastpath packets that match the ACL rule entry. Additionally,
KERN:INFO messages are not logged by default. You must configure the EMS target to log these
messages. See Chapter 5, “Status Monitoring and Statistics,” for information about configuring EMS.
Metering Packets—BlackDiamond 8800 Series Switches, SummitStack, Summit

Family, and BlackDiamond 12800 Series Switches Only
The meter <metername> action modifier associates a rule entry with an ACL meter. See the section,
“Metering Using ACLs—BlackDiamond 8800 Series, SummitStack, the Summit Family of Switches, and
BlackDiamond 12800 R-Series Switch Only” on page 457 for more information.
Mirroring Packets
You must enable port-mirroring on your switch. See the section, “Mirroring” on page 151. If you
attempt to apply a policy that requires port-mirroring, you will receive an error message if port-
mirroring is not enabled.
Redirecting Packets—BlackDiamond 10808, BlackDiamond 12800 series switches,

BlackDiamond 8800 a-series and e-series modules, and Summit X450a, X450e, and
X250e series switches Only
Packets are forwarded to the IPv4 address specified, without modifying the IP header (except the TTL is
decremented and the IP checksum is updated). The IPv4 address must be in the IP ARP cache,
otherwise the packet is forwarded normally. Only fast path traffic can be redirected. This capability can
be used to implement Policy Based Routing.
You may want to create a static ARP entry for the redirection IP address, so that there will always be a
cache entry.

Access Lists (ACLs)
See “Policy Based Routing” on page 421 for more information.
Replacing DSCP or 802.1p Fields

Specify a QoS profile for matching packets. The field values are replaced with the value associated with
that profile. In the following example, DiffServ replacement is configured such that QP8 is mapped to
code point 56. Matching packets are sent to QP8, and the DSCP value in the packet is set to 56:
entry voice_entry {
if {
} then {
qosprofile qp8;
replace-dscp;
}
}
See Chapter 14, “Quality of Service,” for more details about QoS profiles, and 802.1p and DSCP
replacement.
ACL Rule Syntax Details

Table 51 lists the match conditions that can be used with ACLs, and whether the condition can be used
for ingress ACLs only, or with both ingress and egress. The conditions are case-insensitive; for example,
the match condition listed in the table as TCP-flags can also be written as tcp-flags. Within Table 51
are five different data types used in matching packets. Table 52 lists the data types and details on using
them.
Table 51: ACL match conditions

Applicable
IP Protocols/
Match Conditions Description Direction
ethernet-type <number> Ethernet packet type. In place of the numeric value, you can Ethernet/
specify one of the following text synonyms (the field values are Ingress only
also listed): ETHER-P-IP (0x0800), ETHER-P-8021Q (0x8100),
ETHER-P-IPV6 (0x86DD)a.
ethernet-source-address Ethernet source MAC address Ethernet/
<mac-address> Ingress only
ethernet-source-address Ethernet source MAC address and mask. The mask is optional, Ethernet/
<mac-address> mask and is in the same format as the MAC address, for example: Ingress only
<mask>
ethernet-source-address 00:01:02:03:01:01 mask
or ff:ff:ff:ff:00:00
ethernet-source-address or
<mac-address> / <mask>
ethernet-source-address 00:01:02:03:01:01 /
ff:ff:ff:ff:00:00
Only those bits of the MAC address whose corresponding bit in
the mask is set to 1 will be used as match criteria. So, the
example above will match 00:01:02:03:xx:xx.
If the mask is not supplied then it will be assumed to be
ff:ff:ff:ff:ff:ff. In other words, all bits of the MAC address will be
used for matching.

ACL Rule Syntax
Table 51: ACL match conditions (Continued)

Applicable
IP Protocols/
ethernet-destination-address Ethernet destination MAC address Ethernet/
<mac-address> Ingress only
ethernet-destination-address Ethernet destination MAC address and mask. The mask is Ethernet/
<mac-address> mask optional, and is in the same format as the MAC address, for Ingress only
<mask> example:
or ethernet-destination-address 00:01:02:03:01:01
mask ff:ff:ff:ff:00:00
ethernet-destination-address
<mac-address> / <mask> or
ethernet-destination-address 00:01:02:03:01:01
/ ff:ff:ff:ff:00:00
Only those bits of the MAC address whose corresponding bit in
the mask is set to 1 will be used as match criteria. So, the
example above will match 00:01:02:03:xx:xx.
If the mask is not supplied then it will be assumed to be
ff:ff:ff:ff:ff:ff. In other words, all bits of the MAC address will be
used for matching.
source-address <prefix> IP source address and mask. Egress ACLs do not support IPv6 All IP/
addresses, only IPv4 addresses. Use either all IPv4 or all IPv6 Ingress and
addresses in an ACL.b Egress
destination-address <prefix> IP destination address and mask. Egress ACLs do not support All IP/
IPv6 addresses, only IPv4 addresses. Use either all IPv4 or all Ingress and
IPv6 addresses in an ACL.b Egress
protocol <number> IP protocol field. For IPv6c, this matches the Next Header field All IP/
in the packet. In place of the numeric value, you can specify Ingress and
one of the following text synonyms (the field values are also Egress
listed): egp(8), esp(5), gre(47), icmp(1), igmp(2), ipip(4),
ipv6(41), ospf(89), pim(102), rsvp(46), tcp(6), or udp(17).
Source-port {<number> | TCP or UDP source port. You must also specify the protocol TCP, UDP/
<range>} match condition to determine which protocol is being used on Ingress and
the port, any time you use the this match condition. In place of Egress
the numeric value, you can specify one of the text synonyms
listed under destination port. If no source-port is specified, the
default source-port is “any.”
Destination-port {<number> | TCP or UDP destination port. You must also specify the protocol TCP, UDP/
<range>} match condition to determine which protocol is being used on Ingress and
the port, any time you use the this match condition. In place of Egress
the numeric value, you can specify one of the following text
synonyms (the field values are also listed): afs(1483), bgp(179),
biff(512), bootpc(68), bootps(67), cmd(514), cvspserver(2401),
DHCP(67), domain(53), eklogin(2105), ekshell(2106),
exec(512), finger(79), ftp(21), ftp-data(20), http(80),
https(443), ident(113), imap(143), kerberos-sec(88),
klogin(543), kpasswd(761), krb-prop(754), krbupdate(760),
kshell(544), idap(389), login(513), mobileip-agent(434),
mobileip-mn(435), msdp(639), netbios-dgm(138), netbios-
ns(137), netbios-ssn(139), nfsd(2049), nntp(119), ntalk(518),
ntp(123), pop3(110), pptp(1723), printer(515), radacct(1813),
radius(1812), rip(520), rkinit(2108), smtp(25), snmp(161),
snmptrap(162), snpp(444), socks(1080), ssh(22), sunrpc(111),
syslog(514), tacacs-ds(65), talk(517), telnet(23), tftp(69),
timed(525), who(513), xdmcp(177), zephyr-clt(2103), or
zephyr-hm(2104).

Access Lists (ACLs)

Applicable
IP Protocols/
TCP-flags <bitfield> TCP flags. Normally, you specify this match in conjunction with TCP/Ingress
the protocol match statement. In place of the numeric value, and Egress
you can specify one of the following text synonyms (the field
values are also listed): ACK(0x10), FIN(0x01), PUSH(0x08),
RST(0x04), SYN(0x02), URG(0x20), SYN_ACK(0x12).
IGMP-msg-type <number> IGMP message type. Possible values and text synonyms: v1- IGMP/
report(0x12), v2-report(0x16), v3-report(0x22), V2-leave (0x17), Ingress and
or query(0x11). Egress
ICMP-type <number> ICMP type field. Normally, you specify this match in ICMP/Ingress
conjunction with the protocol match statement. In place of the and Egress
numeric value, you can specify one of the following text
synonyms (the field values are also listed): echo-reply(0), echo-
request(8), info-reply(16), info-request(15), mask-request(17),
mask-reply(18), parameter-problem(12), redirect(5), router-
advertisement(9), router-solicit(10), source-quench(4), time-
exceeded(11), timestamp(13), timestamp-reply(14), or
unreachable(3).
source-sap SSAP is a 1 byte field with possible values 0-255 decimal. The Ethernet/
value can be specified in decimal or hexadecimal. The SSAP Ingress Only
field can be found at byte offset 15 in 802.3 SNAP and LLC
formatted packets. (Available on the Summit family,
SummitStack, and BlackDiamond 8800 a- and e-series switches
only.)
destination-sap DSAP is a 1 byte field with possible values 0-255 decimal. The Ethernet/
value can be specified in decimal or hexadecimal. The DSAP Ingress Only
field can be found at byte offset 14 in 802.3 SNAP and LLC
formatted packets. (Available on the Summit family,
only.)
snap-type SNAP type is a 2 byte field with possible values 0-65535 Ethernet/
decimal. The value can be specified in decimal or hexadecimal. Ingress Only
The SNAP type field can be found a byte offset 20 in 802.3
SNAP formatted packets. (Available on the Summit family,
only.)

ACL Rule Syntax

Applicable
IP Protocols/
ICMP-code <number> ICMP code field. This value or keyword provides more specific ICMP/Ingress
information than the icmp-type. Because the value's meaning and Egress
depends upon the associated icmp-type, you must specify the
icmp-type along with the icmp-code. In place of the numeric
value, you can specify one of the following text synonyms (the
field values also listed); the keywords are grouped by the ICMP
type with which they are associated:
Parameter-problem:
ip-header-bad(0), required-option-missing(1)
Redirect:
redirect-for-host (1), redirect-for-network (2), redirect-for-tos-
and-host (3), redirect-for-tos-and-net (2)
Time-exceeded:
ttl-eq-zero-during-reassembly(1), ttl-eq-zero-during-transit(0)
Unreachable:
communication-prohibited-by-filtering(13), destination-host-
prohibited(10), destination-host-unknown(7), destination-
network-prohibited(9), destination-network-unknown(6),
fragmentation-needed(4), host-precedence-violation(14), host-
unreachable(1), host-unreachable-for-TOS(12), network-
unreachable(0), network-unreachable-for-TOS(11), port-
unreachable(3), precedence-cutoff-in-effect(15), protocol-
unreachable(2), source-host-isolated(8), source-route-failed(5)
IP-TOS <number> IP TOS field. In place of the numeric value, you can specify one All IP/
of the following text synonyms (the field values are also listed): Ingress and
minimize-delay 16 (0x10), maximize-reliability 4(0x04), Egress
minimize-cost2 (0x02), and normal-service 0(0x00).
fragments BlackDiamond 10808, BlackDiamond 12800 series switches, All IP, no L4
BlackDiamond 8800 a-series and e-series modules, and Summit rules/Ingress
X450a, X450e, and X250e series switches only—IP fragmented only
packet. FO > 0 (FO = Fragment Offset in IP header)d
first-fragments Non-IP fragmented packet or first fragmented packet. FO==0. All IP/
Ingress only
a. However, packets using the Ethernet type for vMANs, 0x88a8 by default, are handled by vMAN ACLs. See the section, “vMAN
ACLs—BlackDiamond 10808 and BlackDiamond 12800 Series Switches Only” for more information.
b. You can not specify an IPv6 address with a 128-bit mask (host entry) for the Summit family switches.
c. See the section “IPv6 Traffic with L4 Match Conditions” for details about specifying a protocol/port match with IPv6.
d. See the section “Fragmented packet handling” for details.
NOTE
Directed ARP response packets cannot be blocked with ACLs from reaching the CPU and being learned on the
BlackDiamond 8800 a- and e- series switches, SummitStack, and the Summit a- and e- series switches. To block
directed ARP response packets on the original BlackDiamond 8800 and Summit series switches, you must configure
a “deny” ACL and disable the permit-to-CPU option.

Access Lists (ACLs)
Along with the data types described in Table 52, you can use the operators <, <=, >, and >= to specify
match conditions. For example, the match condition, source-port > 190, will match packets with a
source port greater than 190. Be sure to use a space before and after an operator.
NOTE
The BlackDiamond 8800 series switches, SummitStack, and the Summit series switches support 128 rules per
Gigabit Ethernet port and 1024 rules per 10 Gigabit Ethernet port. Certain features also use rules for their
implementation. A single match condition can require the creation of many rules, and may not be supported on
these switches. For example, the match condition source-port 1000 - 3000 requires creating 2000 rules, and is not
supported on these switches. The match condition source-port > 90 will also not work, as it requires a large number
of rules to implement in the hardware. The match condition source-port 100 - 200 could work, as it uses fewer than
120 rules to implement. See also the section, “Conserving ACL Masks and Rules—BlackDiamond 8800 Original
Series Modules and the Summit X450 Series Switches Only”, for further information on ACL limits.
Table 52: ACL match condition data types

Condition Data Type Description
prefix IP source and destination address prefixes. To specify the address prefix, use the
notation prefix/prefix-length. For a host address, prefix-length should be
set to 32.
number Numeric value, such as TCP or UDP source and destination port number, IP protocol
number.
range A range of numeric values. To specify the numeric range, use the notation: number -
number
bit-field Used to match specific bits in an IP packet, such as TCP flags and the fragment flag.
mac-address 6-byte hardware address.
IPv6 Traffic with L4 Match Conditions

If you apply an ACL policy intended to match IPv6 packets using an ACL that specifies L4 conditions,
the traffic will not be matched. For example, the following ACL:
entry destIp {
if {
protocol tcp;
}
then {
permit;
count destIp;
}
}
will not match any IPv6 packets. For IPv6 packets to match, you must add a match condition that
includes all IPv6 L3 addresses. For example, you would change the ACL entry to:
entry destIp {
if {
source-address 0::0/0;
protocol tcp;
}
then {
permit;
count destIp;

ACL Rule Syntax
}
}
NOTE
The match condition shown in the example above can be installed only on the original
BlackDiamond 8800 or Summit X450 modules. It is too complex to be installed on the
BlackDiamond 8800, or Summit X450 a- or e- series modules.
Fragmented packet handling

Two keywords are used to support fragmentation in ACLs:
● fragments—FO field > 0 (FO means the fragment offset field in the IP header.)—
BlackDiamond 10808, BlackDiamond 12800 series switches, BlackDiamond 8800 a-series and e-series
modules, Summit X450a, X450e, and X250e switches only.
● first-fragments—FO == 0.
Policy file syntax checker. The fragments keyword cannot be used in a rule with L4 information. The
syntax checker will reject such policy files.
The following rules are used to evaluate fragmented packets or rules that use the fragments or first-
fragments keywords.
With no keyword specified, processing proceeds as follows:

● An L3-only rule that does not contain either the fragments or first-fragments keyword matches
any IP packets.
● An L4 rule that does not contain either the fragments or first-fragments keyword matches non-
fragmented or initial-fragment packets.
With the fragments keyword specified:

● An L3-only rule with the fragments keyword only matches fragmented packets.
● An L4 rule with the fragments keyword is not valid (see above).
With the first-fragments keyword specified:

● An L3-only rule with the first-fragments keyword matches non-fragmented or initial fragment
packets.
● An L4 rule with the first-fragments keyword matches non-fragmented or initial fragment packets.
IPv6 ACL Address Masks—BlackDiamond 10808 and

The BlackDiamond 10808 and BlackDiamond 12800 series switches use address masks for matching the
128-bit IPv6 addresses in ACLs. The default mask is 0:ffff:ffff:ffff:0:ffff:ffff:ffff, so for purposes of ACL
matching, the switch ignores bits 1 through 16 and 65 through 80 (counting the highest-order bit as
bit 1). There is a separate mask for the IPv6 source and IPv6 destination address.

Access Lists (ACLs)
vMAN ACLs—BlackDiamond 10808 and BlackDiamond

In order to provide more control for the types of traffic transiting vMANs, a group of match conditions
and actions that apply only to vMAN traffic was added in ExtremeXOS release 11.4. Packets arriving on
vMAN ports, that are tagged with the vMAN Ethertype (S-Tag), and that are not IGMP packets are
compared with vMAN ACL entries. By default, the vMAN Ethertype is 0x88a8, but can be changed by
using the configure vman ethertype command. This tag can modified as packets transit the switch
by using the stag-ethertype action modifier for vMAN ACLs.
Any ACL entries that begin with the match conditions in Table 53 will be applied to vMAN packets. If
the entry begins with any other match conditions, the ACL will be applied to non-vMAN packets.
Match conditions that do not belong will be ignored.
You can mix vMAN ACL entries with non-vMAN ACL entries in a policy file. The entries that use a
condition from Table 53 will be applied to vMAN traffic on the interface, and entries that do not will be
applied to non-vMAN traffic.
Table 53: vMAN Only ACL match conditions

ccos <number> C-CoS. Customer or inner Vpri. Can be a single value or a range Ingress and
of values. Range is 0 to 7. Egress
cvid <number> C-VID. Customer or inner VLAN ID. Can be a single value or a Ingress and
range of values. Range is 0 to 4095. Egress
diffserv-codepoint <number> Diffserv Codepoint. Ingress only
ports [ingress | egress] Physical port number. It is a single port. Ingress and
<port> Egress
scos <number> S-CoS. Service VLAN or outer Vpri. Can be a single value or a Ingress and
svid <number> S-VID. Service VLAN or outer ID. Can be a single value or a Ingress and
tos <number> IP TOS field. Ingress only
An additional match condition, ethernet-destination-address, can also be used in vMAN ACLs, but
must not be the first match condition in an entry. This condition is listed in Table 51.
vMAN ACL Actions

The vMAN ACL actions are:
● permit—The packet is forwarded. This is the default action; if no action is specified in a rule entry,
the packet is forwarded.
● deny—The packet is dropped.
● link <port>—The packet is forwarded to this primary port in a trunk port.
● back-up-link <port>—The packet is forwarded to this port in a trunk port if the primary link is
down.

vMAN ACLs—BlackDiamond 10808 and BlackDiamond 12800 Series Switches Only
NOTE
The value for the <port> argument in the link and back-up-link actions should be in the same trunk port.
● link-all—The packet is forwarded to all the ports in a trunk port if the primary link is down.
● link-none—The packet is not forwarded if the primary link is down.
● flood-group [<extended flood group name> no-echo] [<virtual flood-group name> no-
echo]—The packet is forwarded to the set of ports listed in the flood-group.
The following criteria are checked before the flood-group attribute is installed:
■ One of the flood-group names must be present. Refer to the flood-group commands in Chapter
7, “VMAN Commands” in the Command Reference.
■ A flood group may have no port.
■ The flood-group keyword is valid only in an ingress ACL.
■ To use a flood-group, the condition if {svid}; is required in the ingress ACL.
vMAN ACL Action Modifiers

Additional actions can also be specified, independent of whether the packet is dropped or forwarded.
These additional actions are called action modifiers. The action modifiers are:
● count <countername>—Increments the counter named in the action modifier (ingress only).
● cvid <number>—Modifies the C-VID value.
● link-aggregation-hash <number>—Controls which link is used by matching vMAN traffic (egress
only).
● qosprofile <qosprofilename>—Forwards the packet to the specified QoS profile (ingress only).
● scos <number>—Modifies the S-COS value.
● stag-ethertype <ethertype>—Modifies the vMAN Ethertype value, also called the S-Tag value.
● svid <number>—Modifies the S-VID value.
● traffic-queue <traffic-queue>—Places the traffic on the specified traffic-queue (BlackDiamond
12800 series only).
● uplinkport [tagged | untagged] <port_list>—Modifies the uplink port.
● learning-domain <vman name>—The packet's source MAC address is learned on this vMAN.
NOTE
All the ports in the inter-vMAN forwarding should be added to this vMAN as tagged ports.
Counting Packets
When the ACL entry match conditions are met, the specified counter is incremented. The counter value
can be displayed by the command:
{ingress | egress}

Access Lists (ACLs)
Controlling Link Usage

Any matching vMAN traffic that egresses a load sharing (link aggregation) group can be forced to use
the same link in the group for all traffic. The hash value range is 1 to 16. For example, for a link
aggregation hash of three, the matching vMAN traffic will go over link number three. If the hash value
is greater than the number of links, the value rolls over. For example, if the hash value is six, the traffic
will go over link six in an eight link group and over link two in a four link group.
If a member link of a group goes down, the link indices are reassigned, so the traffic still goes over the
same link, but might be assigned to a different member link.
vMAN ACL Examples—BlackDiamond 10808 and BlackDiamond

The following example, vmanex1.pol, allows traffic with an S-VID of 300 and a C-VID of 200, to
destination MAC address 00:00:00:00:01:02 to be forwarded on QP1 with an S-VID of 201, but block all
other traffic with an S-VID of 300:
entry vman_entry1 {
if {
svid 300;
cvid 200;
ethernet-destination-address 00:00:00:00:01:02;
} then {
qosprofile qp1;
svid 201;
}
}
entry vman_entry2 {
if {
svid 300;
} then {
deny;
}
To apply this policy to port 2:2, use the following command:
configure access-list vmanex1 port 2:2
Layer-2 Protocol Tunneling ACLs—BlackDiamond 8800

a-Series modules and e-Series and Summit X250e,
X450a, and X450e Switches Only
Beginning in ExtremeXOS Release 12.0, three new ACL match conditions and one new ACL action are
added to interoperate with vendor-proprietary Layer-2 protocol tunneling.
The following fields within 802.3 Subnetwork Access Protocol (SNAP) and LLC formatted packets can
be matched:
● Destination service access point (SAP)
● Source SAP

Dynamic ACLs
The following field can be matched within Subnetwork Access Protocol (SNAP) packets only:
● SNAP type
The following ACL action is added to the specified switches:

● Replacement of the Ethernet MAC destination address
This action replaces the destination MAC address of any matching Layer-2 forwarded packets on the
supported platforms. This action can be used to effectively tunnel protocol packets, such as STP, across
a network by replacing the well-known protocol MAC address with a different proprietary or otherwise
unique MAC address. After tunnel egress, the MAC destination address can be reverted back to the
well-known MAC address.
NOTE
The “replace-ethernet-destination-address” action applies only to Layer2 forwarded packets.
Dynamic ACLs
Dynamic ACLs are created using the CLI. They use a similar syntax and can accomplish the same
actions as single rule entries used in ACL policy files. More than one dynamic ACL can be applied to an
interface, and the precedence among the dynamic ACLs can be configured. By default, the priority
among dynamic ACLs is established by the order in which they are configured.
NOTE
Dynamic ACLs have a higher precedence than ACLs applied using a policy file.
The two steps to using a dynamic ACL on an interface are:

1 Create the dynamic ACL rule.
2 Configure the ACL rule on the interface.
Creating the Dynamic ACL Rule

Creating a dynamic ACL rule is similar to creating an ACL policy file rule entry. You will specify the
name of the dynamic ACL rule, the match conditions, and the actions and action-modifiers. You can
configure a dynamic ACL to be persistent or non-persistent across system reboots. The match
conditions, actions, and action-modifiers are the same as those that are available for ACL policy files
(see “ACL Rule Syntax” on page 381). In contrast to the ACL policy file entries, dynamic ACLs are
created directly in the CLI. Use the following command to create a dynamic ACL:
create access-list <dynamic-rule> <conditions> <actions> {non-permanent} {application
<appl_name>}

Access Lists (ACLs)
As an example of creating a dynamic ACL rule, let’s compare an ACL policy file entry with the CLI
command that creates the equivalent dynamic ACL rule. The following ACL policy file entry will drop
all ICMP echo-requests:
entry icmp-echo {
if {
protocol icmp;
} then {
deny;
}
}
To create the equivalent dynamic ACL rule, use the following command:
create access-list icmp-echo “protocol icmp;icmp-type echo-request” “deny”
Notice that the conditions parameter is a quoted string that corresponds to the match conditions in the
if { ... } portion of the ACL policy file entry. The individual match conditions are concatenated
into a single string. The actions parameter corresponds to the then { ... } portion of the ACL
policy file entry.
From the command line, you can get a list of match conditions and actions by using the following
command:
check policy attribute {<attr>}
The ACL rule shown in the example will be saved when the save command is executed, because the
optional keyword non-permanent was not configured. This allows the rule to persist across system
reboots.
Note also that the sample ACL rule does not specify an application to which the rule belongs. The
default application is CLI.
Limitations. Dynamic ACL rule names must be unique, but can be the same as used in a policy-file
based ACL. Any dynamic rule counter names must be unique. Clear-FLow rules can be specified only
in policy files and therefore apply only to rules created in a policy file.
Configuring the ACL Rule on the Interface

After a dynamic ACL rule has been created, it can be applied to a port, VLAN, or to the wildcard any
interface. When the ACL is applied, you will specify the precedence of the rule among the dynamic
ACL rules. To configure the dynamic ACL rule on an interface, use the following command:
configure access-list add <dynamic_rule> [ [[first | last]{priority
<p_number>}]|[[before | after] <rule>]|[priority <p_number>]] [any | vlan <vlanname> |
ports <portlist>] {ingress | egress} {zone <zone>} {application <appl_name>}
To remove a dynamic ACL from an interface, use the following command:
configure access-list delete <dynamic_rule> [any | vlan <vlanname> | ports <portlist>

| all] {ingress | egress} {application <appl_name>}

Configuring ACL Priority

In ExtremeXOS 11.6 and later, management of ACLs is more flexible, with configurable priority for
dynamic ACLs. This includes ACLs inserted by internal and external applications, as well as those
inserted via the CLI. The priority will be assigned by a system of zones, and within zones by numeric
codes.
NOTE
The priority of static ACLs is determined by the order they are configured, with the first rule configured having the
highest priority.
Applications will insert ACL's into zones. The zones will consist of default zones and created zones.
Default Zones will group like functions. The administrator will have the ability to create new zones and
configure the priority of default and created zones.
The following zones are configured by default; they are listed in the order of their priority, highest
priority first:
1 DOS
This is the denial of service zone.
2 SYSTEM
This is the zone for applications that require a CPU-copy or mirror, and for redirect ACLs.
3 SECURITY
This is the zone for ACLs installed by security appliances and internal security processes.
NOTE
Default zones cannot be deleted.
There is a configurable process for applications to insert an ACL into a zone according to the priority of
the application within that zone. The following list shows the default assignment and priority of
applications, by zone:
● DOS Zone
Dos (Denial of Service application)
● System Zone
Cli
IpSecurity
Dot1Ag
Dot1AgDefault
MacInMac
NetLogin
● Security Zone
Sentriant
Generic Xml (Allows configuration of one additional external application)

Access Lists (ACLs)
NOTE
The DOT1AG, DOT1AG Default, and MacInMac applications are not supported on the BlackDiamond 8800 switches,
SummitStack, and the Summit family switches.
Applications can occupy multiple zones. For example, you can add the CLI application to the DOS
zone, and assign it a higher priority than the DOS application. The DOS zone then has two applications,
CLI and DOS application, and within the DOS zone, an ACL created by the CLI has a higher priority
than an ACL inserted by the DOS application.
Another way to configure ACL priority is by creating new zones. For example, you might create a zone
called MY_HIGH_ZONE, and assign that zone a priority below the DOS zone and above the System
zone. You can add applications to that zone and assign their priority. The example below shows the
ACL zone priority that would result from adding the MacInMac and CLI applications to
MY_HIGH_ZONE:
1 DOS Zone
Dos
2 MY_HIGH_ZONE
MacInMac
Cli
3 System Zone
Cli
Dot1Ag
Dot1ADefault
MacInMac
Cli
NetLogin)
4 SECURITY
Sentriant
Generic Xml
Applications can insert an ACL into any of the zones to which the application belongs. If an application
attempts to insert an ACL into a zone where the application is not configured, an error message
appears, and the ACL is not installed. Therefore, you have full control of ACL priorities and you can
configure the switch to install ACLs from an application at any priority level. In the example above, the
application CLI can insert an ACL into either MY_HIGH_ZONE or the SYSTEM zone. The location of
the ACL within the zone depends on the priority assigned by the application. An application can assign
priority to an ACL using:
● priority attributes (first or last)

● relative priority
● priority numbers
The priority attributes first (highest priority) and last (lowest priority) can be applied to an ACL to
establish its position within a zone. Relative priority sets the ACL priority relative to another ACL
already installed by the application in the same zone.
Priority numbers allow an application to specify the priority of an ACL within a zone. The priority
numbers are unsigned integers from 0 to 7; a lower number represents a higher priority. This means

that if an application adds an ACL at priority 5 and later adds another ACL at priority 3, the second
ACL has higher priority.
If an application assigns the same priority number to two ACLs, the ACL added most recently has the
higher priority. It will be inserted in the priority map immediately ahead of the older ACL that has the
same priority number. This effectively allows the application to create sub-zones within a zone. The
attributes First and Last can be used in combination with priority numbers to prioritize the ACLs within
a sub-zone. For example, an ACL could be configured with the First attribute, along with the same
priority number as other ACLs in the same zone, effectively assigning that ACL the highest priority
within a sub-zone.
The show configuration command shows the current configuration of the entire switch in the form of
CLI commands which can later be played back to configure the switch.
The show configuration acl command shows the current configuration of the ACL manager.
The new application keyword allows you to specify the application to which the ACL will be bound.
Typically, applications create and insert ACL’s on the switch; however the administrator can install
ACL’s "on behalf" of an application by specifying the application keyword. (This keyword is also
used with the show config acl command to enable CLI playback). If no application is specified, the
default application is CLI.
This means you have the ability to create, delete, and configure ACLs for any application.
To create a zone, use the following command:
create access-list zone <name> zone-priority <number>
To configure the priority of zones, use the following command:
configure access-list zone <name> zone-priority <number>
To add an application to a zone at a particular priority, or to change the priority of an application

within a zone, use the following command:
configure access-list zone <name> {add} application <appl-name> application_priority

<number>
An application must occupy a least one zone.
To move an application within a zone or to another zone use the following command:
configure access-list zone <name> move-application <appl-name> to-zone <name>

application-priority <number>
All applications can be configured to go into any and all zones.
A change in the zone list will result in a change in the order of dynamic ACLs that have been applied,
per interface. The changes in hardware will be achieved by uninstalling and then reinstalling the
dynamic ACLs in the new positions. There is a possibility, due to hardware constraints, that some ACLs
will not be reinstalled. These occurrences will be logged.
To delete an application from a zone, use the following command:
configure access-list zone <name> delete application <appl-name>

Access Lists (ACLs)
When deleting an application from a zone, any ACLs that have been inserted into that zone for the
deleted application will be moved to the next higher zone in which the application appears.
To delete a zone use the following command:
delete access-list zone <name>
You must remove all applications from a zone before you can delete the zone. You cannot delete the
default zones.
ACL Evaluation Precedence—BlackDiamond 10808 and

This section describes the precedence for evaluation among ACL rules for the BlackDiamond 10808 and
BlackDiamond 12800 series switches. In many cases there will be more than one ACL rule entry for an
interface. This section describes how multiple rule entries are evaluated.
Rule Evaluation
When there are multiple rule entries applied to an interface, evaluation proceeds as follows:
● A packet is compared to a rule entry.
● If the packet matches all the match conditions, the action and any action modifiers in the then
statement are taken and evaluation terminates.
● This process continues until either the packet matches all the match conditions in one of the
subsequent rule entries or there are no more entries.
● If a packet passes through all the rule entries in the ACL without matching any of them, it is
permitted.
Often there will be a lowest-precedence rule entry that matches all packets. This entry will match any
packets not otherwise processed, so that user can specify an action to overwrite the default permit
action. This lowest-precedence rule entry is usually the last entry in the ACL policy file applied to the
interface.
Precedence of Dynamic ACLs

Dynamic ACLs have a higher precedence than any ACLs applied using policy files. The precedence
among any dynamic ACLs is determined as they are configured. The precedence of ACL’s applied
using policy files are determined by the rule’s relative order in the policy file. The precedence of ACL’s
applied using policy files is determined by the rule’s relative order in the policy file.
Precedence of L2/L3/L4 ACL Entries on BlackDiamond10k and

12K Switches
An ACL policy file contains one or more rule entries.

ACL Evaluation Precedence—BlackDiamond 8800 Series switches, SummitStack, and the Summit Family Switches Only
When an ACL file contains both L2 and L3/L4 rules:

● L3/L4 rules have higher precedence over L2 rules. L3/L4 rules are evaluated before any L2 rules.
● The precedence among L3/L4 rules is determined by their relative position in the ACL file. Rules are
evaluated sequentially from top to bottom.
● The precedence among L2 rules is determined by their position. Rules are evaluated sequentially
from top to bottom.
● It is recommended that L2 and L3/L4 rules be grouped together for easy debugging.
Precedence Among Interface Types

As an example of precedence among interface types, suppose a physical port 1:2 is member port of a
VLAN yellow. ACLs could be configured on the port, either singly or as part of a port list, on the VLAN
yellow, and on all ports in the switch (the wildcard ACL). For all packets crossing this port, the port-
based ACL has highest precedence, followed by the VLAN-based ACL and then the wildcard ACL.
ACL Evaluation Precedence—BlackDiamond 8800

Series switches, SummitStack, and the Summit Family
Switches Only
This section describes the precedence for evaluation among ACL rules for the BlackDiamond 8800 series
switches, SummitStack, and the Summit family of switches (whether or not included in a SummitStack).
In many cases there will be more than one ACL rule entry for an interface. This section describes how
multiple rule entries are evaluated.
Multiple rule entries do consume hardware resources. If you find that your situation runs up against
those limits, there are steps you can take to conserve resources by modifying the order of the ACL
entries that you create. For details, see “Conserving ACL Masks and Rules—BlackDiamond 8800
Original Series Modules and the Summit X450 Series Switches Only,” and “ACL Mechanisms.”
Rule Evaluation
When there are multiple rule entries applied to an interface, evaluation proceeds as follows:
● A packet is compared to all the rule entry match conditions at the same time.
● For each rule where the packet matches all the match conditions, the action and any action modifiers
in the then statement are taken. If there are any actions or action modifiers that conflict (deny vs.
permit, etc), only the one with higher precedence is taken.
● If a packet matches no rule entries in the ACL, it is permitted.
Often there will be a lowest-precedence rule entry that matches all packets. This entry will match any
packets not otherwise processed, so that user can specify an action to overwrite the default permit
action. This lowest-precedence rule entry is usually the last entry in the ACL policy file applied to the
interface.

Access Lists (ACLs)
NOTE
When a packet matches more than one rule entry, the highest precedence conflicting action is taken, so you can use
the precedence to determine if a packet is permitted or denied. However, incrementing a counter is not a conflicting
action, so a packet that matches more that one rule that increments a common counter, could count the packet
more than once. Don’t use precedence to control counter usage; define different counters for different cases. For
details of this behavior on different platforms, see, “ACL Masks and Rules—BlackDiamond 8800 Series Switches
and the Summit X450 Series Switches Only” and “ACL Slices and Rules—BlackDiamond 8800 a-series and e-series
Modules and Summit X450a, X450e, and X250e Series Switches Only.”
Precedence of Dynamic ACLs

Dynamic ACLs have a higher precedence than any ACLs applied using policy files. The precedence
among any dynamic ACLs is determined as they are configured. The precedence of ACL’s applied
using policy files is determined by the rule’s relative order in the policy file.
Precedence of L2/L3/L4 ACL Entries

Rule precedence is solely determined by the rule’s relative order. L2, L3, and L4 rules are evaluated in
the order found in the file or by dynamic ACL configuration.
Precedence Among Interface Types

As an example of precedence among interface types, suppose a physical port 1:2 is member port of a
VLAN yellow. ACLs could be configured on the port, either singly or as part of a port list, on the VLAN
yellow, and on all ports in the switch (the wildcard ACL). For all packets crossing this port, the port-
based ACL has highest precedence, followed by the VLAN-based ACL and then the wildcard ACL.
Redundant Rules
For the BlackDiamond 8800 series switches and the Summit family of switches, eliminate redundant
rules (any with the EXACT same match criteria) in the policy file. If two rules have identical match
conditions, but different actions, the second rule is rejected by the hardware.
For example, the two following ACL entries are not allowed:
entry DenyNMR {
if {
protocol 17;
destination-port 161;
} then {
deny;
count denyNMR;
}
}
entry DenyNIC {
if {
protocol 17;
} then {
deny;

Applying ACL Policy Files
count denyNIC;
}
}
Applying ACL Policy Files

A policy file intended to be used as an ACL is applied to a port, VLAN, or to all interfaces (the any
keyword). Use the name of the policy file for the <aclname> parameter in the CLI command. To apply
an ACL policy, use the following command:
configure access-list <aclname> [any | ports <portlist> | vlan <vlanname>] {ingress |
egress}
If you use the any keyword, the ACL is applied to all the interfaces and is referred to as the wildcard
ACL. This ACL is evaluated for any ports without specific ACLs, and it is also applied to any packets
that do not match the specific ACLs applied to the interfaces.
If an ACL is already configured on an interface, the command will be rejected and an error message
displayed.
To remove an ACL from an interface, use the following command:
unconfigure access-list {any | ports <portlist> | vlan <vlanname>} {ingress | egress}
To display which interfaces have ACLs configured, and which ACL is on which interface, use the
following command:
Displaying and Clearing ACL Counters

To display the ACL counters, use the following command:
{ingress | egress}
To clear the access list counters, use the following command:
clear access-list {dynamic} counter {<countername>} {any | ports <portlist> | vlan

<vlanname>} {ingress | egress}
Example ACL Rule Entries

The following entry accepts all the UDP packets from the 10.203.134.0/24 subnet that are destined for
the host 140.158.18.16, with source port 190 and a destination port in the range of 1200 to 1250:
entry udpacl {
if {
protocol udp;
source-port 190;

Access Lists (ACLs)
} then {
permit;
}
}
The following rule entry accepts TCP packets from the 10.203.134.0/24 subnet with a source port larger
than 190 and ACK & SYN bits set and also increments the counter tcpcnt. The packets will be forwarded
using QoS profile QP3. This example will only work with the BlackDiamond 10808 and BlackDiamond
12800 series switches, since the match condition source-port > 190 alone will create more than 118 rules
in the hardware:
entry tcpacl {
if {
protocol TCP;
source-port > 190;
tcp-flags syn_ack;
} then {
permit;
count tcpcnt ;
qosprofile qp3;
}
}
The following example denies ICMP echo request (ping) packets originating from the 10.203.134.0/24
subnet, and increments the counter icmpcnt:
entry icmp {
if {
protocol icmp;
} then {
deny;
count icmpcnt;
}
}
The following example prevents TCP connections from being established from the 10.10.20.0/24 subnet,
but allows established connections to continue, and allows TCP connections to be established to that
subnet. A TCP connection is established by sending a TCP packet with the SYN flag set, so this example
blocks TCP SYN packets. This example emulates the behavior of the ExtremeWare permit-established
ACL command:
entry permit-established {
if {
protocol TCP;
tcp-flags syn;
} then {
deny;
}
}
The following entry denies every packet and increments the counter default:
entry default {
if {

ACL Mechanisms
} then {
deny;
count default;
}
}
The following entry permits only those packets with destination MAC addresses whose first 32 bits
match 00:01:02:03:
entry rule1 {
if {
ethernet-destination-address 00:01:02:03:01:01 ff:ff:ff:ff:00:00 ;
} then {
permit ;
}
}
The following entry denies IPv6 packets from source addresses in the 2001:db8:c0a8::/48 subnets and to
destination addresses in the 2001:db8:c0a0:1234::/64 subnets:
entry ipv6entry {
if {
source-address 2001:DB8:C0A8:: / 48;
destination-address 2001:DB8:C0A0:1234:: / 64;
} then {
deny;
}
}
ACL Mechanisms
For many applications of ACLs, it is not necessary to know the details of how ACLs work. However, if
you find that your application of ACLs is constrained by hardware limitations, you can often rearrange
the ACLs to use the hardware more efficiently. The following sections go into some detail about how
the ACLs use hardware resources, and some suggestions about how to reorder or rewrite ACLs to use
fewer resources.
● ACL Masks and Rules—BlackDiamond 8800 Series Switches and the Summit X450 Series Switches
Only on page 405
● ACL Slices and Rules—BlackDiamond 8800 a-series and e-series Modules and Summit X450a, X450e,
and X250e Series Switches Only on page 411
ACL Masks and Rules—BlackDiamond 8800 Series Switches and

the Summit X450 Series Switches Only
The Summit X450 series switches (whether or not included in a SummitStack) and the BlackDiamond
8800 series original modules use a common hardware design to implement ACLs. The same architecture
and guidelines apply to both platforms.
For these platforms, each port has a set of sixteen masks, and memory sufficient to contain 128 rules
and actions (for Gigabit ports) or 1024 rules and actions (for 10 Gigabit ports). The masks determine
which part of a packet header will be compared with the value stored for the rule. As a packet ingresses
a switch port, its header information is placed in all of the sixteen masks and compared with the values

Access Lists (ACLs)
stored in the rules. All the comparisons happen in sync, then the actions that do not conflict are taken.
We’ll discuss actions later, so for now we’ll concentrate on masks and rules only. For example, the
following ACL entry ex_1 will set up a mask that compares only the first 32 bits of the IP source
address in the packet header to the value in the rule:
entry ex_1 {
if {
source-address 10.10.10.10/32 ;
} then {
deny ;
}
}
Figure 44 shows the result of applying this ACL to a port. The first mask is set to compare the first 32
bits in the header source address with the value in the rule, 10.10.10.10. This example is simplified from
any real case, since there are always some ACLs applied to every port by the system. See “Mask and
Rule Use by Feature” for a listing of these ACLs.
Figure 44: ACL Entry ex_1
Masks (16) Rules (128 or 1024)
XM_073
Now apply the following ACL entry ex_2:

entry ex_2 {
if {
} then {
deny ;
}
}

ACL Mechanisms
Since this entry is examining the same parts of the packet header as ex_1, the same ACL mask can be
reused for ex_2. Figure 45 shows that the same mask is used for both rules. If we used a different
netmask, for example /24, or if we had wanted to also match a UDP port value, we would have needed
to use a different mask. Since the only thing different between these two rule entries is the matching
value, we can use the same mask.
Figure 45: ACL Entry ex_1 and ex_2
Masks Rules
XM_074
Now apply the following ACL entry ex_3:
entry ex_3 {
if {
ethernet-source-address 00:e0:2b:11:22:33;
} then {
deny ;
}
}
The entry ex_3 uses different match conditions than the two earlier entries, so a different mask is used
for the comparison to ex_3’s values. Figure 46 shows that ex_3 uses a new mask.
The next entry may, or may not, use a new mask, depending on the match conditions that are chosen.
For example, if the next entry matches an ethernet source address then the mask will be reused. If the
next entry matches an IP source address, then a new mask would be used.
There are two extreme cases that will exhaust the hardware resources for ACLs. In one extreme case,
each entry requires a new mask, so that only 16 rule entries fill up the hardware resources. In the other
extreme, the same mask can be reused, but there are so many entries that the rule memory is exhausted.
For example, if we had 128 entries that each matched ethernet source addresses, the rule memory for a
1 Gigabit port would be filled (ignoring the system created entries).

Access Lists (ACLs)
Figure 46: ACL Entry ex_1, ex_2, and ex_3
Masks Rules
XM_075
Rule Evaluation and Actions

When a packet ingresses the switch, its header is loaded into all the masks, and then the values in the
masks are all compared with the rule values. If the values match, the rule action is taken. Conflicting
actions are resolved by the precedence of the entries. However, incrementing a counter is not a
conflicting action, so a packet that matches more that one rule that increments a common counter, could
count the packet more than once. Don’t use precedence to control counter usage; define different
counters for different cases.
Conserving ACL Masks and Rules—BlackDiamond 8800 Original Series Modules and
the Summit X450 Series Switches Only
The BlackDiamond 8800 series original modules and the Summit X450 switches (whether or not
included in a SummitStack) have a total of sixteen ACL masks per port on the switch. To avoid
exhausting the masks available on the switch, you must carefully plan your use of ACL masks.
Additionally, these switches support 128 rules per Gigabit Ethernet port and 1024 rules per 10 Gigabit
Ethernet port, so you must also plan your use of ACL rules.
An ACL mask defines a unique match criteria and relative rule precedence. Masks are automatically
generated based on the contents of an access-list policy. Only adjacent rules within the policy that have
identical match criteria will utilize the same ACL mask. For this reason, it is advantageous to list all
rules with the same match criteria together unless relative precedence with other policy rules is
required. Using VLAN-based or wildcards ACLs requires that the ACL masks are allocated on every
port in the system. For example, consider the following two policies:
policy1.pol :
entry one {

ACL Mechanisms
if {
} then {
count debug;
}
}
entry two {
if {
protocol tcp;
} then {
permit;
}
}
entry three {
if {
} then {
deny;
}
}
policy2.pol :
entry one {
if {
} then {
count debug;
}
}
entry three {
if {
} then {
deny;
}
}
entry two {
if {
protocol tcp;
} then {
permit;
}
}
In this example, the only difference between policy1.pol and policy2.pol is that rule entries two and
three are swapped. Policy1.pol consumes three masks since there are no adjacent rules with the same
match criteria. Policy2.pol consumes two masks since rules one and three are adjacent and have
identical match criteria. However, policy1.pol and policy2.pol have different meanings because of
precedence.

Access Lists (ACLs)
With this being said, you have to be careful to avoid wasting masks. For example consider the
following policy:
policy3.pol:
entry one {
if {
} then {
count debug;
}
}
entry two {
if {
protocol tcp;
} then {
deny;
}
}
entry three {
if {
} then {
deny;
}
}
Policy3.pol consumes three masks. However, since rule entries two and three have the same action,
their relative precedence doesn't matter, and they could be swapped without affecting the results of the
policy. The following policy accomplishes the same actions, but uses two masks:
policy4.pol:
entry one {
if {
} then {
count debug;
}
}
entry three {
if {
} then {
deny;
}
}
entry two {
if {
protocol tcp;
} then {
deny;
}
}

ACL Mechanisms
The only difference between policy3.pol and policy4.pol is that rule entries two and three are swapped.
The two policies have the same effect, but policy4.pol does not unnecessarily consume an ACL mask.
Mask and Rule Use by Feature

Certain non-ACL features allocate ACL masks and use ACL rules in order to function. Here are is a list
by feature:
● dot1p examination—1 mask, 8 rules (enabled by default)
● DiffServ examination—1 mask, 64 rules for 10G ports; 0 masks, 0 rules for 1G ports (disabled by
default)
● IGMP snooping—2 masks, 2 rules (enabled by default)
● IP interface—2 masks, 2 rules (disabled by default)
● VLAN QoS—1 mask, 1 rule per VLAN (disabled by default)
● port QoS—1 mask, 1 rule (disabled by default)
● VRRP—1 mask, 1 rule
● EAPS—1 master config + 1 transit config masks, 1 + number of transit-mode EAPS domains on the
port rules
● ESRP—1 mask, 1 rule
● LLDP—1 mask, 1 rule
● Netlogin—1 mask, 1 rule
● IPv6—1 mask, 1 rule
● Layer 4 ACL—1 mask, 5 rules1
NOTE
The above list of mask and rule usage changed from that in ExtremeXOS 11.3. In release 11.4, VRRP mask and
rule usage was reduced, dot1p examination could be disabled, and diffserv for 1 G ports did not require any masks
or rules.
To display the number of masks used by a particular port, use the following command:
show access-list usage acl-mask port <port>
To display the number of rules used by a particular port, use the following command:
show access-list usage acl-rule port <port>
ACL Slices and Rules—BlackDiamond 8800 a-series and e-series

Modules and Summit X450a, X450e, and X250e Series Switches
Only
The Summit X450a, X450e, and X250e series switches (whether or not included in a SummitStack) and
the BlackDiamond 8800 a-series and e-series modules use a mechanism different from the earlier
1. When you configure any ACL with Layer 4 lookup, the system consumes one mask and five rules in addi-
tion to the required mask and rules for the associated ACL.

Access Lists (ACLs)
Summit series and BlackDiamond 8800 series switches to implement ACLs. The same architecture and
guidelines apply to both platforms.
Instead of the per port masks used in earlier switches, these platforms use slices that can apply to any
of the supported ports. An ACL applied to a port may be supported by any of the slices.
For the Summit X450a series switch and the BlackDiamond 8800 a-series modules, each group of 24
ports has 16 slices, with each slice having enough memory for 128 rules and actions. For the Summit
X450e and X250e series switches and the BlackDiamond 8800 e-series modules, each group of 24 ports
has 8 slices, and each slice has the memory for 128 rules and actions. Figure 47 shows the 16 slices and
associated rule memory for a Summit X450a series switch or BlackDiamond 8800 a-series module.
Figure 48 shows the 8 slices and associated rule memory for a Summit X450e or X250e series switch or a
BlackDiamond 8800 e-series module.
Figure 47: Slice Support for Summit X450a Series Switch and BlackDiamond 8800 a-Series Modules
Slice 0 Slice 1 Slice 2 Slice 3 Slice 4 Slice 5 Slice 6 Slice 7
The 16 slices support all 24 ports.

Platforms with 48 ports have
another set of 16 slices to support
the additional 24 ports.
XM_076

ACL Mechanisms
Figure 48: Slice Support for Summit X450e, X250e Series Switch and BlackDiamond 8800 e-Series
Modules
The 8 slices support all 24 ports.

Platforms with 48 ports have
another set of 8 slices to support
the additional 24 ports.
XM_077
This architecture also allows a single slice to implement ACLs that are applied to more than one port.
When an ACL entry is applied, if its match conditions don’t conflict with an already existing ACL, the
entry is added to the rule memory of an already populated slice. Because the slices are much more
flexible than masks, a much wider variety of rule entries can use the same slice.
When ACLs are applied, the system programs each slice to select parts of the packet information to be
loaded into it. For example, one possible way a slice can be programmed allows it to hold the
information about a packet’s ingress port, source and destination IP address, IP protocol, source and
destination Layer 4 ports, DSCP value, TCP flag, and if it is a first fragment. Any rule entry that
consists of match conditions drawn from that list is compatible with that slice. This list of conditions is
just one example. A complete description of possible ways to program a slice is discussed in
“Compatible and Conflicting Rules” on page 416.
In the following example, the two rule entries are compatible and only require one slice in hardware
even though they are applied to different ports. The following entry is applied to port 1:
entry ex_A {
if {
destination-port 23 ;
protocol tcp ;
} then {
deny ;
}
}
and the following entry is applied to port 2:
entry ex_B {
if {
destination-address 192.168.0.0/16 ;
source-port 1000 ;
protocol tcp ;
} then {
deny ;

Access Lists (ACLs)
}
}
Both of these ACLs could be supported on the same slice, since the match conditions are taken from the
example list discussed earlier. This example is shown in Figure 49. In the example, we refer to slice A,
even though the slices are numbered. Slice A just means that one slice is used, but does not specify a
particular slice. Some rules require more than one slice, so we use letters to show that different slices
are used, but not which specific slices.
Figure 49: ACL Entry ex_A and ex_B
Slice A Rules (128)
XM_078
There are cases where compatible ACLs require using a different slice. If the memory associated with a
slice is filled with rule entries, then another slice will be used to process any other compatible entries.
For example, consider the following 129 rule entries applied to ports 3-7:
entry one {
if {
protocol tcp ;
} then {
deny ;
}
}
entry two {
if {
source-port 1000 ;
protocol tcp ;
} then {
deny ;
}
}
entry three {
if {
protocol upd ;
source-port 100 ;

ACL Mechanisms
} then {
deny ;
}
}
....
[The 125 intervening entries are not displayed in this example]
....
entry onehundred_twentynine {
if {
protocol udp ;
} then {
deny ;
}
}
Figure 50 shows the result of applying the 129 entries. 128 of the entries is applied to one slice, and the
final entry is applied to a different slice. If another compatible entry is applied from another port, for
example, it will use Slice B.
Figure 50: ACL Entry one through onehundred_twentynine
Slice A Rules (128)
Slice B
Rules (128)
XM_079

Access Lists (ACLs)
As entries are configured on the switch, the slices are programmed to implement the rules, and the rule
memory is filled with the matching values for the rules. If a compatible slice is available, each entry is
added to that slice.
Compatible and Conflicting Rules

The slices can support a variety of different ACL match conditions, but there are some limitations on
how you combine the match conditions in a single slice. A slice is divided up into fields, and each field
uses a single selector. A selector is a combination of match conditions or packet conditions that are used
together. To show all the possible combinations we’ve abbreviated the conditions as shown in Table 54.
Table 54: Abbreviations used in Field Selector Tables

Abbreviation Condition
DIP destination address <prefix> (IPv4 addresses only)
SIP source address <prefix> (IPv4 addresses only)
IP-Proto protocol <number>
L4DP destination-port <number> (a single port)
L4SP source-port <number> (a single port)
DSCP dscp <number>
TCP-Flag TCP-flags <bitfield>
First Fragment first-fragments
L4-Range A Layer 4 port range. For example, if you specify “protocol UDP” and “port 200 -
1200” in an entry, you have used a Layer 4 range. There are a total of sixteen
Layer 4 port ranges. Also, you can have a source port range, or a destination port
range, but not both kinds of ranges together in the same entry.
DIPv6/128 destination address <prefix> (IPv6 address with a prefix length longer than 64)
SIPv6/128 source address <prefix> (IPv6 address with a prefix length longer than 64)
DIPv6/64 destination address <prefix> (IPv6 address with a prefix length up to 64)
SIPv6/64 source address <prefix> (IPv6 address with a prefix length up to 64)
NH IPv6 Next Header field. Use protocol <number> to match.
TC IPv6 Traffic Class field. Use dscp <number>
MACDA ethernet-destination-address <mac-address> <mask>
MACSA ethernet-source-address <mac-address>
Etype ethernet-type <number>
VID This is not a match condition used in ACLs, but is used when an ACL is applied to
VLANs. An ACL applied to a port uses a different field selector than an ACL applied
to a VLAN.
TOS ip-tos <number>
Port-list This is not a match condition used in ACLs, but is used when an ACL is applied to
ports, or to all ports (the wildcard ACL). An ACL applied to a port uses a different
field selector than an ACL applied to a VLAN.
packet-type This selector is used internally and not accessible by users through explicit ACLs.
UDF User-defined field. This selector is used internally and not accessible by users
through explicit ACLs.
The tables that follow list all the combinations of match conditions that are available. Table 55 is a
listing of the possible choices for a single slice for the Summit X450a series switch and BlackDiamond
8800 a-series modules. Table 56 lists the same information for the Summit X450e and X250e series

ACL Mechanisms
switches and BlackDiamond 8800 e-series modules. Any number of match conditions in a single row for
a particular field may be matched. For example if Field 1 has row 1 (Port-list) selected, Field 2 has row 8
(MACDA, MACSA, Etype, VID) selected, and Field 3 has row 6 (Packet-type) selected, any combination
of Port-list, MACDA, MACSA, Etype, VID, and Packet-type may be used as match conditions.
If an ACL requires the use of field selectors from two different rows, it must be implemented on two
different slices.
Table 55: Field Selectors, Summit X450a series and BlackDiamond 8800 a-series
Field 1 Field 2 Field 3
Port-list DIP, SIP, IP-Proto, L4DP, L4SP, DSCP, TCP-Flag, First Fragment Fragments
L4DP, L4SP DIP, SIP, IP-Proto, L4DP, L4-Range, DSCP, TCP-Flag, First Fragment DSCP, TCP-Flag
Etype, VID DIP, SIP, IP-Proto, L4-Range, L4SP, DSCP, TCP-Flag, First Fragment VID
Fragments, VID DIPv6/128 IP-Proto, TOS
Etype, IP-Proto SIPv6/128 L4-range
DIPv6/64, SIPv6/64 packet-type
DIPv6/64, NH, TC, TCP-flag
MACDA, MACSA, Etype, VID
MACDA, DIP, Etype, VID
MACSA, SIP, Etype, VID
Table 56: Field Selectors, Summit X450e and X250e series and BlackDiamond 8800 e-series
Field 1 Field 2 Field 3
Port-list DIP, SIP, IP-Proto, L4DP, L4SP, DSCP, TCP-Flag, First Fragment Fragments,
DSCP, TCP-Flag,
VID
L4DP, L4SP DIP, SIP, IP-Proto, L4DP, L4-Range, DSCP, TCP-Flag, First Fragment packet-type,
DSCP, TCP-Flag,
VID
Etype, VID DIP, SIP, IP-Proto, L4-Range, L4SP, DSCP, TCP-Flag, First Fragment
Etype, IP-Proto DIPv6/128
SIPv6/128
MACDA, MACSA, Etype, VID
MACDA, DIP, Etype, VID
MACSA, SIP, Etype, VID
Use Table 55 and Table 56 to determine which ACL entries are compatible. If the entries are compatible,
they can be on the same slice.
For example, the earlier example entries are applied to ports:
entry ex_A {
if {
protocol tcp ;
} then {
deny ;

Access Lists (ACLs)
}
}
entry ex_B {
if {
source-port 1000 ;
} then {
deny ;
}
}
Entry ex_A consists of the following conditions (using the abbreviations from Table 54), SIP, L4DP, and
IP-Proto. Entry ex_B is DIP, L4SP. Since they are applied to ports, the selector for Field 1 is Port-list (the
first item). The selector for Field 2 would be the first item, and Field 3 could be any item.
Our other example entries are also compatible with the entries ex_A and ex_B:
entry one {
if {
protocol tcp ;
} then {
deny ;
}
}
entry two {
if {
source-port 1000 ;
} then {
deny ;
}
}
entry three {
if {
protocol upd ;
source-port 100 ;
} then {
deny ;
}
}
Entry one is SIP, L4DP, and IP-Proto; entry two is DIP, and L4SP; entry three is SIP, DIP, IP-Proto,
L4SP, and L4DP. All of these examples can use the first item in Field 2 in the tables.
However, if we add the following entry:
entry alpha {
if {
ethernet-destination-address 00:e0:2b:11:22:33 ;
} then {

ACL Mechanisms
deny ;
}
}
this will not be compatible with the earlier one. Entry alpha is MACDA, and there is no MACDA in the
first item for Field 2. Any entry with MACDA will have to use selector 7 or 8 from Table 55 (or 6 or 7
from Table 56, depending on the platform). If an entry requires choosing a different selector from the
table, it is not compatible and must go into a different slice.
Rule Evaluation and Actions

When a packet ingresses the switch, its header is loaded into all the slices, and the header value is
compared with the rule values. If the values match, the rule action is taken. Conflicting actions are
resolved by the precedence of the entries. However, if rule entries are on different slices, then ACL
counters can be incremented on each slice that contains a counter-incrementing rule.
Slice and Rule Use by Feature

A number of slices and rules are used by features present on the switch. You do consume these
resources when the feature is enabled.
● dot1p examination - enabled by default - 1 slice, 8 rules per chip
■ Slice A (F1=Port-list, F2=MACDA, MACSA, Etype, VID, F3=packet-type)
● IGMP snooping - enabled by default - 2 slice, 2rules
■ Slice B (F1=Port-list, F2=MACDA, MACSA, Etype, VID, F3=IP-Proto, TOS)
● IP interface - disabled by default - 2 slices, 3 rules (plus IGMP snooping rules above)
■ Slice C (F1=Port-list, F2=SIP, DIP, IP-proto, L4SP, L4DP, DSCP, F3=packet-type)
● VLAN QoS - disabled by default - 1 slice, n rules (n VLANs)
■ Slice A or B (F1=Port-list, F2=MACDA, MACSA, Etype, VID, F3=anything)
● port QoS - disabled by default - 1 slice, 1 rule
■ Slice D (F1=anything, F2=anything, F3=anything)
● VRRP - 2 slices, 2 rules
● EAPS - 1 slice, 1 rule (master), n rules (transit - n domains)
● ESRP - 2 slices, 2 rules
● IPv6 - 1 slice, 2 rules
● Netlogin - 1 slice, 1 rule

Access Lists (ACLs)
● VLAN Mirroring - 1 slice, n rules (n VLANs)

■ Slice E (F1=Port-list, F2=MACDA, MACSA, Etype, VID, F3=anything)
To display the number of slices used by the ACLs on the slices that support a particular port, use the
following command:
show access-list usage acl-slice port <port>
To display the number of rules used by the ACLs on the slices that support a particular port, use the
following command:
show access-list usage acl-rule port <port>
To display the number of Layer 4 ranges used by the ACLs on the slices that support a particular port,
show access-list usage acl-range port <port>
System Configuration Example

The following example shows incremental configurations and their corresponding ACL resource
consumption taken on a BlackDiamond 8800 switch with an a-Series card.
● Default configuration including: dot1p examination and IGMP snooping:
■ 2 slices, 10 rules
● Add an IP interface to the configuration:
● Add port-based QoS to the configuration:
● Add VLAN-based QoS to the configuration:
● Add VRRP to the configuration:
■ 4 slices 16 rules
● Add EAPS (Master mode) to the configuration:
● Add ESRP to the configuration:
● Add IPv6 routing (slowpath) to the configuration:
● Add Netlogin to the configuration:
ACL Error Messages

Errors may happen when installing an ACL policy on a port, VLAN, or wildcard. Here is a list of the
most common error conditions and their resulting CLI error message:
● Slice resource exceeded: This happens when all slices are allocated for a given chip and an
additional incompatible rule (see “Use Table 55 and Table 56 to determine which ACL entries are

Policy Based Routing
compatible. If the entries are compatible, they can be on the same slice.” on page 417) is installed
which requires allocation of another slice.
Error: ACL install operation failed - slice hardware full for port 3:1
● Rule resource exceeded: This happens when all slices are allocated for a given chip and there is an
attempt to install a compatible rule to the lowest precedence slice which already has 128 rules. This
condition can be triggered with less than the full capacity number of rules installed. For example, if
15 of the slices each have less than 128 rules and there is an attempt to install 129 compatible rules,
this error message will be displayed.
Error: ACL install operation failed - rule hardware full for port 3:1
● Layer-4 port range exceeded: This happens when more than 16 Layer 4 port ranges are installed on a
single chip.
Error: ACL install operation failed - layer-4 port range hardware full for port 3:1
● Incompatible fields selected: This happens when the selected conditions can not be satisfied by the
available single-slice field selections described in “Compatible and Conflicting Rules” on page 416.
Error: ACL install operation failed - conditions specified in rule "r1" cannot be
satisfied by hardware on port 3:1
● UDF exceeded: This happens in the rare case that the two available user-defined fields are exceeded
on a given chip. UDF fields are used to qualify conditions which are not natively supported by the
hardware. Currently, these include: ICMP Type and ICMP Code.
Error: ACL install operation failed - user-defined-field (UDF) hardware full for
port 3:1
NOTE
This feature is available on the BlackDiamond 10808, BlackDiamond 12800 series switches, BlackDiamond 8800
a-series and e-series modules, and Summit X450a, X450e, and X250e series switches only (whether or not included
in a SummitStack).
Layer 3 Policy Based Redirect

Policy Based Routing allows you to bypass standard Layer 3 forwarding decisions for certain flows.
Typically, in a Layer 3 environment, when an IP packet hits an Ethernet switch or router, the Layer 3
processing determines the next hop and outgoing interface for the packet based only on the packet's
destination address. The Layer 3 processing does so by looking up the IP Forwarding Table; this
forwarding table itself is populated either by static routes or by routes learned dynamically from
routing protocols such as OSPF and RIP.
With Policy Based Routing, you can configure policies to use a different next hop than what the routing
lookup would have chosen. The switch first compares packets to the ACL rule entries. If there is a
match, the packet is forwarded to the destination identified by the redirect action modifier. If there is no
match, the packet is forwarded based on normal routing, in other words, by looking up a route in the
IP Forwarding Table.

Access Lists (ACLs)
When there is a match with a redirect ACL rule, the matched traffic will be redirected to the nexthop
specified in the action. Note, the IP packet itself will not be modified, but only redirected to the port
where the nexthop entry resides. The original IP destination address and source address are retained in
the packet. The TTL is decremented and the IP checksum is recalculated.
The applications for Policy Based Routing are quite diverse, since the functionality can be used to set
policies on how flows identified by any Layer 2 to Layer 7 field (bounded by the switch's ACL syntax)
are forwarded. Deployment scenarios include:
● Forwarding flows for certain applications, for example, all HTTP traffic to designated server(s).
● Redirecting flows from certain source IP addresses for security and other applications.
Policy Based Routing is implemented using ACLs, so it inherits the capabilities and limitations of ACL.
All the matching conditions used for ACLs can be used for Policy Based Routing. The destination IP
address must be an IPv4 unicast address.
When a switch finds a matching ACL rule, it forwards the packet to the redirect IP address as specified
in the rule without modifying the packet (except as noted above).
The traffic flow will only be redirected after applying the ACL to the port and redirected only when the
redirect IP address’s adjacency is resolved. If the IP ARP table does not have the information to reach
the redirect IP address, the packet is routed based on the Layer 3 routing table. If the switch does not
know how to reach the redirect IP address in the rule, the rule is installed with a warning, but traffic
won’t get redirected until the address is resolved in the IP ARP table. After the address is resolved, the
traffic will be redirected.
Layer 2 Policy Based Redirect

This feature allows matching packets to override the normal forwarding decision and be Layer 2
switched to the specified physical port. This is accomplished via an additional packet ACL lookup.
While similar to the “Policy Based Redirect” Layer 3 feature described above, it differs in that the
packet will not be modified for Layer 3 routing based on a new IP redirect nexthop. Instead, the packet
will use the packet format based on the forwarding decision. If the packet was Layer 2-switched, the
packet will egress the redirect port unmodified. If the packet was Layer 3-switched, the packet will
egress with the Layer 3 packet modifications of the nexthop found by the normal Layer 3 forwarding
lookups. This feature applies to unicast, multicast, and broadcast traffic.
Layer 2 policy based redirect is available only on BlackDiamond 8800 a- and e- series and Summit X450
a- and e- series switches. The following ACL action is added in support of this feature:
redirect-port <port number>
The <port number> argument must be specified in the correct format for the switch platform. On the
BlackDiamond 8800 a- or e- series switches, this argument must be in the format <slot>:<port> and on
the Summit X450 a- and e- series switches, this argument must be in the format <port>. For example,
consider the following ACL policies.
The policy shown below redirects any TCP traffic with source Layer 4 port 81 to physical port 3:2.
entry one {
if {
protocol tcp;
source-port 81;

} then {
count num_pkts_redirected;
redirect-port 3:2;
}
}
The policy shown below redirects any in-profile traffic as defined by the meter configuration to physical
port 14. The out-of-profile traffic would be subject to the action specified in the meter “out-action”
configuration.
entry one {
if {
}then {
meter redirected_traffic;
count num_pkts_redirected;
redirect-port 14;
}
}
If an incorrect port format is used or if the port number specified is out of range, the following error
message will be displayed:
*BD-8810.68 # check policy l2pbr
Error: Policy l2pbr has syntax errors

Line 7 : 12:3 is not a valid port.
BD-8810.70 # check policy l2pbr
Error: Policy l2pbr has syntax errors
Line 7 : 77 is not a valid port.
On a BlackDiamond 8800 platform, if this action is applied to an original series port, the following error
message will appear:
BD-8810.66 # config access-list l2pbr port 3:1
Error: The redirect-port option is not supported by MSM-G8X I/O ports

and G48T, G48P, G24X, and 10G4X modules, which have been detected in
Slots 3, 5, 6, 8, 9.
If this option was applied to any or to a vlan, retry by
applying the redirect-port option to a port that is
not on unsupported modules.
When this feature is used, the traffic egressing the redirect-port can either be tagged or untagged
depending on the redirect-port VLAN configuration. Table 57 provides the details. The XGS2 row
applies only to BlackDiamond 8800 series mixed mode systems.
Table 57: VLAN Format of Traffic Egressing Redirect-Port

Redirect-Port Not in Egress Redirect-Port Tagged in Redirect-Port Untagged in
Redirect-Port Type VLAN Egress VLAN Egress VLAN
XGS2 Dropped VLAN Tagged Untagged
XGS3 Dropped VLAN Tagged Untagged

Access Lists (ACLs)
Please be aware of the following important implementation notes:

● Using the “redirect-port” action with a disabled port causes traffic to be dropped.
● Using the “redirect-port” action overrides Layer 2 echo kill; the result is that a packet can be made to
egress the ingress port at Layer 2.
● For systems with XGS3 hardware that has the larger table size, packets with IP options will not
match ACLs using the “redirect-port” action. Systems with XGS3 hardware that has the smaller table
size do not have this capability. On these systems, packets with IP options will match ACLs that use
the “redirect-port” action, and be dropped.
NOTE
The BlackDiamond 8800 a-series and e-series modules support Policy Based Routing, but the original series
modules do not. Similarly, within a SummitStack, the Summit X450a, X450e, and X250e series switches support
Policy Based Routing, but the Summit X450 series switches do not. If you apply an ACL that implements Policy
Based Routing in a mixed-module environment, the ACL will be rejected if it applies to original series modules or
switches. Reapply the ACL only to a-series and e-series modules or switches. Since the a-series and e-series
modules have different amounts of hardware resources, an ACL can also be rejected if one module or switch has
exhausted its resources.
Configuring Policy Based Routing
NOTE
This feature is available on the BlackDiamond 10808, BlackDiamond 12800 series switches, BlackDiamond 8800
a-series and e-series modules, and Summit X450a, X450e, and X250e series switches (whether or not included in a
SummitStack) only.
To configure Policy Based Routing, you will configure an ACL on your switch. You can apply an ACL
policy file, or use a dynamic ACL. The following is an example ACL rule entry that redirects any TCP
traffic with a destination port of 81 to the device at IP address 3.3.3.2:
entry redirect_port_81 {
if {
protocol tcp;
} then {
redirect 3.3.3.2;
}
}
Do the following steps:

1 Issue the following command to prevent the redirect IP address from clearing from the IP ARP table
due to a timeout:
enable iparp refresh
2 Configure the ACL, either applying an ACL policy file similar to the example, or a Dynamic ACL.
3 Ping or send traffic so that the redirect IP adjacency is resolved.
You may want to create a static ARP entry for the redirect IP address, so that there will always be a
cache entry.

ACL Troubleshooting—BlackDiamond 8800, SummitStack, and Summit Family of Switches
ACL Troubleshooting—BlackDiamond 8800,

SummitStack, and Summit Family of Switches
The following commands are designed to help troubleshoot and resolve ACL configuration issues.
*switch # show access-list usage [acl-mask | acl-rule | acl-slice | acl-
range] port <port>
show access-list usage <TAB>
acl-mask ACL Mask table resource summary
acl-range ACL Range table resource summary
acl-rule ACL Rule table resource summary
acl-slice ACL slice resource summary
The “acl-mask” keyword is not relevant for the a-Series or e-Series models. If you enter this command
and specify an a-Series or e-Series port, the following error message appears:
This command is not applicable to the specified port.
Use the “acl-rule” keyword to display the total number of ACL rules that are available and consumed
for the specified port. If this keyword is specified on an a-Series or e-Series port, the first part of the
command output details the port list using this resource because the ACL hardware rules are shared by
all ports on a given ASIC (24x1G ports). If you enter the same command and specify any of the listed
ports, the command output is identical.
*switch # show access-list usage acl-rule port 4:1

Ports 4:1-4:12, 4:25-4:36
Total Rules: Used: 46 Available: 2002
The “acl-slice” keyword is used to display ACL resource consumption for each of the independent
TCAMs, or slices, that make up the hardware ACLs. Each slice is a 128-entry TCAM. The command
output displays the number of consumed and available TCAM rules for each slice as follows.
*switch # show access-list usage acl-slice port 4:1

Ports 4:1-4:12, 4:25-4:36
Slices: Used: 8 Available: 8
Slice 0 Rules: Used: 1 Available: 127
Use the “acl-range” keyword to view the Layer-4 port range hardware resource on an a-Series or e-
Series model switch. Each a-Series and e-Series ASIC has 16 Layer-4 port range checkers that are shared
among the 24 1G ports. The first part of the command output lists the ports that utilizes this resource.

Access Lists (ACLs)
The second part of the command output lists the number of range checkers that are consumed and the
number available for use.
switch # show access-list usage acl-range port 4:1

Ports 4:1-4:12, 4:25-4:36
L4 Port Ranges: Used: 0 Available: 16
If the “acl-slice” or “acl-range” keyword is specified with an e-Series port, the following error message
will appear:
This command is not applicable to the specified port.

19 Routing Policies

● Routing Policies on page 427
● Routing Policy File Syntax on page 427
■ Policy Match Conditions on page 429
■ Policy Action Statements on page 431
● Applying Routing Policies on page 432
● Policy Examples on page 432
Routing Policies
Routing policies are used to control the advertisement or recognition of routes communicated by
routing protocols, such as Routing Information Protocol (RIP), Open Shortest Path First (OSPF), or
Border Gateway Protocol (BGP). Routing policies can be used to “hide” entire networks or to trust only
specific sources for routes or ranges of routes. The capabilities of routing policies are specific to the type
of routing protocol involved, but these policies are sometimes more efficient and easier to implement
than access lists.
Routing policies can also modify and filter routing information received and advertised by a switch.
A similar type of policy is an ACL policy, used to control, at the hardware level, the packets accessing
the switch. ACL policy files and routing policy files are both handled by the policy manager and the
syntax for both types of files is checked by the policy manager.
NOTE
Routing Policy File Syntax

A routing policy file contains one or more policy rule entries. Each routing policy entry consists of:
● A policy entry rule name, unique within the same policy.
● Zero or one match type. If no type is specified, the match type is all, so all match conditions must be
satisfied.
● Zero or more match conditions. If no match condition is specified, then every routing entity matches.
● Zero or more actions. If no action is specified, the packet is permitted by default.

Routing Policies
Each policy entry in the file uses the following syntax:
entry <routingrulename>{
if <match-type> {
<match-conditions>;
} then {
<action>;
}
}
Here is an example of a policy entry:
entry ip_entry {
if match any {
nlri 10.203.134.0/24;
nlri 10.204.134.0/24;
} then {
next-hop 192.168.174.92;
origin egp;
}
}
Policy entries are evaluated in order, from the beginning of the file to the end, as follows:
● If a match occurs, the action in the then statement is taken:
■ if the action contains an explicit permit or deny, the evaluation process terminates.
■ if the action does not contain an explicit permit or deny, then the action is an implicit permit, and
the evaluation process terminates.
● If a match does not occur, then the next policy entry is evaluated.
● If no match has occurred after evaluating all policy entries, the default action is deny.
Often a policy will have a rule entry at the end of the policy with no match conditions. This entry will
match anything not otherwise processed, so that user can specify an action to override the default deny
action.
The next sections list detailed information about policy match conditions, about matching BGP AS
paths, and about action statements. For information on those subjects, see the following sections:
● Policy Match Conditions on page 429
● Autonomous system expressions on page 429
● Policy Action Statements on page 431
Policy Match Type

The two possible choices for the match type are:
● match all—All the match conditions must be true for a match to occur. This is the default.
● match any—If any match condition is true, then a match occurs.

Routing Policies
Policy Match Conditions

Table 58 lists the possible policy entry match conditions.
Table 58: Policy match conditions
Match Condition Description
as-path [<as-number> | <as-path-regular- Where <as-number> is a valid Autonomous system number
expression>]; in the range [1 - 65535].
Where <as-path-regular-expression> is a multi-character
regular expression (with 2-byte unsigned Integer being an
Atom). Regular expression will consist of the AS-Numbers
and various regular expression symbols. Regular
expressions must be enclosed in double quotes ("").
community [no-advertise | no-export | no-export- Where no-advertise, no-export and no-export-subconfed are
subconfed | number <community_num> | the standard communities defined by RFC.
<community_regular_expression> | <community_num> is a four byte unsigned integer,
<as_num> : <num>]; <as_num> is a two byte AS-Number and <num> is the
2-bytes community number.
Community regular expression is a multi-character regular
expression (with four byte unsigned integer being an
Atom). Regular expression is enclosed in double quotes
("").
med <number>; Where <number> is a 4-byte unsigned integer.
next-hop [<ipaddress> | <ipaddress-regular- Where <ipaddress> is a valid IP address in dotted decimal
expression>]; format.
nlri [<ipaddress> | any]/<mask-length> {exact}; Where <ipaddress> and <mask> are IP addresses, <mask-
nlri [<ipaddress> | any] mask <mask> {exact}; length> is an integer, and keyword any matches any IP
address with a given (or larger) mask/mask-length.
origin [igp | egp | incomplete]; Where igp, egp and incomplete are the Border Gateway
Protocol (BGP) route origin values.
tag <number>; Where <number> is a 4-byte unsigned number.
route-origin [direct | static | icmp | egp | ggp | hello Matches the origin (different from BGP route origin) of a
| rip | isis | esis | cisco-igrp | ospf | bgp | idrp | route.
dvmrp | mospf | pim-dm | pim-sm | ospf-intra |
A match statement "route-origin bgp" will match routes
ospf-inter | ospf-extern1 | ospf-extern2 | bootp | e-
whose origin are "I-bgp" or "e-bgp" or "I-mbgp" or "e-
bgp | i-bgp | mbgp | i-mbgp | e-mbgp | isis-level-1 |
mbgp". Similarly, the match statement "route-origin ospf"
isis-level-2 | isis-level-1-external | isis-level-2-
will match routes whose origin is "ospf-inta" or "ospf-inter"
external]
or "ospf-as-external" or "ospf-extern-1" or "ospf-extern-2"
Autonomous system expressions. The AS-path keyword uses a regular expression string to match against
the autonomous system (AS) path. Table 59 lists the regular expressions that can be used in the match
conditions for Border Gateway Path (BGP) AS path and community. Table 60 shows examples of regular
expressions and the AS paths they match.
Table 59: AS regular expression notation
Character Definition
N As number
N1 - N2 Range of AS numbers, where N1 and N2 are AS numbers and N1 < N2
[Nx ... Ny] Group of AS numbers, where Nx and Ny are AS numbers or a range of AS numbers
[^Nx ... Ny] Any AS numbers other than the ones in the group
. Matches any number
^ Matches the beginning of the AS path

Routing Policies
Table 59: AS regular expression notation (Continued)

Character Definition
$ Matches the end of the AS path
– Matches the beginning or end, or a space
- Separates the beginning and end of a range of numbers
* Matches 0 or more instances
+ Matches 1 or more instances
? Matches 0 or 1 instance
{ Start of AS SET segment in the AS path
} End of AS SET segment in the AS path
( Start of a confederation segment in the AS path
) End of a confederation segment in the AS path
Table 60: Policy regular expression examples

Attribute Regular Expression Example Matches
AS path is 1234 “1234” 1234
Zero or more occurrences of AS number 1234 “1234*” 1234
1234 1234
Start of As path set “10 12 { 34” 10 12 34 { 99
33 10 12 { 34 37
End of As path set “12 } 34” 12 } 34 56
Path that starts with 99 followed by 34 “^99 34 “ 99 34 45
Path that ends with 99 “99 $” 45 66 99
Path of any length that begins with AS numbers 4, 5, “4 5 6 .*” 456456789
6
Path of any length that ends with AS numbers 4, 5, 6 “.* 4 5 6 $” 456
123456
Here are some additional examples of using regular expressions in the AS-Path statement.
The following AS-Path statement matches AS paths that contain only (begin and end with) AS number
65535:
as-path "^65535$"
The following AS-Path statement matches AS paths beginning with AS number 65535, ending with AS
number 14490, and containing no other AS paths:
as-path "^65535 14490$"
The following AS-Path statement matches AS paths beginning with AS number 1, followed by any AS
number from 2 - 8, and ending with either AS number 11, 13, or 15:
as-path "^1 2-8 [11 13 15]$"
The following AS-Path statement matches AS paths beginning with AS number 111 and ending with
any AS number from 2 - 8:
as-path "111 [2-8]$"

Routing Policies
The following AS-Path statement matches AS paths beginning with AS number 111 and ending with
any additional AS number, or beginning and ending with AS number 111:
as-path "111 .?"
Policy Action Statements

Table 61 lists the possible action statements. These are the actions taken when the policy match
conditions are met in a policy entry.
Table 61: Policy actions

Action Description
as-path "<as_num> {<as_num1> <as_num2> Prepends the entire list of as-numbers to the as-path of
<as_num3> …. <as_numN>}"; the route.
community set [no-advertise | no-export | no-export- Replaces the existing community attribute of a route by
subconfed | <community_num> the communities specified by the action statement.
{<community_num1> <community_num2> …. Communities must be enclosed in double quotes ("").
<community_numN>} | <as_num> :
<community_num> [<as_num1>
<community_num1> <as_num2>
<community_num2> ….}];
community [add | delete] [no-advertise | no-export | Adds/deletes communities to/from a route's community
no-export-subconfed | <community_num> attribute. Communities must be enclosed in double quotes
{<community_num1> <community_num2> …. ("").
<community_numN>} | <as_num> :
<community_num> {<as_num1>
<community_num1> <as_num2>
<community_num2> ….}];
community remove; Strips off the entire community attribute from a route.
Communities must be enclosed in double quotes ("").
cost <cost(0-4261412864)>; Sets the cost/metric for a route.
cost-type {ase-type-1 | ase-type-2 | external | Sets the cost type for a route.
internal};
dampening half-life <minutes (1-45)> reuse-limit Sets the BGP route flap dampening parameters.
<number (1-20000)> suppress-limit <number (1-
20000)> max-suppress <minutes (1-255)>;
deny; Denies the route.
local-preference <number>; Sets the BGP local preference for a route.
med {add | delete} <number>; Performs MED arithmetic. Add means the value of the
MED in the route will be incremented by <number>, and
delete means the value of the MED in the route will be
decremented by <number>.
med {internal | remove}; Internal means that the Interior Gateway Protocol (IGP)
distance to the next hop will be taken as the MED for a
route. Remove means take out the MED attribute from the
route.
med set <number>; Sets the MED attribute for a route.
next-hop <ipaddress>; Sets the next hop attribute for a route.
nlri [<ipaddress> | any]/<mask-length> {exact}; These set statements are used for building a list of IP
nlri [<ipaddress> | any] mask <mask> {exact}; addresses. This is used by PIM to set up the RP list.
origin {igp | egp | incomplete}; Sets the BGP route origin values.
permit; Permits the route.

Routing Policies
Table 61: Policy actions (Continued)

Action Description
tag <number>; Sets the tag number for a route.
weight <number> Sets the weight for a BGP route.
Applying Routing Policies

To apply a routing policy, use the command appropriate to the client. Different protocols support
different ways to apply policies, but there are some generalities. Policies applied with commands that
use the keyword import-policy control the routes imported to the switch routing table by the
protocol. The following are examples for the BGP and RIP protocols:
configure bgp import-policy [<policy-name> | none]
configure rip import-policy [<policy-name> | none]
Commands that use the keyword route-policy control the routes advertised or received by the
protocol. For BGP and RIP, here are some examples:

multicast]} route-policy [in | out] [none | <policy>]
configure bgp peer-group <peer-group-name> route-policy [in | out] [none | <policy>]
configure rip vlan [<vlan-name> | all] route-policy [in | out] [<policy-name> | none]
Other examples of commands that use route policies include:
configure ospf area <area-identifier> external-filter [<policy-map> |none]

configure ospf area <area-identifier> interarea-filter [<policy-map> | none]
configure rip vlan [<vlan-name> | all] trusted-gateway [<policy-name> | none]
To remove a routing policy, use the none option in the command.
Policy Examples
The following sections contain examples of policies. The examples are:
● Translating an access profile to a policy on page 432
● Translating a route map to a policy on page 434
Translating an access profile to a policy

You may be more familiar with using access profiles on other Extreme Networks switches. This
example shows the policy equivalent to an ExtremeWare access profile.
ExtremeWare Access-Profile:
Seq_No Action IP Address IP Mask Exact

5 permit 22.16.0.0 255.252.0.0 No
10 permit 192.168.0.0 255.255.192.0 Yes
15 deny any 255.0.0.0 No
20 permit 10.10.0.0 255.255.192.0 No
25 deny 22.44.66.0 255.255.254.0 Yes

Routing Policies
Equivalent ExtremeXOS policy map definition:
entry entry-5 {
If {
nlri 22.16.0.0/14;
}
then {
permit;
}
}
entry entry-10 {
if {
nlri 192.168.0.0/18 exact;
}
then {
permit;
}
}
entry entry-15 {
if {
nlri any/8;
}
then {
deny;
}
}
entry entry-20 {
if {
nlri 10.10.0.0/18;
}
then {
permit;
}
}
entry entry-25 {
if {
nlri 22.44.66.0/23 exact;
}
then {
deny;
}
}
The policy above can be optimized by combining some of the if statements into a single expression. The
compact form of the policy will look like this:
entry permit_entry {
If match any {
nlri 22.16.0.0/14;
nlri 192.168.0.0/18 exact ;
nlri 10.10.0.0/18;
}

Routing Policies
then {
permit;
}
}
entry deny_entry {
if match any {
nlri any/8;
nlri 22.44.66.0/23 exact;
}
then {
deny;
}
}
Translating a route map to a policy

You may be more familiar with using route maps on other Extreme Networks switches. This example
shows the policy equivalent to an ExtremeWare route map.
ExtremeWare route map:
Route Map : rt
Entry : 10 Action : permit
match origin incomplete
Entry : 20 Action : deny
match community 6553800
match med 30
set next-hop 10.201.23.10
set as-path 20
set as-path 30
set as-path 40
set as-path 40
set local-preference 120
set weight 2
match origin incomplete
match community 19661200
set dampening half-life 20 reuse-limit 1000 suppress-limit 3000 max-suppress
40
match next-hop 192.168.1.5
set community add 949616660
Here is the equivalent policy:
entry entry-10 {
If {
origin incomplete;
}
then {
permit;
}

Routing Policies
entry entry-20 {
if {
community 6553800;
}
then {
deny;
}
}
entry entry-30 {
if {
med 30;
}
then {
next-hop 10.201.23.10;
as-path 20;
as-path 30;
as-path 40;
as-path 40;
permit;
}
}
entry entry-40 {
if {
}
then {
local-preference 120;
weight 2;
permit;
}
}
entry entry-50 match any {

if {
origin incomplete;
community 19661200;
}
then {
dampening half-life 20 reuse-limit 1000 suppress-limit 3000 max-suppress 40
permit;
}
}
entry entry-60 {
if {
next-hop 192.168.1.5;
}
then {
community add 949616660;
permit;
}
}

Routing Policies
entry deny_rest {
if {
}
then {
deny;
}
}

20 Quality of Service

● Overview of Policy-Based Quality of Service on page 437
● Applications and Types of QoS on page 438
● Configuring QoS on page 440
● QoS Profiles on page 441
● Traffic Groupings on page 444
● Verifying QoS Configuration and Performance on page 455
● Guidelines for Configuring QoS on page 457
● Metering Using ACLs—BlackDiamond 8800 Series, SummitStack, the Summit Family of Switches,
and BlackDiamond 12800 R-Series Switch Only on page 457
● Egress Traffic Rate Limiting—BlackDiamond 8800 Series Switches, SummitStack, and the Summit
Family of Switches Only on page 459
● Applying Egress Bandwidth to a Port—BlackDiamond 10808 and 12800 Series Switches Only on
page 460
● Applying Egress Bandwidth to a QoS Queue on page 461
● Bi-Directional Rate Shaping—BlackDiamond 10808 Switch Only on page 461
● Hierarchical QoS—BlackDiamond 12800 R-Series Switch Only on page 464
Policy-based Quality of Service (QoS) is a feature of ExtremeXOS and the Extreme Networks switch
architecture that allows you to specify different service levels for traffic traversing the switch. Policy-
based QoS is an effective control mechanism for networks that have heterogeneous traffic patterns.
Using Policy-based QoS, you can specify the service level that a particular traffic type receives.
NOTE
Specifying the level of service for specified traffic types may be referred to as Class of Service (CoS).
Overview of Policy-Based Quality of Service

Policy-based Quality of Service (QoS) is a feature of ExtremeXOS and the Extreme Networks switch
architecture that allows you to specify different service levels for traffic traversing the switch. Policy-
based QoS is an effective control mechanism for networks that have heterogeneous traffic patterns.
Using Policy-based QoS, you can specify the service level that a particular traffic type receives.
Policy-based QoS allows you to protect bandwidth for important categories of applications or to
specifically limit the bandwidth associated with less critical traffic.
For example, if voice-over-IP (VoIP) traffic requires a reserved amount of bandwidth to function
properly, using policy-based QoS, you can reserve sufficient bandwidth critical to this type of
application. Other applications deemed less critical can be limited so as to not consume excessive
bandwidth.

Quality of Service
On the BlackDiamond 10808 and the 12800 series switches, the switch contains separate hardware
queues on every physical port. On the BlackDiamond 8800 series switches, SummitStack, and the
Summit family of switches, the switch has two default queues (based on flows), and you can configure
up to six additional queues. Each queue is programmed by ExtremeXOS with specific parameters that
modify the forwarding behavior of the switch and affect how the switch transmits traffic for a given
queue on a physical port.
The switch tracks and enforces the specified parameters on every queue for every port. When two or
more queues on the same physical port are contending for transmission, the switch prioritizes use so
long as the respective queue management parameters are satisfied. Up to eight queues per port are
available.
NOTE
Policy-based QoS has no impact on switch performance. Using even the most complex traffic groupings has no cost
in terms of switch performance.
Applications and Types of QoS

Different applications have different QoS requirements. The following applications are ones that you
will most commonly encounter and need to prioritize:
● Voice applications
● Video applications
● Critical database applications
● Web browsing applications
● File server applications
General guidelines for each traffic type are given below and summarized in Table 62. Consider them as
general guidelines and not as strict recommendations. After QoS parameters have been set, you can
monitor the performance of the application to determine if the actual behavior of the applications
matches your expectations. It is very important to understand the needs and behavior of the particular
applications you want to protect or limit. Behavioral aspects to consider include bandwidth needs,
sensitivity to latency and jitter, and sensitivity and impact of packet loss.
Voice Applications
Voice applications, or voice over IP (VoIP), typically demand small amounts of bandwidth. However,
the bandwidth must be constant and predictable because voice applications are typically sensitive to
latency (inter-packet delay) and jitter (variation in inter-packet delay). The most important QoS
parameter to establish for voice applications is minimum bandwidth, followed by priority.
Video Applications
Video applications are similar in needs to voice applications, with the exception that bandwidth
requirements are somewhat larger, depending on the encoding. It is important to understand the
behavior of the video application being used. For example, in the playback of stored video streams,
some applications can transmit large amounts of data for multiple streams in one “spike,” with the

Applications and Types of QoS
expectation that the endstations will buffer significant amounts of video-stream data. This can present a
problem to the network infrastructure, because the network must be capable of buffering the
transmitted spikes where there are speed differences (for example, going from gigabit Ethernet to Fast
Ethernet). Key QoS parameters for video applications include minimum bandwidth and priority, and
possibly buffering (depending upon the behavior of the application).
Critical Database Applications

Database applications, such as those associated with Enterprise Resource Planning (ERP), typically do
not demand significant bandwidth and are tolerant of delay. You can establish a minimum bandwidth
using a priority less than that of delay-sensitive applications.
Web Browsing Applications

QoS needs for Web browsing applications cannot be generalized into a single category. For example,
ERP applications that use a browser front-end may be more important than retrieving daily news
information. Traffic groupings can typically be distinguished from each other by their server source and
destinations. Most browser-based applications are distinguished by the dataflow being asymmetric
(small dataflows from the browser client, large dataflows from the server to the browser client).
An exception to this may be created by some Java™ -based applications. In addition, Web-based
applications are generally tolerant of latency, jitter, and some packet loss; however, small packet loss
may have a large impact on perceived performance because of the nature of TCP. The relevant
parameter for protecting browser applications is minimum bandwidth. The relevant parameter for
preventing non-critical browser applications from overwhelming the network is maximum bandwidth.
File Server Applications

With some dependencies on the network operating system, file serving typically poses the greatest
demand on bandwidth, although file server applications are very tolerant of latency, jitter, and some
packet loss, depending on the network operating system and the use of TCP or UDP.
NOTE
Full-duplex links should be used when deploying policy-based QoS. Half-duplex operation on links can make delivery
of guaranteed minimum bandwidth impossible.

Quality of Service
Table 62 summarizes QoS guidelines for the different types of network traffic.
Table 62: Traffic type and QoS guidelines
Traffic type Key QoS parameters
Voice Minimum bandwidth, priority
Video Minimum bandwidth, priority, buffering (varies)
Database Minimum bandwidth
Web browsing Minimum bandwidth for critical applications, maximum bandwidth for noncritical
applications
File server Minimum bandwidth
Configuring QoS
NOTE
With software version 11.0, you can create access control lists (ACLs) with QoS actions. The QoS forwarding
information you configured in an ACL takes precedence over QoS configuration using the CLI commands.
NOTE
For information on policy files and ACLs, see Chapter 13, “Routing Policies,” and Chapter 12, “Access Lists
(ACLs),” respectively.
To configure QoS, you define how your switch responds to different categories of traffic by creating and
configuring QoS profiles. You then group traffic into categories (according to the needs of the
application, as previously discussed) and assign each category to a QoS profile. Configuring QoS is a
three-step process:
1 Configure the QoS profile.
QoS profile—A class of service that is defined through minimum and maximum bandwidth
parameters and prioritization settings on the BlackDiamond 10808 and 12800 series switches or
through configuration of buffering and scheduling settings on the BlackDiamond 8800 series
switches, SummitStack, and the Summit family of switches. The level of service that a particular type
of traffic or traffic grouping receives is determined by assigning it to a QoS profile. The names of the
QoS profiles are QP1 through QP8; these names are not configurable.
2 Create traffic groupings.
Traffic grouping—A classification or traffic type that has one or more attributes in common. These
can range from a physical port to IP Layer 4 port information. You assign traffic groupings to QoS
profiles to modify switch forwarding behavior. Traffic groupings transmitting out the same port that
are assigned to a particular QoS profile share the assigned characteristics and hence share the class
of service.
3 Monitor the performance of the application with the QoS monitor to determine whether the policies
are meeting the desired results.

QoS Profiles
Configuring QoS on the BlackDiamond 8800 Series Switches,

The BlackDiamond 8800 series switches, SummitStack, and the Summit family of switches allow
dynamic creation and deletion of QoS queues, with Q1 and Q8 always available, rather than the 8 fixed
queues on the BlackDiamond 10808 and 12800 series switches.
NOTE
The sFlow application uses QP2 to sample traffic on the BlackDiamond 8800 series switches, SummitStack, and
the Summit series switch (whether or not included in a SummitStack); any traffic grouping using QP2 may
encounter unexpected results when sFlow is enabled on these specific devices.
The following considerations apply only to QoS on the BlackDiamond 8800 series switches,
SummitStack, and the Summit family of switches:
● The following QoS features share resources on the BlackDiamond 8800 series switches, SummitStack,
and the Summit family of switches:
■ ACLs
■ DiffServ
■ dot1p
■ VLAN-based QoS
■ Port-based QoS
● You may receive an error message when configuring a QoS feature in the above list on the
BlackDiamond 8800 series switches, SummitStack, and the Summit family of switches; it is possible
that the shared resource is depleted. In this case, unconfigure one of the other QoS features and
reconfigure the one you are working on.
● On a SummitStack, you cannot create QoS profiles QP6 and QP7. These are reserved for system
control traffic.
The next sections describe each of these QoS components in detail.
QoS Profiles
QoS profiles are configured differently:
● On the BlackDiamond 8800 series switches, SummitStack, and the Summit family of switches
● On the BlackDiamond 10808 and 12800 series switches.
QoS Profiles on the BlackDiamond 8800 Series Switches,

The BlackDiamond 8800 series switches, SummitStack, and the Summit family of switches have two
default queues, QP1 and QP8, which are based on traffic flows. QP1 has the lowest priority, and QP8
has the highest priority. You can configure up to six additional QoS profiles, or queues, on the switch,
QP2 through QP7. However, on a SummitStack, you cannot create QoS profiles QP6 and QP7; these
profiles are reserved for system control traffic. Creating a queue dynamically will not cause loss of

Quality of Service
traffic. You can also modify the default parameters of each QoS profile. The names of the QoS profiles,
QP1 through QP8, are not configurable.
The parameters that make up a QoS profile on the BlackDiamond 8800 series switches, SummitStack,
and the Summit family of switches include:
● Buffer—This parameter is the maximum amount of packet buffer memory available to all packets
associated with the configured QoS profile within all affected ports. All QoS profiles use 100% of
available packet buffer memory by default. However, in a SummitStack, the scheduling algorithm is
automatically programmed by ExtremeXOS for the stacking links only, and might be different from
the algorithm you select. You can configure the buffer amount from 1 to 100%, in whole integers.
Regardless of the maximum buffer setting, the system does not drop any packets if any packet buffer
memory remains to hold the packet and the current QoS profile buffer use is below the maximum
setting.
NOTE
Use of all eight queues on all ports may result in insufficient buffering to sustain 0 packet loss throughput
during full-mesh connectivity with large packets.
● Weight—This parameter is the relative weighting for each QoS profile; 1 through 16 are the available
weight values. The default value for each QoS profile is 1, giving each queue equal weighting. When
you configure a QoS profile with a weight of 4, that queue is serviced four times as frequently as a
queue with a weight of 1. However, if you configure all QoS profiles with a weight of 16, each queue
is serviced equally but for a longer period of time.
Finally, you configure the scheduling method that the entire switch will use to empty the queues. The
scheduling applies globally to the entire switch, not to each port. However, in a SummitStack, the
scheduling algorithm is automatically programmed by ExtremeXOS for the stacking links only, and will
likely be different than the algorithm you select. You can configure the scheduling to be strict priority,
which is the default, or weighted round robin. In the strict priority method, the switch services the
higher-priority queues first. As long as a queued packet remains in a higher-priority queue, any lower-
priority queues are not serviced. If you configure the switch for weighted-round-robin scheduling, the
system services all queues based on the weight assigned to the QoS profile. The hardware services
higher-weighted queues more frequently, but lower-weighted queues continue to be serviced at all
times.
When configured to do so, the priority of a QoS profile can determine the 802.1p bits used in the
priority field of a transmitted packet (see “Replacing 802.1p priority information” on page 448). The
priority of a QoS profile determines the DiffServ code point value used in an IP packet when the packet
is transmitted (see “Replacing DiffServ code points” on page 451).
A QoS profile switch does not alter the behavior of the switch until it is assigned to a traffic grouping.
The default QoS profiles cannot be deleted. The settings for the default QoS parameters on the
BlackDiamond 8800 series switches, SummitStack, and the Summit family of switches are summarized
in Table 63.
Table 63: Default QoS profile parameters on the BlackDiamond 8800 series switches, SummitStack,
and the Summit family of switches
Profile name Priority Buffer Weight
QP1 Low 100% 1
QP8 High 100% 1

QoS Profiles
Bandwidth on QoS Profiles for the BlackDiamond a-series and e-series Modules and
Summit X450a and X450e series Switch
Beginning with ExtremeXOS 11.5, you can configure egress bandwidth per QoS profile (or queue) on
the following devices:
● BlackDiamond a-series modules
● BlackDiamond e-series modules
● Summit X450a series switch
The egress bandwidth parameters for each QoS profile include the following:
● Minimum bandwidth—The minimum total link bandwidth that is reserved for use by a hardware
queue on a physical port (each physical port has eight hardware queues, corresponding to a QoS
profile). The minimum bandwidth value is configured either as a percentage of the total link
bandwidth or using absolute committed rates in Kbps or Mbps. Bandwidth unused by the queue can
be used by other queues. The minimum bandwidth for all queues should add up to less than 100%.
The default value on all minimum bandwidth parameters is 0%.
● Maximum bandwidth—The maximum total link bandwidth that can be transmitted by a hardware
profile). The default value on all maximum bandwidth parameters is 100%. The maximum
bandwidth value can be configured as either:
■ an absolute percentage of the total maximum link speed, regardless of the currently configured or
negotiated speed
OR
■ an absolute peak rate in Mbps or Kbps
QoS Profiles on the BlackDiamond 10808 and

12800 Series Switches
The BlackDiamond 10808 and 12800 series switches have eight hardware queues for each egress port.
The QoS profiles, QP1 to QP8, map to these hardware queues.
A QoS profile on the BlackDiamond 10808 and 12800 series switches defines a class of service by
specifying traffic behavior attributes, such as bandwidth. The parameters that make up a QoS profile on
the BlackDiamond 10808 and 12800 series switches include:
● Minimum bandwidth—The minimum total link bandwidth that is reserved for use by a hardware
profile). The minimum bandwidth value is configured either as a percentage of the total link
bandwidth or using absolute committed rates in Kbps or Mbps. Bandwidth unused by the queue can
be used by other queues. The minimum bandwidth for all queues should add up to less than 100%.
The default value on all minimum bandwidth parameters is 0%.
● Maximum bandwidth—The maximum total link bandwidth that can be transmitted by a hardware
profile). The maximum bandwidth value is configured either as a percentage of the total link
bandwidth or using absolute peak rates in Kbps or Mbps. The default value on all maximum
bandwidth parameters is 100%.

Quality of Service
● Priority—The level of priority assigned to a hardware egress queue on a physical port. There are
eight different available priority settings and eight different hardware queues. By default, each of the
default QoS profiles is assigned a unique priority. You use prioritization when two or more
hardware queues on the same physical port are contending for transmission on the same physical
port, only after their respective bandwidth management parameters have been satisfied. If two
hardware queues on the same physical port have the same priority, a round-robin algorithm is used
for transmission, depending on the available link bandwidth.
■ When configured to do so, the priority of a QoS profile can determine the 802.1p bits used in the
priority field of a transmitted packet (see “Replacing 802.1p priority information” on page 448).
■ The priority of a QoS profile determines the DiffServ code point value used in an IP packet when
the packet is transmitted (see “Replacing DiffServ code points” on page 451).
A QoS profile does not alter the behavior of the switch until it is assigned to a traffic grouping. Recall
that QoS profiles on the BlackDiamond 10808 and 12800 series switches are linked to hardware queues.
There are multiple hardware queues per physical port. By default, a QoS profile links to the identical
hardware queue across all the physical ports of the switch.
The default QoS profiles cannot be deleted. Also by default, a QoS profile maps directly to a specific
hardware queue across all physical ports. The settings for the default QoS parameters on the
BlackDiamond 10808 and 12800 series switches are summarized in Table 64.
Table 64: Default QoS profile parameters on the BlackDiamond 10808 and 12800 series switches
Profile name Hardware queue Priority Minimum bandwidth Maximum bandwidth
QP1 Q0 Low 0% 100%
QP2 Q1 LowHi 0% 100%
QP3 Q2 Normal 0% 100%
QP4 Q3 NormalHi 0% 100%
QP5 Q4 Medium 0% 100%
QP6 Q50 MediumHi 0% 100%
QP7 Q6 High 0% 100%
QP8 Q7 HighHi 0% 100%
Traffic Groupings
After a QoS profile has been created or modified, you assign a traffic grouping to the profile. A traffic
grouping is a classification of traffic that has one or more attributes in common. Traffic is typically
grouped based on the needs of the applications discussed starting on page 438.
Traffic groupings are separated into the following categories for discussion:
● ACL-based information
● Explicit packet class of service information, such as 802.1p or DiffServ (IP TOS)
● Physical/Logical configuration (physical source port or VLAN association)

Traffic Groupings
Precedence of Traffic Groupings

In the event that a given packet matches two or more grouping criteria, there is a predetermined
precedence for which traffic grouping applies. In general, the more specific traffic grouping takes
precedence. Those groupings listed at the top of the table are evaluated first. By default, all traffic
groupings are placed in the QoS profile QP1. The groupings are listed in order of precedence (highest to
lowest). The three main types of traffic groupings are described in detail on the following pages.
Because the hardware handles DiffServ examination differently on different platforms beginning with
ExtremeXOS version 11.4, the precedence varies by platform, as shown in the following paragraphs.
Precedence of Traffic Groupings on the BlackDiamond 10808 and

12800 Series Switches
The supported traffic groupings, by precedence on the BlackDiamond 10808 and 12800 series switches,
are listed in the following order:
● Access list groupings (ACLs)
■ IP ACL
■ MAC ACL
● Explicit packet class of service groupings
■ DiffServ (IP TOS)
■ 802.1p
● Physical/logical groupings
■ Source port
■ VLAN
NOTE
When 802.1p examination is enabled, the source port and VLAN QoS apply only to untagged packets, and 802.1p
QoS applies only to tagged packets. If you use 802.1p or DiffServ QoS in conjunction with ACLs, you must
configure the 802.1p or DiffServ action within the ACL itself.
Precedence of Traffic Groupings on the BlackDiamond 8800 Series Switches,

SummitStack, and the Summit Family of Switches
Beginning with ExtremeXOS version 11.4, the supported traffic groupings, by precedence on the
BlackDiamond 8800 series switches, SummitStack, and the Summit family of switches, are listed in the
following order:
● Access list groupings (ACLs)
■ IP ACL—Depends on specifications in the ACL rule
■ MAC ACL—Depends on specification in the ACL rule
■ 802.1p
● Physical/logical groupings
■ Source port
■ VLAN

Quality of Service

■ DiffServ (IP TOS)
NOTE
The source port and VLAN QoS apply only to untagged packets, and 802.1p QoS applies only to tagged packets. If
you use 802.1p or DiffServ QoS in conjunction with ACLs, you must configure the 802.1p or DiffServ action within
the ACL itself.
ACL-Based Traffic Groupings

ACL-based traffic groupings are based on any combination of the following items:
● IP source or destination address
● IP protocol
● TCP flag
● TCP/UDP or other Layer 4 protocol
● TCP/UDP port information
● IP fragmentation
● MAC source or destination address
● Ethertype
ACL-based traffic groupings are defined using access lists. Access lists are discussed in detail in Chapter
12, “Access Lists (ACLs).” By supplying a named QoS profile on an ACL rule, you can prescribe the
bandwidth management and priority handling for that traffic grouping. This level of packet filtering has
no impact on performance.
Explicit Class of Service (802.1p and DiffServ) Traffic Groupings

This category of traffic groupings describes what is sometimes referred to as explicit packet marking, and
refers to information contained within a packet intended to explicitly determine a class of service. That
information includes:
● Prioritization bits used in IEEE 802.1p packets
● IP Differentiated Services (DiffServ) code points, formerly known as IP Type of Service (TOS) bits
An advantage of explicit packet marking is that the class of service information can be carried
throughout the network infrastructure, without repeating what can be complex traffic grouping policies
at each switch location. Another advantage is that endstations can perform their own packet marking
on an application-specific basis. Extreme Networks switch products have the capability of observing
and manipulating packet marking information with no performance penalty.
The documented capabilities for 802.1p priority markings or DiffServ capabilities (if supported) are not
impacted by the switching or routing configuration of the switch. For example, 802.1p information can
be preserved across a routed switch boundary and DiffServ code points can be observed or overwritten
across a Layer 2 switch boundary.

● Configuring 802.1p priority
● Configuring DiffServ

Traffic Groupings
Configuring 802.1p Priority

Extreme Networks switches support the standard IEEE 802.1p priority bits that are part of a tagged
Ethernet packet. The 802.1p bits can be used to prioritize the packet and to assign that packet to a
particular QoS profile.
When a tagged packet arrives at the switch, the switch examines the 802.1p priority field and maps the
packet to a specific queue when subsequently transmitting the packet. The 802.1p priority field is
located directly following the 802.1Q type field and preceding the 802.1Q VLAN ID, as shown in
Figure 51.
Figure 51: Ethernet packet encapsulation
802.1Q 802.1p 802.1Q

type priority VLAN ID
8100
Destination Source
IP packet CRC
address address
EW_024

● Observing the 802.1p information
● Changing the default 802.1p mapping
● Replacing 802.1p information
● 802.1p information on the 8800 series switches, SummitStack, and the Summit family of switches
only
● 802.1p information on the BlackDiamond 10808 and 12800 series switches only
Observing 802.1p information. When ingress traffic that contains 802.1p prioritization information is
detected by the switch, that traffic is mapped to various queues on the egress port of the switch. The
BlackDiamond 10808 and 12800 series switches supports 8 hardware queues by default; you can modify
the characteristics of each queue. By default, the BlackDiamond 8800 series switches, SummitStack, and
the Summit family of switches support 2 queues based on flows; you can define up to 6 additional
queues, except for a SummitStack where you can define up to four additional queues. The transmitting
queue determines the characteristics used when transmitting packets.
NOTE
See for Chapter 6, “Virtual LANs,” information regarding vMANs using 802.1p information to direct packets to
appropriate egress QoS queues.

Quality of Service
To control the mapping of 802.1p prioritization values to queues, 802.1p prioritization values can be
mapped to a QoS profile. The default mapping of each 802.1p priority value to QoS profile is shown in
Table 65.
Table 65: Default 802.1p priority value-to-QoS profile mapping

BlackDiamond 10808 and BlackDiamond 8800 series switches, SummitStack,
12800 series switches default and the Summit family of switches
Priority value QoS profile default QoS profile
0 QP1 QP1
1 QP2 QP1
2 QP3 QP1
3 QP4 QP1
4 QP5 QP1
5 QP6 QP1
6 QP7 QP1
7 QP8 QP8
Changing the default 802.1p mapping. By default, a QoS profile is mapped to a queue, and each QoS
profile has configurable parameters. In this way, an 802.1p priority value seen on ingress can be
mapped to a particular QoS profile.
To change the mapping of 802.1p priority value to QoS profile, use the following command:
configure dot1p type <dot1p_priority> {qosprofile} <qosprofile>
Replacing 802.1p priority information. By default, 802.1p priority information is not replaced or
manipulated, and the information observed on ingress is preserved when transmitting the packet. This
behavior is not affected by the switching or routing configuration of the switch.
However, the switch is capable of inserting and/or overwriting 802.1p priority information when it
transmits an 802.1Q tagged frame. If 802.1p replacement is enabled, the 802.1p priority information that
is transmitted is determined by the queue that is used when transmitting the packet. The 802.1p
replacement configuration is based on the ingress port. To replace 802.1p priority information, use the
following command:
enable dot1p replacement ports [<port_list> | all]
NOTE
The port in this command is the ingress port.
To disable this feature, use the following command:
disable dot1p replacement ports [<port_list> | all]
NOTE
On the BlackDiamond 8800 series switches, SummitStack, and the Summit family of switches, only QP1 and QP8
exist by default; you must create QP2 to QP7(QP2 to QP5 in a SummitStack). If you have not created these QPs,
the replacement feature will not take effect.

Traffic Groupings
The 802.1p priority information is replaced according to the queue that is used when transmitting from
the switch. The mapping is described in Table 66. This mapping cannot be changed.
Table 66: Queue-to-802.1p priority replacement value

BlackDiamond 10808 and BlackDiamond 8800 series switches,
802.1p priority 12800 series switches SummitStack, and the Summit family of switches
replacement value hardware queue 802.1p queue
0 Q0 Q1
1 Q1 Q2
2 Q2 Q3
3 Q3 Q4
4 Q4 Q5
5 Q5 Q6 *
6 Q6 Q7 *
7 Q7 Q8
You cannot direct traffic to these queues in a SummitStack.
NOTE
This command affects only that traffic based on explicit packet class of service information and physical/logical
configuration.
802.1p information on the BlackDiamond 8800 series switches, SummitStack, and the Summit family of
switches only. Beginning in ExtremeXOS 11.4, you can disable the 802.1p examination. Although this
feature is enabled by default, you can disable it to conserve ACL resources. (See Chapter 12, “Access
Lists (ACLs),” for information on available ACL resources.) To display the amount of available ACL
resources, use the following command:
show access-list usage port
NOTE
Extreme Networks recommends disabling this feature when you are not using this form of traffic grouping to free
resources.
To disable the 802.1p examination feature, use the following command:
disable dot1p examination ports [<port_list> | all]
NOTE
802.1p examination cannot be disabled for 802.1p priority levels 5 and 6 in a SummitStack. When 802.1p
examination is disabled on a SummitStack, the precedence for the remaining examination becomes lower than that
of all other traffic groupings.
To re-enable the 802.1p examination feature, use the following command:
enable dot1p examination ports [<port_list> | all]
By default, this feature is enabled.

Quality of Service
To display whether the 802.1p examination feature is enabled or disabled, use the following command:
Also, beginning with ExtremeXOS version 11.4 on the 1 Gigabit Ethernet ports, 802.1p replacement
always happens when you configure the DiffServ traffic grouping.
802.1p information on the BlackDiamond 10808 series and 12800 series switches only. If a port is in more
than one virtual router, you cannot use the QoS 802.1p features. The default VLAN DiffServ
examination mappings apply on ports in more than one VR. If you attempt to configure examining or
replacing 802.1p information on a port that is in more than one virtual router, the system returns the
following message:
Warning: Port belongs to more than one VR. Port properties related to diff serv and
code replacement will not take effect.
Configuring DiffServ
Contained in the header of every IP packet is a field for IP Type of Service (TOS), now also called the
Differentiated Services (DiffServ) field. The DiffServ field is used by the switch to determine the type of
service provided to the packet.
Figure 52 shows the encapsulation of an IP packet header.
Figure 52: IP packet header encapsulation
0 1 2 3 4 5 6 7
DiffServ code point

0 bits 31
Version IHL Type-of-service Total length
Identification Flags Fragment offset
Time-to-live Protocol Header checksum
Source address
Destination address
Options (+ padding)
Data (variable)
EW_023
Observing DiffServ code points as a traffic grouping mechanism for defining QoS policies and
overwriting the Diffserv code point fields are supported.
This section describes the following sections:

● Observing DiffServ information
● Changing DiffServ code point (DSCP) mapping

Traffic Groupings
● Replacing DSCP information

● DiffServ information on the BlackDiamond 8800 series switches, SummitStack, and the Summit
family of switches only
● DiffServ information on the BlackDiamond 10808 and 12800 series switches only
● DiffServ example configurations
Observing DiffServ information. When a packet arrives at the switch on an ingress port and this feature is
enabled, the switch examines the first six of eight TOS bits, called the DiffServ code point. The switch can
then assign the QoS profile used to subsequently transmit the packet based on the code point. The QoS
profile controls which queue is used when transmitting the packet out of the switch and determines the
forwarding characteristics of a particular code point. Examining DiffServ information can be enabled or
disabled; by default it is disabled. To enable DiffServ examination, use the following command:
enable diffserv examination port [<port_list> | all]
To disable DiffServ examination, use the following command:
disable diffserv examination port [<port_list> | all]

Because the DiffServ code point uses six bits, it has 64 possible values (26 = 64). By default, the values
are grouped and assigned to the default QoS profiles listed in Table 67.
Table 67: Default DiffServ code point-to-QoS profile mapping

BlackDiamond 10808 and
12800 series switches QoS BlackDiamond 8800 series switches, SummitStack,
Code point profile and the Summit family of switches QoS profile
0-7 QP1 QP1
8-15 QP2 QP1
16-23 QP3 QP1
24-31 QP4 QP1
32-39 QP5 QP1
40-47 QP6 QP1
48-55 QP7 QP1
56-63 QP8 QP8
Changing the default DiffServ code point mapping. You can change the QoS profile assignment for each of
the 64 code points using the following command:
configure diffserv examination code-point <code-point> {qosprofile} <qosprofile>
After they have been assigned, the rest of the switches in the network prioritize the packet using the
characteristics specified by the QoS profile.
Replacing DiffServ code points. The switch can be configured to change the DiffServ code point in the
packet before the packet is transmitted by the switch. This is done with no impact on switch
performance.
The DiffServ code point value used in overwriting the original value in a packet is determined by the
QoS profile. You enter the QoS profile you want to use to determine the replacement DiffServ code
point value.

Quality of Service
To replace DiffServ code points, you must enable DiffServ replacement using the following commands
enable diffserv replacement ports [<port_list> | all]
NOTE
The port in this command is the ingress port. This command affects only that traffic based on explicit packet class
of service information and physical/logical configuration.
disable diffserv replacement port [<port_list> | all]
The default QoS profile to DiffServ code point mapping is shown in Table 68, and the default 802.1p
priority value to code point mapping is described in Table 68.
Table 68: Default 802.1p priority value-to-DiffServ code point mapping
BlackDiamond 10808 and BlackDiamond 8800 series switches,
12800 series switches QoS SummitStack, and the Summit family of
profile switches QoS profile 802.1p priority value Code point
QP1 QP1 0 0
QP2 QP1 1 8
QP3 QP1 2 16
QP4 QP1 3 24
QP5 QP1 4 32
QP61 QP1 5 40
QP7 QP1 6 48
QP8 QP8 7 56
You change the DiffServ code point mapping, using either the QoS profile or the 802.1p value, to any
code point value using the following command:
configure diffserv replacement [{qosprofile} <qosprofile> | priority <value>] code-

point <code_point>
NOTE
Extreme Networks recommends that you use the qosprofile <qosprofile> value to configure this parameter.
By doing so, the queue used to transmit a packet determines the DiffServ value replaced in the IP
packet.
To view currently configured DiffServ information, use the following command:
show diffserv [examination | replacement]
NOTE
You can also use ACLs to replace the DSCP value; to do so you must first configure the values using these CLI
commands.

Traffic Groupings
DiffServ information on the BlackDiamond 8800 series switches, SummitStack, and the Summit family of
switches only. Beginning with ExtremeXOS version 11.4 on the 1 Gigabit Ethernet ports, 802.1p
replacement always happens when you configure the DiffServ traffic grouping.
DiffServ information on the BlackDiamond 10808 and 12800 series switches only. The default VLAN
DiffServ examination mappings apply on ports in more than one VR. If you attempt to configure
examining or replacing DiffServ information on a port that is in more than one virtual router, the
system returns the following message:
Warning: Port belongs to more than one VR. Port properties related to diff serv and
code replacement will not take effect.
DiffServ example. In this example, we use DiffServ to signal a class of service throughput and assign any
traffic coming from network 10.1.2.x with a specific DiffServ code point. This allows all other network
switches to send and observe the Diffserv code point instead of repeating the same QoS configuration
on every network switch.
To configure the switch:

1 Using ACLs, assign a traffic grouping for traffic from network 10.1.2.x to QP3:
configure access-list qp3sub any
The following is a sample policy file example:
#filename: qp3sub.pol
entry QP3-subnet {
if {
source-address 10.1.2.0/24
} then {
Qosprofile qp3;
replace-dscp;
}
2 Configure the switch so that other switches can signal calls of service that this switch should observe
by entering the following:
enable diffserv examination ports all
NOTE
The switch only observes the DiffServ code points if the traffic does not match the configured access list. Otherwise,
the ACL QoS setting overrides the QoS DiffServ configuration.
Physical and Logical Groupings

Two traffic groupings exist in this category:
● Source port
● VLAN

Quality of Service
Source port
A source port traffic grouping implies that any traffic sourced from this physical port uses the indicated
QoS profile when the traffic is transmitted out to any other port. To configure a source port traffic
grouping, use the following command:
configure ports <port_list> {qosprofile} <qosprofile>
In the following modular switch example, all traffic sourced from slot 5 port 7 uses the QoS profile
named QP8 when being transmitted.
configure ports 5:7 qosprofile qp8
NOTE
On the BlackDiamond 10808 and 12800 series switches, this command applies only to untagged packets. On the
BlackDiamond 8800 series switches, SummitStack, and the Summit family of switches, this command applies to all
packets.
VLAN
A VLAN traffic grouping indicates that all intra-VLAN switched traffic and all routed traffic sourced
from the named VLAN uses the indicated QoS profile. To configure a VLAN traffic grouping, use the
following command:
configure vlan <vlan_name> {qosprofile} <qosprofile>
For example, all devices on VLAN servnet require use of the QoS profile QP1. The command to
configure this example is as follows:
configure vlan servnet qosprofile qp1
NOTE
On the BlackDiamond 10808 and 12800 series switches, this command applies only to untagged packets. On the
BlackDiamond 8800 series switches, SummitStack, and the Summit family of switches, this command applies to all
packets.
Verifying Physical and Logical Groupings

You can display QoS settings on the ports or VLANs.
NOTE
On the BlackDiamond 10808 switch, the screen displays both ingress and egress QoS settings. The 10Gbps ports
have 8 ingress queues, and the 1 Gbps ports have 2 ingress queues. (Refer to “Bi-Directional Rate Shaping—
BlackDiamond 10808 Switch Only” on page 461 for more information on ingress queues, or bi-directional rate
shaping.)
To verify settings on ports or VLANs, use the following command:

Verifying QoS Configuration and Performance
BlackDiamond 8800 series switches, SummitStack, and the Summit family of switches display. You display
which QoS profile, if any, is configured on the BlackDiamond 8800 series switches, SummitStack, and
the Summit family of switches using the show ports {mgmt | <port_list>} information
{detail} command.
BlackDiamond 10808 switch display. You display information on the egress QoS profiles and the ingress
QoS profiles (shown as Ingress Rate Shaping), as well as the minimum and maximum available
bandwidth and priority on the BlackDiamond 10808 switch using the show ports {mgmt |
<port_list>} information {detail} command. The display is slightly different for a 1 Gbps port
and for a 10 Gbps port.
BlackDiamond 12800 series switches display. You display information on the QoS profiles and rate
limiting, as well as the minimum and maximum available bandwidth and priority on the BlackDiamond
12800 series switches using the show ports {mgmt | <port_list>} information {detail} detail
command.
BlackDiamond 12800 R-series switch display. You display information on the QoS profiles, hierarchical
QoS, and rate limiting, as well as the minimum and maximum available bandwidth and priority on the
BlackDiamond 12800 series switches using the show ports {mgmt | <port_list>} information
{detail} command.
NOTE
To ensure that you display the QoS information, you must use the detail variable.
Verifying QoS Configuration and Performance

You can display a variety of QoS measures using the CLI.
Monitoring Performance
NOTE
This command is supported on the BlackDiamond 8800 series “a” and “e” switches, Summit X250e, X450a,
X450e series switches (whether or not in a SummitStack). But, only one port per a-series or e-series module or
switch can be monitored at any one time.
After you have created QoS policies that manage the traffic through the switch, you can use the QoS
monitor to determine whether the application performance meets your expectations.
QoS features performance monitoring with a either a real-time or a snapshot display of the monitored
ports. To view switch performance per port, use the following command:
show ports <port_list> qosmonitor {ingress | egress} {bytes | packets} {no-refresh}
NOTE
You must specify ingress to view the ingress rate-shaping performance. By default, this command displays the egress
performance.

Quality of Service
Displaying QoS Profile Information

You can also verify the QoS configuration in place.
Refer to “Verifying Physical and Logical Groupings” on page 454 for additional information on
displaying QoS information for each port.
Displaying QoS Profile Information on the BlackDiamond 8800 Series switches,

To display QoS information on the BlackDiamond 8800 series switches, SummitStack, and the Summit
family of switches, use the following command:
show qosprofile
Displayed information includes:

● QoS profiles configured
● Weight
● Maximum buffer percent
Beginning with ExtremeXOS version 11.5, use this command to display the bandwidth configuration per
egress QoS profile (or queue) on the following devices:
● BlackDiamond a-series modules
● BlackDiamond e-series modules
● Summit X450a series switch
Displaying QoS Profile Information on the BlackDiamond 10808 and

To display QoS information on the BlackDiamond 10808 and 12800 series switches, use the following
command:
show qosprofile [ingress | egress | ports [ all | <port_list>]
Displayed information includes:

● QoS profile name
● Minimum bandwidth
● Maximum bandwidth
● Priority

Guidelines for Configuring QoS
Guidelines for Configuring QoS

The following are useful guidelines for configuring QoS:
● If you are using DiffServ for QoS parameters, Extreme Networks recommends that you also
configure 802.1p or port-based QoS parameters to ensure that high-priority traffic is not dropped
before it reaches the Master Switch Module (MSM) on modular switches.
● The command to replace the 802.1p or DiffServ value affects only those traffic groupings based on
explicit packet class of service and physical/logical groupings.
Metering Using ACLs—BlackDiamond 8800 Series,

SummitStack, the Summit Family of Switches, and
BlackDiamond 12800 R-Series Switch Only
NOTE
On he BlackDiamond 12800 R-series switch, this feature can be used only in conjunction with rate-limiting
commands. See “Hierarchical QoS—BlackDiamond 12800 R-Series Switch Only” on page 464 for more information
on rate limiting on the BlackDiamond 12800 R-series switch.
The BlackDiamond 8800 series switches, SummitStack, and the Summit family of switches provide a
metering capability that can be used to associate an ACL rule to a specified bit-rate and out-of-profile
action. This feature provides a method of ingress rate shaping for these platforms. The rate granularity
is 64 Kbps (up to 1 Gbps for GE ports and up to 10 Gbps for 10 G ports) and the out-of-profile actions
are drop, set the drop precedence, or mark the DSCP with a configured value. Additionally, each meter
has an associated out-of-profile counter that counts the number of packets that were above the
committed-rate (and subject to the out-of-profile-action). These meters are a per-port resource; so, for
example, assigning a 50 Mbps meter to a VLAN means that each port in the VLAN is assigned 50
Mbps.
On the BlackDiamond 8800 a-series and e-series modules and the Summit X450a, X450e, and X250e
series switches with ExtremeXOS version 11.5 (or higher), the meters are a per-chip, per-slice resource
(see Chapter 12, “Access Lists (ACLs),” for complete information.)
With strict priority HQoS on the BlackDiamond 12800 R-series switch, you configure a peak rate for a
specified ingress queue. When you are using bandwidth mode HQoS, you configure a peak rate and a
committed rate for each CoS value pair for a specified ingress queue or for all CoS value pairs for the
specified ingress queue.
NOTE
With bandwidth mode HQoS, you must configure the peak rate to be higher than the committed rate.
To configure ACL metering:

1 Create the meter
2 Configure the meter

Quality of Service
3 Associate the meter with an ACL rule entry (see Chapter 12, “Access Lists (ACLs),” for complete
information on ACLs)
Creating the ACL Meter

To create the ACL meter, use the following command:
create meter <metername>
To delete the meter, use the following command:
delete meter <metername>
Configuring the ACL Meter

You configure the meter differently depending on the platform you are using.
Configuring ACL Meters on the BlackDiamond 8800 Series Switches, SummitStack,

and the Summit Family of Switches
After the ACL meter is created, you configure it. Configuring the ACL meter sets allowable traffic limits
and the actions to take with out of limit traffic. To configure an ACL meter, use the following
command:
configure meter <metername> {max-burst-size <burst-size> [Gb | Kb | Mb]} {committed-
rate <cir> [Gbps | Mbps | Kbps]} {out-actions [drop | set-drop-precedence {dscp [none
| <dscp-value>]}}
Configuring ACL Meters on the BlackDiamond 12800 R-Series Switch
NOTE
This feature can only be used in conjunction with rate-limiting commands on the BlackDiamond 12800 R-series
switch. See “Hierarchical QoS—BlackDiamond 12800 R-Series Switch Only” on page 464 for more information on
rate limiting on the BlackDiamond 12800 R-series switch.
Once you create the meter, use the following command to configure the meter on the BlackDiamond
12800 R-series switch:
configure meter <metername> [peak-rate <pr> [Gbps | Mbps | Kbps]]
Associating the Meter with an ACL

To associate a meter with an ACL, you add the meter <metername> statement to the action modifier of
the ACL rule entry, similar to the count <countername> statement.
NOTE
See Chapter 12, “Access Lists (ACLs),” for complete information on ACLs.

Egress Traffic Rate Limiting—BlackDiamond 8800 Series Switches, SummitStack, and the Summit Family of Switches Only
Displaying Meters
To display the meters that you created you can use either the show-access list or, beginning in
ExtremeXOS 11.4, the show meter command.
Egress Traffic Rate Limiting—BlackDiamond 8800

Series Switches, SummitStack, and the Summit Family
of Switches Only
You can configure the maximum egress traffic allowed per port by specifying the committed rate, or
you can allow the egress traffic to pass an unlimited flow.
You can limit egress traffic on a 1 Gbps port in increments of 64 Kbps; on a 10 Gbps port, you can limit
egress traffic in increments of 1 Mbps. Optionally, you can also configure a maximum burst size, which
is higher than the limit, allowed to egress the specified port(s) for a burst, or short duration.
The default behavior is to have no limit on the egress traffic per port. To configure egress rate limiting,
configure ports <port_list> rate-limit egress [no-limit | <cir-rate> [Kbps | Mbps |

Gbps] {max-burst-size <burst-size> [Kb | Mb]}]
To view the configured egress port rate-limiting behavior, use the following command:
You must use the detail parameter to display the Egress Port Rate configuration and, if configured, the
Max Burst size. Refer to “Displaying Port Configuration Information” for more information on the show
ports information command.
You can also display this information using the following command:
show configuration vlan
The following is sample output from the show configuration vlan command for configured egress
rate limiting:
# Module vlan configuration.

#
configure vlan Default tag 1
config port 3:1 rate-limit egress 128 Kbps max-burst-size 200 Kb
config port 3:2 rate-limit egress 128 Kbps
config port 3:10 rate-limit egress 73 Kbps max-burst-size 128 Kb
configure vlan Default add ports 3:1-48 untagged
NOTE
Refer to Chapter 9, “Forwarding Database,” for more information on limiting broadcast, multicast, or unknown MAC
traffic ingressing the port.

Quality of Service
Applying Egress Bandwidth to a Port—BlackDiamond

10808 and 12800 Series Switches Only
Beginning with ExtremeXOS version 11.4, you can configure aggregate egress committed rate, peak rate,
minimum bandwidth, and maximum bandwidth per port on the BlackDiamond 10808 and 12800
(including the 12800-R Series) series switches only. This feature allows you to enforce a bandwidth
restriction to traffic that exits the switch on specific port or ports. To configure this feature, you directly
assign the aggregate bandwidth limit to specified egress ports.
NOTE
You can use port-based aggregate egress committed rate in conjunction with HQoS ingress and egress rate limiting.
To configure the egress bandwidth per port, use the following CLI command:
configure ports <port_list> rate-limit egress aggregate [{{committed_rate

<committed_bps> [k | m ]} {peak_rate <peak_bps> [k | m]}} | {{minbw <minbw_number>}
{maxbw <maxbw_number>}}]
To remove the limit on egress bandwidth per port, re-issue this command using the default values.
To display the current configuration for the port-based egress rate limiting, use the following command:
show ports <port_list> information detail
NOTE
You must use the detail variable to display this information.
The following sample output from the show port information detail command shows the setting
for aggregate egress bandwidth, which is displayed in the Aggregate Queue line, on the BlackDiamond
12800 R-series switch:
.
.
.
Multicast Flooding: Enabled
Broadcast Flooding: Enabled
Jumbo: Disabled
High Priority COS: 7
Packet Length Adjustment: 0 bytes
QoS Profile: None configured
Aggregate Queue:
QP0 MinBw = 0% MaxBw = 100% Pri = 8
Queue:
.
.
.

Applying Egress Bandwidth to a QoS Queue
Applying Egress Bandwidth to a QoS Queue

Beginning with ExtremeXOS version 11.5, you can apply egress bandwidth to a specified QoS queue (or
QP) using all platforms, except the original BlackDiamond 8800 modules, and the Summit X450 series
switches.
You configure egress committed rate, peak rate, minimum bandwidth, and maximum bandwidth per
QoS queue. Additionally, you can configure the priority level for each queue on the BlackDiamond
10808 and 12800 (including the 12800-R Series) series switches only. This feature allows you to enforce a
bandwidth restriction to traffic that exits the switch on specific queues. To configure this feature, you
directly assign the specified bandwidth limit to specified queues on specified egress ports.
To configure the egress bandwidth per QoS queue per port, use the following CLI command:
configure qosprofile {egress} <qosprofile> [{committed_rate <committed_bps> [k | m]}

{maxbw <maxbw_number>} {minbw <minbw_number>} {peak_rate <peak_bps> [k | m} {priority
[<priority> | <priority_number>]}] ports [<port_list> | all]
NOTE
You cannot configure the priority for the QoS queue on the BlackDiamond 8800 series switches, SummitStack, or
the Summit family of switches.
To remove the limit on egress bandwidth per QoS queue per port, re-issue this command using the
default values.
If you attempt to apply this command to ports on the original BlackDiamond 8800 modules, the system
returns the following error:
ERROR: Setting minbw/maxbw is not supported on MSM-G8X I/O ports and G48T, G48P, G24X,
and 10G4X modules.
To display the current configuration for the QoS queue-based egress rate limiting, use the following
command:
show qosprofile ports [all | <port_list>]
Bi-Directional Rate Shaping—BlackDiamond 10808

Switch Only
NOTE
If you are working with the BlackDiamond 8800 series switches, SummitStack, or the Summit family of switches,
refer to Metering Using ACLs—BlackDiamond 8800 Series, SummitStack, the Summit Family of Switches, and
BlackDiamond 12800 R-Series Switch Only for information on metering the ingressing traffic. if you are working
with the BlackDiamond 12800 R-series modules, refer to Hierarchical QoS—BlackDiamond 12800 R-Series Switch
Only for information on rate limiting.
With software version 11.0, you can configure and display bi-directional rate shaping parameters on the
BlackDiamond 10808 switch. Bi-directional rate shaping allows you to manage bandwidth on Layer 2

Quality of Service
and Layer 3 traffic flowing to each port on the switch and from there to the backplane. You can
configure up to 8 ingress queues, which send traffic to the backplane, per physical port on the I/O
module. By defining minimum and maximum bandwidth for each queue, you define committed and
peak information rates for each queue. You can define different priorities for each queue for each port.
Rate shaping on the ingress port allows the switch to enforce how much traffic from a particular port
can ingress to the system.
Bi-directional rate shaping on the BlackDiamond 10808 switch controls the traffic from the ingress ports
to the backplane and provides guaranteed minimum rates. The number of queues from the ingress port
to the backplane differs between I/O modules. The 1 Gbps I/O module has 2 queues from the ingress
port to the backplane, and the 10 Gbps I/O module has 8 queues from the ingress port to the
backplane.
You set minimum bandwidth, maximum bandwidth, and priority for each queue for each port. Use
prioritization when two or more hardware queues on the same physical port are contending for
transmission, only after their respective bandwidth management parameters have been satisfied. After
the priorities are satisfied, the switch uses a round-robin system to empty the queues to the backplane.
Table 69 displays the mapping of the ingress queues and the priority value for each I/O module.
Table 69: Ingress queue mapping for I/O modules on the BlackDiamond 10808 switch
I/O module Ingress queues Priority value
1 Gbps module IQP1 1 to 4
IQP2 5 to 8
10 Gbps module IQP1 1
IQP2 2
IQP3 3
IQP4 4
IQP5 5
IQP61 6
IQP7 7
IQP8 8
Using bi-directional rate shaping, excess traffic is discarded at the I/O module and does not traverse to
the backplane. You view statistics on the discarded traffic using the show ports qosmonitor or show
ports information command.
The 802.1p value is mapped to the ingress queue. For untagged ports, use port- or VLAN-based QoS to
map traffic to the ingress queue.
Bandwidth Settings
You apply ingress QoS profile (IQP or rate shaping) values on the BlackDiamond 10808 switch as either
a percentage of bandwidth or as an absolute value in Kbps or Mbps. IQP bandwidth settings are in turn
applied to queues on physical ports. The impact of the bandwidth setting is determined by the port
speed (1 or 10 Gbps).

Bi-Directional Rate Shaping—BlackDiamond 10808 Switch Only
NOTE
You may see slightly different bandwidths because the switch supports granularity down to 62.5 Kbps.
Maximum Bandwidth Settings

The maximum bandwidth settings determine the port bandwidth available to each of the ingress port
queues.
Minimum Bandwidth Settings

The minimum bandwidth settings, or maximum committed rate settings, determine the port bandwidth
reserved for each of the ingress port queues.
Table 70 displays the maximum committed rates available for each port on each BlackDiamond 10808
switch I/O module.
Table 70: Maximum committed rates per port for I/O module
on the BlackDiamond 10808 switch
I/O module MSM configuration Maximum committed rate
1 Gbps module Single MSM 200 Mbps
Dual MSM 400 Mbps
10 Gbps module Single MSM 2 Gbps
Dual MSM 4 Gbps
Note that these maximum committed rates vary with the number of active ports on each I/O module.
The rates shown in Table 70 are what you can expect when you all running all ports at traffic level. If
you are using fewer ports, you will have higher committed rates available for each port. And, the
maximum committed rate is reached when you are running traffic on only one port.
NOTE
Cumulative percentages of minimum bandwidth of the queues on a given port should not exceed 100%
If you choose a setting not listed in the tables, the setting is rounded up to the next value. If the actual
bandwidth used is below the minimum bandwidth, the additional bandwidth is not available for other
queues on that physical port.
Configuring Bi-Directional Rate Shaping

The maximum bandwidth or rate defined in the BlackDiamond 10808 switch ingress QoS profile defines
the rate limit for ingress traffic on rate-shaped ports. You set minimum and maximum rates for each
port on the ingress port, using either percentage of total bandwidth or absolute values for committed
and peak rates in Kbps or Mbps. You also set the priority level for each queue.

Quality of Service
To define rate shaping on a port, you assign a minimum and maximum bandwidth or rate plus a
priority value to each queue on the ingress port (see Table 69 for the number of queues available to each
port on the I/O module). To define rate shaping, use the following command:
configure qosprofile ingress <iqp> [{committed_rate <committed_bps> [k | m]} {maxbw

<maxbw_number>} {minbw <minbw_number>} {peak_rate <peak_bps> [k | m} {priority
[<priority> | <priority_number]}] ports [<port_list> | all]
If you choose to use committed rate and peak rate values, be aware of the interactions between the
values and the command line interface (CLI) management system. You can enter any integer from 0 in
the CLI; however, functionally the switch operates only in multiples of 62.5 Kbps. Also note that the
CLI system does not accept decimals.
Rate shaping is disabled by default on all ports; the system does use existing 802.1p, port, and VLAN
values to assign packets to the ingress queue. The rate shaping function is used to assign specific
priorities by absolute rates or percentages of the bandwidth.
To enable this ingress rate shaping feature, use the configuration command. To disable the ingress rate
shaping, use the following command:
unconfigure qosprofile ingress ports all
To display the parameters for ingress rate shaping (the values for the IQPs), use the following
commands:
show qosprofile [ingress | egress | ports [ all | <port_list>]

Additionally, you can monitor the performance on the BlackDiamond 10808 switch by using the
following command:
show ports <port_list> qosmonitor {ingress | egress} {bytes | packets} {no-refresh}
NOTE
You must specify ingress to view ingress rate shaping performance.
Hierarchical QoS—BlackDiamond 12800 R-Series

Switch Only
CAUTION
You cannot mix BlackDiamond 12800 R-series module with non-R-series modules in the BlackDiamond 12800
chassis. If you attempt to mix these, the system returns an error message.
This section discusses the following topics:

● Setting the HQoS Mode on page 466
● HQoS Implementation on page 467
● Guidelines for Using Ingress-Only and Ingress and Egress HQoS on page 472

Hierarchical QoS—BlackDiamond 12800 R-Series Switch Only
● Configuring HQoS Ingress and Egress Queues on page 472

● Displaying HQoS on page 476
● HQoS Examples on page 478
Overview
Hierarchical Quality of Service (HQoS), or rate limiting, is an important enabling feature for the
deployment of residential integrated networks and advanced business services; it is available on the
BlackDiamond 12800 R-series switch. HQoS allows the Internet Service Provider (ISP) to assign a
committed bandwidth to each customer. The customer then assigns priorities to their applications, and
HQoS, or rate limiting, ensures that the highest priority traffic is always serviced first. If the highest
priority applications are idle, the lower priority applications are allowed to send their data. HQoS also
allows a straightforward way to determine network load, easing the network planning burden.
This is accomplished by limiting traffic flows in both directions—ingress and egress—within the
network. Such an efficient, high-performance network supports the diverse requirements of several
service types, including the requirements of real-time voice and video, along with priority and best-
effort data traffic. All this is accomplished with no impact on network performance.
NOTE
You can configure vMAN-based rate limiting using vMAN ACLs. (See Chapter 12, “Access Lists (ACLs),” for
information on vMAN ACLs.).
ExtremeXOS software has the following two modes of HQoS:

● Strict priority mode
● Bandwidth mode—You need an enabled Feature Pack to run the bandwidth mode HQoS on your
BlackDiamond 12800 R-series switch (refer to Chapter 2, “ExtremeXOS Overview” for complete
information on Feature Packs). This mode meets the MEDF-14 standard.
Strict Priority Mode

The strict priority mode of HQoS represents a special traffic queue with the following three set levels of
priority:
● High
● Medium
● Low
This mode supports one rate as an aggregate rate for the traffic queue; that rate is the peak rate. (The
entire traffic queue is referred to as an L2 queue.)
Bandwidth Mode
NOTE
You must have an enabled Feature Pack to run bandwidth mode HQoS.

Quality of Service
Beginning with ExtremeXOS software version 11.6, you can run bandwidth mode HQoS. Bandwidth
mode allows you to set the committed rate and peak rate for traffic, or the bandwidth. The bandwidth
mode is meets the requirements of the Metro Ethernet Forum 14 (MEF-14) certification.
This capability allows ISPs to enforce service level agreements (SLAs) agreed to with their customers.
Part of the SLA enforces bandwidth restrictions on that customer’s use of the ISP network. Each
customer is assigned a bandwidth profile in the SLA; this bandwidth profile determines whether each
frame from that customer arriving at the ISP network is in-profile or out-of-profile. Based on the specific
SLA, out-of-profile traffic is either discarded or marked for a best effort delivery. Also, using bandwidth
mode HQoS, certain types of traffic, such as voice and video traffic, may be marked for expedited
delivery over data traffic. In this way, as single SLA can support both real-time and data-intensive
traffic for the customer.
There are four sub-queues in each traffic queue in bandwidth mode; these sub-queues correspond to
CoS values and can be metered separately. (These sub-queues are referred to as L1 queues.) The four
sub-queues in each traffic queue are as follows:
● COS0-1
● COS2-3
● COS4-5
● COS6-7
The ExtremeXOS software supports the following bandwidth profile attributes for each sub-queue (or
you can configure the bandwidth profile for all four CoS value pairs in one traffic queue at once):
● Committed rate
● Peak rate
The committed rate and peak rate settings for each queue are programmed by associating a
preconfigured meter with each queue.
NOTE
You must configure the peak rate higher than the committed rate.
Setting the HQoS Mode

You set the HQoS mode to either strict priority or bandwidth mode (using committed rate and peak
rate). Once you choose the HQoS mode, any CLI commands, parameters, keywords, and variables that
apply to the mode you did not choose do not appear; that is, only those CLI commands relevant to the
chosen mode appear on your screen.
NOTE
You must have an enabled Feature Pack to run HQoS in bandwidth mode
To set the HQoS mode, use the following command:
configure traffic mode [pr-and-cr | strict-priority]

The default HQoS mode is strict priority mode. You must reboot when you choose the bandwidth mode
(configure traffic mode pr-and-cr) or when you change from one mode to the other. After issuing
this command, the system returns the following message:
Do you want to reboot?
You must also delete all traffic queues prior to changing the HQoS mode.
To check the HQoS traffic mode the switch is currently using, use the following command:
show traffic mode
HQoS Implementation
You configure HQoS in different ways for different streams of traffic: ingress-only and ingress and
egress. To configure HQoS, you create and configure traffic queues, create and apply ACL policy files,
create and configure meters, and associate the traffic queues with the ACL policy files.
You use the meters to configure a peak rate for a specified queue in strict priority mode; in bandwidth
mode, you use the meters to configure a committed rate and a peak rate for each sub-queue (or CoS
value pair) in the specified queue.

● Ingress-Only and Ingress and Egress HQoS on page 467
● CoS Values and HQoS Modes on page 469
● Changing Rates on page 470
● Controlling Flooding, Multicast, and Broadcast Traffic on vMAN Egress Ports with Rate Limiting
Applied on page 471
● Configuring Bytes Used to Calculate Traffic Rates on page 471
Ingress-Only and Ingress and Egress HQoS

You configure HQoS, or rate limiting, as one of the following strict priority styles:
● Ingress HQoS only.
● Ingress and egress HQoS.
● Egress HQoS. (You map the egress queue to an ingress queue and configure the ingress queue as
unlimited rate limiting.)
NOTE
All egress traffic queues must be mapped to an ingress traffic queue, and those ingress queues must be configured
to allow egress shaping on the specified ingress queues.
You can create multiple traffic queues, using the CLI commands, that create different limits for HQoS,
or rate limited, traffic. The traffic is first matched, using ACLs you specify in an ACL policy file, and
then assigned to a particular traffic queue. (See Chapter 12, “Access Lists (ACLs),” for more information
on using specific ACLs to achieve HQoS or rate limiting.) The ACL policy file is then applied to one or
more ingress ports.

Quality of Service
NOTE
All traffic that enters these ports that does not match any ACLs for that port is unlimited.
Ingress HQoS limits the rate of traffic entering the switch on a specified interface or interfaces. The
system classifies the ingressing traffic using the ACLs, and the configured ingress traffic queue defines
the rate limiting parameters.
With ingress and egress HQoS, or rate limiting, you associate the egress traffic queue with an ingress
traffic queue.
NOTE
The egress traffic queue is applied to ports only.
The ingress traffic queue identifies the traffic that is to be rate limited. By associating an ingress traffic
queue with an egress traffic queue, you send the classified traffic (classified at the ingress queue) to a
rate limited HQoS egress queue that limits the traffic as it exits the switch. Egress traffic queues are
associated with one or more ingress traffic queues, and traffic is limited by the egress traffic queue only
if that traffic is not dropped by the ingress traffic queue.
NOTE
Multiple ingress traffic queues can be attached to a single egress traffic queue. Thus, traffic matching different
ingress traffic queues is combined and sent to the same egress traffic queue for HQoS.
You must associate one or more ports to the egress traffic queue with ingress and egress HQoS, or rate
limiting.
To configure ingress-only or ingress and egress HQoS, or rate limiting, you create a traffic queue—
either an ingress-only queue, an ingress queue that allows egress shaping, and/or an egress queue you
associate to ingress queues. For each queue, you create and configure one or more meters, which are
used to enforce rate limits on traffic streams, and configure the meter to be associated with the traffic
queue. ACL rules assign traffic queues to the traffic, using ACL policy files.
NOTE
You must create and configure the traffic queues before applying the ACLs, and you must associate the traffic queue
with one or more ACLs to achieve rate limiting.
The system accomplishes HQoS, or rate limiting, using the following elements:
● ACLs—Classifies the traffic to be limited.
● traffic queues—Specifies how the classified traffic is handled.
● meters—Specifies the rate.
Ingress-only and ingress and egress HQoS are accomplished either by strict priority mode or bandwidth
mode. Strict priority mode HQoS has three levels of priority: high, medium, and low and supports one
rate as an aggregate rate for the traffic queue; this type of HQoS can be associated with one aggregate

peak limit. Bandwidth mode HQoS supports a committed rate and a peak rate for each of the four CoS
sub-queues within a traffic queue.
NOTE
You configure egress-only HQoS by associating the egress queue with an ingress queue that allows unlimited traffic;
you configure the egress rate limits.
CoS Values and HQoS Modes

In strict priority HQoS, the system has three queues: high, medium, and low. You use ACL policies to
match traffic and map each traffic category to a Class of Service (CoS) value, which in turn determines
the queue (high, medium, or low) that the traffic moves to. If your ACL policy file does not assign a
CoS value, the packet is assigned the default CoS value of 0. By default, packets with a CoS value of 7
go to the high queue, those with a value of 6 go to the medium queue, and those packets with a value
of 5 to 0 go to the low queue. (See “Mapping CoS value to high queue in strict priority HQoS mode” on
page 469 for information on changing the mapping of CoS values to the respective queues.)
In bandwidth mode HQoS, each traffic queue has four sub-queues that correspond to the CoS value
pairs, as follows:
● COS0-1
● COS2-3
● COS4-5
● COS6-7
You can apply meters to each CoS value pair independently or to the entire traffic queue (by specifying
all CoS pairs). You configure both a committed rate and a peak rate on the meters that you associate
with each CoS value pair, or sub-queue, in bandwidth mode HQoS.
NOTE
You must configure the peak rate higher than the committed rate.
Mapping CoS value to high queue in strict priority HQoS mode. You can configure which CoS value is in
the high queue, and the remaining CoS values are automatically queued, in strict priority bandwidth
mode only, as follows:
● High queue—Contains packets with CoS value assigned to high queue. By default, the CoS value
assigned to the high queue is 7; you can configure this value from 7 to 2. The high queue contains
packets with a single CoS value.
NOTE
All packets with a CoS value higher than that configured for the high queue automatically go to the low queue. (If
you configure 2 as the CoS value for the high queue, packets with a CoS value of 3 to 7 all go to the low queue.)
● Medium queue—Contains packets with the next-highest CoS value (numerically lower) to that value
configured as the high queue value; the medium queue contains packets with a single CoS value.
You do not need to configure the CoS value for packets in the medium queue; rather the system
automatically places packets with the next-highest value to the high queue CoS value in this queue.

Quality of Service
By default, this queue contains packets with a CoS value of 6 (because default high queue CoS value
is 7).
● Low queue—Contains packets with all other CoS values plus packets with no CoS value. By default,
this queue contains all packets with a CoS value of 5, 4, 3, 2, 1, or 0, as well as packets with no CoS
value. The low queue contains packets with any of six different CoS values and those with no CoS
value. You do not configure the CoS value for this queue; the system automatically puts all packets
not assigned to the high or medium queue to the low queue. Also note that the system automatically
puts all packets with a CoS value higher than that assigned to the high queue into the low queue.
For example, if you configure 5 as the CoS value for the high queue, the queues contain packets with
the following CoS values:
● High contains 5
● Medium contains 4
● Low contains 3, 2, 1, 0, 7, 6, and those with no CoS value
As another example, if you configure 2 as the CoS value for the high queue, the queues contain packets
with the following CoS values:
● High contains 2
● Medium contains 1
● Low contains 0, 7, 6, 5, 4, 3 and those with no CoS value
By default, the system puts packets with a CoS value of 7 into the high queue, packets with a CoS value
of 6 into the medium queue, and all other packets into the low queue.
To configure the CoS value you want as the value for the high queue in strict priority mode HQoS for
the specified ports, use the following CLI command:
configure ports <port_list> rate-limit map cos high <cos_value>
To display the value, use the following command:
NOTE
Changing Rates
To change the rate for a traffic queue, you must disassociate the current ACL policy and delete the
specified traffic queue. Then, you begin again creating and configuring the HQoS by:
● Creating and configuring a new meter with the rate you want.
● Creating a traffic queue again.
● Associating the new meter to the new queue
● Applying the ACL policy files again.
NOTE
See Guidelines for Using Ingress-Only and Ingress and Egress HQoS for more information on deleting or modifying
HQoS parameters.

See “Configuring HQoS Ingress and Egress Queues” on page 472 for information on configuring HQoS
and “Displaying HQoS” on page 476 for information on displaying HQoS.
Controlling Flooding, Multicast, and Broadcast Traffic on vMAN Egress Ports with
Rate Limiting Applied
You apply egress rate limiting to vMAN ports using vMAN ACLs. (See Chapter 12, “Access Lists
(ACLs),” for information on applying egress rate limiting to a vMAN using vMAN ACLs).
After you apply egress rate limiting to a vMAN, you can configure how to control flooding, multicast,
and broadcast traffic on the specified vMAN. You specify one port in that vMAN as the designated
port, and the system applies the egress rate limiting profile of the designated port to send flood,
multicast, and broadcast traffic on all ports on that vMAN.
To assign the port to use the rate limiting profile, use the following command:
configure vman <vman_name> rate-limit flood-traffic designated-port <port>
NOTE
This vMAN feature works only with those vMANs to which you have applied egress rate limiting using vMAN ACLs.
To remove the control on a vMAN with egress rate limiting enabled, use the following command:
unconfigure vman <vman_name> rate-limit flood-traffic designated-port
To display the designated port, use the show vman detail command.
If you do not issue this command, this traffic is not limited at egress.
Configuring Bytes Used to Calculate Traffic Rates

You configure the per packet byte adjustment in each packet the system uses to calculate the ingress
traffic rate, traffic utilization, and traffic statistics. You configure either the number of bytes you want
subtracted from each packet ingressing the specified ports or the number of bytes you want added to
the packet ingressing the specified ports.
You add or subtract bytes from packets ingressing specified ports by using the following command:
configure ports <port_list> rate-limit packet byte-adjustment [increase <add_bytes> |

decrease <sub_bytes>]
By default, all bytes are counted for the ingressing traffic rate. After you issue this command, the
default number of bytes removed is 0; you can add or subtract from 1 to 4 bytes from each ingressing
packet on the specified ports for calculating the ingressing traffic rate.
To display the number of bytes added to or subtracted from the packet to calculate the ingressing traffic
rate, traffic utilization, and traffic statistics, use the following command:
NOTE

Quality of Service
To unconfigure this setting, re-issue the command and enter the value 0.
If you use this command to decrease by 4, the show traffic queue statistics display shows only
bytes; the packets are displayed as 0.
Guidelines for Using Ingress-Only and Ingress and Egress HQoS

The following guidelines apply to HQoS, or rate limiting, on the BlackDiamond 12800 R-series switch:
● All interfaces in a given ingress traffic queue must be on the same one-half of the module (GM--XTR
or XM-2XR module), and you cannot configure cross-module traffic queues. You will receive an
error message (about failure to connect the ACL policy file) if you attempt to apply ingress traffic
queues on interfaces located on more than one-half of each module. All interfaces in a given ingress
traffic queue must be only on the specified ports on the following modules:
■ GM-XTR module—All interfaces (in all VLANs and vMANs) on a given ingress queue must be
either on ports 1 to 10 or on ports 11 to 20.
■ XM-2XR module—All interfaces (in all VLANs and vMANs) on a given ingress queue must be
either on port 1 or on port 2.
● You can apply the same traffic queue to multiple ports (on the same module), keeping in mind the
ingress queues guidelines.
● You can apply multiple traffic queues to the same interface.
● To change the meter attributes, you must delete the meter and create and configure the new meter;
then you configure that new meter with the traffic queue. (See Changing Rates on page 470 for
complete information.)
● After you apply the ACL policy file, you cannot change either the traffic queue attribute or the meter
attribute. To change either the traffic queue or meter attributes, you must unconfigure the ACL
policy file, re-configure the ACL policy file, and rebind the ACL policy file to the interfaces. (See
Changing Rates on page 470 for complete information.)
● You can specify multiple traffic queues in one ACL policy file.
● To delete a traffic queue, you must remove all ACL policy file associations; you cannot delete a
traffic queue that is currently associated with one or more ACL policy files.
● The associated meters are not deleted when you delete any type of traffic queue. Those meters
remain and can be associated with other traffic queues. To display the configured meters, use the
show meters command.
Configuring HQoS Ingress and Egress Queues
NOTE
You must be running the BlackDiamond 12800 R-series modules to use HQoS. You must use the MSM-5R plus
either the GM-20XTR module or the XM-2XR module (or both).
Ingress-only and Ingress and Egress HQoS

This section describes how to configure strict priority HQoS, or rate limiting, for ingress-only and
ingress and egress on the BlackDiamond 12800 R-series switch. You configure HQoS for:
● Ingress-only HQoS.

● Egress HQoS on the specified ingress queue. (You must previously allow egress shaping on the
specified ingress queue.)
NOTE
With ingress and egress HQoS, you are applying separate bandwidth limits to both the specified ingressing traffic
and the specified egressing traffic.
Configuring ingress-only HQoS. To configure ingress-only HQoS:
● Enable jumbo frames.

● If you are working in strict priority mode, you can configure the CoS value that directs the packet to
the high queue.
● Review the section “Guidelines for Using Ingress-Only and Ingress and Egress HQoS” on page 472.
● Create a meter.
● Configure the meter with appropriate bandwidth parameters.
● Create an ingress-only traffic queue.
● Associate the meters to the traffic queues.
● Create ACLs matching rules to traffic queues.
● Configure the ACL policy files on the desired interfaces; you must have configured the traffic
queues before this step.
To create a meter, use the following command:
To configure the meter for strict priority mode ingress-only HQoS, use the following command:
To configure the meter for bandwidth mode ingress-only HQoS, use the following command:
configure meter <metername> [committed-rate <cir> [Gbps | Mbps | Kbps] | peak-rate

<pr>[Gbps | Mbps | Kbps]]
To create an ingress-only traffic queue, use the following CLI command:
create traffic queue <queue_name> ingress-only {pr-and-cr | strict-priority}
To associate the meter to the strict priority mode traffic queue, use the following command:
configure traffic queue <queue_name> aggregate-meter <metername>
To associate the meter to the bandwidth mode traffic queue, use the following command:
configure traffic queue <queue_name> cos [{ all<metername>} | {<level> <metername>} |

bypass]
See Chapter 12, “Access Lists (ACLs),” for information on creating and configuring ACL policy files. To
display the meters already configured on the switch, use the show meters command.
To delete a traffic queue, use the following command:
delete traffic queue <queue_name>

Quality of Service
Before deleting an ingress traffic queue, you must remove all ACL policy file associations; you cannot
delete an ingress traffic queue that is currently associated with one or more ACL policy files.
When you delete an ingress traffic queue, the associated egress queue is also deleted.
Configuring ingress and egress HQoS. To configure ingress and egress HQoS:
● Enable jumbo frames.

● If you are working in strict priority mode, you can configure the CoS value that directs the packet to
the high queue.
● Review the section “Guidelines for Using Ingress-Only and Ingress and Egress HQoS” on page 472.
● Create a meter for the ingress queue that allows egress shaping.
● Create a traffic queue that specifies allowing egress shaping (this creates an ingress queue).
● Associate the meters to the ingress traffic queue that allows egress shaping.
● Create a meter for the egress queue.
● Create an egress traffic queue.
● Associate the meters to the egress traffic queue.
● Associate the egress queue with ingress queue (the specified ingress queue must allow egress
shaping).
● Create ACLs matching rules to traffic queues.
● Configure the ACLs on the desired interfaces; you must have configured the traffic queues before
this step.
To create a meter for the ingress queue that allows egress shaping, use the following command:
To configure the meter for this strict priority mode ingress queue that allows egress shaping, use the
following command:
To configure the meter for this bandwidth mode ingress queue that allows egress shaping, use the
following command:

To create an ingress traffic queue that allows egress shaping, use the following CLI command:
create traffic queue <queue_name> allow-egress-shaping {pr-and-cr | strict-priority}
To associate the meter to the strict priority traffic queue, use the following command:

bypass]

To create a meter, use the following command:
To configure the meter for egress strict priority mode HQoS, use the following command:
To configure the meter for egress bandwidth mode HQoS, use the following command:

To create an egress traffic queue, use the following CLI command:
create traffic queue <queue_name> egress {pr-and-cr | strict-priority}
To associate the meter to the strict priority mode traffic queue, use the following command:

bypass]
To associate the egress queue with an ingress queue that allows egress shaping (and specify interfaces
for the egress queue), use the following command:
configure traffic ingress queue <queue_name> add egress queue <equeue_name> ports [all
| <port_list>]
See Chapter 12, “Access Lists (ACLs),” for information on creating and configuring ACL policy files. To
display the meters already configured on the switch, use the show meters command.
After you associate an egress queue with an ingress queue, in order to delete the association between
the ingress and egress queue, use the following command:
configure traffic ingress queue <queue_name> delete egress queue [all | <equeue_name>]
ports [all | <port_list>]
To delete a queue, use the following command:
delete traffic queue <queue_name>

Before deleting an ingress traffic queue, you must remove all ACL policy file associations; you cannot
delete an ingress traffic queue that is currently associated with one or more ACL policy files.
Changing the traffic queue rate. If you need to change the rate on a traffic queue:
● Use the unconfigure access-list command.

● Delete the traffic queue.
● Create a new meter, and configure it with the rate you want.
● Create the traffic queue again, and associate it with the new meter.
● Apply the ACL policy files again.

Quality of Service
Displaying HQoS
This section describes how to display the following HQoS configurations:
● Traffic Mode on page 476
● Ingress/Egress Queue Configuration and Use on page 476
● CoS to High Queue Mapping in Strict Priority Mode HQoS Only on page 477
● vMAN-based Rate-limited Flooding Control on page 477
● Packet Size Used To Calculate Ingressing Traffic Rate on page 478
Traffic Mode
You can display the HQoS traffic mode currently running on the switch by using the following
command:
show traffic mode
To change the HQoS traffic mode, refer to “Setting the HQoS Mode” on page 466.
NOTE
You must have an enabled MEF Feature Pack to run HQoS bandwidth mode.
Ingress/Egress Queue Configuration and Use

You can display the HQoS statistics, including packets passed and dropped and queue utilization, as
well as the HQoS configuration on the switch.
To display the current ingress egress queues on your switch (HQoS or rate limiting configuration), use
show traffic queue
To display detailed information on all traffic queues or on a specific queue, use the following command:
show traffic queue {detail | <queue_name>}
To display the statistics for your current ingress-only and ingress and egress HQoS configuration, use
show traffic queue [port <port_list> | <queue_name>] statistics
If you issue the command configure ports rate-limit packet byte-adjustment, the system
calculates uses that byte size to calculate the statistics. If you use this command to decrease by 4, the
output of the show traffic queue statistics command displays only bytes; the packets are
displayed as 0.
To display utilization information for your current ingress-only and ingress and egress HQoS
configuration, use the following command:
show traffic queue [port <port_list> | <queue_name>] utilization

You can also clear the counters for these statistics, by using the following command:
clear counters traffic queue [all | <queue_name>]
Finally, to display the meters you configured, use the following command:
show meter
See Chapter 12, “Access Lists (ACLs),” for information on displaying ACL files.
CoS to High Queue Mapping in Strict Priority Mode HQoS Only

To display the CoS value configured to the highest queue in strict priority mode, use the following
command:
show port information detail
NOTE
for this feature, displayed in the High Priority COS line, on the BlackDiamond 12800 R-series switch:
.
.
Multicast Flooding: Enabled
Jumbo: Disabled
.
.
.
vMAN-based Rate-limited Flooding Control

To display the designated port, use the following command:
show vman detail
The following sample output from this command shows the setting for this feature, which is displayed
in the ERL Designated Port line:
.
.
Loopback: Disabled
NetLogin: Disabled
ERL Designated Port: 6:2
QosProfile: QP1
.
.
.

Quality of Service
Packet Size Used To Calculate Ingressing Traffic Rate

To display the number of bytes the system adds to or subtracts from each packet to calculate the
ingressing traffic rate, use the following command:
NOTE
for this feature, which is displayed in the Packet Length Adjustment line, on the BlackDiamond 12800
R-series switch:
.
.
.
Jumbo: Disabled
.
.
.
HQoS Examples
The following example shows the configuration for six ingress-only strict priority traffic queues, each
associated with a separate meter. All traffic that ingresses the interfaces associated with these six traffic
queues that does not match any of the ACL rules, passes through unlimited.
create traffic queue tq1_1 ingress-only strict-priority
create meter m1_1

create meter m2_10
create meter m3_40
create meter m4_2
create meter m5_20
create meter m6_80
configure meter "m1_1" peak-rate 1 Mbps

configure traffic queue "tq1_1" aggregate-meter m1_1


The ACL policy file, named tq1gig is displayed in the following:
NOTE
These are the ACL policy files without vMAN ACLs.
entry mac-rule-1 {
if match all {
ethernet-destination-address 00:00:00:00:00:01 ;
}
then {
permit ;
traffic-queue "tq1_100" ;
count mac-counter-1 ;
qosprofile qp8 ;
}
}
entry mac-rule-2 {
if match all {
}
then {
permit ;
qosprofile qp7 ;
}
}
entry mac-rule-3 {
if match all {
}
then {
permit ;
qosprofile qp6 ;
}
}
entry mac-rule-4 {
if match all {
}
then {
permit ;
qosprofile qp8 ;

Quality of Service
}
}
entry mac-rule-5 {
if match all {
}
then {
permit ;
qosprofile qp7 ;
}
}
entry mac-rule-6 {
if match all {
}
then {
permit ;
qosprofile qp6 ;
}
}
The following bandwidth mode HQoS example shows a user’s traffic rate limited to 10 Mbps
committed rate and 20 Mbps peak rate. (For this example, the user is connected to port 1:5 and
identified by the policy file customer1.pol.)
create meter cir_10_pr_20

configure meter cir_10_pr_20 committed-rate 10 Mbps peak-rate 20 Mbps
create traffic queue itq1
configure traffic queue tq1 ingress-only pr-and-cr
configure traffic queue itq1 cos 0 meter cir_10_pr_20
configure access-list Customer1 ports 1:5

21 Network Login

● Network Login Overview on page 481
● Configuring Network Login on page 485
● Authenticating Users on page 487
● 802.1x Authentication on page 496
● Web-Based Authentication on page 506
● MAC-Based Authentication on page 513
● Additional Network Login Configuration Details on page 516
Network Login Overview

Network login controls the admission of user packets into a network by allowing MAC addresses from
users that are properly authenticated. Network login is controlled on a per port basis. When network
login is enabled on a port, that port does not forward any packets until authentication takes place.
Network login is capable of three types of authentication: web-based, MAC-based, and 802.1x. In
addition, network login has two different modes of operation: Campus mode and ISP mode. The
authentication types and modes of operation can be used in any combination.
When web-based network login is enabled on a switch port, that port is placed into a non-forwarding
state until authentication takes place. To authenticate, a user must open a web browser and provide the
appropriate credentials. These credentials are either approved, in which case the port is placed in
forwarding mode, or not approved, in which case the port remains blocked. You can initiate user logout
by submitting a logout request or closing the logout window.
The following capabilities are included with network login:

● Web-based login using HTTP available on each port
● Web-based login using HTTPS—if you install the SSH software module that includes SSL—available
on each port
● Multiple supplicants for web-based, MAC-based, and 802.1x authentication on each port

● Web-Based, MAC-Based, and 802.1x Authentication on page 482
● Multiple Supplicant Support on page 483
● Campus and ISP Modes on page 484
● Network Login and Hitless Failover—Modular Switches and SummitStack Only on page 484

Network Login
Web-Based, MAC-Based, and 802.1x Authentication

Authentication is handled as a web-based process, MAC-based process, or as described in the
IEEE 802.1x specification. Web-based network login does not require any specific client software and
can work with any HTTP-compliant web browser. By contrast, 802.1x authentication may require
additional software installed on the client workstation, making it less suitable for a user walk-up
situation, such as a cyber-café or coffee shop.1 Extreme Networks supports a smooth transition from
web-based to 802.1x authentication.
MAC-based authentication is used for supplicants that do not support a network login mode, or
supplicants that are not aware of the existence of such security measures, for example an IP phone.
If a MAC address is detected on a MAC-based enabled network login port, an authentication request is
sent once to the AAA application. AAA tries to authenticate the MAC address against the configured
Remote Authentication Dial In User Server (RADIUS) server and its configured parameters (timeout,
retries, and so on) or the configured local database.
The credentials used for this are the supplicant’s MAC address in ASCII representation and a locally
configured password on the switch. If no password is configured the MAC address is also used as the
password. You can also group MAC addresses together using a mask.
Dynamic Host Control Protocol (DHCP) is required for web-based network login because the
underlying protocol used to carry authentication request-response is HTTP. The client requires an IP
address to send and receive HTTP packets. Before the client is authenticated, however, the only
connection that exists is to the authenticator. As a result, the authenticator must be furnished with a
temporary DHCP server to distribute the IP address.
The switch responds to DHCP requests for unauthenticated clients when DHCP parameters such as
dhcp-address-range and dhcp-options are configured on the netlogin VLAN. The switch can also
answer DHCP requests following authentication if DHCP is enabled on the specified VLAN. If netlogin
clients are required to obtain DHCP leases from an external DHCP server elsewhere on the network,
DHCP should not be enabled on the VLAN.
The DHCP allocation for network login has a short time duration of 10 seconds and is intended to
perform web-based network login only. As soon as the client is authenticated, it is deprived of this
address. The client must obtain an operational address from another DHCP server in the network.
DHCP is not required for 802.1x, because 802.1x uses only Layer 2 frames (EAPOL) or MAC-based
network login.
URL redirection (applicable to web-based mode only) is a mechanism to redirect any HTTP request to
the base URL of the authenticator when the port is in unauthenticated mode. In other words, when the
user tries to log in to the network using the browser, the user is first redirected to the network login
page. Only after a successful login is the user connected to the network. URL redirection requires that
the switch is configured with a DNS client.
Web-based, MAC-based, and 802.1x authentication each have advantages and disadvantages, as
summarized next.
1. A workstation running Windows 2000 Service Pack 4 or Windows XP supports 802.1x natively and does not
require additional authentication software.

Network Login Overview
Advantages of Web-Based Authentication:

● Works with any operating system that is capable of obtaining an IP address using DHCP. There is
no need for special client side software; only a web browser is needed.
Disadvantages of Web-Based Authentication:

● The login process involves manipulation of IP addresses and must be done outside the scope of a
normal computer login process. It is not tied to a Windows login. The client must bring up a login
page and initiate a login.
● Supplicants cannot be re-authenticated transparently. They cannot be re-authenticated from the
authenticator side.
● This method is not as effective in maintaining privacy protection.
Advantages of MAC-Based Authentication:

● Works with any operating system or network enabled device.
● Works silently. The user, client, or device does not know that it gets authenticated.
● Ease of management. A set of devices can easily be grouped by the vendor part of the MAC address.
Disadvantages of MAC-Based Authentication:

● There is no re-authentication mechanism. The FDB aging timer determines the logout.
● Security is based on the MAC address of the client, so the network is more vulnerable to spoofing
attacks.
Advantages of 802.1x Authentication:

● In cases where the 802.1x is natively supported, login and authentication happens transparently.
● Authentication happens at Layer 2. It does not involve getting a temporary IP address and
subsequent release of the address to obtain a permanent IP address.
● Allows for periodic, transparent re-authentication of supplicants.
Disadvantages of 802.1x Authentication:

● 802.1x native support is available only on newer operating systems, such as Windows XP.
● 802.1x requires an EAP-capable RADIUS Server. Most current RADIUS servers support EAP, so this
is not a major disadvantage.
● Transport Layer Security (TLS) and Tunneled TLS (TTLS) authentication methods involve Public Key
Infrastructure (PKI), which adds to the administrative requirements.
Multiple Supplicant Support

An important enhancement over the IEEE 802.1x standard is that ExtremeXOS supports multiple clients
(supplicants) to be individually authenticated on the same port. This feature makes it possible for two
or more client stations to be connected to the same port, with some being authenticated while others are
not. A port's authentication state is the logical “OR” of the individual MAC's authentication states. In
other words, a port is authenticated if any of its connected clients is authenticated. Multiple clients can
be connected to a single port of authentication server through a hub or Layer 2 switch.
Multiple supplicants are supported in ISP mode for both web-based and 802.1x authentication. On the
BlackDiamond 10808 switch and the BlackDiamond 12800 series switches, multiple supplicants are
supported in Campus mode only if all supplicants move to the same VLAN. In addition, multiple
supplicants are supported in Campus mode if you configure and enable netlogin MAC-based VLANs.

Network Login
Netlogin MAC-based VLANs are not supported on the BlackDiamond 10808 switch or 10 Gigabit
Ethernet ports. For more information, see “Configuring Netlogin MAC-Based VLANs” on page 517.
The choice of web-based versus 802.1x authentication is again on a per-MAC basis. Among multiple
clients on the same port, it is possible that some clients use web-based mode to authenticate, and some
others use 802.1x, but the restriction is that they must be in the same VLAN. This restriction is not
applicable if you configure netlogin MAC-based VLANs. For more information, see “Configuring
Netlogin MAC-Based VLANs” on page 517.
NOTE
With multiple supplicant support, after the first MAC is authenticated, the port is transitioned to the authenticated
state and other unauthenticated MACs can listen to all data destined for the first MAC. Be aware of this as
unauthenticated MACs can listen to all broadcast and multicast traffic directed to a network login-authenticated
port.
Campus and ISP Modes

Network login supports two modes of operation, Campus and ISP. Campus mode is intended for
mobile users who tend to move from one port to another and connect at various locations in the
network. ISP mode is meant for users who connect through the same port and VLAN each time (the
switch functions as an ISP).
In Campus mode, the clients are placed into a permanent VLAN following authentication with access to
network resources. For wired ports, the port is moved from the temporary to the permanent VLAN.
In ISP mode, the port and VLAN remain constant. Before the supplicant is authenticated, the port is in
an unauthenticated state. After authentication, the port forwards packets.
You do not explicitly configure the mode of operation; rather, the presence of any Extreme Networks
Vendor Specific Attribute (VSA) that has a VLAN name or VLAN ID (any VLAN attribute) in the
RADIUS server determines the mode of operation. If a VLAN attribute is present, it is assumed to be
Campus mode. If a VLAN attribute is not present, it is assumed to be ISP mode.
Network Login and Hitless Failover—Modular Switches and

SummitStack Only
When you install two Management Switch Fabric Module (MSM) modules (nodes) in a BlackDiamond
chassis or when redundancy is available in a SummitStack, one node assumes the role of primary and
the another node assumes the role of backup node. The primary node executes the switch’s
management functions, and the backup node acts in a standby role. Hitless failover transfers switch
management control from the primary node to the backup node.
NOTE
Not all platforms support hitless failover in the same software release. To verify if the software version you are
running supports hitless failover, see Table 11 in Chapter 4, “Managing the Switch.” For more information about
protocol, platform, and MSM support for hitless failover, see “Understanding Hitless Failover Support—Modular
Switches and SummitStack Only” in Chapter 4, “Managing the Switch.”

Configuring Network Login
Network login supports hitless failover by relaying current client authentication information from the
master node to the backup node. For example, if a client moves to the authenticated state, or moves
from an authenticated state to an unauthenticated state, the primary node conveys this information to
the backup node. If failover occurs, your authenticated client continues to operate as before the failover.
NOTE
If you use 802.1x network login, authenticated clients remain authenticated during failover; however, shortly after
failover, all authenticated clients automatically re-authenticate themselves. Re-authentication occurs without user
intervention.
If failover occurs during the authentication or re-authentication of a client, the client must repeat the
authentication process.
NOTE
Before initiating failover, review the section “Synchronizing Nodes—Modular Switches and SummitStack Only” on
page 1020 to confirm that your switch (or SummitStack) and both (or all) nodes are running software that supports
the synchronize command.
To initiate hitless failover on a network that uses network login:

1 Confirm that the nodes are synchronized and have identical software and switch configurations
using the show switch {detail} command. The output displays the status of the primary and
backup nodes, with the primary node showing MASTER and the backup node showing BACKUP
(InSync).
● If the primary and backup nodes, are not synchronized and both nodes are running a version of
ExtremeXOS that supports synchronization, proceed to step 2.
● If the primary and backup nodes, are synchronized, proceed to step 3.
2 If the primary and backup nodes, are not synchronized, use the synchronize command to replicate
all saved images and configurations from the primary to the backup.
After you confirm that the nodes are synchronized, proceed to step 3.
3 If the nodes are synchronized, use the run msm-failover command to initiate failover.
For more detailed information about verifying the status of the nodes and system redundancy, see
“Understanding System Redundancy—Modular Switches and SummitStack Only” on page 77. For more
information about hitless failover, see “Understanding Hitless Failover Support—Modular Switches and
SummitStack Only” on page 82.
Configuring Network Login

This section provides a general overview of the commands used for:
● Enabling or Disabling Network Login on the Switch on page 486
● Enabling or Disabling Network Login on a Specific Port on page 486
● Configuring the Move Fail Action on page 486
● Displaying Network Login Settings on page 487
This section also describes information about the Exclusions and Limitations of network login.

Network Login
For more detailed information about a specific mode of network login, including configuration
examples, refer to the following sections:
● 802.1x Authentication on page 496
● Web-Based Authentication on page 506
● MAC-Based Authentication on page 513
Enabling or Disabling Network Login on the Switch

To enable or disable network login, use one of the following commands and specify the authentication
method:
enable netlogin [{dot1x} {mac} {web-based}]
disable netlogin [{dot1x} {mac} {web-based}]
By default netlogin is disabled.
Enabling or Disabling Network Login on a Specific Port

To enable network login on a port, use the following command to specify the ports and the
authentication method:
enable netlogin ports <ports> [{dot1x} {mac} {web-based}]
By default, all methods of network login are disabled on all ports.
Network login must be disabled on a port before you can delete a VLAN that contains that port. To
disable network login, use the following command:
disable netlogin ports <ports> [{dot1x} {mac} {web-based}]
Configuring the Move Fail Action

If network login fails to perform Campus mode login, you can configure the switch to authenticate the
client in the original VLAN or deny authentication even if the user name and password are correct. For
example, this may occur if a destination VLAN does not exist. To configure the behavior of network
login if a VLAN move fails, use the following command:
configure netlogin move-fail-action [authenticate | deny]
By default, the setting is deny.
The following describes the parameters of this command if two clients want to move to a different
untagged VLAN on the same port:
● authenticate—Network login authenticates the first client that requests a move and moves that
client to the requested VLAN. Network login authenticates the second client but does not move that
client to the requested VLAN. The second client moves to the first client’s authenticated VLAN.
● deny—Network login authenticates the first client that requests a move and moves that client.
Network login does not authenticate the second client.

Displaying Network Login Settings

To display the network login settings and parameters, use the following command:
show netlogin {port <portlist> vlan <vlan_name>} {dot1x {detail}} {mac} {web-based}
Exclusions and Limitations

The following are limitations and exclusions for network login:
● All unauthenticated MACs will be seeing broadcasts and multicasts sent to the port if even a single
MAC is authenticated on that port.
● Network login must be disabled on a port before that port can be deleted from a VLAN.
● In Campus mode on all switches with untagged VLANs and the netlogin ports' mode configured as
port-based-VLAN, after the port moves to the destination VLAN, the original VLAN for that port is
not displayed.
● A network login VLAN port should not be a part of following protocols:
■ Ethernet Automatic Protection Switching (EAPS)
■ Extreme Standby Router Protocol (ESRP)
■ Spanning Tree Protocol (STP)
■ Link Aggregation
Network login uses two methods to authenticate users trying to access the network:
● RADIUS servers
● Local database
All three network login protocols, web-based, MAC-based, and 802.1x netlogin, support RADIUS
authentication. Only web-based and MAC-based netlogin support local database authentication.
NOTE
Beginning with ExtremeXOS 12.0, if you are configuring both a NetLogin Radius server and a Local-User database,
you can control which database is used first for authentication in case authentication fails. For additional details,
see the command reference pages configure netlogin authentication database-order and
unconfigure netlogin authentication database-order in the Extreme XOS 12.0 Command Reference
Guide.
This section describes the following topics in greater detail:

● Creating User Accounts on the RADIUS Server on page 488
● Configuring Local Database Authentication on page 492

Network Login
Creating User Accounts on the RADIUS Server

You can create two types of user accounts on your RADIUS server for authenticating network login
users: netlogin-only enabled and netlogin-only disabled. A netlogin-only disabled user can log in using
network login and can also access the switch using Telnet or SSH. A netlogin-only enabled user can
only log in using network login and cannot access the switch using the same login.
NOTE
For information on how to use and configure your RADIUS server, refer to the documentation that came with your
RADIUS server.
In the following example using FreeRADIUS, you add the configuration to the RADIUS server users
file. The users file determines which attributes are sent back by the RADIUS server to the RADIUS
client (an Extreme Networks switch). Depending on your RADIUS server, where and how you add the
configuration might be different.
Add the following line to the RADIUS server users file for netlogin-only disabled users:
Extreme:Extreme-Netlogin-Only = Disabled
Add the following line to the RADIUS server users file for netlogin-only enabled users:
Extreme:Extreme-Netlogin-Only = Enabled
Table 71 contains the Vendor Specific Attribute (VSA) definitions for web-based, MAC-based, and 802.1x
network login. The Extreme Networks Vendor ID is 1916.
NOTE
For information about the VSA definitions for 802.1x and Network Access Protection (NAP), see “802.1x
Authentication and Network Access Protection” on page 502.
Table 71: VSA definitions for Web-based, MAC-based, and 802.1x network login
Vendor
VSA Type Type Sent-in Description
Extreme: Netlogin- 211 String Access-Accept Name or ID of the destination VLAN after
Extended-VLAN successful authentication (must already exist on
switch).
NOTE: When using this attribute, specify
whether the port should be moved tagged or
untagged to the VLAN. Please see the guidelines
listed on page 490 for more information.
Extreme: Netlogin- 203 String Access-Accept Name of destination VLAN after successful
VLAN-Name authentication (must already exist on switch).
Extreme: Netlogin- 209 Integer Access-Accept ID of destination VLAN after successful
VLAN-ID authentication (must already exist on switch).
Extreme: Netlogin-URL 204 String Access-Accept Destination web page after successful
authentication.
Extreme: Netlogin- 205 String Access-Accept Text description of network login URL attribute.
URL-Desc

Table 71: VSA definitions for Web-based, MAC-based, and 802.1x network login (Continued)
Vendor
Extreme: Netlogin-Only 206 Integer Access-Accept Indication of whether the user can authenticate
using other means, such as telnet, console,
SSH, or Vista. A value of “1” (enabled)
indicates that the user can only authenticate via
network login. A value of “0” (disabled)
indicates that the user can also authenticate via
other methods.
Table 72 contains the standard RADIUS attributes used by network login.
Table 72: Standard RADIUS attributes used by network login

Attribute
Attribute Value Type Sent-in Description
IETF: Tunnel-Type 64 Integer Access-Accept Specifies the tunneling protocol
that is used.
IETF: Tunnel-Medium-Type 65 Integer Access-Accept Specifies the transport medium
used when creating a tunnel for
protocols (for example, VLANs) that
can operate over multiple
transports.
IETF: Tunnel-Private-Group-ID 81 String Access-Accept Specifies the VLAN ID of the
destination VLAN after successful
authentication; used to derive the
VLAN name.
The NetLogin-Url and NetLogin-Url-Desc attributes are used in case of Web-based login as the page to
use for redirection after a successful login. Other authentication methods will ignore these attributes.
The other attributes are used in the following order to determine the destination VLAN to use:
● Extreme: Netlogin-Extended-VLAN (VSA 211)
● Extreme: Netlogin-VLAN-Name (VSA 203)
● Extreme: Netlogin-VLAN-ID (VSA 209)
● IETF: Tunnel-Private-Group-ID representing the VLAN TAG as a string, but only if IETF: Tunnel-
Type == VLAN(13) and IETF: Tunnel-Medium-Type == 802 (6).
If none of the previously described attributes are present ISP mode is assumed, and the client remains
in the configured VLAN.
Guidelines and Examples for Using VSAs

This section contains guidelines and examples for using the Extreme Networks VSAs listed in Table 71.
The examples in this section use FreeRADIUS to modify the VSA. Depending on your RADIUS server,
configuration might be different.
NOTE
RADIUS server.

Network Login
VSA 211—Extreme: Netlogin-Extended-VLAN. The following describes the guidelines for VSA 211:
● For tagged VLAN movement with 802.1x netlogin, you must use VSA 211.
● For untagged VLAN movement with 802.1x netlogin, you can use all current Extreme Networks
VLAN VSAs: VSA 203, VSA 209, and VSA 211.
● To specify the VLAN name or the VLAN ID, use an ASCII string; however, you cannot specify both
the VLAN name and the VLAN ID at the same time. If the string only contains numbers, it is
interpreted as the VLAN ID.
● For tagged VLANs, specify T for tagged before the VLAN name or VLAN ID.
● For untagged VLANs, specify U for untagged before the VLAN name or VLAN ID.
● For movement based on the incoming port’s traffic, specify * before the VLAN name or VLAN ID.
The behavior can be either tagged or untagged, based on the incoming port’s traffic, and mimics the
behavior of VSA 203 and VSA 209, respectively.
VSA 211 Examples. The following examples use FreeRADIUS to modify the VSA to support tagged or
untagged VLANs using either the VLAN name or the VLAN ID.
● Configuring VLAN names
The three options to use when configuring VLAN names are:
■ T—Include before the VLAN name for a tagged VLAN
■ U—Include before the VLAN name for an untagged VLAN
■ *—Include before the VLAN name for movement based on the incoming port’s traffic (mimics
the behavior of VSA 203)
To configure the tagged VLAN voice, add the following line:
Extreme-Netlogin-Extended-VLAN = Tvoice
To configure the untagged VLAN data, add the following line:
Extreme-Netlogin-Extended-VLAN = Udata
To configure the VLAN orange to be tagged or untagged based on the incoming port’s traffic, add
the following line:
Extreme-Netlogin-Extended-VLAN = *orange
● Configuring VLAN IDs
The same is true of configuring the VLAN ID. The three options to use when configuring VLAN IDs
are:
■ T—Include before the VLAN ID for a tagged VLAN
■ U—Include before the VLAN ID for an untagged VLAN
■ *—Include before the VLAN name for movement based on the incoming port’s traffic (mimics
the behavior of VSA 209)
To configure a tagged VLAN with a VLAN ID of 229, add the following line:
Extreme-Netlogin-Extended-VLAN = T229
To configure an untagged VLAN with a VLAN ID of 4091, add the following line:
Extreme-Netlogin-Extended-VLAN = U4091
To configure a VLAN with a VLAN ID of 145 to be tagged or untagged based on the incoming
port’s traffic, add the following line:
Extreme-Netlogin-Extended-VLAN = *145

VSA 203—Extreme: Netlogin-VLAN-Name. The following describes the guidelines for VSA 203:
● To specify the VLAN name, use an ASCII string.
When using this VSA, you do not specify a tagged or untagged VLAN.
VSA 203 Example. The following example modifies the VSA to specify the name of the destination
VLAN. To configure VLAN purple as the destination VLAN, add the following line:
Extreme: Netlogin-VLAN-Name = purple
VSA 209—Extreme: Netlogin-VLAN-ID. The following describes the guidelines for VSA 209:
● To specify the VLAN ID, use an ASCII string.
When using this VSA, you do not specify a tagged or untagged VLAN.
VSA 209 Example. The following example modifies the VSA to specify the VLAN ID (tag) of the
destination VLAN. To configure the VLAN with VLAN ID 234 as the destination VLAN, add the
following line:
Extreme: Netlogin-VLAN-ID = 234
VSA 204—Extreme: Netlogin-URL. The following describes the guidelines for VSA 204:
● To specify the URL to display after authentication, use an ASCII string.
● If you do not specify a URL, the network login infrastructure uses the default redirect page URL,
http://www.extremenetworks.com, or the URL that you configured using the configure netlogin
redirect-page command.
VSA 204 applies only to the web-based authentication mode of Network Login.
VSA 204 Example. The following example modifies the VSA to specify the destination URL after
successful authentication. To configure the redirect URL as http://www.myhomepage.com, add the
following line:
Extreme: Netlogin-URL = http://www.myhomepage.com
VSA 205—Extreme: Netlogin-URL-Desc. The following describes the guidelines for VSA 205:
● To include a description of the redirect page URL (as specified by VSA 204), use an ASCII string.
● To let the user know where they will be redirected to after authentication, use an ASCII string to
provide a brief description of the URL.
VSA 205 applies only to the web-based authentication mode of Network Login.
VSA 205 Example. The following example modifies the VSA to describe the network login URL. To
describe the network login URL as my home page, add the following line:
Extreme: Netlogin-URL-Desc = homepage

Network Login
VSA 206—Extreme: Netlogin-Only. The following describes the guidelines for VSA 206:
● To specify that a user can authenticate only via network login, use a value of “1” (enabled).
● To specify that a user can authenticate via other methods, use a value of “0” (disabled).
VSA 206 Example. See the examples described in the section “Creating User Accounts on the RADIUS
Server” on page 488.
Configuring Local Database Authentication

You can configure the switch to use its local database for web-based and MAC-based network login
authentication. 802.1x network login does not support local database authentication. Local
authentication essentially mimics the functionality of the remote RADIUS server locally. This method of
authentication is useful in the following situations:
● If both the primary and secondary (if configured) RADIUS servers timeout or are unable to respond
to authentication requests
● If no RADIUS servers are configured
● If the RADIUS server used for network login authentication is disabled
If any of the above conditions are met, the switch checks for a local user account and attempts to
authenticate against that local account.
For local authentication to occur, you must configure the switch’s local database with a user name and
password for network login. Extreme Networks recommends a maximum of 64 local accounts. If you
need more than 64 local accounts, Extreme Networks recommends using RADIUS for authentication.
Beginning with ExtremeXOS 11.3 you can also specify the destination VLAN to enter upon a successful
authentication.
You can also use local database authentication in conjunction with netlogin MAC-based VLANs. For
more detailed information about netlogin MAC-based VLANs, see “Configuring Netlogin MAC-Based
VLANs” on page 517.
The following sections describe how to configure your switch for local database authentication:
● Creating a Local Netlogin Account—User Name and Password Only on page 492
● Specifying a Destination VLAN on page 493
● Modifying an Existing Local Netlogin Account on page 494
● Displaying Local Netlogin Accounts on page 495
● Deleting a Local Netlogin Account on page 495
Creating a Local Netlogin Account—User Name and Password Only

Extreme Networks recommends creating a maximum of 64 local accounts. If you need more than 64
local accounts, Extreme Networks recommends using RADIUS for authentication. For information about
RADIUS authentication, see “Creating User Accounts on the RADIUS Server” on page 488.
To create a local netlogin user name and password, use the following command and specify the <user-
name> parameter:
create netlogin local-user <user-name> {encrypted <password>} {vlan-vsa [[{tagged |

untagged} [<vlan_name>] | <vlan_tag>]]}

User names are not case-sensitive; passwords are case-sensitive. User names must have a minimum of 1
character and a maximum of 32 characters. Passwords must have a minimum of 0 characters and a
maximum of 32 characters. If you use RADIUS for authentication, Extreme Networks recommends that
you use the same user name and password for both local authentication and RADIUS authentication.
If you attempt to create a user name with more than 32 characters, the switch displays the following
messages:
%% Invalid name detected at '^' marker.

%% Name cannot exceed 32 characters.
If you attempt to create a password with more than 32 characters, the switch displays the following
message after you re-enter the password:
Password cannot exceed 32 characters
The encrypted option is used by the switch to encrypt the password. Do not use this option through the
command line interface (CLI). After you enter a local netlogin user name, press [Return]. The switch
prompts you twice to enter the password.
The following example:

● Creates a new local netlogin user name
● Creates a password associated with the local netlogin user name
create netlogin local-user megtest
password: <Enter the password. The switch does not display the password.>
Reenter password: <Re-enter the password. The switch does not display the password.>
For information about specifying the destination VLAN, see the next section “Specifying a Destination
VLAN.”
Specifying a Destination VLAN

If you configure a local netlogin account with a destination VLAN, upon successful authentication, the
client transitions to the permanent, destination VLAN. You can specify the destination VLAN when you
initially create the local netlogin account or at a later time.
Adding VLANs when Creating a Local Netlogin Account. To specify the destination VLAN when creating
the local netlogin account, use the following command and specify the vlan-vsa option with the
associated parameters:
create netlogin local-user <user-name> {encrypted <password>} {vlan-vsa [[{tagged |
untagged} [<vlan_name>] | <vlan_tag>]]}

● tagged—Specifies that the client be added as tagged
● untagged—Specifies that the client be added as untagged
● vlan_name—Specifies the name of the destination VLAN
● vlan_tag—Specifies the VLAN ID, tag, of the destination VLAN

Network Login

● Creates a new local netlogin user name
● Creates a password associated with the local netlogin user name
● Adds the VLAN test1 as the destination VLAN
The following is a sample display from this command:
create netlogin local-user megtest vlan-vsa "test1"

password: <Enter the password. The switch does not display the password.>
Reenter password: <Re-enter the password. The switch does not display the password.>
Adding VLANs at a Later Time. To specify the destination VLAN after you created the local netlogin
account, use the following command:
configure netlogin local-user <user-name> {vlan-vsa [[{tagged | untagged}

[<vlan_name>] | <vlan_tag>]] | none]}

● vlan_tag—Specifies the VLAN ID, tag, of the destination VLAN
● none—Specifies that the VSA 211 wildcard (*) is applied, only if you do not specify tagged or
untagged

● Modifies a previously created local netlogin account
● Specifies that clients are added as tagged to the VLAN
● Adds the VLAN blue as the destination VLAN
configure netlogin local-user megtest vlan-vsa tagged "blue"
Modifying an Existing Local Netlogin Account

After you create a local netlogin user name and password, you can update the following attributes of
that account:
● Password of the local netlogin account
● Destination VLAN attributes including: adding clients tagged or untagged, the name of the VLAN,
and the VLAN ID
If you try modifying a local netlogin account that is not present on the switch, or you incorrectly enter
the name of the account, output similar to the following appears:
* SummitX450.14 # configure netlogin local-user purplenet

^
%% Invalid input detected at '^' marker.
To confirm the names of the local netlogin accounts on your switch, use the following command:
show netlogin local-users

Updating the Local Netlogin Password. To update the password of an existing local netlogin account, use
configure netlogin local-user <user_name>
Where user_name specifies the name of the existing local netlogin account. After you enter the local
netlogin user name, press [Return]. The switch prompts you to enter a password. At the prompt enter
the new password and press [Return]. The switch then prompts you to reenter the password.
Passwords are case-sensitive. Passwords must have a minimum of 0 characters and a maximum of 32
characters. If you attempt to create a password with more than 32 characters, the switch displays the
following message after you re-enter the password:
Password cannot exceed 32 characters
The following example modifies the password for the existing local netlogin account megtest. The
following is a sample display from this command:
configure netlogin local-user megtest

password: <Enter the new password. The switch does not display the password.>
Reenter password: <Re-enter the new password. The switch does not display the
password.>
After you complete these steps, the password has been updated.
Updating VLAN Attributes. You can add a destination VLAN, change the destination VLAN, or remove
the destination VLAN from an existing local netlogin account. To make any of these VLAN updates,
configure netlogin local-user <user-name> {vlan-vsa [[{tagged | untagged}
[<vlan_name>] | <vlan_tag>]] | none]}

● user_name—Specifies the name of the existing local netlogin account
● vlan_name_tag—Specifies the VLAN ID, tag, of the destination VLAN
● none—Specifies that the VSA 211 wildcard (*) is applied, only if you do not specify tagged or
untagged
Displaying Local Netlogin Accounts

To display a list of local netlogin accounts on the switch, including VLAN information, use the
following command:
show netlogin local-users
Deleting a Local Netlogin Account

To delete a local netlogin user name and password, use the following command:
delete netlogin local-user <user-name>

Network Login
802.1x Authentication
802.1x authentication methods govern interactions between the supplicant (client) and the
authentication server. The most commonly used methods are Transport Layer Security (TLS); Tunneled
TLS (TTLS), which is a Funk/Certicom standards proposal; and PEAP.
TLS is the most secure of the currently available protocols, although TTLS is advertised to be as strong
as TLS. Both TLS and TTLS are certificate-based and require a Public Key Infrastructure (PKI) that can
issue, renew, and revoke certificates. TTLS is easier to deploy, as it requires only server certificates, by
contrast with TLS, which requires client and server certificates. With TTLS, the client can use the MD5
mode of user name/password authentication.
If you plan to use 802.1x authentication, refer to the documentation for your particular RADIUS server
and 802.1x client on how to set up a PKI configuration.

● Interoperability Requirements on page 496
● Enabling and Disabling 802.1x Network Login on page 497
● 802.1x Network Login Configuration Example on page 497
● Configuring Guest VLANs on page 498
● Post-authentication VLAN Movement on page 501
● 802.1x Authentication and Network Access Protection on page 502
Interoperability Requirements
For network login to operate, the user (supplicant) software and the authentication server must support
common authentication methods. Not all combinations provide the appropriate functionality.
Supplicant Side
The supported 802.1x clients (supplicants) are Windows 2000 SP4 native client, Windows XP native
clients, and Meetinghouse AEGIS.
A Windows XP 802.1x supplicant can be authenticated as a computer or as a user. Computer

authentication requires a certificate installed in the computer certificate store, and user authentication
requires a certificate installed in the individual user's certificate store.
By default, the Windows XP machine performs computer authentication as soon as the computer is
powered on, or at link-up when no user is logged into the machine. User authentication is performed at
link-up when the user is logged in.
Windows XP also supports guest authentication, but this is disabled by default. Refer to relevant
Microsoft documentation for further information. The Windows XP machine can be configured to
perform computer authentication at link-up even if user is logged in.

Authentication Server Side

The RADIUS server used for authentication must be EAP-capable. Consider the following when
choosing a RADIUS server:
● Types of authentication methods supported on RADIUS, as mentioned previously.
● Need to support VSAs. Parameters such as Extreme-Netlogin-Vlan-Name (destination vlan for port
movement after authentication) and Extreme-NetLogin-Only (authorization for network login only)
are brought back as VSAs.
● Need to support both EAP and traditional user name-password authentication. These are used by
network login and switch console login respectively.
NOTE
RADIUS server.
Enabling and Disabling 802.1x Network Login

To enable 802.1x network login on the switch, use the following command:
enable netlogin dot1x
Any combination of types of authentication can be enabled on the same switch. At least one of the
authentication types must be specified on the CLI.
To disable 802.1x network login on the switch, use the following command:
disable netlogin dot1x
To enable 802.1x network login on one or more ports, use the following command:
enable netlogin ports <portlist> dot1x

Network Login must be disabled on a port before you can delete a VLAN that contains that port. To
disable 802.1x network login on one or more ports, use the following command:
disable netlogin ports <portlist> dot1x
802.1x Network Login Configuration Example

The following configuration example shows the Extreme Networks switch configuration needed to
support the 802.1x network login example.
NOTE
In the following sample configuration, any lines marked (Default) represent default settings and do not need to
be explicitly configured.
create vlan “temp”

create vlan “corp”
configure vlan “default” delete ports 4:1-4:4
# Configuration Information for VLAN corp

Network Login
# No VLAN-ID is associated with VLAN corp.

configure vlan “corp” protocol “ANY” (Default)
configure vlan “corp” ipaddress 10.203.0.224 255.255.255.0
# Configuration Information for VLAN Mgmt

configure vlan “Mgmt” ipaddress 10.10.20.30 255.255.255.0
# Network Login Configuration

configure netlogin vlan “temp”
enable netlogin dot1x
enable netlogin ports 1:10-1:14, 4:1-4:4 dot1x
# RADIUS Configuration
configure radius netlogin primary server 10.0.1.2 1812 client-ip 10.10.20.30 vr “VR-
Mgmt”
configure radius netlogin primary shared-secret purple
enable radius
The following example is for the FreeRADIUS server; the configuration might be different for your
RADIUS server:
#RADIUS Server Setting, in this example the user name is eaptest

eaptest Auth-Type := EAP, User-Password == "eaptest"
Termination-Action =1
NOTE
For information about how to use and configure your RADIUS server, refer to the documentation that came with your
RADIUS server.
Configuring Guest VLANs

Ordinarily, a client that does not respond to 802.1x authentication remains disabled and cannot access
the network. 802.1x authentication supports the concept of “guest VLANs” that allow such a supplicant
(client) limited or restricted network access. If a supplicant connected to a port does not respond to the
802.1x authentication requests from the switch, the port moves to the configured guest VLAN. A port
always moves untagged into the guest VLAN.
NOTE
The supplicant does not move to a guest VLAN if it fails authentication after an 802.1x exchange; the supplicant
moves to the guest VLAN only if it does not respond to an 802.1x authentication request.
When the authentication server sends an 802.1x request to the supplicant, there is a specified time
interval for the supplicant to respond. By default, the switch uses the supplicant response timer to
authenticate the supplicant every 30 seconds for a maximum of three tries. If the supplicant does not
respond within the specified time, the authentication server sends another request. After the third
802.1x request without a supplicant response, the port is placed in the guest VLAN, if the guest VLAN
feature has been configured for the port. The number of authentication attempts is not a user-
configured parameter.

If a supplicant on a port in the guest VLAN becomes 802.1x-capable, the switch starts processing the
802.1x responses from the supplicant. If the supplicant is successfully authenticated, the port moves
from the guest VLAN to the destination VLAN specified by the RADIUS server. If the RADIUS server
does not specify a destination VLAN, the port moves to the VLAN it belonged to before it was placed
in the guest VLAN. After a port has been authenticated and moved to a destination VLAN, it is
periodically re-authenticated. If the port fails authentication, it moves to the VLAN to which it belonged
originally.
NOTE
A guest VLAN is not a normal netlogin VLAN. A guest VLAN performs authentication only if authentication is
initiated by the supplicant.

● Using Guest VLANs on page 499
● Guidelines for Configuring Guest VLANs on page 500
● Creating Guest VLANs on page 500
● Enabling Guest VLANs on page 501
● Modifying the Supplicant Response Timer on page 501
● Disabling Guest VLANs on page 501
● Unconfiguring Guest VLANs on page 501
● Displaying Guest VLAN Settings on page 501
Using Guest VLANs

Suppose you have a meeting that includes company employees and visitors from outside the company.
In this scenario, your employees have 802.1x enabled supplicants but your visitors do not. By
configuring a guest VLAN, when your employees log into the network, they are granted network access
(based on their user credentials and 802.1x enabled supplicants). However, when the visitors attempt to
log into the network, they are granted limited network access because they do not have 802.1x enabled
supplicant. The visitors might be able to reach the Internet, but they are unable to access the corporate
network.
For example, in Figure 53 Host A has 802.1x capability and Host B does not. When Host A is
authenticated, it is given full access to the network. Host B does not have 802.1x capability and
therefore does not respond to 802.1x requests from the switch. If port B is configured with the guest
VLAN, port B is moved to the guest VLAN. Then Host B will be able to access the Internet but not the
corporate network. After Host B is equipped with 802.1x capability, it can be authenticated and allowed
to be part of the corporate network.

Network Login
Figure 53: Guest VLAN for netlogin
Host B
No 802.1x
capabilities
Authentication
server
Port B
Extreme switch Internet

Host A
Has 802.1x
capabilities
Port A
Corporate
network
EW_110
EX_173
Guidelines for Configuring Guest VLANs

Keep in mind the following guidelines when configuring guest VLANs:
● You must create a VLAN and configure it as a guest VLAN before enabling the guest VLAN feature.
● Configure guest VLANs only on netlogin ports with 802.1x enabled.
● Movement to guest VLANs is not supported on netlogin ports with MAC-based or web-based
authentication.
● 802.1x must be the only authentication method enabled on the port for movement to guest VLAN.
● No supplicant on the port has 802.1x capability.
Creating Guest VLANs

If you configure a guest VLAN, and a supplicant has 802.1x disabled and does not respond to 802.1x
authentication requests from the switch, the supplicant moves to the guest VLAN. Upon entering the
guest VLAN, the supplicant gains limited network access.
NOTE
Beginning with ExtremeXOS 11.6, you can configure guest VLANs on a per port basis, which allows you to configure
more than one guest VLAN per VR. In ExtremeXOS 11.5 and earlier, you can only configure guest VLANs on a per
VR basis, which allows you to configure only one guest VLAN per VR.
To create a guest VLAN, use the following command:
configure netlogin dot1x guest-vlan <vlan_name> {ports <port_list>}

Enabling Guest VLANs

To enable the guest VLAN, use the following command:
enable netlogin dot1x guest-vlan ports [all | <ports>]
Modifying the Supplicant Response Timer

To modify the supplicant response timer, use the following command and specify the supp-resp-
timeout parameter:
configure netlogin dot1x timers [{server-timeout <server_timeout>} {quiet-period
<quiet_period>} {reauth-period <reauth_period>} {supp-resp-timeout
<supp_resp_timeout>}]
The default supplicant response timeout is 30 seconds, and the range is 1 to 120 seconds. The number of
authentication attempts is not a user-configured parameter.
Disabling Guest VLANs

To disable the guest VLAN, use the following command:
disable netlogin dot1x guest-vlan ports [all | <ports>]
Unconfiguring Guest VLANs

To unconfigure the guest VLAN, use the following command:
unconfigure netlogin dot1x guest-vlan {ports <port_list> | <vlan_name>}
Displaying Guest VLAN Settings

To display the guest VLAN settings, use the following command:
show netlogin guest-vlan {vlan_name}
If you specify the vlan_name, the switch displays information for only that guest VLAN.
The output displays the following information in a tabular format:

● Port—Specifies the 802.1x enabled port configured for the guest VLAN.
● Guest-vlan—Displays guest VLAN name and status: enable/disable.
● Vlan—Specifies the name of the guest VLAN.
Post-authentication VLAN Movement

After the supplicant has been successfully authenticated and the port has been moved to a VLAN, the
supplicant can move to a VLAN other than the one it was authenticated on. This occurs when the
RADIUS server sends a message to the suppliant telling it of the new VLAN during 802.1x re-
authentication. The supplicant remains authenticated during this transition. This occurs on both
untagged and tagged VLANs.

Network Login
For example, suppose a supplicant submits the required credentials for network access; however, it is
not running the current, approved anti-virus software or it does not have the appropriate software
updates installed. If this occurs, the supplicant is authenticated but has limited network access until the
problem is resolved. After you update the supplicant’s anti-virus software, or install the software
updates, the RADIUS server re-authenticates the supplicant by sending ACCESS-ACCEPT messages
with the accompanying VLAN attributes, thereby allowing the supplicant to enter its permanent VLAN
with full network access.
This is normal and expected behavior; no configuration is necessary.
802.1x Authentication and Network Access Protection

Beginning with ExtremeXOS 11.6, 802.1x authentication in combination with Microsoft’s Network
Access Protection (NAP) provide additional integrity checks for end users and supplicants that attempt
to access the network. NAP allows network administrators to create system health policies to ensure
supplicants that access or communicate with the network meet administrator-defined system health
requirements. For example, if a supplicant has the appropriate software updates or anti-virus software
installed, the supplicant is deemed healthy and granted network access. On the other hand, if a
supplicant does not have the appropriate software updates or anti-virus software installed, the
supplicant is deemed unhealthy and is placed in a quarantine VLAN until the appropriate update or
anti-virus software is installed. After the supplicant is healthy, it is granted network access. For more
information about NAP, please refer to the documentation that came with your Microsoft Windows or
Microsoft Server software.
To configure your network for NAP, the minimum required components are:
● Extreme Networks switches running ExtremeXOS 11.6 or later.
● RADIUS server that supports NAP (Microsoft Windows Vista operating system refers to this as a
network policy server (NPS), formerly known as the internet authentication server (IAS)).
● Remediation servers that receive unhealthy supplicants. The remediation servers contain the
appropriate software updates, anti-virus software, and so on to make a supplicant healthy.
In addition to the required hardware and software, you must configure NAP-specific VSAs on your
RADIUS server. By configuring these VSAs, you ensure supplicant authentication and authorization to
the network and the switch creates dynamic Access Control Lists (ACLs) to move unhealthy supplicants
to the quarantine VLAN for remediation. For more information see, “Using NAP-Specific VSAs to
Authenticate 802.1x Supplicants” on page 504.
Figure 54 displays a sample network that uses NAP to protect the network.

Figure 54: Sample network using NAP to provide enhanced security
Domain controller running

Active Directory
IAS/NPS
EpiCenter
Remediation
server
Production VLAN
(Corporate network)
Remediation
server Quarantine
Workstation VLAN EX_174
Example Scenarios Using NAP

Using Figure 54, the following two scenarios describe some sample actions taken when an 802.1x-
enabled supplicant initiates a connection to the network. The scenarios assume the following:
● Scenario 1 has a healthy 802.1x-enabled supplicant.
● Scenario 2 has an unhealthy 802.1x-enabled supplicant.
● 802.1x netlogin has been configured and enabled on the switch.
● The RADIUS server has been configured using the NAP-specific VSAs for authenticating
supplicants.
● The remediation servers have been configured with the appropriate software updates, anti-virus
software, and so on.
● The EPICenter server has been configured to receive traps from the switch. The traps sent from the
switch inform EPICenter of the state of the supplicant. In these scenarios, you configure EPICenter as
the syslog target.
● VLANs Production and Quarantine have already been created and configured.
NOTE
You can dynamically create the quarantine VLAN if you configure dynamic VLAN creation on the switch. For
more information see, “Configuring Dynamic VLANs for Netlogin” on page 519.

Network Login
Scenario 1—Healthy Supplicant. The steps to authenticate a healthy supplicant are:

1 The 802.1x supplicant initiates a connection to the 802.1x network access server (NAS), which in this
scenario is the Extreme Networks switch.
2 The supplicant passes its authentication credentials to the switch using PEAP and an inner
authentication method such as MS-CHAPv2.
3 The RADIUS server requests a statement of health (SoH) from the supplicant.
Only NAP-capable supplicants create an SoH, which contains information about whether or not the
supplicant is compliant with the system health requirements defined by the network administrator.
4 If the SoH indicates that the supplicant is healthy, the RADIUS server sends an Access-Accept
message with a RADIUS VSA indicating which VLAN the healthy supplicant is moved to (in this
example, the Production VLAN).
5 The switch authenticates the supplicant and moves it into the Production VLAN.
6 The switch sends a trap to EPICenter indicating that the supplicant has been successfully
authenticated and the VLAN into which it has been moved.
Scenario 2—Unhealthy Supplicant. The steps to authenticate an unhealthy supplicant are:

1 The 802.1x supplicant initiates a connection to the 802.1x network access server (NAS), which in this
scenario is the Extreme Networks switch.
2 The supplicant passes its authentication credentials to the switch using PEAP and an inner
authentication method such as MS-CHAPv2.
3 The RADIUS server requests a statement of health (SoH) from the supplicant.
Only NAP-capable supplicants create an SoH, which contains information about whether or not the
supplicant is compliant with the system health requirements defined by the network administrator.
4 If the SoH indicates that the supplicant is unhealthy, the RADIUS server sends an Access-Accept
message with RADIUS VSAs indicating which:
● VLAN the unhealthy supplicant is moved to (in this example, the Quarantine VLAN)
● Remediation server(s) the supplicant can get software updates, anti-virus software and so on to
remediate itself
5 When the switch receives the VLAN and remediation server information from the RADIUS server,
the switch:
● Moves the supplicant into the Quarantine VLAN.
● Applies ACLs to ensure the supplicant in the Quarantine VLAN can access only the remediation
servers. All other traffic not originating/destined from/to the remediation servers is dropped.
● Sends a trap to EPICenter indicating that the supplicant has been authenticated but has restricted
access in the Quarantine VLAN for remediation.
6 The supplicant connects to the remediation server to get software updates, anti-virus software, and
so on to get healthy.
7 After the supplicant is healthy, it re-starts the authentication process and is moved to the Production
VLAN, as a healthy supplicant with full network access.
Using NAP-Specific VSAs to Authenticate 802.1x Supplicants

Table 73 contains the VSA definitions for 802.1x network login in conjunction with devices and servers
that support NAP. The Microsoft Vendor ID is 311.

NOTE
For more information about NAP and the VSAs supported by NAP, please refer to the documentation that came with
your Microsoft operating system or server.
Table 73: NAP-specific VSA definitions for 802.1x network login

Vendor
MS-Quarantine-State 45 Integer Access-Accept Indicates the network access level that the
RADIUS server authorizes the user. The network
access server (the switch) also enforces the
network access level. A value of “0” gives the
user full network access. A value of “1” gives
the user limited network access. A value of “2”
gives the user full network access within a
specified time period.
MS-IPv4-Remediation- 52 Integer Access-Accept Indicates the IP address(es) of the remediation
Servers server(s) that an unhealthy supplicant moves to
in order to get healthy.
ACLS for Remediation Servers

The NAP VSA, MS-IPv4-Remediation-Servers, contains a list of IP addresses that an unhealthy and
therefore quarantined supplicant should be allowed access to so that it can remediate itself and become
healthy.
The way a quarantine is implemented on the switch is simply by moving the client/port to a user-
designated 'quarantine' VLAN whose VLANID/Name is sent in the Access-Accept message. It is up to
the user to ensure that the quarantine VLAN does indeed have limited access to the rest of the network.
Typically, this can be done by disabling IP forwarding on that VLAN so no routed traffic can get out of
that VLAN. Also, with dynamic VLAN creation, the quarantine VLAN being supplied by RADIUS
could be dynamically created on the switch, once dynamic VLAN creation is enabled on it. The
remediation server(s) would need to be accessible via the uplink port, regardless of whether the
quarantine VLAN is pre-configured or dynamically created, since IP forwarding is not enabled on it.
To get around this restriction, netlogin has been enhanced so when a MS-Quarantine-State attribute is
present in the Access-Accept message with extremeSessionStatus being either 'Quarantined' or 'On
Probation,' then a 'deny all traffic' dynamic ACL will be applied on the VLAN. If such an ACL is
already present on that VLAN, then no new ACL will be applied.
When the last authenticated client has been removed from the quarantine VLAN, then the above ACL
will be removed.
Additionally, if the MS-IPv4-Remediation-Servers VSA is present in the Access-Accept message, for

each IP address present in the VSA a 'permit all traffic to/from this IP address' ACL will be applied on
the quarantine VLAN. This will allow traffic to/from the remediation servers to pass unhindered in the
Quarantine VLAN while all other traffic will be dropped.

Network Login
Web-Based Authentication
This section describes web-based network login. For web-based authentication, you need to configure
the switch DNS name, default redirect page, session refresh, and logout-privilege. URL redirection
requires the switch to be assigned a DNS name. The default name is network-access.net. Any DNS
query coming to the switch to resolve switch DNS name in unauthenticated mode is resolved by the
DNS server on the switch in terms of the interface (to which the network login port is connected to)
IP address.

● Enabling and Disabling Web-Based Network Login on page 506
● Configuring the Base URL on page 506
● Configuring the Redirect Page on page 507
● Configuring Session Refresh on page 507
● Configuring Logout Privilege on page 507
● Configuring the Login Page on page 507
● Customizable Authentication Failure Response on page 509
● Web-Based Network Login Configuration Example on page 510
● Web-Based Authentication User Login on page 511
Enabling and Disabling Web-Based Network Login

To enable web-based network login on the switch, use the following command:
enable netlogin web-based
To disable web-based network login on the switch, use the following command:
disable netlogin web-based
To enable web-based network login on one or more ports, use the following command:
enable netlogin ports <portlist> web-based

disable web-based network login on one or more ports, use the following command:
disable netlogin ports <portlist> dot1x
Configuring the Base URL

To configure the network login base URL, use the following command:
configure netlogin base-url <url>
Where <url> is the DNS name of the switch. For example, configure netlogin base-url network-
access.net makes the switch send DNS responses back to the netlogin clients when a DNS query is
made for network-access.net.

Configuring the Redirect Page

To configure the network login redirect page, use the following command:
configure netlogin redirect-page <url>
Where <url> defines the redirection information for the users after they have logged in. You must
configure a complete URL starting with http:// or https://
By default, the redirect URL value is “http://www.extremenetworks.com”.
This redirection information is used only in case the redirection info is missing from RADIUS server.
For example, configure netlogin base-url http://www.extremenetworks.com redirects all users
to this URL after they get logged in.
To support https, you must first download and install the separate Extreme Networks SSH software
module (ssh.xmod). This additional module allows you to configure both SSH2 and SSL on the switch.
For more information about SSH2, see Chapter 16, “Security.” For information about installing the SSH
module, see Appendix A, “Software Upgrade and Boot Options.”
Configuring Session Refresh

To enable or disable the network login session refresh, use one of the following commands:
enable netlogin session-refresh {<minutes>}
disable netlogin session-refresh
Where <minutes> ranges from 1 - 255. The default setting is 3 minutes. enable netlogin session-
refresh makes the logout window refresh itself at every configured time interval. Session refresh is
disabled by default. When you configure the network login session refresh for the logout window,
ensure that the FDB aging timer is greater than the network login session refresh timer.
Configuring Logout Privilege

To enable or disable network login logout privilege, use one of the following commands:
enable netlogin logout-privilege
disable netlogin logout-privilege
These commands turn the privilege for netlogin users to logout by popping up (or not popping up) the
logout window. Logout-privilege is enabled by default.
Configuring the Login Page

Beginning with ExtremeWare XOS Release 12.0, you can fully customize the HTML login page and also
add custom embedded graphical images to it. This page and the associated graphics must be uploaded
to the switch so that they can be served up as the initial login page at the base URL. Both HTTP and
HTTPS are supported as a means of authenticating the user via the custom page.
In general, the steps for setting up a custom login page and graphical images (if any) are as follows:
1 Write the custom web-page.

2 TFTP the page and any embedded JPEG or GIF graphical images that it references onto the switch.

Network Login
3 Enable and configure web-based Network Login on the switch. When the custom page is present on
the switch, it will over-ride the configured banner.
Login Page Contents

The customized web-page must have the file name netlogin_login_page.html. While the contents of the
page are left up to the customer, they must contain the following elements:
● An HTML submit form with action="/hello" and method="post" that is used to send the Network
Login username and password to the switch. The form must contain the following:
■ A username input field with name="extremenetloginuser"
■ A password input field with name="extremenetloginpassword"
● Optionally, one or more graphical images embedded using the tags
<img src="netlogin_<xxx>.jpg"> or
<img src="netlogin_<xxx>.jpeg"> or
<img src="<netlogin_<xxx>.gif">
where <xxx> is user-configurable.
The following is a sample custom page, where the embedded graphical image is named
netlogin_welcome.jpg:
<html lang="en">
<head>
<title>Network Login Page</title>
</head>
<body>
<form action="/hello" method="post">
<img src="netlogin_welcome.jpg">
<br/>
Please log in:
<br/>
User:
<input type="text" name="extremenetloginuser" />
<br/>
Password:
<input type="password" name="extremenetloginpassword" />
<br/>
<input type="submit" value="Submit" />
</form>
</body>
</html>
Uploading the Login File to the Switch

To upload the page and the JPEG/GIF files, the switch TFTP command must be used. For example,
assuming the page resides on a TFTP server with IP address 10.255.49.19, the command used would be
BD-8810.1 # tftp get 10.255.49.19 netlogin_login_page.html

General Guidelines
The following general guidelines are applicable to the login page:
● When the custom web page is not present on the switch, the system falls back to the using the
default banner. The web page may be added (or removed) from the switch at any time, at which
point the switch will stop (or start) using the banner.
● The graphical image file names referenced in the web page must not have any path information
prepended.
● Both uppercase and lowercase names (or a mixture) for the graphical image filenames are supported,
but the user and password tag names should be either all uppercase or all lowercase, not a mixture
of the two.
● More than one form may exist on the page. This can be useful when, for example, in addition to the
main username and password that is typed in by the user, an additional special username and
password needs to be auto-filled and sent. For example, this could be used when end users without
a valid username or password need to be given restricted access to the network.
Limitations
The following limitations apply to the login page:
● When the client is in the unauthenticated state, any embedded URLs in the custom page are
inaccessible to it.
● Only JPEG and GIF graphical images are supported.
● It is the web page writer's responsibility to write the HTML page correctly and without errors.
● Only TFTP is supported as a method to upload the web-page and graphical images to the switch.
Customizable Authentication Failure Response

In the event of web-based netlogin authentication failure, you can use a custom authentication failure
page to recover. When a customized login page is in effect, 'by default, any authentication failure results
in the following failure response being delivered to the browser:
Login Incorrect. Click here to try again.
Clicking on the indicated link will bring the user back to the initial custom login page.
You may choose to over ride the above default response with a custom one. This custom failure
response page must be uploaded to the switch using TFTP with the name
netlogin_login_fail_page.html. When authentication fails, the switch responds with this page. If the
page is deleted from the switch, the response reverts back to the default.
The same graphical images that are uploaded to the switch for the custom login page can also be
embedded in the custom authentication failure page.
NOTE
The custom authentication failure page can be used only when authentication is being done via the custom login
page.

Network Login
Web-Based Network Login Configuration Example

The following configuration example shows both the Extreme Networks switch configuration and the
Radius server entries needed to support the example. VLAN corp is assumed to be a corporate subnet
which has connections to DNS, WINS servers, network routers, and so on. VLAN temp is a temporary
VLAN and is created to provide connections to unauthenticated network login clients. Unauthenticated
ports belong to the VLAN temp. This kind of configuration provides better security as unauthenticated
clients do not connect to the corporate subnet and will not be able to send or receive any data. They
have to get authenticated in order to have access to the network.
● ISP Mode—Network login clients connected to ports 1:10 - 1:14, VLAN corp, will be logged into the
network in ISP mode. This is controlled by the fact that the VLAN in which they reside in
unauthenticated mode and the RADIUS server Vendor Specific Attributes (VSA), Extreme-
Netlogin-Vlan, are the same, corp. So there will be no port movement. Also if this VSA is missing
from RADIUS server, it is assumed to be ISP Mode.
● Campus Mode—On the other hand, clients connected to ports 4:1 - 4:4, VLAN temp, will be logged
into the network in Campus mode since the port will move to the VLAN corp after getting
authenticated. A port moves back and forth from one VLAN to the other as its authentication state
changes.
Both ISP and Campus mode are not tied to ports but to a user profile. In other words, if the VSA
Extreme:Extreme-Netlogin-Vlan represents a VLAN different from the one in which the user
currently resides, then VLAN movement will occur after login and after logout. In following example, it
is assumed that campus users are connected to ports 4:1-4:4, while ISP users are logged in through ports
1:10-1:14.
NOTE
In the following sample configuration, any lines marked (Default) represent default settings and do not need to
be explicitly configured.

enable ipforwarding
# Configuration Information for VLAN temp

# No VLAN-ID is associated with VLAN temp.
configure vlan “temp” ipaddress 198.162.32.10 255.255.255.0

configure vlan “corp” add port 1:10 untagged

configure vlan “temp” dhcp-address-range 198.162.32.20 - 198.162.32.80
configure vlan “temp” dhcp-options default-gateway 198.162.32.1
configure vlan “temp” dhcp-options dns-server 10.0.1.1
configure vlan “temp” dhcp-options wins-server 10.0.1.85


enable netlogin web-based
enable netlogin ports 1:10-1:14,4:1-4:4 web-based
configure netlogin base-url "network-access.net" (Default)
configure netlogin redirect-page http://www.extremenetworks.com (Default)
enable netlogin logout-privilege (Default)
disable netlogin session-refresh 3 (Default)
# DNS Client Configuration

configure dns-client add name-server 10.0.1.1
configure dns-client add name-server 10.0.1.85
#RADIUS Configuration
configure radius netlogin primary server 10.0.1.2 1812 client-ip 10.10.20.30 vr “Vr-
Mgmt”
enable radius
RADIUS server:
#RADIUS Server Setting (VSAs)(optional)

Extreme:Extreme-Netlogin-Only = Enabled (if no CLI authorization)
Extreme:Extreme-Netlogin-Vlan = "corp" (destination vlan for CAMPUS mode network
login)
NOTE
RADIUS server.
Web-Based Authentication User Login

To use web-based authentication:
1 Set up the Windows IP configuration for DHCP.
2 Plug into the port that has web-based network login enabled.
3 Log in to Windows.
4 Release any old IP settings and renew the DHCP lease.
This is done differently depending on the version of Windows the user is running:
■ Windows 9x—Use the winipcfg tool. Choose the Ethernet adapter that is connected to the port
on which network login is enabled. Use the buttons to release the IP configuration and renew the
DHCP lease.
■ Windows NT/2000/XP—Use the ipconfig command line utility. Use the command ipconfig/
release to release the IP configuration and ipconfig/renew to get the temporary IP address
from the switch. If you have more than one Ethernet adapter, specify the adapter by using a
number for the adapter following the ipconfig command. You can find the adapter number using
the command ipconfig/all.

Network Login
NOTE
The idea of explicit release/renew is required to bring the network login client machine in the same subnet as
the connected VLAN. When using we-based authentication, this requirement is mandatory after every logout and
before login again as the port moves back and forth between the temporary and permanent VLANs.
At this point, the client will have its temporary IP address. In this example, the client should have
obtained the an IP address in the range 198.162.32.20 - 198.162.32.80.
5 Bring up the browser and enter any URL as http://www.123.net or http://1.2.3.4 or switch IP
address as http://<IP address>/login (where IP address could be either temporary or Permanent
VLAN Interface for Campus Mode). URL redirection redirects any URL and IP address to the
network login page. This is significant where security matters most, as no knowledge of VLAN
interfaces is required to be provided to network login users, as they can login using a URL or IP
address.
NOTE
URL redirection requires that the switch is configured with a DNS client.
A page opens with a link for Network Login.

6 Click the Network Login link.
A dialog box opens requesting a user name and password.
7 Enter the user name and password configured on the RADIUS server.
After the user has successfully logged in, the user will be redirected to the URL configured on the
RADIUS server.
During the user login process, the following takes place:

● Authentication is done through the RADIUS server.
● After successful authentication, the connection information configured on the RADIUS server is
returned to the switch:
■ The permanent VLAN
■ The URL to be redirected to (optional)
■ The URL description (optional)
● The port is moved to the permanent VLAN.
You can verify this using the show vlan command. For more information on the show vlan
command, see “Displaying VLAN Settings” on page 311.
After a successful login has been achieved, there are several ways that a port can return to a non-
authenticated, non-forwarding state:
● The user successfully logs out using the logout web browser window.
● The link from the user to the switch’s port is lost.
● There is no activity on the port for 20 minutes.
● An administrator changes the port state.

MAC-Based Authentication
NOTE
Because network login is sensitive to state changes during the authentication process, Extreme Networks
recommends that you do not log out until the login process is complete. The login process is complete when you
receive a permanent address.
MAC-based authentication is used for supplicants that do not support a network login mode, or
supplicants that are not aware of the existence of such security measure, for example an IP phone.
If a MAC address is detected on a MAC-Based enabled netlogin port, an authentication request is sent
once to the AAA application. AAA tries to authenticate the MAC address against the configured radius
server and its configured parameters (timeout, retries, and so on) or the local database.
In a MAC-based authentication environment the authentication verification is done only once at MAC
address detection. However, beginning with ExtremeXOS 11.6, forced reauthentication is allowed
through the Session-Timeout VSA supplied by Radius. When this VSA is present the switch re-
authenticates the client based on the value supplied by the VSA. If no VSA is present, there is no re-
authentication.
The credentials used for this are the supplicants MAC address in ASCII representation, and a locally
configured password on the switch. If no password is configured, the MAC address is used as the
password. You can also group MAC addresses together using a mask.
You can configure a MAC list or a table of MAC entries to filter and authenticate clients based on their
MAC addresses. If there a match is found in the table of MAC entries, authentication occurs. If no
match is found in the table of MAC entries, and a default entry exists, the default will be used to
authenticate the client. All entries in the list are automatically sorted in longest prefix order. All
passwords are stored and showed encrypted.
Beginning with ExtremeXOS 11.3, you can associate a MAC address with one or more ports. By
learning a MAC address, the port confirms the supplicant before sending an authorization request to
the RADIUS server. This additional step protects your network against unauthorized supplicants
because the port accepts only authorization requests from the MAC address learned on that port. The
port blocks all other requests that do not have a matching entry.

● Enabling and Disabling MAC-Based Network Login on page 514
● Associating a MAC Address to a Specific Port on page 514
● Adding and Deleting MAC Addresses on page 514
● Displaying the MAC Address List on page 515
● Secure MAC Configuration Example on page 515
● MAC-Based Network Login Configuration Example on page 516

Network Login
Enabling and Disabling MAC-Based Network Login

To enable MAC-based network login on the switch, use the following command:
enable netlogin mac
To disable MAC-based network login on the switch, use the following command:
disable netlogin mac
To enable MAC-based network login on one or more ports, use the following command:
enable netlogin ports <portlist> mac
disable MAC-based network login on one or more ports, use the following command:
disable netlogin ports <portlist> mac
Associating a MAC Address to a Specific Port

You can configure the switch to accept and authenticate a client with a specific MAC address. Only
MAC addresses that have a match for the specific ports are sent for authentication. For example, if you
associate a MAC address with one or more ports, only authentication requests for that MAC address
received on the port(s) are sent to the configured RADIUS server or local database. The port(s) block all
other authentication requests that do not have a matching entry. This is also known as secure MAC.
To associate a MAC address with one or more ports, specify the ports option when using the following
command:
configure netlogin add mac-list [<mac> {<mask>} | default] {encrypted} {<password>}

{ports <port_list>}
You must enable MAC-based netlogin on the switch and the specified ports. If MAC-based netlogin is
not enabled on the specified port(s), the switch displays a warning message similar to the following:
WARNING: Not all specified ports have MAC-Based NetLogin enabled.
For a sample configuration, see “Secure MAC Configuration Example” on page 515.
Adding and Deleting MAC Addresses

To add a MAC address to the table, use the following command:
configure netlogin add mac-list [<mac> {<mask>} | default] {encrypted} {<password>}
{ports <port_list>}
To remove a MAC address from the table, use the following command:
configure netlogin delete mac-list [<mac> {<mask>} | default]

Displaying the MAC Address List

To display the MAC address table, use the following command:
show netlogin mac-list
When a client needs authentication the best match will be used to authenticate to the server.
MAC-based authentication is VR aware, so there is one MAC list per VR.
Assume we have a supplicant with MAC address 00:04:96:05:40:00, and the switch has the following
table:
MAC Address/Mask Password (encrypted) Port(s)
-------------------- ---------------------- --------------
00:00:00:00:00:10/48 <not configured> 1:1-1:5
00:00:00:00:00:11/48 <not configured> 1:6-1:10
00:00:00:00:00:12/48 <not configured> any
00:01:30:70:0C:00/48 yaqu any
00:01:30:32:7D:00/48 ravdqsr any
00:04:96:00:00:00/24 <not configured> any
The user name used to authenticate against the Radius server would be “000496000000”, as this is the
supplicants MAC address with the configured mask applied.
Note that the commands are VR aware, and therefore one MAC list table exists per VR.
Secure MAC Configuration Example

The following configuration example shows how to configure secure MAC on your Extreme Networks
switch. To configure secure MAC:
● Create a VLAN used for netlogin.
● Configure the VLAN for netlogin.
● Enable MAC-based netlogin on the switch.
● Enable MAC-based netlogin on the ports used for authentication.
● Specify one or more ports to accept authentication requests from a specific MAC address.
In the following example, authentication requests from MAC address:

● 00:00:00:00:00:10 are only accepted on ports 1:1 through 1:5
● 00:00:00:00:00:11 are only accepted on ports 1:6 through 1:10
● 00:00:00:00:00:12 are accepted on all other ports
create vlan nlvlan
configure netlogin vlan nlvlan
enable netlogin mac
enable netlogin ports 1:1-1:10 mac
configure netlogin add mac-list 00:00:00:00:00:10 ports 1:1-1:5
configure netlogin add mac-list 00:00:00:00:00:11 ports 1:6-1:10
configure netlogin add mac-list 00:00:00:00:00:12

Network Login
To view your network login configuration, use one of the following commands:
● show netlogin {port <portlist> vlan <vlan_name>} {dot1x {detail}} {mac} {web-
based}
● show netlogin mac-list
MAC-Based Network Login Configuration Example

The following configuration example shows the Extreme Networks switch configuration needed to
support the MAC-based network login example.


enable netlogin mac
configure netlogin add mac-list default <password>
# RADIUS Configuration
configure radius netlogin primary server 10.0.1.2 1812 client-ip 10.10.20.30 vr “VR-
Mgmt”
enable radius
RADIUS server:
#RADIUS Server Setting

00E018A8C540 Auth-Type := Local, User-Password == "00E018A8C540"
NOTE
RADIUS server.
Additional Network Login Configuration Details

This section describes additional, optional network login configurations. These configurations are not
required to run network login; however, depending on your network settings and environment, you
can use the commands described in this section to enhance your network login settings.
Review the earlier sections of this chapter for general information about network login and information
about MAC-based, web-based, and 802.1x authentication methods.


● Configuring Netlogin MAC-Based VLANs on page 517
● Configuring Dynamic VLANs for Netlogin on page 519
● Configuring Netlogin Port Restart on page 521
Configuring Netlogin MAC-Based VLANs

Currently, network login allows only a single, untagged VLAN to exist on a port. This limits the
flexibility for untagged supplicants because they must be in the same VLAN.
Beginning with ExtremeXOS 11.3, the BlackDiamond 8800 series switches support netlogin MAC-based
VLANs; support for BlackDiamond 12000 series switches was added in ExtremeXOS 12.0. Netlogin
MAC-based VLANs allow a port assigned to a VLAN to operate in a MAC-based fashion. This means
that each individual untagged supplicant, identified by its MAC address, can be in different VLANs.
Netlogin MAC-based VLAN utilizes VSA information from both the netlogin local database and the
RADIUS server. After successfully performing the Campus mode of operation, the supplicant is added
untagged to the destination VLAN.
To support this feature, you must configure the netlogin port’s mode of operation. The following
sections describe the following topics:
● Netlogin MAC-Based VLANs Rules and Restrictions on page 517
● Configuring the Port Mode on page 518
● Displaying Netlogin MAC-Based VLAN Information on page 518
● Netlogin MAC-Based VLAN Example on page 519
Netlogin MAC-Based VLANs Rules and Restrictions

This section summarizes the rules and restrictions for configuring netlogin MAC-based VLANs:
● You must configure and enable netlogin on the switch and before you configure netlogin MAC-
based VLANs.
If you attempt to configure the port’s mode of operation before enabling netlogin, the switch
displays an error message similar to the following:
ERROR: The following ports do not have NetLogin enabled; 1
● On ExtremeXOS versions prior to 12.0 on switches other than the Summit family and the
BlackDiamond 12800 series, 10 Gigabit Ethernet ports such as those on the 10G4X I/O module and
the uplink ports on the Summit family of switches do not support netlogin MAC-based VLANs.
If you attempt to configure netlogin MAC-based VLANs on 10 Gigabit Ethernet ports, the switch
displays an error message similar to the following:
ERROR: The following ports do not support the MAC-Based VLAN mode; 1, 2, 10
In ExtremeXOS version 12.0, on the BlackDiamond 12800 series, SummitStack, and Summit family
switches, you can configure mac-based-VLANs on 10 Gigabit Ethernet ports.
● You can have a maximum of 1,024 MAC addresses per I/O module or per Summit family switch.

Network Login
Configuring the Port Mode

To support netlogin MAC-based VLANs on a netlogin port, you must configure that port’s mode of
operation. To specify MAC-based operation, use the following command and specify mac-based-vlans:
configure netlogin ports [all | <port_list>] mode [mac-based-vlans | port-based-vlans]
By default, the netlogin port’s mode of operation is port-based-vlans. If you modify the mode of
operation to mac-based-vlans and later disable all netlogin protocols on that port, the mode of
operation automatically returns to port-based-vlans.
When you change the netlogin port’s mode of operation, the switch deletes all currently known
supplicants from the port and restores all VLANs associated with that port to their original state. In
addition, by selecting mac-based-vlans, you are unable to manually add or delete untagged VLANs
from this port. Netlogin now controls these VLANs.
With netlogin MAC-based operation, every authenticated client has an additional FDB flag that
indicates a translation MAC address. If the supplicant’s requested VLAN does not exist on the port, the
switch adds the requested VLAN.
Displaying Netlogin MAC-Based VLAN Information

The following commands display important information for netlogin MAC-based VLANs.
FDB Information. To view FDB entries, use the following command:

show fdb netlogin [all | mac-based-vlans]
By specifying netlogin, you see only FDB entries related to netlogin or netlogin MAC-based VLANs.
The flags associated with netlogin include:
● v—Indicates the FDB entry was added because the port is part of a MAC-Based virtual port/VLAN
combination
● n—Indicates the FDB entry was added by network login
VLAN and Port Information. To view the VLANs that netlogin adds temporarily in MAC-based mode, use
By specifying information and detail, the output displays the temporarily added VLANs in netlogin
MAC-based mode. To confirm this, review the following output of this command:
● VLAN cfg—The term MAC-based appears next to the tag number.
● Netlogin port mode—This output was added to display the port mode of operation. Mac based
appears and the network login port mode of operation.
To view information about the ports that are temporarily added in MAC-based mode for netlogin, due
to discovered MAC addresses, use the following command:
show vlan detail
By specifying detail, the output displays detailed information including the ports associated with the
VLAN. The flags associated with netlogin include:
● a—Indicates an authenticated network login port.
● u—Indicates an unauthenticated network login port.
● m—Indicates that the netlogin port operates in MAC-based mode.

Netlogin MAC-Based VLAN Example

The following example configures the netlogin MAC-based VLAN feature:
create vlan users12
create vlan nlvlan
configure netlogin vlan nlvlan
enable netlogin mac
configure netlogin ports 1:1-1:10 mode mac-based-vlans
configure netlogin add mac-list default MySecretPassword
Expanding upon the previous example, you can also utilize the local database for authentication rather
than the RADIUS server:
create netlogin local-user 000000000012 vlan-vsa untagged default

create netlogin local-user 000000000010 vlan-vsa untagged users12
For more information about local database authentication, see “Configuring Local Database
Authentication” on page 492.
Configuring Dynamic VLANs for Netlogin

During an authentication request, netlogin receives a destination VLAN (if configured on the RADIUS
server) to put the authenticated user in. The VLAN must exist on the switch for netlogin to authenticate
the client on that VLAN.
Beginning with ExtremeXOS 11.6, you can configure the switch to dynamically create a VLAN after
receiving an authentication response from a supplicant. A dynamically created VLAN is only a Layer 2
bridging mechanism; this VLAN does not work with routing protocols to forward traffic. If configured
for dynamic VLAN creation, the switch automatically creates a supplicant VLAN that contains both the
supplicant’s physical port and one or more uplink ports. After the switch unauthenticates all of the
supplicants from the dynamically created VLAN, the switch deletes that VLAN.
NOTE
Dynamically created VLANs do not support the session refresh feature of web-based netlogin because dynamically
created VLANs do not have an IP address.
By dynamically creating and deleting VLANs, you minimize the number of active VLANs configured
on your edge switches. In addition, the RADIUS server forwards VSA information to dynamically
create the VLAN thereby simplifying switch management. A key difference between dynamically
created VLANs and other VLANs is that the switch does not save dynamically created VLANs. Even if
you use the save command, the switch does not save a dynamically created VLAN.
After you configure netlogin on the switch, the two steps to configure dynamic VLANs are:
● Specifying the tagged uplink port(s) to be added to each dynamically created VLAN
● Enabling the switch to create dynamic VLANs

Network Login
Specifying the Uplink Ports

The uplink ports send traffic to and from the supplicants from the core of the network. Uplink ports
should not be configured for netlogin (netlogin is disabled on uplink ports).
To specify one or more ports as tagged uplink ports that are added to the dynamically created VLAN,
configure netlogin dynamic-vlan uplink-ports [<port_list> | none]
By default, the setting is none.
If you specify an uplink port with netlogin enabled, the configuration fails and the switch displays an
error message similar to the following:
ERROR: The following ports have NetLogin enabled: 1, 2
If this occurs, select a port with netlogin disabled.
Enabling Dynamic VLANs for Netlogin

To enable the switch to create dynamic VLANs, use the following command:
configure netlogin dynamic-vlan [disable | enable]
By default, the setting is disabled. When enabled, the switch dynamically creates VLANs. Remember,
dynamically created VLANs are not permanent or user-created VLANs. The switch uses the VLAN ID
supplied by the RADIUS attributes (as described below) to create the VLAN. The switch only creates a
dynamic VLAN if the requested VLAN, indicated by the VLAN ID, does not currently exist on the
switch.
To prevent conflicts with existing VLANs on the switch, the RADIUS server uses VSAs to forward
VLAN information, including the VLAN ID. The following list specifies the supported VSAs for
configuring dynamic VLANs:
● Extreme: Netlogin-VLAN-ID (VSA 209)
● Extreme: Netlogin-Extended-VLAN (VSA 211)
NOTE
If the ASCII string only contains numbers, it is interpreted as the VLAN ID. Dynamic VLANS only support
numerical VLAN IDs; VLAN names are not supported.
● IETF: Tunnel-Private-Group-ID (VSA 81)
For more information about the VSAs, see “Creating User Accounts on the RADIUS Server” on
page 488.
The switch automatically generates the VLAN name in the following format: NLD_<TAG> where <TAG>
specifies the VLAN ID. For example, a dynamic VLAN with an ID of 10 has the name NLD_0010.
NOTE
Like all VLAN names, dynamic VLAN names are unique. If you create a VLAN and use the name of an existing
dynamic VLAN, the switch now sees the dynamic VLAN as a user-created VLAN and will save this VLAN to the

switch configuration. If this occurs, the switch does not delete the VLAN after the supplicants are authenticated and
moved to the permanent VLAN.
Dynamic VLAN Example with Web-Based Netlogin

After you finish the web-based netlogin configuration as described in “Web-Based Network Login
Configuration Example” on page 510, complete the dynamic VLAN configuration by:
● Assigning one or more non-netlogin ports as uplink ports
NOTE
Do not enable netlogin on uplink ports. If you specify an uplink port with netlogin enabled, the configuration
fails and the switch displays an error message.
● Enabling the switch to dynamically create VLANs
Whether you have MAC-based, web-based, or 802.1x authentication, you use the same two commands
to configure dynamic VLANs on the switch.
The following example configures dynamic VLANs on the switch:
configure netlogin dynamic-vlan uplink ports 2:1-2:2

configure netlogin dynamic-vlan enable
Displaying Dynamic VLAN Information

To display summary information about all of the VLANs on the switch, including any dynamically
VLANs currently operating on the switch, use the following command:
show vlan
If the switch has dynamically created VLANs, the VLAN name begins with SYS_NLD_.
To display the status of dynamic VLAN configuration on the switch, use the following command:
show netlogin
The switch displays the current state of dynamic VLAN creation (enabled or disabled) and the uplink
port(s) associated with the dynamic VLAN.
Configuring Netlogin Port Restart

You can configure netlogin to restart specific netlogin-enabled ports when the last authenticated
supplicant unauthenticates, regardless of the configured authentication methods on the port. This
feature, known as netlogin port restart, is available with all netlogin authentication methods although is
most practical with web-based netlogin. This section describes how this feature behaves with web-based
netlogin; MAC-based and 802.1x netlogin do not experience any differences in behavior if you enable
netlogin port restart.
Currently with web-based netlogin, if you have an authenticated supplicant and log out of the network,
you must manually release the IP address allocated to you by the Dynamic Host Control Protocol
(DHCP) server. The DHCP server dynamically manages and allocates IP addresses to supplicants.

Network Login
When a supplicant accesses the network, the DHCP server provides an IP address to that supplicant.
DHCP cannot renegotiate their leases, which is why you must manually release the IP address.
For example, if the idle timer expires on the switch, the switch disconnects your network session. If this
occurs, it may be unclear why you are unable to access the network. After you manually renew the IP
address, you are redirected to the netlogin login page and can log back into the network. To solve this
situation in a single supplicant per port environment, port restart triggers the DHCP client on the PC to
restart the DHCP address assignment process.
Guidelines for Using Netlogin Port Restart

Configure netlogin port restart on ports with directly attached supplicants. If you use a hub to connect
multiple supplicants, only the last unauthenticated supplicant causes the port to restart. Although the
hub does not inflict harm to your network, in this situation, the previously unauthenticated supplicants
do not get the benefit of the port restart configuration.
Enabling Netlogin Port Restart

To enable netlogin port restart, use the following command:
configure netlogin ports [all | <port_list>] restart
Disabling Netlogin Port Restart

To disable netlogin port restart, use the following command:
configure netlogin ports [all | <port_list>] no-restart
Displaying the Port Restart Configuration

To display the netlogin settings on the port, including the configuration for port restart, use the
following command:
show netlogin port <port_list>
Output from this command includes the enable/disable state for netlogin port restart.

22 Security

● Security Overview on page 523
● Safe Defaults Mode on page 525
● MAC Security on page 525
● Denial of Service Protection on page 544
● IP Security on page 535
● Authenticating Users Using RADIUS or TACACS+ on page 547
● Hyptertext Transfer Protocol on page 560
● Secure Shell 2 on page 560
● Secure Socket Layer on page 569
Security Overview
Security is a term that covers several different aspects of network use and operation. One general type
of security is control of the devices or users that can access the network. Ways of doing this include
authenticating the user at the point of logging in. You can also control access by defining limits on
certain types of traffic. Another general type of security operates to protect the operation of the switch
itself. Security measures in this category include routing policies that can limit the visibility of parts of
the network or denial of service protection that prevents the CPU from being overloaded. Finally,
management functions for the switch can be protected from unauthorized use. This type of protection
uses various types of user authentication.
ExtremeXOS 11.3 introduces enhanced security features designed to protect, rapidly detect, and correct
anomalies in your network. Extreme Networks products incorporate a number of features designed to
enhance the security of your network while resolving issues with minimal network disruption. No one
feature can ensure security, but by using a number of features in concert, you can substantially improve
the security of your network.
The following list provides a brief overview of some of the available security features:
● Access Control Lists—Access Control Lists (ACLs) are policy files used by the ACL application to
perform packet filtering and forwarding decisions on incoming traffic and packets. Each packet
arriving on an ingress port is compared to the ACL applied to that port and is either permitted or
denied.
For more information about using ACLs to control and limit network access, see Chapter 12, “Access
Lists (ACLs).”
● CLEAR-Flow—CLEAR-Flow is a security rules engine available only on the BlackDiamond 10808
and the BlackDiamond 12800 series switches. CLEAR-Flow inspects Layer 2 and Layer 3 packets,
isolates suspicious traffic, and enforces policy-based mitigation actions. Policy-based mitigation
actions include the switch taking an immediate, predetermined action or sending a copy of the
traffic off-switch for analysis. Working together, CLEAR-Flow and Sentriant™ provide a rapid
response to network threats. For off-switch analysis, CLEAR-Flow sends the suspicious traffic to
Sentriant and Sentriant stops the threats.

Security
For more information about CLEAR-Flow, see Chapter 17, “CLEAR-Flow.” For more information
about Sentriant, contact your Extreme Networks representative.
● Denial of Service Protection—Denial of Service (DoS) protection is a dynamic response mechanism
used by the switch to prevent critical network or computing resources from being overwhelmed and
rendered inoperative. In essence, DoS protection protects the switch, CPU, and memory from attacks
and attempts to characterize the attack (or problem) and filter out the offending traffic so that other
functions can continue. If the switch determines it is under attack, the switch reviews the packets in
the input buffer and assembles ACLs that automatically stop the offending packets from reaching
the CPU. For increased security, you can turn on DoS protection and establish CLEAR-Flow rules at
the same time.
For more information about DoS attacks and DoS protection, see “Denial of Service Protection” on
page 544. For more information about CLEAR-Flow, see Chapter 17, “CLEAR-Flow.”
● Network Login—Network login controls the admission of user packets and access rights thereby
preventing unauthorized access to the network. Network login is controlled on a per port basis.
When network login is enabled on a port in a VLAN, that port does not forward any packets until
authentication takes place. Network login is capable of three types of authentication: web-based,
MAC-based, and 802.1x.
For more information about network login, see Chapter 15, “Network Login.”
● Policy Files—Policy files are text files that contain a series of rule entries describing match conditions
and actions to take. Policy files are used by both routing protocol applications (routing policies) and
the ACL application (ACLs).
For more information about policy files, see Chapter 11, “Policy Manager.”
● Routing Policies—Routing policies are policy files used by routing protocol applications to control
the advertisement, reception, and use of routing information by the switch. By using policies, a set of
routes can be selectively permitted or denied based on their attributes for advertisements in the
routing domain. Routing policies can be used to “hide” entire networks or to trust only specific
sources for routes or ranges of routes.
For more information about using routing policies to control and limit network access, see Chapter
13, “Routing Policies.”
● Sentriant—Sentriant is an external security appliance used by the BlackDiamond 10808 and the
BlackDiamond 12800 series switches to detect and defend against threats without interfering with
network traffic. Sentriant can actively engage, determine, and terminate malicious behavior
occurring in your network. Sentriant and CLEAR-Flow provide a rapid response to network threats.
Sentriant can add to or modify the BlackDiamond 10808 or BlackDiamond 12800 series switches’
CLEAR-Flow rules and ACLs in real-time to inspect additional traffic or change inspection
thresholds.
For more information about Sentriant, contact your Extreme Networks representative. For more
information about CLEAR-Flow, see Chapter 17, “CLEAR-Flow.”
● sFlow—sFlow® is a technology designed to monitor network traffic by using a statistical sampling of
packets received on each port. sFlow also uses IP headers to gather information about the network.
By gathering statistics about the network, sFlow becomes an early warning system notifying you
when there is a spike in traffic activity. Upon analysis, common response mechanisms include
applying an ACL, changing Quality of Service (QoS) parameters, or modifying VLAN settings.
For more information about sFlow, see the section “Using sFlow” in Chapter 5, “Status Monitoring
and Statistics.”

Safe Defaults Mode
Safe Defaults Mode

Beginning with ExtremeXOS 11.2, when you set up your switch for the first time, you must connect to
the console port to access the switch. After logging in to the switch, you enter safe defaults mode.
Although SNMP, Telnet, and switch ports are enabled by default, the script prompts you to confirm
those settings. By answering N (No) to each question, you keep the default settings.
Would you like to disable Telnet? [y/N]: No
Would you like to disable SNMP [y/N]: No
Would you like unconfigured ports to be turned off by default [y/N]: No
Would you like to change the failsafe account username and password now? [y/N]: No
Would you like to permit failsafe account access via the management port? [y/N]: No
In addition, if you keep the default settings for SNMP and Telnet, the switch returns the following
interactive script:
* change your failsafe account username and password
MAC Security
The switch maintains a database of all media access control (MAC) addresses received on all of its
ports. The switch uses the information in this database to decide whether a frame should be forwarded
or filtered. MAC security (formerly known as MAC address security) allows you to control the way the
Forwarding Database (FDB) is learned and populated. For more information about the FDB, see Chapter
9, “Forwarding Database.”
MAC security includes several types of control. You can:

● Limit the number of dynamically-learned MAC addresses allowed per virtual port. For more
information, see “Limiting Dynamic MAC Addresses” on page 526.
● “Lock” the FDB entries for a virtual port, so that the current entries will not change, and no
additional addresses can be learned on the port. For information, see “MAC Address Lockdown” on
page 528.
NOTE
You can either limit dynamic MAC FDB entries or lockdown the current MAC FDB entries, but not both.
● Set a timer on the learned addresses that limits the length of time the learned addresses will be
maintained if the devices are disconnected or become inactive. For more information, see “MAC
Address Lockdown with Timeout” on page 529.
● Use ACLS to prioritize or stop packet flows based on the source MAC address of the ingress
virtual LAN (VLAN) or the destination MAC address of the egress VLAN. For more information
about ACL policies, see Chapter 12, “Access Lists (ACLs).”

Security
● Enhance security, depending on your network configuration, by disabling Layer 2 flooding. For
more information about enabling and disabling Layer 2 flooding, see the section, “Disabling Egress
Flooding” in Chapter 9, “Forwarding Database.”
Limiting Dynamic MAC Addresses

You can set a predefined limit on the number of dynamic MAC addresses that can participate in the
network. After the FDB reaches the MAC limit, all new source MAC addresses are blackholed at both
the ingress and egress points. These dynamic blackhole entries prevent the MAC addresses from
learning and responding to Internet Control Message Protocol (ICMP) and address resolution protocol
(ARP) packets.
NOTE
Blackhole FDB entries added due to MAC security violations on the BlackDiamond 8800 series switches,
SummitStack, and the Summit family of switches are removed after each FDB aging period regardless of whether
the MAC addresses in question are still sending traffic. If the MAC addresses are still sending traffic, the blackhole
entries will be re-added after they have been deleted.
Configuring Limit Learning

To limit the number of dynamic MAC addresses that can participate in the network, use the limit-
learning option in following command:
configure ports <portlist> vlan <vlan_name> [limit-learning <number> | lock-learning |

unlimited-learning | unlock-learning]
This command specifies the number of dynamically-learned MAC entries allowed for these ports in this
VLAN. The range is 0 to 500,000 addresses.
When the learned limit is reached, all new source MAC addresses are blackholed at the ingress and
egress points. This prevents these MAC addresses from learning and responding to ICMP and ARP
packets.
Dynamically learned entries still get aged and can be cleared. If entries are cleared or aged out after the
learning limit has been reached, new entries will then be able to be learned until the limit is reached
again.
Permanent static and permanent dynamic entries can still be added and deleted using the create
fdbentry and disable flooding port commands. These override any dynamically learned entries.
For ports that have a learning limit in place, the following traffic still flows to the port:
● Packets destined for permanent MAC addresses and other non-blackholed MAC addresses
● Broadcast traffic
● EDP traffic
Traffic from the permanent MAC and any other non-blackholed MAC addresses still flows from the
virtual port.

MAC Security
To remove the learning limit, use the unlimited-learning option in the following command:

Displaying Limit Learning Information

To verify the configuration, use the following commands:
show vlan <vlan name> security
This command displays the MAC security information for the specified VLAN.
show ports {mgmt | <portlist>} info {detail}
This command displays detailed information, including MAC security information, for the specified
port.
Example of Limit Learning

In Figure 55, three devices are connected through a hub to a single port on the Extreme Networks
device. If a learning limit of 3 is set for that port, and you connect a fourth device to the same port, the
switch does not learn the MAC address of the new device; rather, the switch blackholes the address.
Figure 55: Switch configured for limit learning
Device A
Hub
Device B
EW_110
Device C EW75003 EX_175
Limiting MAC Addresses with ESRP Enabled

If you configure a MAC address limit on VLANS that participate in an Extreme Standby Router
Protocol (ESRP) domain, you should add an additional back-to-back link (that has no MAC address
limit on these ports) between the ESRP-enabled switches. Doing so prevents ESRP protocol data units
(PDUs) from being dropped due to MAC address limit settings.
Figure 56 is an example of configuring a MAC address limit on a VLAN participating in an ESRP

domain.

Security
Figure 56: MAC address limits and VLANs participating in ESRP
ESRP 10.1.2.1 20.1.1.1

vlan
S1 S2 20.1.2.2
192.10.1.1
S4
10.1.2.2
10.1.2.100 192.10.1.100
S3
30.1.1.2
10.1.2.1
30.1.1.1
EX_036
In Figure 56, S2 and S3 are ESRP-enabled switches, while S1 is an ESRP-aware (regular Layer 2) switch.
Configuring a MAC address limit on all S1 ports might prevent ESRP communication between S2 and
S3. To resolve this, you should add a back-to-back link between S2 and S3. This link is not needed if
MAC address limiting is configured only on S2 and S3, but not on S1.
MAC Address Lockdown

In contrast to limiting learning on virtual ports, you can lockdown the existing dynamic FDB entries
and prevent any additional learning using the lock-learning option from the following command:
This command causes all dynamic FDB entries associated with the specified VLAN and ports to be
converted to locked static entries. It also sets the learning limit to 0, so that no new entries can be
learned. All new source MAC addresses are blackholed.
NOTE
Blackhole FDB entries added due to MAC security violations on the BlackDiamond 8800 series switches,
SummitStack, and the Summit family of switches are removed after each FDB aging period regardless of whether
the MAC addresses in question are still sending traffic. If the MAC addresses are still sending traffic, the blackhole
entries will be re-added after they have been deleted.
Locked entries do not get aged, but can be deleted like a regular permanent entry.
For ports that have lock-down in effect, the following traffic still flows to the port:
● Packets destined for the permanent MAC and other non-blackholed MAC addresses
● Broadcast traffic
● EDP traffic

MAC Security
Traffic from the permanent MAC still flows from the virtual port.
To remove MAC address lockdown, use the unlock-learning option from the following command:
When you remove the lockdown using the unlock-learning option, the learning-limit is reset to
unlimited, and all associated entries in the FDB are flushed.
To display the locked entries on the switch, use the following command:
show fdb
Locked MAC address entries have the “l” flag.
MAC Address Lockdown with Timeout

The MAC address lockdown with timeout feature provides a timer for aging out MAC addresses on a
per port basis and overrides the FDB aging time. That is, when this feature is enabled on a port, MAC
addresses learned on that port age out based on the MAC lockdown timeout corresponding to the port,
not based on the FDB aging time. By default, the MAC address lockdown timer is disabled.
When this feature is enabled on a port, MAC addresses learned on that port remain locked for the MAC
lockdown timeout duration corresponding to the port, even when the port goes down. As a result,
when a device is directly connected to the switch and then disconnected, the MAC address
corresponding to the device will be locked up for the MAC lockdown timeout duration corresponding
to that port. If the same device reconnects to the port before the MAC lockdown timer expires and
sends traffic, the stored MAC address becomes active and the MAC lockdown timer is restarted. If the
device is not reconnected for the MAC lockdown timeout duration, the MAC entry is removed.
MAC lockdown timeout entries are dynamically learned by the switch, which means these entries are
not saved or restored during a switch reboot. If the switch reboots, the local MAC entry table is empty,
and the switch needs to relearn the MAC addresses.
MAC address lockdown with timeout is configured by individual ports. The lockdown timer and
address learning limits are configured separately for a port.
NOTE
You cannot enable the lockdown timeout feature on a port that already has MAC address lockdown enabled. For
more information about MAC address lockdown, see “MAC Address Lockdown” on page 528.
MAC address learning limits and the lockdown timer work together in the following ways:
● When the learning limit has been reached on a port, a new device attempting to connect to the port
has its MAC address blackholed.
● As long as the timer is still running for a MAC entry, a new device cannot connect in place of the
device that entry represents. That is, if a device has disconnected from a port, a new device cannot
replace it until the lockdown timer for the first device has expired. This condition is true if the limit
on the port is set to 1 or if the limit (greater than 1) on the port has been reached.
● If a learning limit is already configured on a port when you enable the lockdown timeout feature,
the configured limit will continue to apply. Existing blackholed entries are therefore not affected. If

Security
you enable this feature on a port with no configured learning limit, the default maximum learning
limit (unlimited learning) is used.

● Understanding the Lockdown Timer on page 530
● Examples of Active and Inactive Devices on page 530
● Examples of Disconnecting and Reconnecting Devices on page 531
● Example of Port Movement on page 532
● Configuring MAC Address Lockdown with Timeout on page 532
● Enabling and Disabling MAC Address Lockdown with Timeout on page 532
● Displaying MAC Address Lockdown Information on page 533
Understanding the Lockdown Timer

The lockdown timer works in the following ways:
● When you enable this feature on a port, existing MAC entries for the port begin aging out based on
the configured MAC lockdown timer value.
● If you move a device from one port to another, its MAC address entry is updated with the new port
information, including the lockdown timer value configured for that port.
● If this feature is enabled on a port and you decrease the lockdown timer value for that port, all of
the MAC FDB entries for that port will time out and be removed at the next polling interval.
● When you disable the lockdown timer on a port, existing MAC address entries for the port will time
out based on the FDB aging period.
Examples of Active and Inactive Devices

Figure 57 shows three devices (A, B, and C) connected through a hub to an Extreme Networks device
with MAC lockdown timeout configured on the ports. When each device starts sending traffic, the
source MAC address of the device is learned and FDB entries are created. The MAC lockdown timer is
set at 100 seconds.
Figure 57: Devices using MAC address lockdown
Device A
Hub
Device B
EW_110
Device C EW75003 EX_175

MAC Security
Device Inactivity for Less than the MAC Lockdown Timer. As long as a device continues to send traffic, the
MAC entry for that device is refreshed, and the MAC lockdown timer corresponding to the MAC entry
is refreshed. Therefore, as long as the device is active, the timer does not expire. The traffic can be
continuous or can occur in bursts within the MAC lockdown timeout duration for the port.
In this example, Device A starts sending traffic. When the MAC address of Device A is learned and
added to the FDB, the MAC lockdown timer is started for this entry.
Device A stops sending traffic and resumes sending traffic after 50 seconds have elapsed. At this point
the MAC entry for Device A is refreshed and the MAC lockdown timer is restarted.
Device Inactivity for Longer than the MAC Lockdown Timer. When a device stops sending traffic and does
not resume within the MAC lockdown timer interval for the port, the MAC lockdown timer expires,
and the MAC entry is removed from the FDB.
In this example, Devices A, B, and C start sending traffic. As each MAC address is learned, the MAC
lockdown timer is started for each entry.
Device A stops sending traffic; Devices B and C continue sending traffic. After 100 seconds, the MAC
lockdown timer for the Device A entry is removed from the FDB. Because Devices B and C have
continued to send traffic, their MAC entries continue to be refreshed and their MAC lockdown timers
continue to be restarted.
Examples of Disconnecting and Reconnecting Devices

Figure 58 shows Device A connected to an Extreme Networks device with MAC lockdown timeout
configured for the ports. When Device A starts sending traffic, the source MAC address is learned on
the port, the FDB entry is created, and the MAC lockdown timer is started for the entry. The MAC
lockdown timer is set at 3,000 seconds.
Figure 58: Single device with MAC lockdown timeout
EW75004
Device A EX_176
Disconnecting a Device. In this example, Device A is disconnected from the port, triggering a port-down
action. The MAC entry for Device A is removed from the hardware FDB; however, the MAC entry for
the device is maintained in the software. The MAC lockdown timer for this entry starts when the port
goes down.
After 3,000 seconds, the MAC entry for Device A is removed from the software.
Disconnecting and Reconnecting a Device. When Device A is disconnected from the port, the resulting
port-down action causes the MAC entry for Device A to be removed from the hardware FDB. The MAC
entry in software is maintained, and the MAC lockdown timer is started for the port.
After only 1,000 seconds have elapsed, Device A is reconnected to the same port and starts sending
traffic. A MAC entry is created in the hardware FDB, and the MAC lockdown timer is restarted for the
MAC entry in the software.

Security
If Device A is reconnected but does not send any traffic for 3,000 seconds, no MAC entry is created in
the hardware FDB, and the MAC lockdown timer will expire after reaching 3,000 seconds.
Disconnecting and Reconnecting Devices with MAC Limit Learning. In this example, a MAC learning limit
of 1 has also been configured on the ports in addition to the MAC lockdown timer of 3000 seconds.
When Device A is disconnected, the resulting port-down action removes the MAC entry for Device A
from the hardware FDB. The MAC entry for Device A is maintained in the software, and the MAC
lockdown timer for this entry is restarted when the port goes down.
After 1000 seconds, a different device is connected to the same port and starts sending traffic. Because
the MAC learning limit is set to 1 and the MAC lockdown timer is still running, the MAC address of
the new device is not learned. Instead, the new MAC address is blackholed in the hardware.
When the MAC lockdown timer for Device A expires, its MAC entry is removed from the software. If
the new device is still connected to the same port and sends traffic, the MAC address for the new
device is learned and added to the FDB. The MAC lockdown timer for the new device is started, and
the blackhole entry that was created for this device is deleted.
Example of Port Movement

Figure 59 shows Device A connected to port X. Port X has a MAC lockdown timer setting of 100
seconds, and port Y has a MAC lockdown timer setting of 200 seconds.
Figure 59: Port movement with MAC lockdown timeout
Device A EW75005
Device X Device Y EX_177
Device A starts sending traffic on port X. The MAC address for Device A is learned and added to the
FDB, and the MAC lockdown timer (100 seconds) is started for this entry.
After 50 seconds, Device A is disconnected from port X and connected to port Y where it begins
sending traffic. When Device A starts sending traffic on port Y, the existing MAC entry for Device A is
refreshed, and port X in the entry is replaced with port Y. At the same time, the MAC lockdown timer
for the entry is restarted for a duration of 200 seconds (the configured MAC lockdown timer setting for
port Y).
Configuring MAC Address Lockdown with Timeout

To configure the MAC lockdown timeout value on one or more specified ports, or on all ports, use the
following command:
configure mac-lockdown-timeout ports [all | <port_list>] aging-time <seconds>
Enabling and Disabling MAC Address Lockdown with Timeout

To enable the MAC lockdown timeout feature on one or more specified ports, or on all ports, use the
following command:
enable mac-lockdown-timeout ports [all | <port_list>]

DHCP Server
To disable the MAC lockdown timeout feature on one or more specified ports, or on all ports, use the
following command:
disable mac-lockdown-timeout ports [all | <port_list>]
Displaying MAC Address Lockdown Information

To display configuration information about the MAC lockdown timeout feature, use the following
command:
show mac-lockdown-timeout ports [all | <port_list>]
Output from this command includes the configured timeout value and whether the feature is enabled
or disabled.
To display the MAC entries learned on one or more ports, or on all ports, use the following command:
show mac-lockdown-timeout fdb ports [all | <port_list>]
Output from this command also lists the aging time of the port.
DHCP Server
Dynamic Host Configuration Protocol (DHCP) support was introduced into ExtremeXOS in release 11.0.
In simple terms, a DHCP server dynamically manages and allocates IP addresses to clients. When a
client accesses the network, the DHCP server provides an IP address to that client. The client is not
required to receive the same IP address each time it accesses the network. A DHCP server with limited
configuration capabilities is included in the switch to provide IP addresses to clients.

● Enabling and Disabling DHCP on page 533
● Configuring the DHCP Server on page 533
● Displaying DHCP Information on page 534
Enabling and Disabling DHCP

DHCP is enabled on a per port, per VLAN basis. To enable or disable DHCP on a port in a VLAN, use
one of the following commands:
enable dhcp ports <portlist> vlan <vlan_name>
disable dhcp ports <portlist> vlan <vlan_name>
Configuring the DHCP Server

The following commands allow you to configure the DHCP server included in the switch. The
parameters available to configure include the IP address range, IP address lease, and multiple DHCP
options.

Security
To configure the range of IP addresses assigned by the DHCP server, use the following command:
configure vlan <vlan_name> dhcp-address-range <ipaddress1> - <ipaddress2>
To remove the address range information, use the following command:
unconfigure vlan <vlan_name> dhcp-address-range
To set how long the IP address lease assigned by the server exists, use the following command:
configure vlan <vlan_name> dhcp-lease-timer <lease-timer>
To set the default gateway, Domain Name Servers (DNS) addresses, or Windows Internet Naming
Service (WINS) server, use the following command:
configure vlan <vlan_name> dhcp-options [default-gateway | dns-server | wins-server]

<ipaddress>
To remove the default gateway, DNS server addresses, and WINS server information for a particular
VLAN, use the following command:
unconfigure vlan <vlan_name> dhcp-options
To remove all the DHCP information for a particular VLAN, use the following command:
unconfigure vlan <vlan_name> dhcp
You can clear the DHCP address allocation table selected entries, or all entries. You would use this
command to troubleshoot IP address allocation on the VLAN. To clear entries, use the following
command:
clear vlan <vlan_name> dhcp-address-allocation [[all {offered | assigned | declined |

expired}] | <ipaddress>]
Displaying DHCP Information

To display the DHCP configuration, including the DHCP range, DHCP lease timer, network login lease
timer, DHCP-enabled ports, IP address, MAC address, and time assigned to each end device, use the
following command:
show dhcp-server {vlan <vlan_name>}
The next two commands were retained for compatibility with earlier versions of ExtremeWare. To view
only the address allocation of the DHCP server on a VLAN, use the following command:
show vlan <vlan_name> dhcp-address-allocation
To view only the configuration of the DHCP server on a VLAN, use the following command:
show vlan <vlan_name> dhcp-config

IP Security
IP Security
This section describes a collection of IP security features implemented in ExtremeXOS. If you configure
any of the features described in this section, you can enhance your network security by controlling
which hosts are granted or not granted access to your network.
The IP security features described in this section are:

● DHCP Snooping and Trusted DHCP Server on page 535
● Source IP Lockdown on page 538
● ARP Learning on page 539
● Gratuitous ARP Protection on page 541
● ARP Validation on page 543
Figure 60 displays the dependencies of IP security. Any feature that appears directly above another
feature depends on it. For example, to configure ARP validation, you must configure DHCP snooping
and trusted DHCP server.
Figure 60: IP security dependencies
Gratuitous ARP
DHCP ARP Gratuitous ARP Source IP inspection—
Disable secured validation inspection—protect lockdown protect only
ARP ARP hosts and switch switch
learning
DHCP snooping and trusted DHCP server
EX_178
DHCP Snooping and Trusted DHCP Server

A fundamental requirement for most of the IP security features described in this section is to configure
DHCP snooping and trusted DHCP server. DHCP snooping enhances security by filtering untrusted
DHCP messages and by building and maintaining a DHCP bindings database. Trusted DHCP server
also enhances security by forwarding DHCP packets from only configured trusted servers within your
network.
The DHCP bindings database contains the IP address, MAC Address, VLAN ID, and port number of
the untrusted interface or client. If the switch receives a DHCP ACK message and the IP address does
not exist in the DHCP bindings database, the switch creates an entry in the DHCP bindings database. If
the switch receives a DHCP RELEASE, NAK or DECLINE message and the IP address exists in the
DHCP bindings database, the switch removes the entry.
You can enable DHCP snooping on a per port, per-VLAN basis and trusted DHCP server on a per-vlan
basis. If configured for DHCP snooping, the switch snoops DHCP packets on the indicated ports and
builds a DHCP bindings database of IP address and MAC address bindings from the received packets.
If configured for trusted DHCP server, the switch forwards only DHCP packets from the trusted
servers. The switch drops DHCP packets from other DHCP snooping-enabled ports.
In addition, to prevent rogue DHCP servers from farming out IP addresses, you can optionally
configure a specific port or set of ports as trusted ports. Trusted ports do not block traffic; rather, the
switch forwards any DHCP server packets that appear on trusted ports. When configured to do so, the

Security
switch drops packets from DHCP snooping-enabled ports and causes one of the following user-
configurable actions: disables the port temporarily, disables the port permanently, blocks the violating
MAC address temporarily, blocks the violating MAC address permanently, and so on.
Configuring DHCP Snooping

By default DHCP snooping is disabled on the switch. To enable DHCP snooping on the switch, use the
following command:
enable ip-security dhcp-snooping {vlan} <vlan_name> ports [all | <ports>] violation-

action [drop-packet {[block-mac | block-port] [duration <duration_in_seconds> |
permanently] | none]}] {snmp-trap}
NOTE
Snooping IP fragmented DHCP packets is not supported.
The violation action setting determines what action(s) the switch takes when a rogue DHCP server
packet is seen on an untrusted port or the IP address of the originating server is not among those of the
configured trusted DHCP servers. The DHCP server packets are DHCP OFFER, ACK and NAK. The
following list describes the violation actions:
● block-mac—The switch automatically generates an ACL to block the MAC address on that port. The
switch does not blackhole that MAC address in the FDB. The switch can either temporarily or
permanently block the MAC address.
● block-port—The switch blocks all traffic on that port by disabling the port either temporarily or
permanently.
● none—The switch takes no action to drop the rogue DHCP packet or block the port, and so on. In
this case, DHCP snooping continues to build and manage the DHCP bindings database and DHCP
forwarding will continue in hardware as before. This option can be used when the intent is only to
monitor the IP addresses being assigned by the DHCP server.
NOTE
You must enable DHCP snooping on both the DHCP server port as well as on the client port. The latter ensures that
DHCP client packets (DHCP Request, DHCP Release etc.) are processed appropriately.
Any violation that occurs causes the switch to generate an Event Management System (EMS) log
message. You can configure to suppress the log messages by configuring EMS log filters. For more
information about EMS, see the section “Using the Event Management System/Logging” in Chapter
5, “Status Monitoring and Statistics.”
To disable DHCP snooping on the switch, use the following command:
disable ip-security dhcp-snooping {vlan} <vlan_name> ports [all | <ports>]
Configuring Trusted DHCP Server

To configure a trusted DHCP server on the switch, use the following command:
configure trusted-servers {vlan} <vlan_name> add server <ip_address> trust-for dhcp-
server

IP Security
You can configure a maximum of eight trusted DHCP servers on the switch.
If you configure one or more trusted ports, the switch assumes that all DHCP server packets on the
trusted port are valid. For more information about configuring trusted ports, see the next section
“Configuring Trusted DHCP Ports.”
To delete a trusted DHCP server, use the following command:
configure trusted-servers vlan <vlan_name> delete server <ip_address> trust-for dhcp-

server
Configuring Trusted DHCP Ports

To enable DHCP snooping, use the following command:
For more information about DHCP snooping see, “Configuring DHCP Snooping” on page 536.
Trusted ports do not block traffic; rather, the switch forwards any DHCP server packets that appear on
trusted ports. Depending on your DHCP snooping configuration, the switch drops packets and can
disable the port temporarily, disable the port permanently, block the MAC address temporarily, block
the MAC address permanently, and so on.
To enable trusted ports on the switch, use the following command:
configure trusted-ports [<ports>|all] trust-for dhcp-server
To disable trusted ports on the switch, use the following command:
unconfigure trusted-ports [<ports>|all] trust-for dhcp-server
Displaying DHCP Snooping and Trusted Server Information

To display the DHCP snooping configuration settings, use the following command:
show ip-security dhcp-snooping {vlan} <vlan_name>
DHCP Snooping enabled on ports: 1:2, 1:3, 1:4, 1:7, 1:9

Trusted Ports: 1:7
Trusted DHCP Servers: None
--------------------------------------------
Port Violation-action
--------------------------------------------
1:2 none
1:3 drop-packet
1:4 drop-packet, block-mac permanently
1:7 none
1:9 drop-packet, snmp-trap
To display the DHCP bindings database, use the following command:
show ip-security dhcp-snooping entries {vlan} <vlan_name>

Security
--------------------------------------------
Vlan: dhcpVlan
--------------------------------------------
Server Client
IP Addr MAC Addr Port Port
------- -------- ------ ------
172.16.100.9 00:90:27:c6:b7:65 1:1 1:2
Clearing DHCP Snooping Entries

Existing DHCP snooping entries can be cleared by using the following command. (Note that this will
also clear out any associated Source IP Lockdown and DHCP Secured ARP entries.)
clear ip-security dhcp-snooping entries {vlan} <vlan_name>
Source IP Lockdown
Another type of IP security prevents IP address spoofing by automatically placing source IP address
filters on specified ports. This feature, called source IP lockdown, allows only traffic from a valid
DHCP-assigned address obtained by a DHCP snooping-enabled port to enter the network. In this way,
the network is protected from attacks that use random source addresses for their traffic. With source IP
lockdown enabled, end systems that have a DHCP address assigned by a trusted DHCP server can
access the network, but traffic from others, including those with static IP addresses is dropped at the
switch.
Source IP lockdown is linked to the “DHCP snooping” feature. The same DHCP bindings database
created when you enable DHCP snooping is also used by source IP lockdown to create ACLs that
permit traffic from DHCP clients. All other traffic is dropped. In addition, the DHCP snooping violation
action setting determines what action(s) the switch takes when a rouge DHCP server packet is seen on
an untrusted port.
When source IP lockdown is enabled on a port, a default ACL is created to deny all IP traffic on that
port. Then an ACL is created to permit DHCP traffic on specified ports. Each time source IP lockdown
is enabled on another port, the switch creates ACLs to allow DHCP packets and to deny all IP traffic for
that particular port.
Source IP lockdown is enabled on a per-port basis; it is not available at the VLAN level. If source IP
lockdown is enabled on a port, the feature is active on the port for all VLANS to which the port
belongs.
Configuring Source IP Lockdown

To configure source IP lockdown, you must enable DHCP snooping on the ports connected to the
DHCP server and DHCP client before you enable source IP lockdown. You must enable source IP
lockdown on the ports connected to the DHCP client, not on the ports connected to the DHCP server.
To enable DHCP snooping, use the following command:

IP Security
By default, source IP lockdown is disabled on the switch. To enable source IP lockdown, use the
following command:
enable ip-security source-ip-lockdown ports [all | <ports>]
To disable source IP lockdown, use the following command:
disable ip-security source-ip-lockdown ports [all | <ports>]
Displaying Source IP Lockdown Information

To display the source IP lockdown configuration on the switch, use the following command:
show ip-security source-ip-lockdown
Ports Locked IP Address

23 10.0.0.101
Clearing Source IP Lockdown Information

To remove existing source IP lockdown entries on the switch, use the following command:
clear ip-security source-ip-lockdown entries ports [<ports> | all]
ARP Learning
The address resolution protocol (ARP) is part of the TCP/IP suite used to dynamically associate a
device’s physical address (MAC address) with its logical address (IP address). The switch broadcasts an
ARP request that contains the IP address, and the device with that IP address sends back its MAC
address so that traffic can be transmitted across the network. The switch maintains an ARP table (also
known as an ARP cache) that displays each MAC address and its corresponding IP address.
By default, the switch builds its ARP table by tracking ARP requests and replies, which is known as
ARP learning. Beginning with ExtremeXOS 11.6, you can disable ARP learning so that the only entries
in the ARP table are either manually added or those created by DHCP secured ARP; the switch does
not add entries by tracking ARP requests and replies. By disabling ARP learning and adding a
permanent entry or configuring DHCP secured ARP, you can centrally manage and allocate client IP
addresses and prevent duplicate IP addresses from interrupting network operation.

● Configuring ARP Learning on page 539
● Adding a Permanent Entry to the ARP Table on page 540
● Configuring DHCP Secured ARP on page 540
● Displaying ARP Information on page 541
Configuring ARP Learning

As previously described, ARP learning is enabled by default. The switch builds its ARP table by
tracking ARP requests and replies.

Security
To disable ARP learning on one or more ports in a VLAN, use the following command:
disable ip-security arp learning learn-from-arp {vlan} <vlan_name> ports [all |

<ports>]
To re-enable ARP learning on one or more ports in a VLAN, use the following command:
enable ip-security arp learning learn-from-arp {vlan} <vlan_name> ports [all |

<ports>]
Adding a Permanent Entry to the ARP Table

If you disable ARP learning, you must either manually add a permanent entry to the ARP table or
configure DHCP secured ARP to populate the ARP table.
To manually add a permanent entry to the ARP table, use the following command:
configure iparp add <ip_addr> {vr <vr_name>} <mac>
For more detailed information about this command and IP routing, see Chapter 6, “IPv4 Unicast
Routing.”
Configuring DHCP Secured ARP

Another method available to populate the ARP table is DHCP secured ARP. DHCP secured ARP
requires that ARP entries be added to or deleted from the ARP table only when the DHCP server
assigns or re-assigns an IP address. These entries are known as a secure ARP entry. If configured, the
switch adds the MAC address and its corresponding IP address to the ARP table as a permanent ARP
entry. Regardless of other ARP requests and replies seen by the switch, the switch does not update
secure ARP entries. DHCP secured ARP is linked to the “DHCP snooping” feature. The same DHCP
bindings database created when you enabled DHCP snooping is also used by DHCP secured ARP to
create secure ARP entries. The switch only removes secure ARP entries when the corresponding DHCP
entry is removed from the trusted DHCP bindings database.
NOTE
If you enable DHCP secured ARP on the switch without disabling ARP learning, ARP learning continues which
allows insecure entries to be added to the ARP table.
Before you configure DHCP secured ARP, you must enable DHCP snooping on the switch. To enable
DHCP snooping, use the following command:

By default, DHCP secured ARP learning is disabled. To enable DHCP secured ARP, use the following
command:
enable ip-security arp learning learn-from-dhcp {vlan} <vlan_name> ports [all |
<ports>]

IP Security
DHCP Secured ARP must be enabled on the DHCP server port as well as the DHCP client ports. To
disable DHCP secured ARP, use the following command:
disable ip-security arp learning learn-from-dhcp {vlan} <vlan_name> ports [all |

<ports>]
NOTE
You must enable DHCP secured ARP on the DHCP server as well as on the client ports. DHCP snooping, as always,
must also be enabled on both the server and client ports.
Displaying ARP Information

To display how the switch builds an ARP table and learns MAC addresses for devices on a specific
VLAN and associated member ports, use the following command:
show ip-security arp learning {vlan} <vlan_name>
Port Learn-from
----------------------------------
2:1 ARP
2:2 DHCP
2:3 ARP
2:4 None
2:5 ARP
2:6 ARP
2:7 ARP
2:8 ARP
To view the ARP table, including permanent and DHCP secured ARP entries, use the following
command:
show iparp {<ip_addre> | <mac> | vlan <vlan_name> | permanent} {vr <vr_name>}
NOTE
DHCP secured ARP entries are stored as static entries in the ARP table.
Gratuitous ARP Protection

When a host sends an ARP request to resolve its own IP address it is called gratuitous ARP. A
gratuitous ARP request is sent with the following parameters:
● Destination MAC address—FF:FF:FF:FF:FF:FF (broadcast)
● Source MAC address—Host's MAC address
● Source IP address = Destination IP address—IP address to be resolved
In a network, gratuitous ARP is used to:

● Detect duplicate IP address
In a properly configured network, there is no ARP reply for a gratuitous ARP request. However, if
another host in the network is configured with the same IP address as the source host, then the
source host receives an ARP reply.

Security
● Announce that an IP address has moved or bonded to a new network interface card (NIC)
If you change a system NIC, the MAC address to its IP address mapping also changes. When you
reboot the host, it sends an ARP request packet for its own IP address. All of the hosts in the
network receive and process this packet. Each host updates their old mapping in the ARP table with
this new mapping
● Notify a Layer 2 switch that a host has moved from one port to another port
However, hosts can launch man-in-the-middle attacks by sending out gratuitous ARP requests for the
router's IP address. This results in hosts sending their router traffic to the attacker, and the attacker
forwarding that data to the router. This allows passwords, keys, and other information to be
intercepted.
To protect against this type of attack, the router will send out its own gratuitous ARP request to
override the attacker whenever a gratuitous ARP broadcast with the router's IP address as the source is
received on the network. Also, the switch will add a MAC ACL to blackhole the offending host.
Beginning in ExtremeXOS 11.6 if you enable both DHCP secured ARP and gratuitous ARP protection,
the switch protects its own IP address and those of the hosts that appear as secure entries in the ARP
table.
Configuring Gratuitous ARP

You enable the gratuitous ARP feature on a per VLAN basis, not on a per port basis. The validation is
done for all gratuitous ARP packets received on a VLAN in which this feature is enabled irrespective of
the port in which the packet is received.
When enabled, the switch generates gratuitous ARP packets in the following scenarios:
● When the sender IP address is the same as the switch IP address.
● When the sender MAC address is the same as the switch MAC address.
● When the sender IP address of another host is matching an entry in the static ARP table.
When the switch generates an ARP packet, the switch generates logs and traps.
Beginning with ExtremeXOS 11.6, you enable gratuitous ARP protection with the following command:
enable ip-security arp gratuitous-protection {vlan} [all | <vlan_name>]
In addition, to protect the IP addresses of the hosts that appear as secure entries in the ARP table, use
the following commands to enable DHCP snooping, DHCP secured ARP, and gratuitous ARP on the
switch:
● enable ip-security dhcp-snooping {vlan} <vlan_name> ports [all | <ports>]
violation-action [drop-packet {[block-mac | block-port] [duration
<duration_in_seconds> | permanently] | none]}] {snmp-trap}
● enable ip-security arp learning learn-from-dhcp {vlan} <vlan_name> ports [all |
<ports>]
● enable ip-security arp gratuitous-protection {vlan} [all | <vlan_name>]
To disable gratuitous ARP protection, use the following command:
disable ip-security arp gratuitous-protection {vlan} [all | <vlan_name>]

IP Security
In ExtremeXOS 11.5 and earlier, you enable gratuitous ARP protection with the following command:
enable iparp gratuitous protect vlan <vlan-name>
And you disable gratuitous ARP protection with the following command:
disable iparp gratuitous protect vlan <vlan-name>
Displaying Gratuitous ARP Information

To display information about gratuitous ARP, use the following command:
show ip-security arp gratuitous-protection
Gratuitous ARP Protection enabled on following VLANs:

Default, test
ARP Validation
ARP validation is also linked to the “DHCP snooping” feature. The same DHCP bindings database
created when you enabled DHCP snooping is also used to validate ARP entries arriving on the
specified ports.
Validation Option ARP Request Packet Type ARP Response Packet Type
DHCP Source IP is not present in the DHCP
snooping database OR is present but Source
Hardware Address doesn't match the MAC in
the DHCP bindings entry.
IP Source IP == Mcast OR Source IP == Mcast OR
Target IP == Mcast OR Target IP == Mcast
Source IP exists in the DHCP
bindings database but Source
Hardware Address doesn't match the
MAC in the DHCP bindings entry.
Source-MAC Ethernet source MAC does not match Ethernet source MAC does not match the
the Source Hardware Address. Source Hardware Address.
Destination-MAC Ethernet destination MAC does not match
the Target Hardware Address.
Depending on the options specified when enabling ARP validation, the following validations are done.
Note that the 'DHCP' option does not have to be specified explicitly, it is always implied when ARP
validation is enabled.
Configuring ARP Validation

Before you configure ARP validation, you must enable DHCP snooping on the switch. To enable DHCP
snooping, use the following command:

Security
By default, ARP validation is disabled. To enable and configure ARP validation, use the following
command:
enable ip-security arp validation {destination-mac} {source-mac} {ip} {vlan}
<vlan_name> [all | <ports>] violation-action [drop-packet {[block-port] [duration
<duration_in_seconds> | permanently]}] {snmp-trap}
The violation action setting determines what action(s) the switch takes when an invalid ARP is received.
Any violation that occurs causes the switch to generate an Event Management System (EMS) log
message. You can configure to suppress the log messages by configuring EMS log filters. For more
information about EMS, see the section “Using the Event Management System/Logging” in Chapter
5, “Status Monitoring and Statistics.”
To disable ARP validation, use the following command:
disable ip-security arp validation {vlan} <vlan_name> [all | <ports>]
Displaying ARP Validation Information

To display information about ARP validation, use the following command:
show ip-security arp validation {vlan} <vlan_name>
----------------------------------------------------------------
Port Validation Violation-action
----------------------------------------------------------------
7 DHCP drop-packet, block-port for 120 seconds, snmp-trap
23 DHCP drop-packet, block-port for 120 seconds, snmp-trap
Denial of Service Protection

A Denial-of-Service (DoS) attack occurs when a critical network or computing resource is overwhelmed
and rendered inoperative in a way that legitimate requests for service cannot succeed. In its simplest
form, a Denial of Service attack is indistinguishable from normal heavy traffic. There are some
operations in any switch or router that are more costly than others, and although normal traffic is not a
problem, exception traffic must be handled by the switch’s CPU in software.
Some packets that the switch processes in the CPU software include:
● Learning new traffic (only the BlackDiamond 10808 switch, the BlackDiamond 8800 series switches,
and the Summit family of switches learn in hardware)
● Routing and control protocols including ICMP, BGP, OSPF, STP, EAPS, ESRP, and so forth
● Switch management traffic (switch access by Telnet, SSH, HTTP, SNMP, and so forth)
● Other packets directed to the switch that must be discarded by the CPU
If any one of these functions is overwhelmed, the CPU may be too busy to service other functions and
switch performance will suffer. Even with very fast CPUs, there will always be ways to overwhelm the
CPU with packets that require costly processing.

Denial of Service Protection
DoS Protection is designed to help prevent this degraded performance by attempting to characterize the
problem and filter out the offending traffic so that other functions can continue. When a flood of CPU
bound packets reach the switch, DoS Protection will count these packets. When the packet count nears
the alert threshold, packets headers will be saved. If the threshold is reached, then these headers are
analyzed, and a hardware access control list (ACL) is created to limit the flow of these packets to the
CPU. This ACL will remain in place to provide relief to the CPU. Periodically, the ACL will expire, and
if the attack is still occurring, it will be re-enabled. With the ACL in place, the CPU will have the cycles
to process legitimate traffic and continue other services.
NOTE
User-created ACLs take precedence over the automatically applied DoS protect ACLs.
DoS Protection will send a notification when the notify threshold is reached.
You can also specify some ports as trusted ports, so that DoS protection will not be applied to those
ports.
Configuring Simulated Denial of Service Protection

The conservative way to deploy DoS protection is to use the simulated mode first. In simulated mode,
DoS protection is enabled, but no ACLs are generated. To enable the simulated mode, use the following
command:
enable dos-protect simulated
This mode is useful to gather information about normal traffic levels on the switch. This will assist in
configuring denial of service protection so that legitimate traffic is not blocked.
The remainder of this section describes how to configure DoS protection, including alert thresholds,
notify thresholds, ACL expiration time, and so on.
Configuring Denial of Service Protection

To enable or disable DoS protection, use the following commands:
enable dos-protect
disable dos-protect
After enabling DoS protection, the switch will count the packets handled by the CPU and periodically
evaluate whether to send a notification and/or create an ACL to block offending traffic. You can
configure a number of the values used by DoS protection if the default values are not appropriate for
your situation.
The values that you can configure are:

● interval—How often, in seconds, the switch evaluates the DoS counter (default: 1 second)
● alert threshold—The number of packets received in an interval that will generate an ACL (default:
4000 packets)
● notify threshold—The number of packets received in an interval that will generate a notice (default:
3500 packets)

Security
● ACL expiration time—The amount of time, in seconds, that the ACL will remain in place (default: 5
seconds)
To configure the interval at which the switch checks for DoS attacks, use the following command:
configure dos-protect interval <seconds>
To configure the alert threshold, use the following command:
configure dos-protect type l3-protect alert-threshold <packets>
To configure the notification threshold, use the following command:
configure dos-protect type l3-protect notify-threshold <packets>
To configure the ACL expiration time, use the following command:
configure dos-protect acl-expire <seconds>
Configuring Trusted Ports

Traffic from trusted ports will be ignored when DoS protect counts the packets to the CPU. If we know
that a machine connected to a certain port on the switch is a safe "trusted" machine, and we know that
we will not get a DoS attack from that machine, the port where this machine is connected to can be
configured as a trusted port, even though a large amount of traffic is going through this port.
To configure the trusted ports list, use the following command:
configure dos-protect trusted-ports [ports [<ports> | all] | add-ports [<ports-to-add>

| all] | delete-ports [<ports-to-delete> | all] ]
Displaying DoS Protection Settings

To display the DoS protection settings, use the following command:
show dos-protect {detail}
Protocol Anomaly Protection

Beginning with ExtremeXOS 12.0, the new Extreme chipsets contain built-in hardware protocol checkers
that support port security features for security applications, such as stateless DoS protection. The
protocol checkers allow users to drop the packets based on the following conditions, which are checked
for ingress packets prior to the L2/L3 entry table:
● SIP = DIP for IPv4/IPv6 packets.
● TCP_SYN Flag = 0 for Ipv4/Ipv6 packets
● TCP Packets with control flags = 0 and sequence number = 0 for Ipv4/Ipv6 packets
● TCP Packets with FIN, URG & PSH bits set & seq. number = 0 for Ipv4/Ipv6 packets
● TCP Packets with SYN & FIN bits are set for Ipv4/Ipv6 packets
● TCP Source Port number = TCP Destination Port number for Ipv4/Ipv6 packets
● First TCP fragment does not have the full TCP header (less than 20bytes) for Ipv4/Ipv6 packets

Authenticating Users Using RADIUS or TACACS+
● TCP header has fragment offset value as 1 for Ipv4/Ipv6 packets

● UDP Source Port number = UDP Destination Port number for Ipv4/Ipv6 packets
● CMP ping packets payload is larger than programmed value of ICMP max size for Ipv4/Ipv6
packets
● Fragmented ICMP packets for Ipv4/Ipv6 packets
The protocol anomaly detection security functionality is supported by a set of anomaly-protection

enable, disable, configure, clear, and show CLI commands. For further details, see Chapter
13, “Security Commands,” in the Extreme XOS 12.0 Command Reference Guide.

ExtremeXOS provides three methods to authenticate users who login to the switch:
● RADIUS
● TACACS+
● Local database of accounts and passwords
RADIUS, TACACS+, local database of accounts and passwords, and SSH are management access
security features that control access to the management functions available on the switch. These features
help ensure that any configuration changes to the switch can be done only by authorized users.
This section describes RADIUS and TACACS+. For a detailed description of the local database of
accounts and passwords (the two levels of management accounts), see Chapter 3, “Accessing the
Switch.” For information about SSH, see “Secure Shell 2” on page 560.
RADIUS
Remote Authentication Dial In User Service (RADIUS), in RFC 2138, is a mechanism for authenticating
and centrally administrating access to network nodes. The ExtremeXOS RADIUS implementation allows
authentication for Telnet or console access to the switch.
NOTE
You cannot enable RADIUS and TACACS+ at the same time.
You define a primary and secondary RADIUS server for the switch to contact. When a user attempts to
log in using Telnet, HTTP, or the console, the request is relayed to the primary RADIUS server and then
to the secondary RADIUS server, if the primary does not respond. If the RADIUS client is enabled, but
access to the RADIUS primary and secondary server fails, the switch uses its local database for
authentication. Beginning with ExtremeXOS 11.2, you can specify one pair of RADIUS servers for
switch management and another pair for network login.
The privileges assigned to the user (admin versus nonadmin) at the RADIUS server take precedence
over the configuration in the local switch database.

● Configuring the RADIUS Servers on page 548
● Configuring the RADIUS Timeout Value on page 548

Security
● Configuring the Shared Secret Password for RADIUS Servers on page 548
● Enabling and Disabling RADIUS on page 549
● Configuring RADIUS Accounting on page 549
● Configuring the RADIUS Accounting Timeout Value on page 549
● Configuring the Shared Secret Password for RADIUS Accounting Servers on page 549
● Enabling and Disabling RADIUS Accounting on page 550
Configuring the RADIUS Servers

To configure the RADIUS servers, use the following command:
configure radius {mgmt-access | netlogin} [primary | secondary] server [<ipaddress> |
<hostname>] {<udp_port>} client-ip [<ipaddress>] {vr <vr_name>}
To configure the primary RADIUS server, specify primary. To configure the secondary RADIUS server,
specify secondary.
By default, switch management and network login use the same primary and secondary RADIUS
servers for authentication. To specify one pair of RADIUS servers for switch management and another
pair for network login, make sure to specify the mgmt-access or netlogin keywords.
Configuring the RADIUS Timeout Value

To configure the timeout if a server fails to respond, use the following command:
configure radius {mgmt-access | netlogin} timeout <seconds>
If the timeout expires, another authentication attempt will be made. After three failed attempts to
authenticate, the alternate server will be used. After six failed attempts, local user authentication will be
used.
If you do not specify the mgmt-access or netlogin keywords, the timeout interval applies to both
switch management and netlogin RADIUS servers.
Configuring the Shared Secret Password for RADIUS Servers

In addition to specifying the RADIUS server IP information, RADIUS also contains a means to verify
communication between network devices and the server. The shared secret is a password configured on
the network device and RADIUS server, used by each to verify communication.
To configure the shared secret for RADIUS servers, use the following command:
configure radius {mgmt-access | netlogin} [primary | secondary] shared-secret

{encrypted} <string>
To configure the primary RADIUS server, specify primary. To configure the secondary RADIUS server,
specify secondary.
If you do not specify the mgmt-access or netlogin keywords, the secret applies to both the primary or
secondary switch management and netlogin RADIUS servers.

Do not use the encrypted keyword to set the shared secret. The encrypted keyword is primarily for
the output of the show configuration command, so the shared secret is not revealed in the command
output.
Enabling and Disabling RADIUS

After server information is entered, you can start and stop RADIUS authentication as many times as
necessary without needing to reconfigure server information.
To enable RADIUS authentication, use the following command:
enable radius {mgmt-access | netlogin}
If you do not specify the mgmt-access or netlogin keywords, RADIUS authentication is enabled on the
switch for both management and network login.
To disable RADIUS authentication, use the following command:
disable radius {mgmt-access | netlogin}
If you do not specify the mgmt-access or netlogin keywords, RADIUS authentication is disabled on
the switch for both management and network login.
Configuring RADIUS Accounting

Extreme Networks switches are capable of sending RADIUS accounting information. As with RADIUS
authentication, you can specify two servers for receipt of accounting information.
To specify RADIUS accounting servers, use the following command:
configure radius-accounting {mgmt-access | netlogin} [primary | secondary] server

[<ipaddress> | <hostname>] {<tcp_port>} client-ip [<ipaddress>] {vr <vr_name>}
To configure the primary RADIUS accounting server, specify primary. To configure the secondary
RADIUS accounting server, specify secondary.
By default, switch management and network login use the same primary and secondary RADIUS
servers for accounting. To specify one pair of RADIUS accounting servers for switch management and
another pair for network login, make sure to specify the mgmt-access or netlogin keywords.
Configuring the RADIUS Accounting Timeout Value

configure radius-accounting {mgmt-access | netlogin} timeout <seconds>
If the timeout expires, another authentication attempt will be made. After three failed attempts to
authenticate, the alternate server will be used.
Configuring the Shared Secret Password for RADIUS Accounting Servers

RADIUS accounting also uses the shared secret password mechanism to validate communication
between network access devices and RADIUS accounting servers.

Security
To specify shared secret passwords for RADIUS accounting servers, use the following command:
configure radius-accounting {mgmt-access | netlogin} [primary | secondary] shared-

secret {encrypted} <string>
To configure the primary RADIUS accounting server, specify primary. To configure the secondary
RADIUS accounting server, specify secondary.
If you do not specify the mgmt-access or netlogin keywords, the secret applies to both the primary or
secondary switch management and netlogin RADIUS accounting servers.
output.
Enabling and Disabling RADIUS Accounting

After you configure RADIUS accounting server information, you must enable accounting before the
switch begins transmitting the information. You must enable RADIUS authentication for accounting
information to be generated. You can enable and disable accounting without affecting the current state
of RADIUS authentication.
To enable RADIUS accounting, use the following command:
enable radius-accounting {mgmt-access | netlogin}
If you do not specify a keyword, RADIUS accounting is enabled on the switch for both management
and network login.
To disable RADIUS accounting, use the following command:
disable radius-accounting {mgmt-access | netlogin}
If you do not specify a keyword, RADIUS accounting is disabled on the switch for both management
and network login.
Per Command Authentication Using RADIUS

You can use the RADIUS implementation to perform per command authentication. Per command
authentication allows you to define several levels of user capabilities by controlling the permitted
command sets based on the RADIUS user name and password.
You do not need to configure any additional switch parameters to take advantage of this capability. The
RADIUS server implementation automatically negotiates the per command authentication capability
with the switch. For examples on per-command RADIUS configurations, see the next section.
Configuring RADIUS
You can define primary and secondary server communication information and, for each RADIUS
server, the RADIUS port number to use when talking to the RADIUS server. The default port value is
1812 for authentication and 1813 for accounting. The client IP address is the IP address used by the
RADIUS server for communicating back to the switch.

NOTE
RADIUS server.
RADIUS RFC 2138 Attributes

The RADIUS RFC 2138 optional attributes supported are as follows:
● User-Name
● User-Password
● Service-Type
● Login-IP-Host
RADIUS RFC 3580 Attributes

The RFC 3580 attributes for Netlogin 802.1x supported are as follows:
● EAP-Message
● Message-Authenticator
● State
● Termination-Action
● Session-Timeout
● NAS-Port-Type
● Calling-Station-ID
Using RADIUS Servers with Extreme Networks Switches

Extreme Networks switches have two levels of user privilege:
● Read-only
● Read-write
Because no command line interface (CLI) commands are available to modify the privilege level, access
rights are determined when you log in. For a RADIUS server to identify the administrative privileges of
a user, Extreme Networks switches expect a RADIUS server to transmit the Service-Type attribute in the
Access-Accept packet, after successfully authenticating the user.
Extreme Networks switches grant a RADIUS-authenticated user read-write privilege if a Service-Type

value of 6 is transmitted as part of the Access-Accept message from the RADIUS server. Other Service-
Type values or no value, result in the switch granting read-only access to the user. Different
implementations of RADIUS handle attribute transmission differently. You should consult the
documentation for your specific implementation of RADIUS when you configure users for read-write
access.
Extreme RADIUS
Extreme Networks provides its users, free of charge, a radius server based on Merit RADIUS. Extreme
RADIUS provides per-command authentication capabilities in addition to the standard set of radius

Security
features. Source code for Extreme RADIUS can be obtained from the Extreme Networks Technical
Assistance Center and has been tested on Red Hat Linux.
When Extreme RADIUS is up and running, the two most commonly changed files will be users and
profiles. The users file contains entries specifying login names and the profiles used for per-command
authentication after they have logged in. Sending a HUP signal to the RADIUS process is sufficient to
get changes in the users file to take place. Extreme RADIUS uses the file named profiles to specify
command lists that are either permitted or denied to a user based on their login identity. Changes to the
profiles file require the RADIUS server to be shutdown and restarted. Sending a HUP signal to the
RADIUS process is not enough to force changes to the profiles file to take effect.
When you create command profiles, you can use an asterisk to indicate any possible ending to any
particular command. The asterisk cannot be used as the beginning of a command. Reserved words for
commands are matched exactly to those in the profiles file. Due to the exact match, it is not enough to
simply enter “sh” for “show” in the profiles file, the complete word must be used. Commands can still
be entered in the switch in partial format.
When you use per-command authentication, you must ensure that communication between the
switch(es) and radius server(s) is not lost. If the RADIUS server crashes while users are logged in, they
will have full administrative access to the switch until they log out. Using two RADIUS servers and
enabling idle timeouts on all switches will greatly reduce the chance of a user gaining elevated access
due to RADIUS server problems.
Cistron RADIUS
Cistron RADIUS is a popular server, distributed under GPL. Cistron RADIUS can be found at:
http://www.radius.cistron.nl/
When you configure the Cistron server for use with Extreme switches, you must pay close attention to
the users file setup. The Cistron RADIUS dictionary associates the word Administrative-User with
Service-Type value 6, and expects the Service-Type entry to appear alone on one line with a leading tab
character.
The following is a user file example for read-write access:
adminuser Auth-Type = System
Service-Type = Administrative-User,
Filter-Id = “unlim”
RSA Ace
For users of their SecureID product, RSA offers RADIUS capability as part of their ACE server software.
With some versions of ACE, the RADIUS shared-secret is incorrectly sent to the switch resulting in an
inability to authenticate. As a work around, do not configure a shared-secret for RADIUS accounting
and authentication servers on the switch.
Limiting Max-Concurrent Sessions with Funk Software’s Steel Belted Radius

For users who have Funk Software’s Steel Belted Radius (SBR) server, it is possible to limit the number
of concurrent login sessions using the same user account. This feature allows the use of shared user

accounts, but limits the number of simultaneous logins to a defined value. Using this feature requires
Funk Software Steel-Belted-Radius for Radius Authentication & Accounting.
To limit the maximum concurrent login sessions under the same user account:
1 Configure Radius and Radius-Accounting on the switch.
The Radius and Radius-Accounting servers used for this feature must reside on the same physical
Radius server. Standard Radius and Radius-Accounting configuration is required as described earlier
in this chapter.
2 Modify the Funk SBR vendor.ini file and user accounts.
To configure the Funk SBR server, the file vendor.ini must be modified to change the Extreme
Networks configuration value of ignore-ports to yes as shown in the example below:
vendor-product = Extreme Networks
dictionary = Extreme
ignore-ports = yes
port-number-usage = per-port-type
help-id = 2000
After modifying the vendor.ini file, the desired user accounts must be configured for the Max-
Concurrent connections. Using the SBR Administrator application, enable the check box for Max-
Concurrent connections and fill in the desired number of maximum sessions.
RADIUS Server Configuration Example (Merit)

Many implementations of RADIUS server use the Merit© AAA server application.
Included below are excerpts from relevant portions of a sample Merit RADIUS server implementation.
The example shows excerpts from the client and user configuration files. The client configuration file
(ClientCfg.txt) defines the authorized source machine, source name, and access level. The user
configuration file (users) defines username, password, and service type information.
ClientCfg.txt
#Client Name Key [type] [version] [prefix]
#---------------- --------------- -------------- --------- --------
#10.1.2.3:256 test type = nas v2 pfx
#pm1 %^$%#*(&!(*&)+ type=nas pm1.
#pm2 :-):-(;^):-}! type nas pm2.
#merit.edu/homeless hmoemreilte.ses
#homeless testing type proxy v1
#xyz.merit.edu moretesting type=Ascend:NAS v1
#anyoldthing:1234 whoknows? type=NAS+RAD_RFC+ACCT_RFC
10.202.1.3 andrew-linux type=nas
10.203.1.41 eric type=nas
10.203.1.42 eric type=nas
10.0.52.14 samf type=nas
users
user Password = ""

Filter-Id = "unlim"
admin Password = "", Service-Type = Administrative
Filter-Id = "unlim"
eric Password = "", Service-Type = Administrative

Filter-Id = "unlim"

Security
albert Password = "password", Service-Type = Administrative

Filter-Id = "unlim"
samuel Password = "password", Service-Type = Administrative

Filter-Id = "unlim"
RADIUS Per-Command Configuration Example

Building on this example configuration, you can use RADIUS to perform per-command authentication
to differentiate user capabilities. To do so, use the Extreme-modified RADIUS Merit software that is
available from the Extreme Networks by contacting Extreme Networks technical support. The software
is available in compiled format for Solaris™ or Linux™ operating systems, as well as in source code
format. For all clients that use RADIUS per-command authentication, you must add the following type
to the client file:
type:extreme:nas + RAD_RFC + ACCT_RFC
Within the users configuration file, additional keywords are available for Profile-Name and Extreme-
CLI-Authorization. To use per-command authentication, enable the CLI authorization function and
indicate a profile name for that user. If authorization is enabled without specifying a valid profile, the
user is unable to perform any commands.
Next, define the desired profiles in an ASCII configuration file called profiles. This file contains
named profiles of exact or partial strings of CLI commands. A named profile is linked with a user
through the users file. A profile with the permit on keywords allows use of only the listed commands.
A profile with the deny keyword allows use of all commands except the listed commands.
CLI commands can be defined easily in a hierarchal manner by using an asterisk (*) to indicate any
possible subsequent entry. The parser performs exact string matches on other text to validate
commands. Commands are separated by a comma (,) or newline.
Looking at the following example content in profiles for the profile named PROFILE1, which uses the
deny keyword, the following attributes are associated with the user of this profile:
● Cannot use any command starting with enable.
● Cannot use the disable ipforwarding command.
● Cannot use a show switch command.
● Can perform all other commands.
We know from the users file that this applies to the users albert and lulu. We also know that eric is
able to log in, but is unable to perform any commands, because he has no valid profile assigned.
In PROFILE2, a user associated with this profile can use any enable command, the clear counters
command and the show management command, but can perform no other functions on the switch. We
also know from the users file that gerald has these capabilities.
The following lists the contents of the file users with support for per-command authentication:
user Password = ""

Filter-Id = "unlim"
admin Password = "", Service-Type = Administrative

Filter-Id = "unlim"

eric Password = "", Service-Type = Administrative, Profile-Name = ""

Filter-Id = "unlim"
Extreme:Extreme-CLI-Authorization = Enabled
albert Password = "", Service-Type = Administrative, Profile-Name =

"Profile1"
Filter-Id = "unlim"
lulu Password = "", Service-Type = Administrative, Profile-Name =

"Profile1"
Filter-Id = "unlim"
gerald Password = "", Service-Type = Administrative, Profile-Name "Profile2"

Filter-Id = "unlim"
Contents of the file “profiles”:
PROFILE1 deny
{
enable *, disable ipforwarding
show switch
}
PROFILE2
{
enable *, clear counters
show management
}
PROFILE3 deny
{
create vlan *, configure iproute *, disable *, show fdb
delete *, configure rip add
}
TACACS+
Terminal Access Controller Access Control System Plus (TACACS+) is a mechanism for providing
authentication, authorization, and accounting on a centralized server, similar in function to RADIUS.
The ExtremeXOS version of TACACS+ is used to authenticate prospective users who are attempting to
administer the switch. TACACS+ is used to communicate between the switch and an authentication
database.
NOTE
You cannot use RADIUS and TACACS+ at the same time.
You can configure two TACACS+ servers, specifying the primary server address, secondary server
address, and Transmission Control Protocol (TCP) port number to be used for TACACS+ sessions.

Security

● Configuring the TACACS+ Servers on page 556
● Configuring the TACACS+ Timeout Value on page 556
● Configuring the Shared Secret Password for TACACS+ Servers on page 556
● Enabling and Disabling TACACS+ on page 557
● TACACS+ Configuration Example on page 557
● Configuring TACACS+ Accounting on page 558
● Configuring the TACACS+ Accounting Timeout Value on page 558
● Configuring the Shared Secret Password for TACACS+ Accounting Servers on page 558
● Enabling and Disabling TACACS+ Accounting on page 559
● TACACS+ Accounting Configuration Example on page 559
Configuring the TACACS+ Servers

To configure the TACACS+ servers, use the following command:
configure tacacs [primary | secondary] server [<ipaddress> | <hostname>] {<tcp_port>}
client-ip <ipaddress> {vr <vr_name>}
To configure the primary TACACS+ server, specify primary. To configure the secondary TACACS+
server, specify secondary.
Configuring the TACACS+ Timeout Value

configure tacacs timeout <seconds>
To detect and recover from a TACACS+ server failure when the timeout has expired, the switch makes
one authentication attempt before trying the next designated TACACS+ server or reverting to the local
database for authentication. In the event that the switch still has IP connectivity to the TACACS+ server,
but a TCP session cannot be established, (such as a failed TACACS+ daemon on the server), fail over
happens immediately regardless of the configured timeout value.
For example, if the timeout value is set for 3 seconds (the default value), it will take 3 seconds to fail
over from the primary TACACS+ server to the secondary TACACS+ server. If both the primary and the
secondary servers fail or are unavailable, it takes approximately 6 seconds to revert to the local database
for authentication.
Configuring the Shared Secret Password for TACACS+ Servers

In addition to specifying the TACACS+ server IP information, TACACS+ also contains a means to
verify communication between network devices and the server. The shared secret is a password
configured on the network device and TACACS+ server, used by each to verify communication.
To configure the shared secret for TACACS+ servers, use the following command:
configure tacacs [primary | secondary] shared-secret {encrypted} <string>
To configure the primary TACACS+ server, specify primary. To configure the secondary TACACS+
server, specify secondary.

output.
Enabling and Disabling TACACS+

After server information is entered, you can start and stop TACACS+ authentication as many times as
necessary without needing to reconfigure server information.
To enable TACACS+ authentication, use the following command:
enable tacacs
To disable TACACS+ authentication, use the following command:
disable tacacs
TACACS+ Configuration Example

This section provides a sample TACACS+ server configuration.

● Configures the primary TACACS+ server
● Configures the shared secret for the primary TACACS+ server
● Configures the secondary TACACS+ server
● Configures the shared secret for the secondary TACACS+ server
● Enables TACACS+ on the switch
All other settings use the default settings as described earlier in this section or in the ExtremeXOS
configure tacacs primary server 10.201.31.238 client-ip 10.201.31.85 vr "VR-Default"

configure tacacs primary shared-secret purple
configure tacacs secondary server 10.201.31.235 client-ip 10.201.31.85 vr "VR-Default"
configure tacacs secondary shared-secret purple
enable tacacs
To display the TACACS+ server settings, use the show tacacs command. The following is sample
output from this command:
TACACS+: enabled
TACACS+ Authorization: disabled
TACACS+ Accounting : disabled
TACACS+ Server Connect Timeout sec: 3
Primary TACACS+ Server:
Server name :
IP address : 10.201.31.238
Server IP Port: 49
Client address: 10.201.31.85 (VR-Default)
Shared secret : purple
Secondary TACACS+ Server:
Server name :
IP address : 10.201.31.235

Security
Server IP Port: 49
TACACS+ Acct Server Connect Timeout sec: 3
Primary TACACS+ Accounting Server:Not configured
Secondary TACACS+ Accounting Server:Not configured
Configuring TACACS+ Accounting

Extreme Networks switches are capable of sending TACACS+ accounting information. As with
TACACS+ authentication, you can specify two servers for receipt of accounting information.
To specify TACACS+ accounting servers, use the following command:
configure tacacs-accounting [primary | secondary] server [<ipaddress> | <hostname>]

{<udp_port>} client-ip <ipaddress> {vr <vr_name>}
To configure the primary TACACS+ accounting server, specify primary. To configure the secondary
TACACS+ accounting server, specify secondary.
Configuring the TACACS+ Accounting Timeout Value

configure tacacs-accounting timeout <seconds>
To detect and recover from a TACACS+ accounting server failure when the timeout has expired, the
switch makes one authentication attempt before trying the next designated TACACS+ accounting server
or reverting to the local database for authentication. In the event that the switch still has IP connectivity
to the TACACS+ accounting server, but a TCP session cannot be established, (such as a failed
TACACS+ daemon on the accounting server), fail over happens immediately regardless of the
configured timeout value.
For example, if the timeout value is set for 3 seconds (the default value), it takes 3 seconds to fail over
from the primary TACACS+ accounting server to the secondary TACACS+ accounting server. If both
the primary and the secondary servers fail or are unavailable, it takes approximately 6 seconds to revert
to the local database for authentication.
Configuring the Shared Secret Password for TACACS+ Accounting Servers

TACACS+ accounting also uses the shared secret password mechanism to validate communication
between network access devices and TACACS+ accounting servers.
To specify shared secret passwords for TACACS+ accounting servers, use the following command:
configure tacacs-accounting [primary | secondary] shared-secret {encrypted} <string>
To configure the primary TACACS+ accounting server, specify primary. To configure the secondary
TACACS+ accounting server, specify secondary.
output.

Enabling and Disabling TACACS+ Accounting

After you configure TACACS+ accounting server information, you must enable accounting before the
switch begins transmitting the information. You must enable TACACS+ authentication for accounting
information to be generated. You can enable and disable accounting without affecting the current state
of TACACS+ authentication.
To enable TACACS+ accounting, use the following command:
enable tacacs-accounting
To disable TACACS+ accounting, use the following command:
disable tacacs-accounting
TACACS+ Accounting Configuration Example

This section provides a sample TACACS+ accounting server configuration.

● Configures the primary TACACS+ accounting server
● Configures the shared secret for the primary TACACS+ accounting server
● Configures the secondary TACACS+ accounting server
● Configures the shared secret for the secondary TACACS+ accounting server
● Enables TACACS+ accounting on the switch
All other settings use the default settings as described earlier in this section or in the ExtremeXOS
configure tacacs-accounting primary server 10.201.31.238 client-ip 10.201.31.85 vr

"VR-Default"
configure tacacs-accounting primary shared-secret purple
configure tacacs-accounting secondary server 10.201.31.235 client-ip 10.201.31.85 vr
"VR-Default"
config tacacs-accounting secondary shared-secret purple
enable tacacs-accounting
To display the TACACS+ accounting server settings, use the show tacacs or the show tacacs-
accounting command. The following is sample output from the show tacacs command:
TACACS+: enabled
TACACS+ Authorization: enabled
TACACS+ Accounting : enabled
TACACS+ Server Connect Timeout sec: 3
Primary TACACS+ Server:
Server name :
IP address : 10.201.31.238
Server IP Port: 49
Secondary TACACS+ Server:
Server name :
IP address : 10.201.31.235
Server IP Port: 49

Security

TACACS+ Acct Server Connect Timeout sec: 3
Primary TACACS+ Accounting Server:
Server name :
IP address : 10.201.31.238
Server IP Port: 49
Secondary TACACS+ Accounting Server:
Server name :
IP address : 10.201.31.235
Server IP Port: 49
Hyptertext Transfer Protocol

The Hyptertext Transfer Protocol (HTTP) is a set of rules for transferring and exchanging information
(data, voice, images, and so on) on the World Wide Web. HTTP is based on a request-response model.
An HTTP client initiates requests by establishing a TCP connection to a port on a remote host (port 80
by default). An HTTP server listening on that port waits for and then responds to the request; in many
instances, the client is requesting a specific URL or IP address. Upon receiving a request, the destination
server sends back the associated file or files and then closes the connection.
The web server in ExtremeXOS allows HTTP clients to access the switch on port 80 (by default) as well
as the network login page without additional encryption or security measures. For information about
secure HTTP transmission, including Secure Socket Layer (SSL), see “Secure Socket Layer” on page 569.
By default, HTTP is enabled on the switch. If you disabled HTTP access, you can re-enable HTTP access
on the default port (80) using the following command:
enable web http
To disable HTTP, use the following command:
disable web http
Secure Shell 2
Secure Shell 2 (SSH2) is a feature of ExtremeXOS that allows you to encrypt session data between a
network administrator using SSH2 client software and the switch or to send encrypted data from the
switch to an SSH2 client on a remote system. Configuration, image, and public key policy files may also
be transferred to the switch using the Secure Copy Protocol 2 (SCP2) or the Secure File Transfer
Protocol (SFTP). Beginning with ExtremeXOS 11.2, you can also use SSH2 to connect to other devices
from the switch. The ExtremeXOS CLI provides a command that enables the switch to function as an
SSH2 client, sending commands to a remote system via an SSH2 session.
The ExtremeXOS SSH2 switch application also works with putty client, SSH2 client (version 2.x or later)
from SSH Communication Security, and with (version 2.5 or later) from OpenSSH. The SFTP file
transfer protocol is required for file transfer using SCP2 OpenSSH SCP uses the RCP protocol. This has,

Secure Shell 2
for security reasons, been disabled on ExtremeXOS. Hence OpenSSH SCP will not work with EXOS SSH
implementation; however OpenSSH SFTP can be used.
The section describes the following topics:

● Enabling SSH2 for Inbound Switch Access on page 561
● Using ACLs to Control SSH2 Access on page 563
● Using SCP2 from an External SSH2 Client on page 565
● Understanding the SSH2 Client Functions on the Switch on page 566
● Using SFTP from an External SSH2 Client on page 567
Enabling SSH2 for Inbound Switch Access

SSH2 functionality is not present in the base ExtremeXOS software image, but is available as an
additional, installable module. Before you can access any SSH2 commands, you must install this
additional software module. Without the software module, the commands do not appear on the
command line. To install the software module, see the instructions in Appendix A, “Software Upgrade
and Boot Options.”
NOTE
Do not terminate the SSH process (exsshd) that was installed since the last reboot unless you have saved your
configuration. If you have installed a software module and you terminate the newly installed process without saving
your configuration, your module may not be loaded when you attempt to restart the process with the start
process command.
Because SSH2 is currently under U.S. export restrictions, you must first obtain and install the ssh.xmod
software module from Extreme Networks before you can enable SSH2.
You must enable SSH2 on the switch before you can connect to the switch using an external SSH2
client. Enabling SSH2 involves two steps:
● Generating or specifying an authentication key for the SSH2 sessions.
● Enabling SSH2 access by specifying a TCP port to be used for communication and specifying on
which virtual router SSH2 is enabled.
After it has enabled, by default, SSH2 uses TCP port 22 and is available on all virtual routers.
Standard Key Authentication

An authentication key must be generated before the switch can accept incoming SSH2 sessions. This can
be done automatically by the switch, or you can enter a previously generated key. To have the key
generated by the switch, use the following command:
configure ssh2 key
The key generation process can take up to ten minutes. After the key has been generated, you should
save your configuration to preserve the key.
To use a key that has been previously created, use the following command:
configure ssh2 key {pregenerated}

Security
The switch prompts you to enter the pregenerated key.
NOTE
The pregenerated key must be one that was generated by the switch. To get such key, you can use the command
show configuration exsshd to display the key on the console. Copy the key to a text editor and remove the carriage
return/line feeds from the key. Finally, copy and paste the key into the command line. The key must be entered as
one line.
The key generation process generates the SSH2 private host key. The SSH2 public host key is derived
from the private host key and is automatically transmitted to the SSH2 client at the beginning of an
SSH2 session.
User Key Based Authentication

Public-Key authentication is an alternative method to password authentication that SSH uses to verify
identity. You can generate a key pair consisting of a private key and a public-key. The public-key is
used by the ExtremeXOS SSH server to authenticate the user.
In ExtremeXOS, user public keys are stored in the switch’s configuration file; these keys are then
associated (or bound) to a user. The keys are configured on the switch in one of two ways:
● by copying the key to the switch using scp2/sftp2 with the switch acting as the server
● by configuring the key using the CLI
RSA and DSA encryption keys are both supported.
The public key can be loaded onto the switch using SCP or SFTP, where the switch is the server. The
administrator can do this by using the SCP2 or SFTP2 client software to connect to and copy the key file
to the switch. The public key file must have the extension ssh; for example id_dsa_2048.ssh. When the
.ssh file is copied to the switch, the key is loaded into the memory. The loaded public keys are saved to
the configuration file (*.cfg) when the save command is issued via the CLI.
The key name is derived from the file name. For example, the key name for the file id_dsa_2048.ssh will
be id_dsa_2048.
The key is associated with a user either implicitly, by pre-pending the user name to the file or explicitly,
using the CLI.
In order for a key to be bound or associated to a user, the user must be known. In other words, that
user must have an entry in the local database on the switch. Once the user is authenticated the user’s
rights (read-only or read/write) are obtained from the database.
The key can be associated with a user by pre-pending the user name to the file name. For example,
admin.id_dsa_2048.ssh.
If the user specified in the filename does not exist on the switch, the key is still accepted, but will not be
associated to any user. Once the user is added, the key can be associated with the user via the CLI. If
the user name is not pre-pended to the filename, the key is accepted by the switch but is not associated
with any user. The key can be then be associated with the user via the CLI.

Secure Shell 2
You can also enter or paste the key using the CLI. There cannot be any carriage returns or new lines in
the key. See the appropriate reference page in the Extreme XOS 12.0 Command Reference Guide for
additional details.
The host and user public keys can be written to a file in the config directory using the create sshd2
key-file command. This enables the administrator to copy the public key to an outside server.
Enabling SSH2
To enable SSH2, use the following command:
enable ssh2 {access-profile [<access_profile> | none]} {port <tcp_port_number>} {vr

[<vr_name> | all | default]}
You can also specify a TCP port number to be used for SSH2 communication. By default the TCP port
number is 22. Beginning with ExtremeXOS 11.2, the switch accepts IPv6 connections.
Before you initiate a session from an SSH2 client, ensure that the client is configured for any non-
default access list or TCP port information that you have configured on the switch. After these tasks are
accomplished, you may establish an SSH2-encrypted session with the switch. Clients must have a valid
user name and password on the switch in order to log in to the switch after the SSH2 session has been
established.
Up to eight active SSH2 sessions can run on the switch concurrently. If you enable the idle timer using
the enable idletimeout command, the SSH2 connection times out after 20 minutes of inactivity by
default. If you disable the idle timer using the disable idletimeout command, the SSH2 connection
times out after 61 minutes of inactivity. If a connection to an SSH2 session is lost inadvertently, the
switch terminates the session within 61 minutes.
For additional information on the SSH protocol refer to Federal Information Processing Standards
Publication (FIPSPUB) 186, Digital Signature Standard, 18 May 1994. This can be download from: ftp://
ftp.cs.hut.fi/pub/ssh. General technical information is also available from:
http://www.ssh.fi
Viewing SSH2 Information

To view the status of SSH2 sessions on the switch, use the following command:
show management
The show management command displays information about the switch including the enable/disable
state for SSH2 sessions and whether a valid key is present.
Using ACLs to Control SSH2 Access

You can restrict SSH2 access by creating and implementing an ACL policy. You configure an ACL
policy to permit or deny a specific list of IP addresses and subnet masks for the SSH2 port.
The two methods to load ACL policies to the switch are:

● Use the edit policy command to launch a VI-like editor on the switch. You can create the policy
directly on the switch.

Security
● Use the tftp command to transfer a policy that you created using a text editor on another system to
the switch.
For more information about creating and implementing ACLs and policies, see Chapter 11, “Policy
Manager,” and Chapter 12, “Access Lists (ACLs).”
Sample SSH2 Policies

The following are sample policies that you can apply to restrict SSH2 access.
In the following example named MyAccessProfile.pol, the switch permits connections from the subnet
10.203.133.0/24 and denies connections from all other addresses:
MyAccessProfile.pol
if {
}
then
{
permit;
}
}
In the following example named MyAccessProfile.pol, the switch permits connections from the subnets
10.203.133.0/24 or 10.203.135.0/24 and denies connections from all other addresses:
MyAccessProfile.pol
if {
}
then
{
permit;
}
}
the subnet 10.203.133.0/24 but accepts connections from all other addresses:
if {
}
then
{
deny;
}
}
If {

Secure Shell 2
; #none specified
}
then
{
permit;
}
}
the subnets 10.203.133.0/24 or 10.203.135.0/24 but accepts connections from all other addresses:
if {
source-address 10.203.135.0 /24
}
then
{
deny;
}
}
If {
; #none specified
}
then
{
permit;
}
}
Configuring SSH2 to Use ACL Policies

This section assumes that you have already loaded the policy on the switch. For more information
about creating and implementing ACLs and policies, see Chapter 11, “Policy Manager,” and Chapter
12, “Access Lists (ACLs).”
To configure SSH2 to use an ACL policy to restrict access, use the following command:
enable ssh2 {access-profile [<access_profile> | none]} {port <tcp_port_number>} {vr

[<vr_name> | all | default]}
Use the none option to remove a previously configured ACL.
Using SCP2 from an External SSH2 Client

In ExtremeXOS version 11.0 or later, the SCP2 protocol is supported for transferring configuration,
image and public key policy files to the switch from the SCP2 client.
The user must have administrator-level access to the switch. The switch can be specified by its switch
name or IP address.

Security
ExtremeXOS only allows SCP2 to transfer to the switch files named as follows:
● *.cfg—ExtremeXOS configuration files
● *.pol—ExtremeXOS policy files
● *.xos—ExtremeXOS core image files
● *.xmod—ExtremeXOS modular package files
● *.ssh—Public key files
In the following examples, you are using a Linux system to move files to and from the switch at
192.168.0.120, using the switch administrator account admin.You are logged into your Linux system as
user.
To transfer the primary configuration file from the switch to your current Linux directory using SCP2,
[user@linux-server]# scp2 [email protected]:primary.cfg primary.cfg
To copy the policy filename test.pol from your Linux system to the switch, use the following command:
[user@linux-server]# scp2 test.pol [email protected]:test.pol
To copy the image file test.xos from your Linux system to the switch, use the following command:
[user@linux-server]# scp2 test.xos [email protected]:test.xos
Now you can use the command install image test.xos to install the image in the switch.
To copy the SSH image file test.xmod from your Linux system to the switch, use the following
command:
[user@linux-server]# scp2 test.xmod [email protected]:test.xmod
Now you can use the command install image test.xmod to install the image in the switch
To load the public key id_rsa.pub from your Linux system to the switch, use the following command:
[user@linux-server]# scp2 id_rsa.pub [email protected]:test.ssh
This command loads the key into memory, which can be viewed with the command show sshd2
user-key.
Understanding the SSH2 Client Functions on the Switch

Beginning with ExtremeXOS 11.2, an Extreme Networks switch can function as an SSH2 client. This
means you can connect from the switch to a remote device running an SSH2 server and send
commands to that device. You can also use SCP2 to transfer files to and from the remote device.
You do not need to enable SSH2 or generate an authentication key to use the SSH2 and SCP2
commands from the ExtremeXOS CLI.
NOTE
The BlackDiamond 8800 series switches, SummitStack, and the Summit family of switches do not support user-
created VRs.

Secure Shell 2
To send commands to a remote system using SSH2, use the following command:
ssh2 {cipher [3des | blowfish]} {port <portnum>} {compression [on | off]} {user
<username>} {<username>@} [<host> | <ipaddress>] {<remote command>} {vr <vr_name>}
The remote commands can be any command acceptable by the remote system. You can specify the login
user name as a separate argument or as part of the user@host specification. If the login user name for
the remote system is the same as your user name on the switch, you can omit the username parameter
entirely.
For example, to obtain a directory listing from a remote Linux system with IP address 10.10.0.2 using
SSH2, enter the following command:
ssh2 [email protected] ls
To initiate a file copy from a remote system to the switch using SCP2, use the following command:
scp2 {cipher [3des | blowfish]} {port <portnum>} <user>@ [<hostname> |

<ipaddress>]:<remote_file> <local_file> {vr <vr_name>}
For example, to copy the configuration file test.cfg on host system1 to the switch, enter the following
command:
scp2 admin@system1:test.cfg localtest.cfg
To initiate a file copy to a remote system from the switch using SCP2, use the following command:
scp2 {cipher [3des | blowfish]} {port <portnum>} <local_file> <user>@ [<hostname> |

<ipaddress>]:<remote_file> {vr <vr_name>}
For example, to copy the configuration file engineering.cfg from the switch to host system1, enter the
following command:
scp2 engineering.cfg admin@system1:engineering.cfg
Using SFTP from an External SSH2 Client

In ExtremeXOS version 11.2 or later, the SFTP protocol is supported for transferring configuration, and
policy files to the switch from the SFTP client. You must have administrator-level access to the switch.
The switch can be specified by its switch name or IP address.
ExtremeXOS requires that SFTP transfer to the switch files named as follows:
● *.cfg—ExtremeXOS configuration files

● *.pol—ExtremeXOS policy files
● *.xos—ExtremeXOS core image file
● *.xmod—ExtremeXOS modular package file
● *.ssh—Public key files
In the following examples, you are using a Linux system to move files to and from the switch at
192.168.0.120, using the switch administrator account admin. You are logged into your Linux system as
account user.

Security
To transfer the primary configuration file from the switch to your current Linux directory using SCP2,
[user@linux-server]# sftp [email protected]

password: <Enter password>
sftp> put primary.cfg
To copy the policy filename test.pol from your Linux system to the switch, use the following command:

sftp> put test.pol
To copy the image file test.xos from your Linux system to the switch, use the following command:

sftp> put test.xos
To copy the SSH image file test-ssh.xmod from your Linux system to the switch, use the following
command:

sftp> put test-ssh.xmod
To load the public keyed_rsa.pub from your Linux system to the switch, use the following command:

sftp> put id_rsa.pub id_rsa.ssh
For image file transfers, only one image file at a time can be available for installation. In other words, if
test.xos and test-ssh.xmod both need to be installed, you must follow these steps:
1 Transfer test.xos into the switch using scp/sftp

2 Install the test.xos image using the "install image" command
3 Transfer test-ssh.xmod into the switch using scp/sftp
4 Install the test-ssh.xmod modular package file using "install image" command.
For image file transfers using SFTP or SCP (with the switch acting as the server), once the image is
copied to the switch, validation of image is done by the switch, as indicated by the following log
message:
<Info:AAA.LogSsh> Validating Image file, this could take approximately 30 seconds..

test.xos
You must receive the following log message before you can proceed with the installation of the image:
<Info:AAA.LogSsh> Image file test-ssh.xmod successfully validated

Secure Socket Layer
In stacking switches, you must receive the following log message from all slots before proceeding with
the installation. For example, in a four-switch stack, the installation can be proceed only after the
following log messages are received:
04/19/2007 17:41:09.71 <Info:AAA.LogSsh> Slot-1: Sent file "test.xos" info to backup

04/19/2007 17:41:09.71 <Info:AAA.LogSsh> Slot-1: Sent file "test.xos" info to standby
slot 3
04/19/2007 17:41:09.71 <Info:AAA.LogSsh> Slot-1: Sent file "test-12.0.0.13.xos" info
to standby slot 4
Secure Socket Layer

Secure Socket Layer (SSLv3) is a feature of ExtremeXOS that allows you to authenticate and encrypt
data over an SSL connection to provide secure communication. The existing web server in ExtremeXOS
allows HTTP clients to access the network login page. By using HTTPS on the web server, clients
securely access the network login page using an HTTPS enabled web browser. Since SSL encrypts the
data exchanged between the server and the client, you protect your data, including network login
credentials, from unwanted exposure.
HTTPS access is provided through SSL and the Transport Layer Security (TLS1.0). These protocols
enable clients to verify the authenticity of the server to which they are connecting, thereby ensuring that
users are not compromised by intruders.
Similar to SSH2, before you can use any SSL commands, you must first download and install the
separate Extreme Networks SSH software module (ssh.xmod). This additional module allows you to
configure both SSH2 and SSL on the switch. SSL is packaged with the SSH module; therefore, if you do
not install the module, you are unable to configure SSL. If you try to execute SSL commands without
installing the module first, the switch notifies you to download and install the module. To install the
module, see the instructions in Appendix A, “Software Upgrade and Boot Options.”
You must upload or generate a certificate for SSL server use. Before you can upload a certificate, you
must purchase and obtain an SSL certificate from an Internet security vendor. The following security
algorithms are supported:
● RSA for public key cryptography (generation of certificate and public-private key pair, certificate
signing). RSA key size between 1024 and 4096 bits.
● Symmetric ciphers (for data encryption): RC4, DES, and 3DES.
● Message Authentication Code (MAC) algorithms: MD5 and SHA.
The Converged Network Analyzer (CNA) Agent requires SSL to encrypt communication between the
CNA Agent and the CNA Server. For more information about the CNA Agent, see Appendix C, “CNA
Agent.”

● Enabling and Disabling SSL on page 570
● Creating Certificates and Private Keys on page 570
● Displaying SSL Information on page 572

Security
Enabling and Disabling SSL

This section describes how to enable and disable SSL on your switch.
NOTE
Before ExtremeXOS 11.2, the Extreme Networks SSH module did not include SSL. To use SSL for secure HTTPS
web-based login, you must upgrade your core software image to ExtremeXOS 11.2 or later, install the SSH module
that works in concert with that core software image, and reboot the switch.
Keep in mind the following guidelines when using SSL:

● To use SSL with web-based login (secure HTTP access, HTTPS) you must specify the HTTPS
protocol when configuring the redirect URL.
● If you are downloading the SSH module for the first time and want to immediately use SSL for
secure HTTPS web-based login, restart the thttpd process after installing the SSH module. For more
detailed information about activating the SSH module, see “Guidelines for Activating SSL” in
Appendix A, “Software Upgrade and Boot Options.”
To enable SSL and allow secure HTTP (HTTPS) access on the default port (443), use the following
command:
enable web https
To disable SSL and HTTPS, use the following command:
disable web https
Creating Certificates and Private Keys

When you generate a certificate, the certificate is stored in the configuration file, and the private key is
stored in the EEPROM. The certificate generated is in PEM format.
To create a self-signed certificate and private key that can be saved in the EEPROM, use the following
command:
configure ssl certificate privkeylen <length> country <code> organization <org_name>

common-name <name>
Make sure to specify the following:

● Country code (maximum size of 2 characters)
● Organization name (maximum size of 64 characters)
● Common name (maximum size of 64)
Any existing certificate and private key is overwritten.
The size of the certificate depends on the RSA key length (privkeylen) and the length of the other
parameters (country, organization name, and so forth) supplied by the user. If the RSA key length is
1024, then the certificate is approximately 1 kb. For an RSA key length of 4096, the certificate length is
approximately 2 kb, and the private key length is approximately 3 kb.

Secure Socket Layer
Downloading a Certificate Key from a TFTP Server

You can download a certificate key from files stored in a TFTP server. If the operation is successful, any
existing certificate is overwritten. After a successful download, the software attempts to match the
public key in the certificate against the private key stored. If the private and public keys do not match,
the switch displays a warning message similar to the following: Warning: The Private Key does
not match with the Public Key in the certificate. This warning acts as a reminder to also
download the private key.
Downloaded certificates and keys are not saved across switch reboots unless you save your current
switch configuration. After you use the save command, the downloaded certificate is stored in the
configuration file and the private key is stored in the EEPROM.
To download a certificate key from files stored in a TFTP server, use the following command:
download ssl <ip_address> certificate <cert file>
NOTE
For security measures, you can only download a certificate key in the VR-Mgmt virtual router.
To see whether the private key matches with the public key stored in the certificate, use the following
command:
show ssl
This command also displays:

● HTTPS port configured. This is the port on which the clients will connect.
● Length of the RSA key (the number of bits used to generate the private key).
● Basic information about the stored certificate.
Downloading a Private Key from a TFTP Server

To download a private key from files stored in a TFTP server, use the following command:
download ssl <ip_address> privkey <key file>
If the operation is successful, the existing private key is overwritten. After the download is successful, a
check is performed to find out whether the private key downloaded matches the public key stored in
the certificate. If the private and public keys do not match, the switch displays a warning message
similar to the following: Warning: The Private Key does not match with the Public Key in
the certificate. This warning acts as a reminder to also download the corresponding certificate.
For security reasons, when downloading private keys, Extreme Networks recommends obtaining a pre-
generated key rather than downloading a private key from a TFTP server. See “Configuring
Pregenerated Certificates and Keys” on page 572 for more information.
Downloaded certificates and keys are not saved across switch reboots unless you save your current
switch configuration. After you use the save command, the downloaded certificate is stored in the
configuration file and the private key is stored in the EEPROM.

Security
Configuring Pregenerated Certificates and Keys

To get the pregenerated certificate from the user, use the following command:
configure ssl certificate pregenerated
You can copy and paste the certificate into the command line followed by a blank line to end the
command.
This command is also used when downloading or uploading the configuration. Do not modify the
certificate stored in the uploaded configuration file because the certificate is signed using the issuer’s
private key.
The certificate and private key file should be in PEM format and generated using RSA as the
cryptography algorithm.
To get the pregenerated private key from the user, use the following command:
configure ssl privkey pregenerated
You can copy and paste the key into the command line followed by a blank line to end the command.
This command is also used when downloading or uploading the configuration. The private key is
stored in the EEPROM.
The certificate and private key file should be in PEM format and generated using RSA as the
cryptography algorithm.
Displaying SSL Information

To display whether the switch has a valid private and public key pair and the state of HTTPS access,
show ssl

23 CLEAR-Flow

● Configuring CLEAR-Flow on page 573
● Adding CLEAR-Flow Rules to ACLs on page 574
● CLEAR-Flow Rule Examples on page 587
Overview
CLEAR-Flow is a broad framework for implementing security, monitoring, and anomaly detection in
ExtremeXOS software. Instead of simply looking at the source and destination of traffic,
CLEAR-Flow allows you to specify certain types of traffic that require more attention. After certain
criteria for this traffic are met, the switch can either take an immediate, predetermined action, or send a
copy of the traffic off-switch for analysis.
CLEAR-Flow is an extension to Access Control Lists (ACLs). You create ACL policy rules to count
packets of interest. CLEAR-Flow rules are added to the policy to monitor these ACL counter statistics.
The CLEAR-Flow agent monitors the counters for the situations of interest to you and your network.
You can monitor the cumulative value of a counter, the change to a counter over a sampling interval,
the ratio of two counters, or even the ratio of the changes of two counters over an interval. For example,
you can monitor the ratio between TCP SYN and TCP packets. An abnormally large ratio may indicate
a SYN attack.
The counters used in CLEAR-Flow are either defined by you in an ACL entry, or can be a predefined
counter. See “Predefined CLEAR-Flow Counters” on page 585 for a list and description of these
counters.
If the rule conditions are met, the CLEAR-Flow actions configured in the rule are executed. The switch
can respond by modifying an ACL that will block, prioritize, or mirror the traffic, executing a set of CLI
commands, or sending a report using a SNMP trap or EMS log message.
NOTE
CLEAR-Flow is available only on the BlackDiamond 10808 family and BlackDiamond 12000 series switches.
Configuring CLEAR-Flow
CLEAR-Flow is an extension to ACLs, so you must be familiar with configuring ACLs before you add
CLEAR-Flow rules to your ACL policies. Creating ACLs is described in detail in Chapter 12, “Access
Lists (ACLs).” Chapter 12 describes how to create ACL policies, the syntax of an ACL policy file, and
how to apply ACL policies to the switch. In this current chapter, you will find information about the
CLEAR-Flow rules that you add to ACL policies, including the CLEAR-Flow rules’ syntax and
behavior.

CLEAR-Flow
After creating the ACLs that contain CLEAR-Flow rules, and after applying the ACLs to the appropriate
interface, you will enable CLEAR-Flow on the switch. When CLEAR-Flow is enabled, the rules will be
evaluated by the CLEAR-Flow agent on the switch, and if any rules are triggered, the CLEAR-Flow
actions are executed.
To enable CLEAR-Flow, use the following command:
enable clear-flow
When you disable the CLEAR-Flow agent on the switch, CLEAR-Flow sampling stops, and all rules are
left in the current state. To disable CLEAR-Flow, use the following command:
disable clear-flow
NOTE
Any actions triggered while CLEAR-Flow is enabled will continue when CLEAR-Flow is disabled, unless explicitly
stopped.
Displaying CLEAR-Flow Configuration and Activity

To display the state of the CLEAR-Flow agent, any CLEAR-Flow policies on each interface, and the
number of CLEAR-Flow rules, use the following command:
show clear-flow
To display the CLEAR-Flow rules and configuration, use the following command:
show clear-flow rule
Or to display all the rules, use the following command:
show clear-flow rule-all
When CLEAR-Flow is enabled, any rules that satisfy the threshold will trigger and take action. To
display the CLEAR-Flow rules that have been triggered, use the following command:
show clear-flow rule-triggered
To display which ACLs have been modified by CLEAR-Flow rules, use the following command:
show clear-flow acl-modified
Adding CLEAR-Flow Rules to ACLs

As described in the chapter about ACLs, each ACL policy file consists of a number of named entries.
Each entry consists of match conditions and actions to take if the entry is matched. CLEAR-Flow builds
on the ACL concept to include rules that are periodically checked, and actions to take if a rule is
triggered. The CLEAR-Flow entries are similar to the ACL entries.

The syntax of a CLEAR-Flow rule entry is:
entry <CLFrulename> <match-type> {

if { <match-conditions>;
}
then {
<actions>;
}
}
Or you can specify an optional else clause:

if { <match-conditions>;
}
then {
<actions>;
} else {
<actions>;
}
}
In the CLEAR-Flow rule syntax, the <CLFrulename> is the name of the rule (maximum of 31
characters). The <match-type> specifies whether the rule is triggered when any of the expressions that
make up the conditions are true (logical OR), or only when all of the expressions are true (logical AND).
The <match-type> is an optional element. The <match-conditions> specifies the conditions that will
trigger the rule, and how often to evaluate the rule. The <actions> in the then clause is the list of
actions to take when the rule is triggered, and the optional else clause <actions> is the list of actions to
take after the rule is triggered, and when the <match-conditions> later become false.
NOTE
When you create an ACL policy file that contains CLEAR-Flow rules, the CLEAR-Flow rules do not have any
precedence, unlike the ACL entries. Each CLEAR-Flow rule specifies how often it should be evaluated. The order of
evaluation depends on the sampling time and when the CLEAR-Flow agent receives the counter statistics. The order
of the CLEAR-Flow rules in the policy file does not have any significance.
The rule match type, rule match conditions, and rule actions are discussed in these sections:
● CLEAR-Flow Rule Match Type on page 575
● CLEAR-Flow Rule Match Conditions on page 576
● CLEAR-Flow Rule Actions on page 582
CLEAR-Flow Rule Match Type

The match type is optional. The two possible choices for the match type are:
● match all—All the match expressions must be true for a match to occur. This is the default.
● match any—If any match expression is true, then a match occurs.

CLEAR-Flow
NOTE
The match type was first available for CLEAR-Flow in ExtremeXOS 11.2.
CLEAR-Flow Rule Match Conditions

In a CLEAR-Flow rule, the <match-conditions> portion consists of one to four expressions, an
optional global-rule statement, and an optional period statement:
if { <expression>;
<expression>;
<expression>;
<expression>;
global-rule;
period <interval>;
}
then {
<actions>;
} else {
<actions>;
}
}
In the following example, the CLEAR-Flow rule cflow_count_rule_example will be evaluated every ten
seconds. The actions statements will be triggered if the value of counter1 (defined earlier in the ACL
policy file) is greater than 1,000,000:
entry cflow_count_rule_example {
if { count counter1 > 1000000 ;
period 10 ;
}
then {
<actions>;
}
}
The global-rule statement is optional and affects how the counters are treated. An ACL that defines
counters can be applied to more than one interface. In the original release of CLEAR-Flow, however,
any counters used in an expression were only evaluated for that particular interface that the CLEAR-
Flow rule was applied to. Beginning with the ExtremeXOS 11.2 release, you can specify the global-rule
statement so that counters are evaluated for all the applied interfaces. For example, if a policy that
defines a counter is applied to port 1:1 and 2:1, a CLEAR-Flow rule that used the global-rule statement
would sum up the counts from both ports. Without the global-rule statement, the CLEAR-Flow rule
would only look at the counts received on one port at a time.
The period <interval> statement is optional and sets the sampling interval, in seconds. This
statement specifies how often the rule is evaluated by the CLEAR-Flow agent. If not specified, the
default value is 5 seconds.
NOTE
In ExtremeXOS 11.1 only one expression was allowed per CLEAR-Flow rule.

The five CLEAR-Flow rule expressions are: count, delta, ratio, delta-ratio, and rule. All of these
expressions check the values of counters to evaluate if an action should be taken. The counters are
either defined in the ACL entries that are defined on the switch, or are the predefined CLEAR-Flow
counters. When you use a counter statement in an ACL, you are defining the counter used by CLEAR-
Flow to monitor your system.
The following sections discuss the CLEAR-Flow rule expressions in detail:

● Count Expression on page 577
● Delta Expression on page 578
● Ratio Expression on page 579
● Delta-Ratio Expression on page 580
● Rule-True-Count Expression on page 582
Count Expression
A CLEAR-Flow count expression compares a counter with the threshold value. The following is the
syntax for a CLEAR-Flow count expression:
count <counterName> REL_OPER <countThreshold> ;

hysteresis <hysteresis> ;
Beginning in ExtremeXOS release 11.4, the value of <countThreshold> and <hysteresis> can be
specified as floating point numbers. The count statement specifies how to compare a counter with its
threshold. The <counterName> is the name of an ACL counter referred to by an ACL rule entry and the
<countThreshold> is the value compared with the counter. The REL_OPER is selected from the
relational operators for greater than, great than or equal to, less than, or less than or equal to (>, >=, <,
<=).
The hysteresis <hysteresis> statement is optional, and sets a hysteresis value for the threshold.
After the count statement is true, the value of the threshold is adjusted so that a change smaller than
the hysteresis value will not cause the statement to become false. For statements using the REL_OPER >
or >=, the hysteresis value is subtracted from the threshold; for < or <=, the hysteresis value is added to
the threshold.
Here is an example of a count expression used in a CLEAR-Flow rule:
period 10 ;
}
then {
<actions>;
}
}
The actions will be discussed in the section, “CLEAR-Flow Rule Actions” on page 582.

CLEAR-Flow
In Table 74 is an example of evaluating the CLEAR-Flow count expression above multiple times. Notice
that the rule is not triggered until evaluation 3, when the value of the counter is greater than 1,000,000.
Table 74: Count Expression Evaluation Example

Evaluation counter1 value Rule triggered?
1 384625 No
2 769250 No
3 1153875 Yes
4 1538500 Yes
See “Count Expression Example” on page 587 for a full example of an ACL and a CLEAR-Flow rule
using a count expression.
Delta Expression
A CLEAR-Flow delta expression computes the difference from one sample to the next of a counter
value. This difference is compared with the threshold value. The following is the syntax for a CLEAR-
Flow delta expression:
delta <counterName> REL_OPER <countThreshold> ;
specified as floating point numbers. The delta expression specifies how to compare the difference in a
counter value from one sample to the next with its threshold. The <counterName> is the name of an
ACL counter referred to by an ACL rule entry and the <countThreshold> is the value compared with
the difference in the counter from one sample to the next. The REL_OPER is selected from the relational
operators for greater than, great than or equal to, less than, or less than or equal to (>, >=, <, <=).
After the delta statement is true, the value of the threshold is adjusted so that a change smaller than
the threshold.
For example, the following delta expression:
delta counter1 >= 100 ;

hysteresis 10 ;
will only be true after the delta of the counter reaches at least 100. At the time it becomes true, the
hysteresis value is subtracted from the threshold (setting the threshold to 90). With the threshold now at
90, the condition would stay true until the delta of the counter becomes less than 90.
If the expression becomes false, the threshold is reset to its original value. You would use the hysteresis
value to prevent the expression from vacillating between the true and false states if the difference
between the counter values is near the threshold. If the hysteresis value is greater than the threshold
value, the hysteresis value will be set to 0.

In Table 75 is an example of evaluating the CLEAR-Flow delta expression above multiple times. Notice
that the rule is not triggered until evaluation 4, when the delta value (the change in the counter value
from one evaluation to the next) is greater than or equal to 100. After the rule is triggered, it remains
triggered until the delta value is less than 90 (the original threshold minus the hysteresis), at evaluation
7. At evaluation 9, the rule is again triggered when the delta reaches 100. The rule will remain triggered
until the delta drops below 90.
Table 75: Delta Expression Evaluation Example

Evaluation counter1 value Delta value Rule triggered?
1 397 N/A No
2 467 70 No
3 547 80 No
4 657 110 Yes
5 757 100 Yes
6 852 95 Yes
7 932 80 No
8 1031 99 No
9 1131 100 Yes
10 1230 99 Yes
See “Delta Expression Example” on page 588 for a full example of an ACL and a CLEAR-Flow rule
using a delta expression.
Ratio Expression
A CLEAR-Flow ratio expression compares the ratio of two counter values with the threshold value. The
following is the syntax for a CLEAR-Flow ratio expression:
ratio <counterNameA> <counterNameB> REL_OPER <countThreshold> ;
min-value <min-value> ;
specified as floating point numbers, and the ratio is computed as a floating point number. The ratio
statement specifies how to compare the ratio of two counters with its threshold. The value of
<counterNameA> is divided by the value of <counterNameB>, to compute the ratio. That ratio is
compared with the <countThreshold>. The REL_OPER is selected from the relational operators for
greater than, great than or equal to, less than, or less than or equal to (>, >=, <, <=).
The min-value statement is optional, and sets a minimum value for the counters. If either counter is less
than the minimum value, the expression evaluates to false. If not specified, the minimum value is 1.
After the ratio statement is true, the value of the threshold is adjusted so that a change smaller than
the threshold.

CLEAR-Flow
For example, the following ratio expression:
ratio counter1 counter2 >= 5 ;

min-value 100;
hysteresis 1 ;
will only be true after the ratio of the counters reached at least 5 and the counter values are at least 100.
At the time it became true, the hysteresis value would be subtracted from the threshold (setting the
threshold to 4). With the threshold now at 4, the condition would stay true until the ratio of the
counters became less than 4.
If the statement becomes false, the threshold is reset to its original value. You would use the hysteresis
value to prevent the rule from vacillating between the true and false states if the ratio between the
counter values is near the threshold. If the hysteresis value is greater than the threshold value, the
hysteresis value will be set to 0.
In Table 76 is an example of evaluating the CLEAR-Flow ratio expression above multiple times. Notice
that the rule is not triggered at the first evaluation because both counters have not yet reached the min-
value of 100. The rule first triggers at evaluation 3, when ratio of the two counters exceeds 5. After the
rule is triggered, it remains triggered until the ratio value is less than 4 (the original threshold minus
the hysteresis), at evaluation 5. At evaluation 7, the rule is again triggered when the ratio reaches 5. The
rule will remain triggered until the ratio drops below 4.
Table 76: Ratio Expression Evaluation Example

Evaluation counter1 value counter2 value ratio Rule triggered?
1 427 70 6 No
2 941 235 4 No
3 2475 412 6 Yes
4 2308 570 4 Yes
5 2313 771 3 No
6 3597 899 4 No
7 5340 1065 5 Yes
See “Ratio Expression Example” on page 589 for a full example of an ACL and a CLEAR-Flow rule
using a ratio expression.
Delta-Ratio Expression
A CLEAR-Flow delta-ratio expression is a combination of the delta and ratio expressions. The CLEAR-
Flow agent computes the difference from one sample to the next for each of the two counters. The ratio
of the differences is then compared to the threshold value. The following is the syntax for a CLEAR-
Flow delta-ratio expression (note the similarity to the delta expression):
delta-ratio <counterNameA> <counterNameB> REL_OPER <countThreshold> ;
min-value <min-value> ;
specified as floating point numbers, and the delta-ratio is computed as a floating point number. The
delta-ratio statement specifies how to compare the ratio of the counter differences with its threshold.
The difference of the sample values of <counterNameA> is divided by the difference of the sample
values of <counterNameB>, to compute the ratio that is compared with the <countThreshold>. The

REL_OPER is selected from the relational operators for greater than, great than or equal to, less than, or
less than or equal to (>, >=, <, <=).
The min-value statement is optional, and sets a minimum value for the counters. If either counter is less
than the minimum value, the expression evaluates to false. If not specified, the minimum value is 1.
After the ratio statement is true, the value of the threshold is adjusted so that a change smaller than
the threshold.
For example, the following delta-ratio expression:
delta-ratio counter1 counter2 >= 5 ;

min-value 100 ;
hysteresis 1 ;
will only be true after the ratio of the deltas of the counters reached at least 5. At the time it became
true, the hysteresis value would be subtracted from the threshold (setting the threshold to 4). With the
threshold now at 4, the condition would stay true until the ratio of the deltas of the counters became
less than 4.
If the statement becomes false, the threshold is reset to its original value. You would use the hysteresis
value to prevent the rule from vacillating between the true and false states if the ratio of the deltas of
the counters is near the threshold. If the hysteresis value is greater than the threshold value, the
hysteresis value will be set to 0.
In Table 77 is an example of evaluating the CLEAR-Flow delta-ratio expression above multiple times.
Notice that the rule is not triggered at the second evaluation because both counters have not yet reached
the min-value of 100. The rule first triggers at evaluation 4, when ratio of the two counters exceeds 5.
After the rule is triggered, it remains triggered until the ratio value is less than 4 (the original threshold
minus the hysteresis), at evaluation 6. At evaluation 8, the rule is again triggered when the ratio reaches
5. The rule will remain triggered until the ratio drops below 4.
Table 77: Delta-Ratio Expression Evaluation Example

counter1 counter1 counter2 counter2 Rule
Evaluation value delta value delta Ratio triggered?
1 110 N/A 20 N/A N/A No
2 537 427 90 70 6 No
3 1478 941 325 235 4 No
4 3953 2475 737 412 6 Yes
5 6261 2308 1307 570 4 Yes
6 8574 2313 2078 771 3 No
7 12171 3597 2977 899 4 No
8 17511 5340 4042 1065 5 Yes
See “Delta-Ratio Expression Example” on page 590 for a full example of an ACL and a CLEAR-Flow
rule using a delta-ratio expression.

CLEAR-Flow
Rule-True-Count Expression
A CLEAR-Flow rule-true-count expression compares how many times a CLEAR-Flow rule is true with a
threshold value. One use is to combine multiple rules together into a complex rule. The following is the
syntax for a CLEAR-Flow rule-true-count expression:
rule-true-count <ruleName> REL_OPER <countThreshold> ;
The rule-true-count statement specifies how to compare how many times a CLEAR-Flow rule is true
with the expression threshold. The <ruleName> is the name of the CLEAR-Flow rule to monitor and the
<countThreshold> is the value compared with the number of times the rule is true. The REL_OPER is
selected from the relational operators for greater than, great than or equal to, less than, or less than or
equal to (>, >=, <, <=).
For example, the following delta-ratio expression:
rule-true-count cflow_count_rule_example >= 5 ;
will only be true after the CLEAR-Flow rule cflow_count_rule_example has been true at least five times. If
the rule cflow_count_rule_example becomes true and remains true, and the period for
cflow_count_rule_example is the default five seconds, the rule would have to be true for at least 20
seconds before the rule-true-count expression will become true. If the period of the rule
cflow_count_rule_example is 10 seconds, it will need to be true for at least 40 seconds before the rule-
true_count expression becomes true.
CLEAR-Flow Rule Actions

CLEAR-Flow rules specify an action to take when the rule is triggered and can optionally specify an
action to take when the expression is false. Because more than one action can be taken in a single rule,
the collection of actions is referred to as an action list.
The actions that can be taken are:

● Permit/Deny
● QoS Profile
● Mirror
● SNMP Trap
● Syslog
● CLI
Additionally, the SNMP trap, syslog, and CLI rule actions can use keyword substitution to make the
rule actions more flexible. The keyword substitutions are described at the end of the rule action
descriptions. See “Keyword Substitution” on page 584 for more information.
The following sections describe the different rule actions.
Permit/Deny
This action modifies an existing ACL rule to permit or block traffic that matches that rule.
To change an ACL to permit, use the following syntax:

permit <ACLRuleName>

To change an ACL to deny, use the following syntax:

deny <ACLRuleName>
QoS Profile
This action modifies an existing ACL rule to set the QoS profile for traffic that matches that rule.
To change the ACL to forward to QoS profile <QPx>, use the following syntax:
qosprofile <ACLRuleName> <QPx>
For example:
qosprofile acl_rule_1 QP3
Mirror
This action modifies an existing ACL rule to mirror traffic that matches that rule, or to stop mirroring
that traffic. The mirroring port must be enabled when mirroring on an ACL rule is turned on. This
could be configured earlier, or use the CLI action to execute CLI commands to configure mirroring at
the same time.
To change the ACL to mirror traffic, use the following syntax:

mirror [add|delete] <ACLRuleName>
For example (enabling mirroring from within CLEAR-Flow rule):

cli “enable mirroring to port 7:4 tagged”
mirror add acl_rule_1
SNMP Trap
This action sends an SNMP trap message to the trap server, with a configurable ID and message string,
when the rule is triggered.
The message is sent periodically with interval <period> seconds. If <period> is 0, or if this optional
parameter is not present, the message is sent only once when the rule is triggered. The interval must be
a multiple of the rule sampling/evaluation interval, or the value will be rounded down to a multiple of
the rule sampling/evaluation interval.
To send an SNMP trap, use the following syntax:

snmptrap <id> <message> <period>
Syslog
This action sends log messages to the ExtremeXOS EMS sever. The possible values for message level
are: DEBU, INFO, NOTI, WARN, ERRO, and CRIT.
The message is sent periodically with interval <period> seconds. If <period> is 0, or if this optional
parameter is not present, the message is sent only once when the rule is triggered. The interval must be
a multiple of the rule sampling/evaluation interval, or the value will be rounded down to a multiple of
the rule sampling/evaluation interval.

CLEAR-Flow
The messages are logged on both MSMs, so if the backup log is sent to the primary MSM, then the
primary MSM will have duplicate log messages.
To send a log message, use the following syntax:

syslog <message> <level> <period>
CLI
This action executes a CLI command. There is no authentication or checking the validity of each
command. If a command fails, the CLI will log a message in the EMS log.
To execute a CLI command, use the following syntax:
cli <cliCommand>
where <cliCommand> is a quoted string.
Keyword Substitution
To make the SNMP trap, syslog, and CLI actions more flexible, keyword substitutions are supported in
the syslog and SNMP trap message strings, as well as in the CLI command strings. Table 78 lists the
keywords and their substitutions.
If a keyword is not supported, or a counter name is not found, a string of

“unknownKeyword[$keyword]” will be substituted.
For the $vlanName and $port keyword, the keyword all will be substituted for those rules in the
wildcard ACL Some CLI commands do not support the all keyword, so caution must be used with CLI
commands that use this feature.
A maximum of 10 different counter substitutions can be used per rule, including counters used in
expressions. For example, if a rule uses four counters in its expressions, then we can use six more
different counters in keyword substitutions, for a total of 10.
Table 78: Keyword Substitutions

Keyword Substitution
$policyName Replace with the policy name.
$ruleName Replace with the CLEAR-Flow rule name.
$<counterName> Replace with counter value for the indicated counter name.
$ruleValue Replace with the current expression value.
$ruleThreshold Replace with the expression threshold value.
$ruleInterval Replace with the rule sampling/evaluation interval.
$vlanName Replace with the interface VLAN name.
$port Replace with the interface port number.

Predefined CLEAR-Flow Counters

A number of packet statistics are gathered by the XOS kernel. To allow you to use these statistics in
CLEAR-Flow expressions, these kernel counters are now available for use with CLEAR-Flow. Most of
the counter names are based directly on well known names from common kernel structures and MIBs.
The names are modified from their familiar form by prepending the characters sys_ to the counter
names.
Table 79: Predefined CLEAR-Flow Counters

Counter Name Descriptiona
sys_IpInReceives The total number of input IP packets received from interfaces, including those
received in error.
sys_IpInHdrErrors The number of input IP packets discarded due to errors in their IP headers,
including bad checksums, version number mismatch, other format errors, time-
to-live exceeded, errors discovered in processing their IP options, etc.
sys_IpInAddrErrors The number of input IP packets discarded because the IP address in their IP
header's destination field was not a valid address to be received at this entity.
This count includes invalid addresses (for example, 0.0.0.0) and addresses of
unsupported Classes (for example, Class E).
sys_IpForwDatagrams The number of input IP packets for which this entity was not their final IP
destination, as a result of which an attempt was made to find a route to forward
them to that final destination.
sys_IpInUnknownProtos The number of locally-addressed IP packets received successfully but discarded
because of an unknown or unsupported protocol.
sys_IpInDiscards The number of input IP packets for which no problems were encountered to
prevent their continued processing, but which were discarded (for example, for
lack of buffer space). Note that this counter does not include any IP packets
discarded while awaiting re-assembly.
sys_IpInDelivers The total number of input IP packets successfully delivered to IP user-protocols
(including ICMP).
sys_IpOutRequests The total number of IP packets which local IP user-protocols (including ICMP)
supplied to IP in requests for transmission. Note that this counter does not
include any IP packets counted in ipForwDatagrams.
sys_IpOutDiscards The number of output IP packets for which no problem was encountered to
prevent their transmission to their destination, but which were discarded (for
example, for lack of buffer space). Note that this counter would include IP
packets counted in ipForwDatagrams if any such packets met this (discretionary)
discard criterion.
sys_IpOutNoRoutes The number of IP packets discarded because no route could be found to
transmit them to their destination. Note that this counter includes any packets
counted in ipForwDatagrams which meet this `no-route' criterion.
sys_IpReasmTimeout The maximum number of seconds which received fragments are held while they
are awaiting reassembly at this entity.
sys_IpReasmReqds The number of IP fragments received which needed to be reassembled at this
entity.
sys_IpReasmOKs The number of IP packets successfully re-assembled.
sys_IpReasmFails The number of failures detected by the IP re-assembly algorithm (for whatever
reason: timed out, errors, etc.). Note that this is not necessarily a count of
discarded IP fragments since some algorithms (notably the algorithm in RFC
815) can lose track of the number of fragments by combining them as they are
received.
sys_IpFragOKs The number of IP packets that have been successfully fragmented at this entity.

CLEAR-Flow
Table 79: Predefined CLEAR-Flow Counters (Continued)

sys_IpFragFails The number of IP packets that have been discarded because they needed to be
fragmented at this entity but could not be, for example, because their Don't
Fragment flag was set.
sys_IpFragCreates The number of IP packet fragments that have been generated as a result of
fragmentation at this entity.
sys_IcmpInMsgs The total number of ICMP messages which the entity received. Note that this
counter includes all those counted by icmpInErrors.
sys_IcmpInErrors The number of ICMP messages which the entity received but determined as
having ICMP-specific errors (bad ICMP checksums, bad length, etc.).
sys_IcmpInDestUnreachs The number of ICMP Destination Unreachable messages received.
sys_IcmpInTimeExcds The number of ICMP Time Exceeded messages received.
sys_IcmpInParmProbs The number of ICMP Parameter Problem messages received.
sys_IcmpInSrcQuenchs The number of ICMP Source Quench messages received.
sys_IcmpInRedirects The number of ICMP Redirect messages received.
sys_IcmpInEchos The number of ICMP Echo (request) messages received.
sys_IcmpInEchoReps The number of ICMP Echo Reply messages received.
sys_IcmpInTimestamps The number of ICMP Timestamp (request) messages received.
sys_IcmpInTimestampReps The number of ICMP Timestamp Reply messages received.
sys_IcmpInAddrMasks The number of ICMP Address Mask Request messages received.
sys_IcmpInAddrMaskReps The number of ICMP Address Mask Reply messages received.
sys_IcmpOutMsgs The total number of ICMP messages which this entity attempted to send. Note
that this counter includes all those counted by icmpOutErrors.
sys_IcmpOutErrors The number of ICMP messages which this entity did not send due to problems
discovered within ICMP such as a lack of buffers. This value should not include
errors discovered outside the ICMP layer such as the inability of IP to route the
resultant datagram. In some implementations there may be no types of error
which contribute to this counter's value.
sys_IcmpOutDestUnreachs The number of ICMP Destination Unreachable messages sent.
sys_IcmpOutTimeExcds The number of ICMP Time Exceeded messages sent.
sys_IcmpOutParmProbs The number of ICMP Parameter Problem messages sent.
sys_IcmpOutSrcQuenchs The number of ICMP Source Quench messages sent.
sys_IcmpOutRedirects The number of ICMP Redirect messages sent.
sys_IcmpOutEchos The number of ICMP Echo (request) messages sent.
sys_IcmpOutEchoReps The number of ICMP Echo Reply messages sent.
sys_IcmpOutTimestamps The number of ICMP Timestamp (request) messages sent.
sys_IcmpOutTimestampReps The number of ICMP Timestamp Reply messages sent.
sys_IcmpOutAddrMasks The number of ICMP Address Mask Request messages sent.
sys_IcmpOutAddrMaskReps The number of ICMP Address Mask Reply messages sent.
sys_IcmpInProtoUnreachs The number of incoming ICMP packets addressed to a not-in-use/unreachable/
invalid protocol. This message is in the general category of ICMP destination
unreachable error messages.
sys_IcmpInBadLen The number of incoming bad ICMP length packets.b
sys_IcmpInBadCode The number of incoming ICMP packets with a bad code field value.
sys_IcmpInTooShort The number of incoming short ICMP packets.

CLEAR-Flow Rule Examples
Table 79: Predefined CLEAR-Flow Counters (Continued)

sys_IcmpInBadChksum The number of incoming ICMP packets with bad checksums.
sys_IcmpInRouterAdv The number of incoming ICMP router advertisements. Router advertisements are
used by IP hosts to discover addresses of neighboring routers.
sys_IcmpOutProtoUnreachs The number of outgoing ICMP packets addressed to a not-in-use/unreachable/
invalid protocol. This message is in the general category of ICMP destination
unreachable error messages.
sys_IcmpOutRouterAdv The number of outgoing ICMP router advertisements. Router advertisements are
used by IP hosts to discover addresses of neighboring routers.
sys_IgmpInQueries The number of Host Membership Query messages that have been received on
this interface.
sys_IgmpInReports The number of Host Membership Report messages that have been received on
this interface for this group address.
sys_IgmpInLeaves The number of incoming IGMP leave requests.
sys_IgmpInErrors The number of incoming IGMP errors.
sys_IgmpOutQueries The number of Host Membership Query messages that have been sent on this
interface
sys_IgmpOutReports The number of Host Membership Report messages that have been sent on this
interface for this group address.
sys_IgmpOutLeaves The number of outgoing IGMP leave requests.
a. Most of these descriptions can be found in RFC 2011, SNMPv2 Management Information Base for the In-
ternet Protocol using SMIv2.
b. The length of an ICMP packet depends on the type and code field.

In the examples that follow, one to two ACL rule entries are followed by a CLEAR-Flow rule entry. The
examples illustrate the four CLEAR-Flow rule expressions: count, delta, ratio, and delta-ratio.
Count Expression Example

In the following example, every 10 seconds the CLEAR-Flow agent will request the counter1 statistics
from the hardware. After it receives the counter value, it will evaluate the CLEAR-Flow rule. If the
value of counter1 is greater than 1000000 packets, the CLEAR-Flow agent will send a trap message to
the SNMP master, and change the ACL acl_rule1 to block traffic (acl_rule1 is modified to a deny rule).
Since there is no period configured for the snmptrap statement, the message is sent only once.
entry acl_rule1 {
if {
protocol tcp;
} then {
count counter1;
}

CLEAR-Flow
period 10 ;
}
then {
snmptrap 123 "Traffic on acl_rule1 exceeds threshold";
deny acl_rule1;
}
}
Delta Expression Example

In this example, every 10 seconds the CLEAR-Flow agent will request the counter1 statistics from the
hardware. After it receives the counter value, it will then evaluate the rule. If the delta (change) of the
counter1 value from the last sampled value 10 seconds ago is greater than or equal to 1000 packets, the
CLEAR-Flow agent will send a trap message to the SNMP master, and change the ACL acl_rule1 to
move the traffic to QP3. In addition, reduce the peak rate to 5 Kbps on QP3. As long as the delta
continues to be greater than or equal to 1000 packets, the CLEAR-Flow agent will repeatedly send a
trap message every 120 seconds. When the delta falls below the threshold, the agent will execute the
two actions in the else portion; it will send a single SNMP trap message, return the traffic to QP1, and
reset QP3 to its original bandwidth.
entry acl_rule1 {
if {
protocol tcp;
} then {
count counter1;
}
}
entry cflow_delta_rule_example {
if { delta counter1 >= 100000 ;
period 10 ;
} then {
snmptrap 123 "Traffic to 192.168.16.0/24 exceed rate limit" 120;
qosprofile acl_rule1 QP3;
cli "configure qosprofile qp3 peak_rate 5 K ports all" ;
} else {
snmptrap 123 "Traffic to 192.168.16.0/24 falls below rate limit";
qosprofile acl_rule1 QP1;
cli "configure qosprofile qp3 maxbw 100 ports all" ;
}
}

Ratio Expression Example

In this example, every 2 seconds the CLEAR-Flow agent will request the counter1 and counter2 statistics
from the hardware. After it receives the two counter values, it will then check each counter value
against its minimum valid threshold, which is 1000. If both of the counter values is greater then 1000, it
then calculates the ratio of counter1 and counter2. If the ratio is greater than 5, then the agent will
execute the actions in the then clause, which consists of logging a message to the syslog server. Before
logging the syslog string, the agent will replace the $ruleName keyword with the string
cflow_ratio_rule_example, the $ruleValue keyword with the calculated ratio value, and the
$ruleThreshold keyword with a value of 5. If either of the counter values is below the minimum value
of 1000, or the ratio is below the threshold of 5, the expression is false and no action is taken.
entry acl_rule1 {
if {
protocol udp;
} then {
count counter1;
}
}
entry acl_rule2 {
if {
protocol tcp;
} then {
count counter2;
}
}
entry cflow_ratio_rule_example {
if { ratio counter1 counter2 > 5 ;
period 2;
min-value 1000;
}
then {
syslog "Rule $ruleName threshold ratio $ruleValue exceeds limit
$ruleThreshold";
}
}

CLEAR-Flow
Delta-Ratio Expression Example

In this example, every 2 seconds, the CLEAR-Flow agent will request the tcpSynCounter and tcpCounter
values from the hardware. After it receives the two counter values, it will first calculate the delta for
each of the counters and then check each counter’s delta value for its minimum value, which is 100. If
both of the counters’ delta values are greater then 100, it then calculates the ratio of the delta of two
counters. If the ratio is greater than 10, then the agent will log a warning message and deny all SYN
traffic on the interface. No period value for the syslog message is given, so the message will be logged
once when the expression first becomes true. When the expression transitions from true to false, a
different message will be logged and the SYN traffic on the interface will be permitted again. The delta-
ratio value has to fall below a threshold of 8 for the expression to be evaluated to be false.
entry acl_syn {
if {
protocol tcp_flags SYN;
} then {
count tcpSynCounter;
}
}
entry acl_tcp {
if {
protocol tcp;
} then {
count tcpCounter;
}
}
entry cflow_delta_ratio_rule_example {
if { delta-ratio tcpSynCounter tcpCounter > 10 ;
period 2;
min-value 100;
threshold 8;
} then {
syslog "Syn attack on port $port is detected" WARN;
deny acl_syn;
} else {
syslog "Syn attack on port $port is no longer detected" WARN;
permit acl_syn;
}
}

24 Configuring Stacked Switches

● SummitStack Compatible Switches on page 598
● SummitStack Topologies on page 599
● Logging into a SummitStack on page 603
● Managing Licenses on a SummitStack on page 604
● Configuring a New Stack on page 607
● SummitStack Configuration Tasks on page 611
● Managing an Operating SummitStack
● Managing a Dual Master Situation on page 645
● Configuration Examples on page 626
● Troubleshooting a Stack on page 644
● FAQs on SummitStack on page 650
Overview
SummitStack allows you to physically connect up to eight individual Summit switches together as a
single logical unit. This logical unit behaves as a single switch with a single IP address and a single
point of authentication.
In ExtremeXOS, a stack is controlled by a master switch, called the master. The master switch runs full
ExtremeXOS and is responsible for maintaining all of the software tables for all the switches in the
stack. There can only be one master switch in a stack of switches. All switches in the stack, including
the master switch, are called nodes.
A SummitStack can be thought of as a virtual chassis. Each node acts as if it was occupying a slot in a
chassis and is controlled by the master. The high-speed stacking links function like the backplane links
of a chassis.
The master switch stores any configuration information for the stack in its primary and secondary flash
memory. Since the master switch has the knowledge of the state and the configuration of all the other
switches in the stack, it can respond to all external requests for those switches. For example, the master
switch can respond to a request for SNMP information from all ports within the stack.
NOTE
All participants in a stack must run compatible image versions.

Configuring Stacked Switches
This section introduces the following SummitStack topics:
● Understanding SummitStack Terms on page 592

● Required Software Release on page 594
● Understanding SummitStack Configuration Parameters, Configuration Files, and Port Numbering on
page 594
● Understanding Stacking Link Overcommitment on page 595
● About SummitStack Logging Messages on page 595
● About QoS in Stacking on page 596
● About Power Management and Power Over Ethernet on Stacking on page 597
● About Stacking Node Roles, Redundancy, and Failover on page 597
● About the Failsafe Account on SummitStack Nodes on page 598
Understanding SummitStack Terms

This section describes the terms used in SummitStack.
Table 80: List of Stacking Terms

Term Description
Stackable Switch A switch that provides two stacking ports and can participate in a stack.
Stacking Port A physical interface of a stackable switch that is used to allow the connection
of a stacking link. Stacking ports are point-to-point links that are dedicated for
the purpose of forming a stack.
Stacking Link A wire that connects a stacking port of one stackable to a stacking port of
another stackable, plus the stacking ports themselves.
Node A node is a stackable switch that runs the ExtremeXOS operating system. The
ExtremeXOS version must be the initial stacking release (ExtremeXOS 12.0) or
greater. The terms node and stackable are used interchangeably in this chapter.
Stack A stack is a set of stackable switches and their connected stacking links made
with the intentions that: (1) all switches are reachable through their common
connections; (2) a single stackable can manage the entire stack; and (3)
configurable entities such as VLANs and link trunk groups can have members
on multiple stackables. A stack consists of all connected nodes regardless of
the state of these nodes.
Control Path A data path that is formed over the stacking links that are dedicated to carrying
stacking control traffic, such as commands to program hardware or software
image data for software upgrade.
Active Node This is a node in a multi-node stack that has joined the control path. The
active node can forward the control path messages or can process the control
path messages. It can also forward data traffic. Only an active node can appear
as a card inserted into a slot when the show slot command is executed on
the master node of the stack.
Active Topology The set of active nodes in a stack topology plus the set of stacking links that
connect them form the active topology. When an active topology consists of
more than one node, each node in the active topology is directly and physically
connected to at least one other node in the active topology. Thus, the active
topology within a stack topology is a set of physically contiguous nodes.
Candidate Node A node that is a potential member of an active topology is called a candidate
node. An active node is also a candidate node. Unlike an active node, a
candidate node need not necessarily join the OCP.

Overview
Table 80: List of Stacking Terms (Continued)

Term Description
Node Role A node in the active topology plays a role in the stack. There are three node
roles: master, backup, and standby.
Master Node Role A node that is elected as the master runs all of the configured control protocols
such as OSPF, RIP, Spanning Tree, EAPS, and so forth.
The master node controls all data ports on itself, the backup node, and all
standby nodes. The master node issues specific programming commands to the
backup or standby nodes to accomplish this purpose.
Backup Node Role The node that takes over the master node role if the master node fails. The
master node keeps the backup node databases in synchronization with its own
database in preparation for this event. Upon transfer of role, the backup node
becomes the master node and begins operating with the databases it has
previously received. This allows all other nodes in the stack to continue
operating even after the master node fails.
Standby Node Role A standby node is prepared to become a backup node in the event that the
backup node becomes the master node. When becoming a backup node, the
new master node synchronizes all of its databases to the new backup node. As
a standby node, most databases are not synchronized, except for those few that
directly relate to hardware programming.
Acquired Node A standby or backup node is normally acquired by a master node. This means
the master node has used its databases to program the hardware of the standby
or backup node. The standby or backup node has acted as a hardware
programming proxy, accepting the instructions of the master node to do so. An
acquired standby node does not maintain the databases needed to reflect why
the hardware is programmed as it is; however, a backup node does. An
acquired node can only be re-acquired (without a reboot) by the backup node
when that backup node becomes a master node, and only if both the backup
and standby nodes were already acquired by the same master node at the time
of its failure.
Data Ports This is the set of ports provided by a stackable switch that are available to you
for connection to your data networks. Such ports can be members of a user
configured VLAN or trunk group, and can be used for layer 2 and 3 forwarding
of user data traffic or for mirroring, or other features you can configure. This
term does not refer to stacking ports.
Failover When a node that is executing the master node role in a stack fails, a failover
is initiated. If there is a node that is executing the backup node role, and if the
node has completed its initial synchronization with the master node before it
failed, the backup node takes on the master node role. The standby nodes
continue their operation, and their data ports do not fail.
Hitless Failover A failover whereby all data ports in the stack, except those of the failing master
node, continue normal operation when the master node fails.
Hitless Upgrade This is an operation where an upgrade of the software image and the
commencement of the new image execution is possible without interrupting
data traffic or forcing any network reconvergence. This version of SummitStack
does not support hitless upgrade.
Node ID A unique identifier for a node. The identifier is the factory assigned MAC
address of the node.
Node Role Election This is the process that determines the role for each node. The election takes
place during initial stack startup and elects a master and a backup node, The
election also takes place after a master node failover and elects a new backup
node.
Node Role Election Priority For each node, the stack computes a priority to be used in node role election.
The node with the highest node role election priority during a role election
becomes the master node. The node with the second highest priority becomes
the backup.

Table 80: List of Stacking Terms (Continued)

Term Description
Operational Node This is a node that has achieved operational state as a card in a slot. The
operational state can be displayed using the show slot command.
System UpTime In a stack, this is the amount of time that has passed since a stack segment
first elected a master node after a stack segment reboot. The time can be
displayed on a master node by entering the show switch command.
Stack Segment This is a collection of nodes that form a stack topology. The term is useful
when a stack is severed. Each severed portion of the stack is referred to as a
stack segment.
Stack State A state assigned by the stack to a node. This can be displayed using the
command show stacking.
Easy-Setup Easy-setup is a procedure that configures the essential stack parameters of
every node for initial stack deployment, and automatically reboots the stack to
put the parameters into effect. The choice to run easy-setup is offered when
the enable stacking command is run and the essential stacking parameters
are unconfigured or inconsistent. It can also be invoked directly by running the
configure stacking easy-setup command.
Required Software Release

Stacking is first supported in ExtremeXOS 12.0. You cannot run any of the stacking features in releases
earlier than ExtremeXOS 12.0. All stack members should run compatible versions of ExtremeXOS.
Understanding SummitStack Configuration Parameters,

Configuration Files, and Port Numbering
The stacking configurations are stored in the NVRAM of each node. Some of these configurations take
effect only during the next node restart.
Table 81: Stacking Configuration Items, Time of Effect and Default Value
Configuration Item Takes effect Default Value
Stacking Mode at boot time Disabled
Slot Number at boot time 1
Master-Capable at boot time Yes
License Restriction at boot time Not configured
Priority at the next master Automatic
election
Alternate IP Address immediately Not configured
Stack MAC at boot time Not configured
Stacking parameters, such as mode, slot number, etc., can be configured from a single unit in the stack
topology. You can change the stacking-specific configuration even when a node is not in stacking mode
but is connected to the stack. The target node for the configuration must be powered on and running a
version of ExtremeXOS that supports stacking. The version of ExtremeXOS on the target node need not
be the same version as the one running on the unit from which you are issuing the configuration
commands. Further, the node need not be in stacking mode and can be in any node role.

Overview
Most ExtremeXOS configuration parameters are not stored in NVRAM, but are instead stored in a
configuration file. Configurations stored in NVRAM are those that are needed when the configuration
file is not available. The configuration file chosen for the stack is the one selected on the master node
that is first elected after a stack restart.
The data (non-stacking) port numbers, in the existing configuration files, are simple integer quantities.
On a stack, the data port numbers are expressed as slot:port; where the slot is an integer representing
the slot and port is an integer representing the port. For example: 1:2. The configuration file contains an
indication that it was created on a stackable in stacking mode. The indication is the stacking platform
ID.
Thus when in stacking mode, the ports are referenced with the slot:port notation and when not in
stacking mode, the ports are referenced as simple integers.
When the stack restarts, if a switch becomes the master and its selected configuration file was not
created in stacking mode, the configuration file is de-selected, and the stack completes its restart using a
default configuration. In addition, if the previously selected file was named with one of the default
names (primary.cfg or secondary.cfg), the file is renamed to old_non_stack.cfg.
Similarly, if a switch is configured not to operate in stacking mode and the selected configuration file
was created in stacking mode, the configuration file is de-selected, and the switch boots with a default
configuration. In addition, if the file was named with one of the default names (primary.cfg or
secondary.cfg), the file is renamed to old_stack.cfg.
The renamed file replaces any file that exists with the same name; the existing file is deleted.
Understanding Stacking Link Overcommitment

The stack is formed by each node supplying a pair of full-duplex 10Gbps stacking ports. Each node can
operate on a stack with 20 Gbps full duplex throughput.
Even though two links are available, the links might not be fully utilized. For example, suppose there is
a ring of eight nodes and the nodes are numbered clockwise from 1 to 8. Suppose node 1 wants to send
10Gbps of unicast traffic to each of node 2 and node 3. The shortest path topology forces all traffic from
node 1 over the link to node 2. Traffic from node 1 to node 3 passes through node 2. Thus, there is only
10Gbps available. However, if node 1 wanted to send 10Gbps to node 2 and node 8, there would be
20Gbps available because both links connected to node 1 would be used.
In a ring of eight nodes, between any two nodes (with one exception), only one link is used. If the
devices provide 48 1Gbps Ethernet ports, the overcommitment ratio between two such nodes is
approximately 5:1. The exception is if there is an equal distance between the nodes. In this case, if both
nodes are 48 port nodes, the nodes are grouped into two groups of 24 ports (by the hardware
architecture), and thus it is possible to use both directions around the stack.
About SummitStack Logging Messages

Each node might generate log messages through the usual logging mechanism.
On backup and standby nodes, a log target and related filter is automatically installed. The log target is
Master Node. The filter allows all messages that have a log level of warning, error, or critical to be
saved in the log file of the master node.

If the master node changes, the log target is updated on all the remaining nodes.
You can also log in to any node in the active topology and see the complete log of the node.
About QoS in Stacking

This section describes the differences in QoS while using it in SummitStack.
QoS Profile Restrictions

In stacking mode, since the CoS levels 6 and 5 are reserved for stacking you cannot create the quality
profiles QP7 and QP6. Since these QoS profiles cannot be created, you will not be able to create a packet
classification that assigns CoS levels 6 or 5 to a packet.
NOTE
This restriction is applicable only when the stackable is operating in stacking mode.
Processing of Packets Received With 802.1p Priority 6 or 5

By default, 802.1p examination is turned on. Priority 7 is mapped to QoS profile QP8, and priorities 6
through 0 are mapped to QoS profile QP1. You can create other QoS profiles and may change this
mapping as needed. Since you cannot create QP7 or QP6 in stacking mode, 802.1p examination will
always map packets with priorities 6 and 5 to other CoS levels.
Effects on 802.1p Examination

You can turn off 802.1p examination. When stacking is enabled, the examination remains turned on for
the priorities 6 and 5. However, the examination happens at a lower precedence than that of all other
traffic groupings.
The mapping you have configured for these priorities will remain in effect, and will change accordingly
if you subsequently change the mapping.
When stacking is not enabled, all 802.1p examination is disabled when the feature is turned off.
Effects on DiffServ Examination

When DiffServ Examination and 802.1p Examination are both turned off, the 802.1p examination for
packets arriving at 802.1p priority levels 5 and 6 remain on at the lowered precedence. In addition, the
examination is adjusted to apply to all packets. The actual priority levels that will be used for such
packets are the defaults (QP1), or the values last configured using the configure dot1p type
command.
Effects on Port QoS and VLAN QoS

Port QoS and VLAN QoS has a higher precedence than the 802.1p priority examination performed
when 802.1p Examination feature is turned off, and is therefore unaffected.

Overview
About Power Management and Power Over Ethernet on Stacking

The power management for Power over Ethernet (PoE) is applicable only if there are one or more
X450e-XXp or X250e-XXp switches on the stack.
Each X450e-XXp or X250e-XXp switch is equipped with its own independent power supply that
provides power for the PoE ports on that switch. Power is not shared with other switches in the stack.
PoE configuration and status are maintained on the master node. Configuration information is sent by
the master to the hardware on each PoE capable switch to be controlled by the local PoE hardware on
that switch. Status is gathered on the master by querying the PoE hardware on each switch.
The power supply for each X450e-24p switch is capable of providing a full 15.4 watts per PoE port for
all 24 ports. The following power management CLI commands are not supported on a X450e-24p:
● configure inline-power priority
● configure inline-power disconnect-precedence
The X450e-48p contains an optional external modular Power Supply Unit (PSU) that can provide
redundant PoE power or full PoE power to all ports depending on the EPS-C/EPS600LS configuration.
When using the EPS-C/EPS600LS, the PoE capability of the X450e-48p varies depending on the number
of power modules present.
The following stacking CLI commands are applicable only to Summit X450e-48p:
● configure inline-power budget slot

● configure inline-power disconnect-precedence
● configure inline-power priority ports
● unconfigure inline-power disconnect-precedence
● unconfigure inline-power priority ports
● show power slot
These commands are available in stacking mode and will only function on a slot that contains an X450e-
48p.
About Stacking Node Roles, Redundancy, and Failover

EXOS supports control plane redundancy and Hitless Failover. A stack supports control plane
redundancy and hitless failover. “Hitless” failover is supported to the extent that the failing master
node and all of its ports are operationally lost, including the loss of supplied power on any PoE ports
that the node provided, but all other nodes and their provided ports continue to operate. After the
failover, the backup node becomes the master node.
At failover time, a new backup node is selected from the remaining standby nodes that are configured
to be master capable. All operational databases are then synchronized from the new master node to the
new backup node. Another hitless failover is possible only after the initial synchronization to the new
backup node has completed. This can be seen using the show switch command and noting that the
new backup node is In Sync.
When a backup node transitions to the master node role, it activates the Management IP interface that is
common to the whole stack. If you have correctly configured an alternate management IP address, the
IP address remains reachable.
When a standby node is acquired by a master node, the master node does not synchronize most of its
databases with the standby nodes. The standby node also learns the identity of its backup node.

Once a standby node loses the contact with both its acquiring master and backup nodes, it reboots.
A master node that detects the loss of an acquired standby node will indicate that the “slot” the standby
node occupied is now “empty” and flushes its dynamic databases of all information previously learned
about the lost standby node.
A backup node restarts if the backup node has not completed its initial synchronization with the master
node before the master node is lost.
Reboot or Failure of a Non-Master Node

If a backup node fails, a standby node configured as master-capable is elected as the new backup. That
new backup node is then synchronized to the databases of the master node.
For all non-master nodes, a node that reboots or is power cycled loses all of its connections to all
networks for the duration of the reboot cycle. Any PoE ports that were providing power prior to the
event do not supply power.
When a non-master node fails, the master node marks the related slot as Empty. All other nodes exclude
the failed node from any data paths including the topology protocol path, the control path, and any
customer-configured VLANs, trunk group ports, mirroring ports, and so forth.
About the Failsafe Account on SummitStack Nodes

The failsafe account is a special user account that is set up in the default configuration (see the chapter
on Accessing the Switch). The failsafe account functions even when there is no master node in the stack.
By default, the failsafe account can only be accessed through the console port of a node. The failsafe
account cannot be deleted, but you can modify the user ID and password (see “Configuring the Failsafe
Account on a Stack” on page 619).
If you fail to configure the failsafe account, the default failsafe user name and password are in effect.
SummitStack Compatible Switches

The following summit switches are supported in this release:
● X450e-24p
● X450a-24t
● X450a-24tDC
● X450a-48t
● X450a-24x
● X450a-24xDC
● X450e-48p
● X450a-48tDC
● X450-24t
● X450-24x
● X250e-24p

SummitStack Topologies
● X250e-24t
● X250e-48p
● X250e-48t
The following images provides you a graphical representation of a stack and the topologies.
Figure 61: Stack and Topologies
A
Active
topology
A
Stack
topology A
Stack
F
P
BD_162
A Active node
D Stacking is disabled on this node.
F Node failed to join the stack because of duplicate slot numbers.
P Node is powered off.
The image shows a graphical representation of a stack, stack topology and an active topology. In this
case, the stack consists of 8 nodes.
This section introduces the following topics:
● Ring Topology on page 600

● Daisy Chain Topology on page 601
● Stack Depth on page 602

Ring Topology
SummitStack nodes should be connected to each other in a ring topology. In a ring topology, one link is
used to connect to a node and the other link is used to connect to another node. The result forms a
physical ring connection. This topology is highly recommended for normal operation. Figure 62 shows a
maximal ring topology of eight active nodes.
Figure 62: Graphical Representation of a Ring Topology
A
A
A A
A A
BD_163
While a physical ring connection may be present, a ring active topology only exists if all nodes in the
stack are active nodes.

Figure 63: Summit X450 Series in a Ring Topology
BD_159A
Daisy Chain Topology

The stackables may be connected in a daisy-chain topology. This is a ring topology with one of the links
disconnected, inoperative, or disabled. A daisy chain can be created when a link fails or a node reboots
in a ring topology, but the daisy chain topology is not recommended for normal operation. In Figure 64,
the nodes delineated as the active topology are operating in a daisy-chain configuration, even though
there is physically a ring connection in the stack.
NOTE
The daisy chain topology is not recommended for normal operation.

Figure 64: X250 Series in Daisy-Chain Topology
BD_153A
You might need to use a daisy chain topology while adding a new node, removing a node, or while
joining two stacks.
If you are using a daisy chain topology, the possibility of a dual master situation increases. So before
you create a daisy chain topology, read “Managing a Dual Master Situation” on page 645.
NOTE
The maximum cable length supported by ExtremeXOS is 5 m.
Stack Depth
A maximum of eight (8) nodes are supported in the active topology. The slot number configuration will
assign only numbers from one (1) to eight (8).
The stack tolerates an accidental connection of up to 17 nodes. Because only eight nodes can join an
active topology, there should never be an accidental connection of two stacks resulting in more than 16
nodes. If you have more than 17 nodes in a stack topology, all nodes enter an overflow state and all
stacking links enter a link overflow state. While in an overflow state, the active topology does not

Logging into a SummitStack
function. All slots containing active nodes show a Failed state. The overflow state is maintained until
the overflow is cleared by manually disconnecting a sufficient number of nodes. After the overflow is
cleared, all nodes in the stack topology reboot.
To see all the nodes in a stack topology, use the show stacking command.
Logging into a SummitStack

You can log into any node in a SummitStack. However you can control more stack features when you
log into the master. The following guidelines describe the options available to you when you log into
different nodes:
● On master nodes, all features supported by the switch license operate correctly.
● On backup nodes, most show commands show correct data for the active stack. For example, show
vlan will show all configured VLANs.
● On all non-master nodes, most of the configuration commands are rejected. The failsafe account,
enable license, and stacking configuration commands work on any node.
● On standby nodes, most “show” commands will not show correct data for the current stack
operation. However, the show switch, show licenses, and all stacking show commands will show
correct data.
● If a node is connected to the stack and stacking is not enabled, you can still configure stacking
features on that node.
The login security that is configured on the master node applies when logging into any node in the
active topology. This includes any active node that is present in a slot. A node that is disabled for
stacking is its own master, and uses its own security configuration.
You can log in to a SummitStack node using the following methods:

● Console connection to any node
● Management connection to the master
● Management connection to a standby node
● Telnet session from any node to any other node over the stack
Logging in Through the Console Port

You can use the console port on any switch to manage the SummitStack. If you connect to the master
node, you can configure and manage the stack. If you connect a nonmaster node, you can view node
status and configure only a few options on the node to which you are connected. However, you can use
the telnet feature to connect to another node and manage that node as if you were connected to it (see
“Logging Into a Node From Another Node” on page 604).
Logging in from the Management Network

The management network is an Ethernet network to which the management port of each switch
connects. The primary management IP address is assigned to the master node. You can use a terminal
emulation program and this IP address to connect to the master for configuration and management.

The alternate management IP addresses allow you to connect to individual nodes. During normal
operation, you connect to the stack using the primary management IP address. However, if the stack is
split, you can use the alternate management IP address to connect to the other half of the stack. For
more information, see “Configuring an Alternate Management IP Address” on page 618.
After you log in to a master or standby node through the management network, you can telnet to any
other node and control that node as if you were directly connected to it. For more information, see
“Logging Into a Node From Another Node” on page 604).
Logging Into a Node From Another Node

You may log into any node in the active topology from any other node in the same active topology. If
you do not know the slot number of the node to which you want to connect, enter the show slot
command. You can telnet to any switch that appears in the show slot command display.
NOTE
If the node to which you want to connect does not appear in the show slot command display, you can connect to
the node through the its console port or management port.
You have the most control over the stack when you log in to the master. To determine which node is
the master, use the command show stacking.
To telnet to another node, enter the command:
telnet slot <slot-number>

When prompted, log in normally. The switches must be active in the stack for this command to
function.
The telnet command accepts a slot number in stacking mode. When the telnet program accepts a
connection from another node in the stack, it performs security validation. The master node validates all
login security information (except for the failsafe account), regardless of the node into which you are
attempting to establish a login. If you are not able to login using your user credentials, use the failsafe
account to login.
Managing Licenses on a SummitStack

The SummitStack feature is not licensed separately. You can use the SummitStack feature with an Edge
license.
The stack must operate on one license level because the master node runs all of the licensed software
while all other nodes only assist the master node in controlling their hardware. The backup node
additionally prepares to be the master node at failover time. Regardless of the configured license level
of the backup node, once the license level is adopted for the stack, the backup node must use this
license level if the master node fails and the backup becomes the master node.
NOTE
For successful operation, all nodes in the stack must use the same license level.

Managing Licenses on a SummitStack
Although the stack must operate at one license level, nodes with different license levels are supported
in the same stack. Nodes with higher licenses levels than other nodes can be restricted to operate at a
lower or effective license level. The rules for licensing are:
● At startup, the license level the stack uses is the effective license level of the elected master node.
● If the stack is using the Advanced Edge license and you attempt to add a node that is using an Edge
license, the node does not become operational and shows as Failed with a License Mismatch reason
when using the show slot <slot> command.
● License mismatch detection is continually checked. If nodes of different license levels are operational
in the stack and there is a failover to a backup node that has a level that is greater than that of the
failed master, the stack operating license level changes to the effective license level of the new
master. If any other node is using an effective license level that is less than that of the new master,
the node fails with a license mismatch.
The following sections describe license management in a SummitStack:
● Viewing Switch Licenses on page 605

● Enabling a Switch License on page 605
● Restricting a Switch License Level on page 606
● Upgrading Licenses on page 606
Viewing Switch Licenses

To examine the current purchased license level of a node, log into that node and use the command
show license
The procedure for upgrading the license depends on whether or not the stacking license level restriction
is already configured. To determine whether or not there is a license level restriction configured and/or
in effect, use the command
show stacking configuration
The result shows the configured and effective license restriction settings on all nodes.
Enabling a Switch License

The purchased license level of a node can be enabled only after you log in to that node. You cannot log
in to one node and enable licenses on other nodes. However, you can log in to a node, telnet to a target
node, and then enable a license on the target node (see “Logging Into a Node From Another Node” on
page 604). To enable a license on a node, use the enable license command as described in Chapter 1,
“ExtremeXOS Overview.”
NOTE
All nodes must have a purchased license level at least equal to the license level of the master node in order to
successfully join the stack.

Restricting a Switch License Level

If the nodes in a SummitStack have different license levels and you want to operate a stack at the
minimum license level, you can apply a license level restriction. The restriction is stored in the NVRAM
of each node. It forces the node to reduce its license level below its purchased level at node restart time
for the life of the restart. This reduced license level is called the effective license level and can be
displayed by entering the show license command on the node you want to evaluate.
To restrict a node to operate at a license level that is lower than the one purchased for the node, use the
command:
configure stacking {node-address <node-address> | slot <slot-number>} license-level
[core | advanced-edge | edge]
You must reboot the specified nodes for the command to take effect. The command forces the specified
node(s) to operate at the specified license level. The specified license level must be less than or equal to
the level of the switch with the lowest purchased level. To avoid stack reboots when future license level
upgrades are purchased, during initial deployment you should purchase the same license level for
every node in the stack, and the license level restriction should not be configured.
Upgrading Licenses
You can purchase license level upgrades of Advanced Edge or Core for X450 and X450a series Summit
switches. License level upgrade of Advanced Edge can be purchased for X250e and X450e series
Summit switches. If you want to set up a SummitStack to operate at the Core level, you need to
purchase and install the Core level license on all switches in the SummitStack. Since the Core level
license is not available for purchase for X250e and X450e series Summit switches, you cannot include
these switches in a SummitStack that needs to run at the Core license level.
When you are ready to upgrade switch licenses, the procedure you use depends on whether or not any
license level restrictions are configured in the stack.
Upgrading the License When the License Level Restriction is In Use

To see the license level restriction on every node in the active topology, use the command show
stacking configuration. To upgrade a license, do the following:
1 Install the required license level in each switch by logging into the switch and using the command
enable license <key>
2 Log into the master node and remove the stack license level restriction on all stack nodes using the
command:
unconfigure stacking license-level
This command removes the restriction on all nodes.
3 Reboot the stack to remove the license level restriction using the command:
reboot
The new license level is now in effect.

Configuring a New Stack
Upgrading the License When the License Level Restriction is Not In Use
You need to log in to each switch to upgrade the license of that particular switch.
1 Use the command:

show stacking
to see the node role (master, backup, or standby) of each node.
2 Install the required license level in each standby node by logging into each node and using the
command:
enable license <key>
3 In the same manner, install the required license level on the backup node.
4 In the same manner, install the required license level on the master node.
The new license level can now be used.
Configuring a New Stack

Before deploying a new stack, consider the following guidelines:
● Plan to use the stack as if it were a single multi-slot switch. You need to decide the number and type
of stackables in the stack and how the stack ports will be connected to the network.
● Connect the intended master and backup nodes directly (adjacent to each other), so that the
ExtremeXOS application synchronization traffic is localized to a single stack link.
● Physically wire the stack as a ring (see “SummitStack Topologies” on page 599), and only wire nodes
together that are intended to be active in the stack. To see the recommended procedures for installing
and wiring a stack, see the “Installing Summit Family Switches” chapter of the Extreme Networks
Consolidated ExtremeXOS Hardware Installation Guide.
● You can physically connect the stack to your networks before the nodes are configured. However,
the default configuration on a non-stacking mode switch assumes a default untagged VLAN that
contains all switch ports. When powered on, the switch acts as a layer 2 switch, possibly resulting in
network loops.
● Make sure all nodes are running the minimum ExtremeXOS release needed for stacking
(ExtremeXOS 12.0. or later). To make sure you have the right version of ExtremeXOS, restart the
node and run the command show version. If any of the nodes do not have the right version, install
ExtremeXOS 12.0 on that switch. Extreme Networks recommends that you use the same image
partition on all nodes. Once the stacking is enabled, image upgrade from the stack is possible only if
the same image is selected on all nodes.
● When you enable the stacking feature, the affected node restarts. If you intend to deploy new units
that might be part of a stack in the future, you might want to turn on stacking mode during initial
deployment to avoid a future restart. The only disadvantages of stacking mode are the loss of the
two QoS profiles QP6 and QP7 and the reservation of some of the packet buffer space for stacking
control traffic.
● You can configure the SummitStack by logging into the master or any of the other nodes. For more
information, see “Logging into a SummitStack” on page 603.
● If the stackables have different purchased license levels, you might need to configure license
restriction on some nodes before those nodes can join the stack (see “Managing Licenses on a
SummitStack” on page 604).

● Most stacking specific configurations are effective only after a restart. This includes stacking enable,
slot number, stack MAC address, master-capability, and license level restriction. But the normal
configuration commands take effect immediately and require no restart.
● A basic stack configuration can be achieved by using the procedure described in “About Easy Setup”
on page 608.
NOTE
If EAPS, Spanning Tree, or any Layer 2 redundancy protocol is not running on the network, you need to make sure
that your network connections do not form a network loop.
About Easy Setup

Using the Easy Setup, you can configure a stack without entering many of the stacking CLI commands.
The Easy Setup provides you an easy way to configure the required stacking parameters for all nodes.
The Easy Setup procedure creates a stack with a master and a backup. The remaining nodes are
configured with the master capability disabled. Extreme Networks recommends that you configure the
stacking license restriction, if needed, before invoking Easy Setup. Otherwise, an additional stack reboot
might be needed.
The configuration procedure described in the next section starts Easy Setup. You can also start Easy
Setup by entering the configure stacking easy-setup command.
Easy Setup performs the functions of the following five commands required to configure the stack:
enable stacking
configure stacking slot-number automatic
configure stacking mac-address
configure stacking redundancy minimal
reboot stack-topology
In a daisy chain topology (which is not recommended), Easy Setup designates the node at the beginning
of the chain as the master, and executes the command configure stacking redundancy none.
Configuration Procedure
To configure a new stack:
1 Physically connect the nodes using the stacking ports. Instructions for setting up the stacking
hardware are provided in the “Building a SummitStack Configuration” section of the Extreme
Networks Consolidated ExtremeXOS Hardware Installation Guide.
2 Power on the nodes.
3 Log in to any of the nodes through the console port, preferably the one you want to use as the
master. If you plan to use Easy Setup, log into the intended master node.
If the stack is a new stack, the default parameters are in effect.
4 Run show stacking command to verify the stack. The show stacking command should display all
nodes in the stack. All nodes are in a disabled state and all nodes appear as master nodes.
5 If necessary, configure a license level restriction (see “Managing Licenses on a SummitStack” on
page 604).

Converting a Standalone Node Deployment to a Stack
6 Enable stacking on all nodes. To enable stacking on all nodes, run the command enable stacking
from the master. This command presents you the option of using the Easy Setup procedure, which is
described in “About Easy Setup” on page 608. If you choose this option, skip steps 7-11.
7 Assign slot numbers to all nodes (see “Configuring Slot Numbers” on page 615).
8 Assign a MAC address to the stack (see “Assigning a MAC address for the Stack” on page 616).
9 (Optional) Configure node priorities on each slot (see “Configuring Node Priority” on page 615).
10 (Optional) Disable the master capability on selected nodes (see “Configuring Master-Capability” on
page 617).
11 Restart the stack using the command reboot stack-topology.
The configuration file is set to default values while entering the stacking mode, so all previously
entered configuration information (except for the NVRAM-based stacking parameters, selected
image, and failsafe account information) are not available.
12 Log in to the intended master node and verify the stack using show stacking, show slot, and show
stacking configuration commands. If the stack configuration is successful:
● All nodes are visible in the stack.
● All nodes move to the active state.
● Some time after the nodes become active, each node is present in the configured slot.
● After the roles are finalized, you can see one master node, one backup, and a set of standby
nodes.
13 Verify that the master node is the one you intended to be the master.
14 (Optional) Configure an alternate management IP address on each node.
15 Configure a management IP network.
16 Configure other normal parameters such as VLANs, IP subnetworks, trunk groups, and so forth.
17 Save the configuration.
Converting a Standalone Node Deployment to a Stack

This section explains how to add a node to a currently deployed standalone (non-stacking) node for
adding ports and centralizing management.
Node 1 is the currently deployed node, and Node 2 is the new node to be used to form a stack of two
nodes.
Before you begin:
● Verify the EXOS version running on both the stackables and confirm that the EXOS version is 12.0 or
later. Both the nodes must be running the same EXOS release.
● Verify the purchased license levels of both nodes using the show license command.
● (Only for nodes on which you have not yet deployed SummitStack) If you want to preserve the
EXOS configuration in use on Node 1, use the upload configuration command to retrieve the
configuration in the CLI command format. The file may be used to restore the EXOS configuration to
the stack after the stacking configuration is complete.

To add a node to currently deployed stack:
1 Connect the stacking ports of the two nodes together to form a ring topology. You can power on
Node 2 before, during, or after the connection.
2 Log into Node 1 (which will be the master node and slot 1 in the stack).
3 If necessary, configure the stacking license level restriction.
4 (Optional) Configure node priorities on each slot.
5 Enable stacking on both nodes by using the command enable stacking {node-address <node-
address>}. The choice to run Easy Setup is offered. If you choose Easy Setup, skip steps 6-9 below.
6 Assign slot numbers to the new node.
● You can specify a number for each node manually.
Or
● You can use the automatic slot number assignment.
7 Assign a MAC address to the stack.
8 (Optional) Configure stacking redundancy or master-capability as desired.
9 Restart the stack using the command reboot stack-topology.
10 After the stack reboots, log in to the console port of the master node. At this time, by default, the
user ID is admin and there is no password.
11 Configure the desired safe-default-script parameters when prompted. Configuration of failsafe
account parameters will be pushed to the nonvolatile memories of both nodes.
12 Use the show stacking and “show stacking configuration” commands to confirm that the stack
is now configured and operating as expected.
13 (Optional) Configure alternate management IP address on each node.
To restore the EXOS configuration, you must first edit the file created during configuration upload. All
port numbers in the file are simple numeric values. You must replace the port number with slot:port
format with slot number set to one (1). Once the file is ready, you can:
● Make sure the file has the extension .xsf (rename if necessary).
● Use TFTP to put the file onto the master node.
● Use the load script <script-name> command to run the commands in the file.
● Use the save configuration command.
If you intend to deploy new units that are not to be stacked should consider at that time whether or not
you want to eventually convert to stacking. If so, you should turn on stacking mode during initial
deployment. If a new node is subsequently added, there will be no need to switch the existing node to
stacking mode, and hence the configuration file will continue to be used. The only disadvantages of
deployment in stacking mode are the inability to use QoS profiles QP6 and QP7 for your traffic and the
reservation of some of the packet buffer space for stacking control traffic.

SummitStack Configuration Tasks

This section describes how to perform the following configuration tasks:
● Enabling the Stack on page 611
● Verifying the Configuration on page 611
● Configuring Slot Numbers on page 615
● Configuring Node Priority on page 615
● Assigning a MAC address for the Stack on page 616
● Configuring Master-Capability on page 617
● Configuring an Alternate Management IP Address on page 618
● Saving the Configuration on page 619
Enabling the Stack

You can enable stacking through the command line interface (CLI).
Use the following command to enable SummitStack on a node:
enable stacking {node-address <node-address>}
● If no parameters are specified, stacking is enabled or disabled on all nodes in the stack topology.
● If the node-address parameter is present, stacking is enabled on the node with the specified node-
address. This is the MAC address assigned to the stackable by the factory. The enable stacking
command takes effect only after you restart the node.
● A node that is booted with stacking enabled is said to be running in stacking mode.
Use show stacking configuration command to see the current configuration of this parameter as
well as the value currently in use.
A node that is running in stacking mode attempts to join the active topology. If successful, it then
negotiates a node role with the other nodes in the stack and becomes an operational node in the stack
according to its role. The master node's configuration is applied to the node.
To disable stacking on a member of stack, use the following command:
disable stacking {node-address <node-address>}
Rebooting the node with stacking disabled causes it to run in standalone mode.
A node that is running in standalone mode becomes its own master and processes its own
configuration.
By default, stacking is disabled on all nodes.
Verifying the Configuration

The show slot and show stacking commands contain stacking configuration information, including
the state of the slot. These commands are also helpful when debugging stacking problems.

The show slot command shows the states of the nodes as they move from the empty to operational
state. Use the show slot command and Table 82 to determine a slot state:
show slot {<slot-number>} {detail}
Slot-1 Stack.25 # show slot

Slots Type Configured State Ports
--------------------------------------------------------------------
Slot-1 X450e-24p X450e-24p Operational 26
Slot-2 X450a-24t X450a-24t Operational 26
Slot-3 X450a-24tDC X450a-24tDC Operational 26
Slot-5 X450a-24x X450a-24x Operational 26
Slot-6 X450a-24xDC X450a-24xDC Operational 26
Slot-8 X450-24t X450-24t Operational 26
Slot-1 Stack.26 #
* Slot-1 Stack.1 # show stacking

Stack Topology is a Ring
Active Topology is a Ring
Node MAC Address Slot Stack State Role Flags
------------------ ---- ----------- ------- ---
*00:04:96:26:60:DD 1 Active Master CA-
00:04:96:26:60:EE 2 Active Backup CA-
00:04:96:26:60:FF 3 Active Standby CA-
00:04:96:26:60:AA 4 Active Standby CA-
00:04:96:26:60:88 5 Active Standby CA-
00:04:96:26:60:BB 7 Active Standby CA-
00:04:96:26:60:CC 8 Active Standby CA-
* - Indicates this node
Flags: (C) Candidate for this active topology, (A) Active Node
(O) node may be in Other active topology
The asterisk (*) that precedes the node MAC address indicates the node to which you are logged in.
The node MAC address is the address that is factory assigned to the stackable.
The slot number shown is the number currently in use by the related node. Since slot number
configuration only takes effect during node initialization, a change in configured value alone does not
cause a change to the slot number in use.
If a node role has not yet been determined, the node role indicates <none>. In a ring topology, the node
on which this command is executed is always the first node displayed. In a daisy chain, the ends of the
daisy chain are the first and last nodes displayed.
Even though the stack topology could be a ring, the active topology could be a daisy chain because it
does not contain every node in the stack topology. If the node on which this command is being
executed is not active, the line
Active Topology is a ___
is replaced by the line
This node is not in an Active Topology.

NOTE
It is possible for a node to be in Stabilizing or Waiting state and still be in the active topology.
Use the following command to get a summary of the stacking configuration for all nodes in the stack:
show stacking configuration
Slot-1 Stack.33 # show stacking configuration

Stack MAC in use: 02:04:96:26:60:DD
Node Slot Alternate Alternate
MAC Address Cfg Cur Prio Mgmt IP / Mask Gateway Flags Lic
------------------ --- --- ---- ------------------ --------------- ------- ---
*00:04:96:26:60:DD 1 1 Auto 192.168.130.101/24 192.168.130.1 CcEeMm- Ee
00:04:96:26:60:EE 2 2 Auto 192.168.130.102/24 192.168.130.1 CcEeMm- Ee
00:04:96:26:60:FF 3 3 Auto 192.168.130.103/24 192.168.130.1 --EeMm- Ee
00:04:96:26:60:AA 4 4 Auto 192.168.130.104/24 192.168.130.1 --EeMm- Ee
00:04:96:26:60:88 5 5 Auto 192.168.130.105/24 192.168.130.1 --EeMm- Ee
00:04:96:26:60:99 6 6 Auto 192.168.130.106/24 192.168.130.1 --EeMm- Ee
00:04:96:26:60:BB 7 7 Auto 192.168.130.107/24 192.168.130.1 --EeMm- Ee
00:04:96:26:60:CC 8 8 Auto 192.168.130.108/24 192.168.130.1 --EeMm- Ee
Flags: (C) master-Capable in use, (c) master-capable is configured,
(E) Stacking is currently Enabled, (e) Stacking is configured Enabled,
(M) Stack MAC in use, (m) Stack MACs configured and in use are the same,
(i) Stack MACs configured and in use are not the same or unknown,
(-) Not in use or not configured
License level restrictions: (C) Core, (A) Advanced edge, or (E) Edge in use,
(c) Core, (a) Advanced edge, or (e) Edge configured,
Use the following command to get a full report from the stacking database:
show stacking {node-address <node-address> | slot <slot-number>} detail
Slot-1 Stack.33 # show stacking slot 1 detail

Stacking Node Slot 1 information:
Current:
Stacking : Enabled
Role : Master
Priority : Auto
Slot number : 1
Stack state : Active
Master capable : Yes
License level restriction : Edge
In active topology? : Yes
Factory MAC address : 00:04:96:26:60:DD
Stack MAC address : 00:04:96:26:60:DD
Alternate IP address : 192.168.130.101/24
Alternate gateway : 192.168.130.1
Stack port 1:
State : Operational
Blocked? : No
Control path active? : Yes
Stack port 2:

State : Operational
Blocked? : No
Control path active? : Yes
Configured:
Stacking : Enabled
Master capable : Yes
Slot number : 1
Stack MAC address : 00:04:96:26:60:DD
License level restriction : Edge
Slot-1 Stack.34 #
If you do not specify any node, the output is generated for all nodes in the stack topology. If the
specified node does not exist, an error message appears.
The slot parameter is available only in stacking mode. The node-address parameter is always available.
Current information represents stacking states and configured values that are currently in effect.
Configured information is that which takes effect at node reboot only.
The roles values are Master, Backup, Standby, and <none>. License level restrictions are Edge,
Advanced Edge, or Core.
To verify the port states of each node in the stack topology use the command show stacking stack-
ports.
Setting the Command Prompt

When stacking is enabled, the nodes inherit the SNMP sysname from the master.
The config snmp sysName <name> command affects the command prompt. The default setting on this
command assigns the model name to the command prompt. When stacking is enabled, the current slot
number is appended to the string, and the sysname is defaulted to Stack.
The command prompt looks similar to:

* Slot-6 Stack.21 #
The * indicates a changed and unsaved ExtremeXOS configuration. Slot-6 indicates that the node is in
stacking mode and is currently using slot number 6 in the active topology. The system name is the
default Stack. The command to be executed is the 21st command entered since login, and you have
logged in as the administrator on the master node (#).
The backup and the standby nodes show > instead of #. For example:
* Slot-6 Stack.23 >
If you have configured a sysName for the stack, each node in the active topology displays sysName in
its command prompt.
There is no specific prompt to indicate the node role. To discover the identities of the master and
backup nodes, use the show switch or show stacking command.
Use the show slot command to verify the local switch type.

Configuring Slot Numbers

Each node in a stack must be assigned a slot number. The slot number must be unique to each node.
You can assign the slot number only through configuration. The stack does not dynamically assign a
slot number. The available slot numbers are 1 through 8.
You can specify a slot number for each node manually, or you can have the system assign the slot
numbers using a single command.
NOTE
Slot numbers take effect only during a restart. If you change a slot number, the unit continues to
operate with the slot number with which it was last restarted.
To manually add a slot number to a node, use the command configure stacking node-address
<node-address> slot-number <slot-number>.
To configure the system to choose slot numbers for all nodes, enter the command configure stacking
slot-number automatic.
Automatic slot number assignment is performed in the order of appearance of the nodes in the show
stacking display. In the case of a ring topology, the first node in the display is the intended master
node into which you have logged in.
Use the show stacking or show stacking configuration command to view the ordering and the
assigned slot numbers.
NOTE
A node that boots in standalone mode does not use a slot number.
Configuring Node Priority

The node priority configuration influences the node role election priority. The node with the highest
node role election priority becomes the master as a result of the first node role election, and the node
with the second highest node role election priority becomes the backup node. All other nodes become
standby nodes.
During subsequent node role elections that occur when a master node fails, the node priority
configuration helps determine the node that becomes the replacement backup node.
Node priority configuration takes effect at the next node role election. A change in node priority
configuration does not cause a new election. Once an active topology has elected a master node, that
node retains the master node role until it fails or loses a dual master resolution.
You can configure one of the following election priority algorithms:
● Priority algorithm - If any node has a numeric priority value configured.
● Automatic algorithm - If all nodes participating in a role election have the automatic priority value
configured.
The priority algorithm is selected if any node has a numeric priority value configured. You can specify
an integer priority value between 1 and 100. The higher the value, the greater the node role election
priority. If any node participating in a role election has a priority value configured, all nodes use the

priority algorithm. A node configured with the automatic algorithm uses a priority value of zero (the
lowest priority) in the priority algorithm if another node has a priority value configured.
The automatic algorithm is selected if no node participating in a role election has a numeric priority
value configured. In automatic mode, the stack determines the highest role election priority based on
factors such as available processing power, maintenance level of ExtremeXOS, and so forth.
In both algorithms, if the highest computed node role election priority is shared among multiple nodes,
the slot number is used to adjust the node role election priority. A numerically lower slot number
results in a higher role election priority than a numerically higher slot number. If you wish to use the
slot number as the sole determining factor in node role election priority calculation, you should
configure every node with the same priority value, and not automatic.
NOTE
Extreme Networks may change the behavior of the automatic priority algorithm in future ExtremeXOS releases.
Nodes that are configured as not master-capable do not participate in the role election. Priority
configuration is not relevant on such nodes.
A dual master resolution does not use the configured node priority in most cases. Instead it uses the
oldest time that a node became a master in the current active topology.
Use the following command to set the stacking node priority:
configure stacking {node-address <node-address> | slot <slot-number>} priority [<node-

pri> | automatic]
Assigning a MAC address for the Stack

The stack must use a single MAC address. When the master node fails over to the backup node, the
backup node must continue to use the same MAC address that the master node was using.
Each stackable is assigned a single unique MAC address during production. By default, no stack MAC
address is configured. You can choose any node to supply its factory assigned MAC address to form the
stack MAC address.
For the stack, one of the stackables is designated as the node whose factory assigned MAC address is
used to form the stack MAC address. Once this is done, all nodes receive and store this formed MAC
address in their own NVRAM. Whenever the stack boots up, this MAC address is used, regardless of
which node is the master node.
Before being stored as the stack MAC address, the chosen node’s factory assigned MAC address is
converted to a locally administered MAC address. This prevents duplicate MAC address problems that
would result in a dual master situation if the MAC address was the exact address assigned to the
selected node. The chosen MAC address is put into effect only at node boot time. If the address needs
to be changed on a single node, rebooting that node results in usage of the same address stack-wide.
If you do not configure the stack MAC address or it is not the same on all nodes, a warning message
appears in the log.
Each node operates with whatever address is available (the configured stack MAC address or the
node’s factory assigned MAC address). If a master node fails over to the backup node, and the backup
node’s address is different than the one the former master node was using, the address is inconsistent
with the addresses programmed into the packet forwarding hardware. The MAC address related to the
management IP address changes to the one in use by the new master, but no gratuitous ARP requests

are sent. In this case, it takes some time for hosts on the management network to flush the related ARP
entry.
The stack MAC address being different on standby nodes does not have an effect since the standby
nodes do not use the stack MAC address. Backup nodes use the stack MAC address only when
becoming master nodes.
NOTE
If the node whose MAC address is chosen is removed from the stack with the intention of using the node elsewhere
in the network, and that node is selected to supply the stack MAC in its new stack, the stack MAC of the original
stack must be reconfigured or there will be a duplicate MAC address in the network.
If new nodes are added to the stack, the new nodes must be configured with the MAC address.
Use the following command to configure the stack mac-address:
configure stacking {node-address <node-address> | slot <slot-number>} mac-address
Configuring Master-Capability
Each node is configurable to be master-capable. This means that a node can be allowed to take on any
node role, or be restricted to executing the standby node role only. The default is that a node can take
on any role. The restriction is used to avoid the dual master condition. A master-capability
configuration change takes effect at the next restart.
You can use any of the following commands to configure the master-capability:
● configure stacking [node-address <node-address> | slot <slot-number>] master-
capability [on | off]
● configure stacking redundancy [none | minimal | maximal]
Using these commands, you can configure one or more nodes to be allowed to operate either as a
master or a backup.
The configure stacking master-capability command allows you to set the master-capability of
specific nodes, while configure stacking redundancy allows you to set the master-capability on all
nodes in the stack.
The commands do not allow you to disable master-capability on all nodes in a stack topology.
NOTE
If the entire stack is restarted in stacking mode without any node having master capability, you need to know the
failsafe account and password to log into any node in the stack.
If you do not know the failsafe account information, you might need to rescue the stack. See “Rescuing a
Stack That Has No Master-Capable Node” on page 648.

Configuring an Alternate Management IP Address

The stack has a primary IP address and subnetwork that is configured with the configure Mgmt
ipaddress command, and there may also be static or default routes associated to it. For each node in
the stack, you can configure an alternate management IP address, subnetwork mask, and default
gateway. The address is restricted to being a member of the primary IP subnetwork that is configured
on the management (Mgmt) VLAN, and thus the alternate IP subnetwork must exactly match the
primary IP Mgmt subnetwork. A subnetwork match is exact if the subnetwork portion of the IP
addresses match exactly. For example, 10.11.12.1/24 and 10.11.12.2/24 are an exact subnetwork match
(i.e., both represent the subnet 10.11.12.0/24).
Standby nodes always install their configured alternate Management IP address and gateway on the
management interface. A standby node does not have the ability to verify whether the configured
alternate IP address matches the primary management IP subnetwork of the stack.
The backup and master nodes have the ability to verify the configured alternate IP address. The master
and backup nodes compare the primary IP subnetwork information to the alternate IP subnetwork. If
there is a match, the backup node installs the Primary IP Mgmt subnetwork’s default routes and installs
only the alternate management IP address (not the primary IP address). The master node installs both
the configured Mgmt subnetwork with specific IP address and the alternate IP address and not the
primary IP address. In this case, the alternate gateway is not used, expecting that primary Mgmt routes
are configured or will be configured. In either case, if the alternate IP subnetwork does not match the
configured Mgmt subnetwork, the alternate IP address is not installed on the management interface.
Each node in the stack normally installs its alternate IP address on the Mgmt subnetwork. When an
ARP request for the alternate IP address is satisfied, the stackable supplies its factory assigned MAC
address and not the MAC address chosen to represent the stack. Only the master node installs the
primary IP address. An ARP request for the configured Mgmt IP address returns the configured
stacking MAC address. Because of the above behavior, all nodes are reachable over their management
ports even during a dual master situation. The VLAN used is the “Mgmt” VLAN (VID 4095) and is
untagged.
The alternate gateway is only installed on a master or backup node when the primary management IP
subnetwork is not configured. Once the primary IP subnetwork is installed, the alternate gateway is
removed. The alternate gateway is always installed on a standby node.
If a dual master situation occurs because of a stack severance, the alternate IP addresses and associated
MAC addresses are unique, and it is possible to use telnet or ssh to reach any node. Any node on the
segment with the incorrect master can then be used to reboot the entire stack segment into standby
mode if you want to rejoin the stack segments later.
If a node is operating in stacking mode, the alternate management IP address configuration takes effect
immediately.
To configure an alternate management IP address for the stack, use the command:
configure stacking [node-address <node-address> | slot <slot-number>] alternate-ip-
address [<ipaddress> <netmask> | <ipNetmask>] <gateway>
You can also specify a block of IP addresses to be assigned to the stack topology. To assign a block of IP
addresses to the stack topology, use the command:
configure stacking alternate-ip-address [<ipaddress> <netmask> | <ipNetmask>]
<gateway> automatic
NOTE
Only IPv4 alternate management IP addresses are supported in this release.

Managing an Operating SummitStack
Configuring the Failsafe Account on a Stack

The failsafe account information is stored in each node's local NVRAM.
To change the failsafe account, use the command configure failsafe-account from the master node.
This command changes the account information in the NVRAM of every active node in the same active
topology. If a new node is added later, you can use the synchronize stacking command to copy the
failsafe account information from the master to the new node (see “Synchronize Command Operation”
on page 624).
Saving the Configuration

The ExtremeXOS configuration file is saved to every active node when you use the save
configuration command on the master.
The stacking specific configuration parameters for a node are saved in the NVRAM of the node when
you run the configuration commands. Stacking configuration parameters are not saved in the
ExtremeXOS configuration file.

This section describes the following topics and tasks:
● Stacking LEDs on page 619
● Configuring and Viewing Stacking Port Status on page 620
● Adding a Node to an Active Topology on page 620
● Replacing a Node with the Same Switch Type on page 621
● Replacing a Node with a Different Switch Type on page 622
● Merging Two Stacks on page 623
● Synchronize Command Operation on page 624
● Upgrading ExtremeXOS on a Stack on page 624
● Rebooting a Stack on page 626
Stacking LEDs
All stackables have a seven segment LED. When the stackable is not in stacking mode, this LED is dark.
When in stacking mode, the LED displays the slot number currently being used by the node. The
number is displayed shortly after the node begins initializing and remains in the display until a restart
occurs.

The stacking ports have LEDs that behave the same as the data port LEDs. The stacking port LEDs can
be in the following states, even if the unit is not in a stacking mode.
Table 82: Stacking LED State

State Description
Off No signal
Solid Green Signal present
Flickering Green Traffic through the port
While in a stack, the remaining LEDs (Mgmt, Fan, PSU-I, and PSU-E) on the unit operate normally.
Configuring and Viewing Stacking Port Status

To view the status of any stacking port, use the show ports stack-ports <stacking-port-list>
command variations.
The variations include stack port:
● Statistics
● Utilization
● Transmit errors
● Receive errors
The commands accept stacking port ranges that span multiple nodes. There are no stack port
configuration options.
There is no way to disable a stacking port. These ports are always enabled.
Adding a Node to an Active Topology

From the perspective of a new node, adding a node to an active topology is similar to bringing up a
new stack. To add a node to a stack, use the following procedure.
NOTE
If the node being added is actually a replacement node for one that was previously removed, see
“Replacing a Node with the Same Switch Type” on page 621 or “Replacing a Node with a Different
Switch Type” on page 622.

1 Connect the stacking links to the new node.

If the node has already been powered on, it is added into the stack as follows:
● If stacking is not enabled on the node, the node does not come up in the stack. It is, however,
visible using the show stacking command on the master node. Its stacking parameters are
configurable and the node is rebootable from the stack.
● If stacking was already enabled on the node, and if the node’s slot number duplicates that of
another node in the stack, the added link status is Inhibited. The node is visible from the master
node using the show stacking command.
● If the node is already enabled for stacking, its slot number is unique, and the node is already a
master node, the result is the same as described in “Merging Two Stacks” on page 623.
● If the node was already enabled for stacking, and its slot number is unique, and the node has
either not yet determined its role or become a standby because it is not configured to be master
capable, the node becomes operational.
If the node is powered on after being connected to the stack, the node responds as follows:
● If stacking is not enabled on the node, the node does not come up in the stack. It is, however,
visible using the show stacking command on the master node. You can configure the stacking
parameters and you can restart the node from the stack.
● If stacking is enabled on the node, and if the node’s slot number duplicates that of another node
in the stack, the node enters the Failed state. The node is visible from the master node using the
show stacking command.
● If stacking is enabled on the node, and its slot number is unique, the node joins the stack. The
node becomes a standby node or a backup node. If there is no backup node already and the node
is master capable, the node becomes a backup.
2 Run the synchronize stacking {node-address <node-address> | slot <slot-number>}
command for the new node.
3 Make sure the node has a unique slot number.
4 Make sure the node's master-capability is properly set.
5 Set a license level restriction, if required.
6 Configure a priority if needed.
7 Run the show stacking configuration command to see the configuration of all nodes.
If node parameters require changing, the changes can be issued from the master node. This is done by
using the individual stacking parameter configuration commands. The synchronize stacking {node-
address <node-address> | slot <slot-number>} command can be helpful because it copies some
of the parameters needed for stacking operation. The node can then be rebooted from the master node
without rebooting the entire stack.
Replacing a Node with the Same Switch Type

When you replace a node with the same switch type, for example when you replace a a Summit X450a-
48t with a Summit X450a-48t, you can continue to use the same stack configuration. The procedure in
this section works only when the old and new nodes have identical switch types.
NOTE
If you are replacing a node with a different switch type, you must change the stack configuration before the new
node can operate. For more information, see “Replacing a Node with a Different Switch Type” on page 622.

To replace a node with an identical switch type:
1 Note the slot number of the node you are replacing.

2 Remove the stack links from the node.
3 Replace the node with the same type of node.
4 Connect the stack links and power on the node. The switch joins the stack topology but might not
join the active topology.
5 Enable stacking on the node, if needed. (See “Enabling the Stack” on page 611.)
6 Configure the slot number for the new node using the slot number noted in Step 1. (See
“Configuring Slot Numbers” on page 615.)
7 Configure the stack MAC address for the stack. You can select any node to provide its factory
assigned MAC address to form the stack MAC address. The command always configures the
address on every node in the stack topology. (See “Assigning a MAC address for the Stack” on
page 616.)
8 Verify that the node license is at the same level or higher than the stack license level. Upgrade the
node’s license level, if needed, or configure a stacking license level restriction on the node if the
node's default or purchased license level is greater than the stack license level. For more information,
see “Managing Licenses on a SummitStack” on page 604.
9 If the master node was replaced, reboot the stack. Otherwise, reboot only the replacement node.
The stack or node restarts. If you replaced the master, the replacement node starts up as the master.
NOTE
If the new node does not become operational, enter the show slot {<slot-number>} {detail}
command. If the slot shows a mismatch, the node was replaced with a different type of switch. (See
“Replacing a Node with a Different Switch Type” on page 622.)
Replacing a Node with a Different Switch Type

When you replace a node with the different switch type, for example when you replace a a Summit
X450a-48t with a Summit X450e-48p, you cannot continue to use the same stack configuration. The slot
configuration for the replaced node must change to reflect the new switch type.
NOTE
If you are replacing a node with the same switch type, you can continue to use the existing stack configuration. For
more information, see “Replacing a Node with the Same Switch Type” on page 621.
To replace a node with a different switch type:
1 Enter the unconfigure slot command to remove the configuration for the node to be replaced.
All configuration parameters (except for the related node's NVRAM-based configurations such as
stacking parameters, image to be used, and failsafe account) for the slot are erased.
2 Remove the stack links from the node.
3 Replace the node.
4 Connect the stack links and power on the node. The switch joins the stack topology but may not be
in the active topology.

5 Enable stacking on the node, if needed. (See “Enabling the Stack” on page 611.)
6 Configure the slot number for the new node using the slot number noted in Step 1. (See
“Configuring Slot Numbers” on page 615.)
7 Configure the stack MAC address for the stack. You can select any node to provide its factory
assigned MAC address to form the stack MAC address. The command always configures the
address on every node in the stack topology. (See “Assigning a MAC address for the Stack” on
page 616.)
8 Verify that the node license is at the same level or higher than the stack license level. Upgrade the
node’s license level, if needed, or configure a stacking license level restriction on the node if the
node's default or purchased license level is greater than the stack license level. For more information,
see “Managing Licenses on a SummitStack” on page 604.
9 If the master node was replaced, reboot the stack. Otherwise, reboot only the replacement node.
The stack or node restarts. If you replaced the master, the replacement node starts up as the master.
Merging Two Stacks

You can join or merge two stacks to create one large stack. However, the maximum number of nodes in
an active topology is still eight.
The operation performed when two stack segments are joined together depends on the following
factors:
● Whether a slot number is duplicated

● Whether both stacks have master nodes
● The states of the nodes in each stack.
If the nodes are configured with stacking enabled, one of the following occurs:
● If two segments are joined, both have master operational nodes, and at least one of the nodes in one
of the stacks duplicates a slot number of a node in the other stack, the join is allowed. The link that
has just connected the two stacks shows as Inhibited. This prevents accidental stack joins. Once in this
condition, the nodes on the joined segment can still be reconfigured for stacking.
● If two segments are joined, both have master operational nodes, and all nodes have assigned slot
numbers that are unique in both stacks, the dual master situation is automatically resolved.
● If two segments are joined, there are no duplicate slot numbers, one of the segments has a master
and a backup node, and the other segment does not have either a master or a backup node, the
nodes in this segment are acquired by the master node. These nodes become standby nodes in the
stack.
The nodes that are not configured for stacking, do not attempt to join the active topology but
nevertheless join the stack.
Any nodes enabled for stacking that are isolated between nodes that are not enabled for stacking
attempt to form an isolated active topology.
If one of the nodes that is not configured for stacking is then configured for stacking and restarted, the
behavior is as if two active stacks were joined. To see an example of how two stacks can be merged, see
“Joining Two Stacks” on page 630.

Synchronize Command Operation

In SummitStack, synchronize means copying the content of the internal compact flash card, which
contains the software images and all configuration files, and the NVRAM contents from one node to
another node.
During synchronization, you cannot use the node that is being synchronized. After synchronization, the
node restarts.
To specify the target node for synchronization enter the synchronize stacking {slot <slot-
number>} command with the target slot number.
If no parameters are provided, the backup node is synchronized with the master node. You cannot
synchronize the entire stack simultaneously.
The target node is not usable during synchronization and reboots before and after synchronization is
complete.
Even though the synchronize command copies the contents of NVRAM from the master node to the
synchronized node, the following stacking-specific parameters are preserved:
● slot number
● master-capable configuration
● alternate IP address, subnetwork mask, and default gateway
● priority
● license restriction
After the synchronization is complete, the stack identity of the node is unchanged and hence the node
can immediately rejoin the stack.
In SummitStack, synchronize stacking means copying certain NVRAM-based parameters from the local
node to the specified target node or all nodes. The parameters are:
● stacking enabled
● stack MAC address
● boot image selection
● failsafe account and password
● failsafe access point permissions
The command is useful when adding or replacing a node in the stack. Since only these few parameters
are copied, unlike the synchronize command, there is no reboot and the command completes quickly.
However, stacking enable, stack MAC address, and boot image selection take effect only at the next
reboot.
To perform this function, use the command synchronize stacking {node-address <node-address>
| slot <slot-number>}.
Upgrading ExtremeXOS on a Stack
Upgrading the Software on All Active Nodes

You can centrally upgrade the software on all active nodes in a stack. To upgrade all nodes in the stack,
all nodes must be running an ExtremeXOS release that supports stacking (ExtremeXOS release 12.0 or
greater).

Use the command download image to download a new ExtremeXOS software release and install it on
all nodes on the active topology. If necessary, use the use image command to select the image partition
(primary or secondary) into which the software was saved. Use the reboot command to restart all
nodes in the new release:
download image [<hostname> | <ipaddress>] <filename> {vr <vr-name>} {primary |

secondary}
use image {primary | secondary}
reboot
Before you upgrade a SummitStack, make sure that the active image partition is same across all nodes.
You can install the image only on the alternate image partition and not on the active image partition. To
run the upgraded software, you must reboot the stack after installation with the image partition that
received the software being selected.
Use the command use image to specify the same image on all active nodes. You can run the
synchronize stacking command to copy the image selection on the local node to all nodes in the
stack topology, but you can only download and install images to active nodes. The image specified
must not be the one in use as the active image partition.
NOTE
Hitless upgrade is not supported in SummitStack.
Upgrading the Software on a Single Node

You can upgrade the software on a single active node. Enter the following commands to download an
image to a node:
download image [<hostname> | <ipaddress>] <filename> {vr <vr-name>} {primary |
secondary} slot <slot-number>
use image {primary | secondary} slot <slot-number>
reboot slot <slot number>
The slot number is the one in use by the active node that is to be upgraded.
Be sure that you keep the same image versions on both the master and on the nodes.
Upgrading the Bootrom

The SummitStack feature does not require a bootrom upgrade. You should not upgrade the bootrom of
any node unless there are other reasons to do so. However, SummitStack does allow centralized
bootrom upgrade.
You can download and install the bootrom to a specific slot using the slot parameter. The slot
parameter is available only on stackables.
If you do not provide a slot number, the stack attempts to download the bootrom image and install it
on all stackables.

Rebooting a Stack
You can reboot a stack by entering the command reboot from the master.
You can:
● Reboot all the nodes in the stack topology

● A specific node
● Reboot all nodes in the active topology
● Move a node to a standby node
● Reboot the stack topology so that every node comes up in standby role
To reboot all the nodes in an active topology or stack topology, enter
reboot stack-topology
To reboot a specific node, enter:
reboot node-address <node-address>
Configuration Examples
This section provides you configuration examples for the following situations:
● Building and Deploying a New Stack on page 626
● Joining Two Stacks on page 630
● Adding a Node to a Stack on page 636
● Configuring an Alternate IP Address on page 638
● Configuring a MAC address on page 639
● Configuring the License Level Restriction on page 642
Building and Deploying a New Stack

This section provides an example of deploying a new stack with 8 nodes.
The new stack contains 8 nodes, node 1 through node 8. Node 1 will be assigned to slot 1 and will be
the master. Node 2 will assigned to slot 2 and will take the role of the backup node. Nodes 3 - 8 will be
assigned slots 3 to 8 respectively and will be standby nodes.
Before you begin the configuration, log in to the nodes and get the information if you do not already
have it:
● The software release installed on the node, using the command show version
● The image selected (all nodes need to be operating from the same selected image)
● The purchased license information, using the command show license
NOTE
Be sure to install ExtremeXOS 12.0 or later on the same image partition and select and boot into this partition on
all nodes. By default, new nodes have the primary image selected.

For this example, assume that all nodes except node 8 have a purchased Advanced Edge license level,
and that node 8 has a purchased license level of Edge.
To Deploy the stack:
1 Power up all nodes, if you have not already done so.

2 Log in to Node 1. The safe-default-script may be displayed at this time. If so, for now, accept the
default answer to each question.
3 Run show stacking command.
* X450a-24x.1 # show stack
This node is not in an Active Topology
------------------ ---- ----------- ------- ---
*00:04:96:26:6c:df - Disabled Master ---
00:04:96:26:6c:e3 - Disabled Master ---
00:04:96:26:6b:e4 - Disabled Master ---
00:04:96:26:6b:f7 - Disabled Master ---
00:04:96:26:6b:ed - Disabled Master ---
00:04:96:26:6b:ec - Disabled Master ---
00:04:96:26:6d:1f - Disabled Master ---
00:04:96:26:6a:e9 - Disabled Master ---
* X450a-24x.2 #
The stack topology is a ring and all the nodes are present in the stack. Node 1 is at the top and Node 8
at the bottom. The asterisk (*) before a node in the above display, indicates the node to which you have
logged in.
4 Display a summary of the configurations of all nodes in the stack using the command show
stacking configuration:
* X450a-24x.2 # show stacking configuration
Stack MAC in use: <none>
------------------ --- --- ---- ------------------ --------------- ------- ---
*00:04:96:26:6c:df 1 - Auto <none> <none> -c----- --
00:04:96:26:6c:e3 1 - Auto <none> <none> -c----- --
00:04:96:26:6b:e4 1 - Auto <none> <none> -c----- --
00:04:96:26:6b:f7 1 - Auto <none> <none> -c----- --
00:04:96:26:6b:ed 1 - Auto <none> <none> -c----- --
00:04:96:26:6b:ec 1 - Auto <none> <none> -c----- --
00:04:96:26:6d:1f 1 - Auto <none> <none> -c----- --
00:04:96:26:6a:e9 1 - Auto <none> <none> -c----- --
License Level Restrictions: (C) Core, (A) Advanced edge, or (E) Edge in use,


* X450a-24x.3 #
Since this example uses new nodes, the factory defaults are displayed
5 Configure a license restriction of Edge so that node 8 will be able to come up in the stack.
* X450a-24x.7 # configure stacking license-level edge
This command will take effect at the next reboot of the specified node(s).
Note that it is preferable to upgrade the license of node 8 instead of restricting the license level of
the entire stack as is shown here.
6 From the master, use the Easy Setup option to enable stacking on all nodes.
* X450a-24x.3 # enable stacking
You have not yet configured all required stacking parameters.
Would you like to perform an easy setup for stacking operation? (y/N) Yes
Executing "configure stacking easy-setup" command...

For every node in the 8-node stack, this command will:
- enable stacking
- configure a stack MAC address
- choose and configure a slot number (this node will be assigned to slot 1)
- configure redundancy to minimal (slot 1 will be the master node)
Upon completion, the stack will automatically be rebooted into the new configuration.
Warning: If stacking is already configured, this command will alter that
configuration.
Do you wish to proceed? (y/N) Yes
Stacking configuration is complete. Rebooting...
* X450a-24x.4 #
After a time the following message appears:
Authentication Service (AAA) on the master node is now available for login.
7 Log in to Node 1.
The safe-default-script will now run.
Select the values for normal operation. You may configure the failsafe account now. The failsafe
account user id, password, and other related values are saved in non-volatile storage in all active
nodes.
8 Run show stacking to verify the configuration.
* Slot-1 Stack.1 # show stacking
------------------ ---- ----------- ------- ---
*00:04:96:26:6c:df 1 Active Master CA-
00:04:96:26:6c:e3 2 Active Backup CA-
00:04:96:26:6b:e4 3 Active Standby CA-
00:04:96:26:6b:f7 4 Active Standby CA-
00:04:96:26:6b:ed 5 Active Standby CA-
00:04:96:26:6b:ec 6 Active Standby CA-
00:04:96:26:6d:1f 7 Active Standby CA-

00:04:96:26:6a:e9 8 Active Standby CA-

* Slot-1 Stack.2 # show stacking configuration
Stack MAC in use: 02:04:96:26:6c:df
------------------ --- --- ---- ------------------ --------------- ------- ---
*00:04:96:26:6c:df 1 1 Auto <none> <none> CcEeMm- Ee
00:04:96:26:6c:e3 2 2 Auto <none> <none> CcEeMm- Ee
00:04:96:26:6b:e4 3 3 Auto <none> <none> --EeMm- Ee
00:04:96:26:6b:f7 4 4 Auto <none> <none> --EeMm- Ee
00:04:96:26:6b:ed 5 5 Auto <none> <none> --EeMm- Ee
00:04:96:26:6b:ec 6 6 Auto <none> <none> --EeMm- Ee
00:04:96:26:6d:1f 7 7 Auto <none> <none> --EeMm- Ee
00:04:96:26:6a:e9 8 8 Auto <none> <none> --EeMm- Ee
* Slot-1 Stack.3 #
The user prompt contains the slot number on which the console session is running. Also notice that the
platform has changed from X450a-24x to Stack. The nodes in the stack have become Active and have
been assigned node roles. The configured slot numbers have become current, and the other stacking
parameters have also taken effect.
To see the ExtremeXOS state of each node, run the command show slot on the master:
* Slot-1 Stack.3 # show slot

--------------------------------------------------------------------
Slot-1 X450a-24x Operational 26
Slot-2 X450a-24xdc Operational 26
Slot-3 X450a-24tdc Operational 26
Slot-7 X450a-24xdc Operational 26
Slot-8 X450e-48p Operational 50
* Slot-1 Stack.4 #
9 Configure a block of IP addresses and a gateway for the alternate management IP functionality.
* X450a-24x.8 # config stacking alternate-ip-address 10.66.13.200/24 10.66.13.1
automatic
Choose the block as a subset of addresses in the intended primary management subnet that will be
configured later. Arrange the stack so that the alternate IP addresses assigned to each node are easily
calculated so you can easily find the address to use in order to log into the severed stack.

There are two methods you can follow:

Choose the stack IP address, and then allocate a consecutive block of addresses that immediately
follow the stack IP address. For example, if the Mgmt VLAN is configured with the address
10.4.73.8, and if there are three master-capable nodes in the stack, then their alternate IP addresses
could be 10.4.73.9, 10.4.73.10, and 10.4.73.11.
Or
Use configured Mgmt VLAN address and the slot number to form the alternate IP address. For
example, if 10.4.73.10 is the Mgmt VLAN address, and we are configuring the alternate IP address
for slot 1, the alternate IP address could be 10.4.73.11 and for slot 8 it could be 10.4.73.18.
10 Configure the management IP address, subnetwork, and gateway (VLAN Mgmt). You can configure
non stacking parameters such as VLANs, load aggregation, spanning tree, and routing protocols
now.
11 Save the configuration.
Joining Two Stacks

This example assumes two stacks named StackA and StackB. The joined stack will assume the name
StackA. Here are displays taken from the original StackA:
Slot-1 StackA.8 # show stacking

------------------ ---- ----------- ------- ---
(*) Indicates This Node
Flags: (C) Candidate for this active topology, (A) Active node,
Slot-1 StackA.9 #
Slot-1 StackA.9 # show stacking configuration

MAC Address Cfg Curr Prio Mgmt IP / Mask Gateway Flags Lic
------------------ --- ---- ---- ------------------ --------------- ------- ---
*00:04:96:26:60:DD 1 1 Auto 192.168.130.101/24 192.168.130.1 CcEeMm- Ee
00:04:96:26:60:EE 2 2 Auto 192.168.130.102/24 192.168.130.1 CcEeMm- Ee
00:04:96:26:60:FF 3 3 Auto 192.168.130.103/24 192.168.130.1 –-EeMm- Ee
? - Cached information, node is not responding
License Level Restrictions : (C) Core, (A) Advanced edge, or (E) Edge in use
(c) Core, (a) Advanced edge, or (e) Edge configured
Slot-1 StackA.10 #

Slot-1 StackA.10 # show stacking stack-ports

Slot Port Node MAC Address Port State Flags
---- ---- ----------------- ----------- -----
*1 1 00:04:96:26:60:DD Operational CB
*1 2 00:04:96:26:60:DD Operational C-
2 1 00:04:96:26:60:EE Operational C-
2 2 00:04:96:26:60:EE Operational C-
3 1 00:04:96:26:60:FF Operational C-
3 2 00:04:96:26:60:FF Operational CB
Flags: (C) Control path is active, (B) Port is Blocked
Slot-1 StackA.11 #
Slot-1 StackA.3 # show slot

--------------------------------------------------------------------
Slot-4 Empty 0
Slot-5 Empty 0
Slot-6 Empty 0
Slot-7 Empty 0
Slot-8 Empty 0
Slot-1 StackA.4 #
Here are displays taken from StackB:
Slot-1 StackB.3 # show stacking

------------------ ---- ----------- ------- ---
00:04:96:26:60:AA 1 Active Master CA-
00:04:96:26:60:88 2 Active Backup CA-
Slot-1 StackB.4 #
Slot-1 StackB.4 # show stacking configuration

Stack MAC in use: 02:04:96:26:60:AA
------------------ --- ---- ---- ------------------ --------------- ------- ---
*00:04:96:26:60:AA 1 1 Auto 192.168.131.101/24 192.168.131.1 CcEeMm- Ee
00:04:96:26:60:88 2 2 Auto 192.168.131.102/24 192.168.131.1 CcEeMm- Ee
00:04:96:26:60:99 3 3 Auto 192.168.131.103/24 192.168.131.1 –-EeMm- Ee

Slot-1 StackB.5 #
Slot-1 StackB.5 # show stacking stack-ports

---- ---- ----------------- ----------- -----
1 1 00:04:96:26:6b:AA Operational C-
1 2 00:04:96:26:6b:AA Operational CB
2 1 00:04:96:26:60:88 Operational CB
2 2 00:04:96:26:60:88 Operational C-
3 1 00:04:96:26:6b:99 Operational C-
3 2 00:04:96:26:6b:99 Operational C-
Slot-1 StackB.6 #
Slot-1 StackB.6 # show slot

--------------------------------------------------------------------
Slot-3 X450a-24xDC X450a-24xDC Operational 26
Slot-4 Empty 0
Slot-5 Empty 0
Slot-6 Empty 0
Slot-7 Empty 0
Slot-8 Empty 0
Slot-1 StackB.7 #
Form the new stack. Assuming both stacks are rings, break one link in each stack as follows:
● For StackA, break the link between node 00:04:96:26:60:FF port 2 and node 00:04:96:26:60:DD port 1.
● For StackB, break the link between node 00:04:96:26:6b:99 port 2 and node 00:04:96:26:6b:AA port 1.
Then connect the broken links between the two stacks to form a ring as follows:
● Connect node 00:04:96:26:60:FF port 2 to node 00:04:96:26:6b:AA port 1.

● Connect node 00:04:96:26:6b:99 port 2 to node 00:04:96:26:60:DD port 1.
Since both are active stacks with duplicate slot numbers, the links between the two stacks will be in
Inhibited state. This can be seen using the show stacking stack-ports command as shown below in
step 1.

Assume that the master of stack A is to be the master node of the joined stack. Log into the intended
master node.
1 Verify the details of the new stack using the commands show stacking, show slot, and show
stacking configuration.
Active Topology is a Daisy-Chain
------------------ ---- ----------- ------- ---
00:04:96:26:60:AA 1 Active Master --O
00:04:96:26:60:88 2 Active Backup --O
00:04:96:26:60:99 3 Active Standby --O
Slot-1 StackA.12 #

------------------ --- ---- ---- ------------------ --------------- ------- ---
*00:04:96:26:60:DD 1 1 Auto 192.168.130.101/24 192.168.130.1 CcEeMm- Ee
00:04:96:26:60:EE 2 2 Auto 192.168.130.102/24 192.168.130.1 CcEeMm- Ee
00:04:96:26:60:FF 3 3 Auto 192.168.130.103/24 192.168.130.1 –-EeMm- Ee
00:04:96:26:60:AA 1 1 Auto 192.168.131.101/24 192.168.131.1 CcEe--i --
00:04:96:26:60:88 2 2 Auto 192.168.131.102/24 192.168.131.1 CcEe--i --
00:04:96:26:60:99 3 3 Auto 192.168.131.103/24 192.168.131.1 --Ee--i --
Slot-1 StackA.13 #

---- ---- ----------------- ----------- -----
*1 1 00:04:96:26:60:DD Inhibited --
*1 2 00:04:96:26:60:DD Operational C-
2 1 00:04:96:26:60:EE Operational C-
2 2 00:04:96:26:60:EE Operational C-
3 1 00:04:96:26:60:FF Operational C-
3 2 00:04:96:26:60:FF Inhibited --
1 1 00:04:96:26:6b:AA Inhibited --

2 1 00:04:96:26:60:88 Operational C-
2 2 00:04:96:26:60:88 Operational C-
3 1 00:04:96:26:6b:99 Operational C-
3 2 00:04:96:26:6b:99 Inhibited --
Slot-1 StackA.14 #
2 Configure the nodes such that they all have unique slot numbers. Since you originally configured
your stack with automatic slot assignment, you may perform automatic slot assignment now:
configure stacking slot-number automatic.
3 Configure the stack MAC address with the command: configure stacking mac-address.
4 Configure stacking redundancy so that only slots 1 and 2 are master-capable with the command:
configure stacking redundancy minimal.
5 Configure new alternate IP addresses for nodes from original StackB. Assume that the block of
addresses allocated to StackA can be extended, and use the automatic form of the command as
follows:
configure stacking alternate-ip-address 192.168.130.101/24 192.168.130.1 automatic
6 Configure a license restriction to be the minimum of the two original values on all nodes.
Alternatively, you may purchase license upgrades from Extreme if necessary. In this case, use the
command:
configure stacking license-level edge
7 Either reboot the entire stack topology using the reboot stack-topology command, or
individually reboot the three nodes formerly from stack B. The latter requires the following
commands:
reboot node 00:04:96:26:60:AA
reboot node 00:04:96:26:60:88
reboot node 00:04:96:26:60:99
Because none of these nodes is master-capable, there are no temporary dual master situations as a
result of these separate node reboots.
8 When the rebooted nodes come back up, run the following commands to see the resulting stack. You
can verify the joined stack came up as expected, that is, all nodes should have unique slot numbers,
a common stack MAC address, and so forth:
------------------ ---- ----------- ------- ---
00:04:96:26:60:AA 4 Active Standby CA-
Slot-1 StackA.12 #


------------------ --- ---- ---- ------------------ --------------- ------- ---
*00:04:96:26:60:DD 1 1 Auto 192.168.130.101/24 192.168.130.1 CcEeMm- Ee
00:04:96:26:60:EE 2 2 Auto 192.168.130.102/24 192.168.130.1 CcEeMm- Ee
00:04:96:26:60:FF 3 3 Auto 192.168.130.103/24 192.168.130.1 –-EeMm- Ee
00:04:96:26:60:AA 4 4 Auto 192.168.130.104/24 192.168.130.1 –-EeMm- Ee
00:04:96:26:60:88 5 5 Auto 192.168.130.105/24 192.168.130.1 –-EeMm- Ee
00:04:96:26:60:99 6 6 Auto 192.168.130.106/24 192.168.130.1 –-EeMm- Ee
Slot-1 StackA.13 #

---- ---- ----------------- ----------- -----
*1 1 00:04:96:26:60:DD Operational C-
*1 2 00:04:96:26:60:DD Operational C-
2 1 00:04:96:26:60:EE Operational C-
2 2 00:04:96:26:60:EE Operational C-
3 1 00:04:96:26:60:FF Operational C-
3 2 00:04:96:26:60:FF Operational C-
4 2 00:04:96:26:6b:AA Operational CB
5 1 00:04:96:26:60:88 Operational CB
5 2 00:04:96:26:60:88 Operational C-
6 1 00:04:96:26:6b:99 Operational C-
6 2 00:04:96:26:6b:99 Operational C-
Slot-1 StackA.14 #
Slot-1 StackA.3 # show slot

--------------------------------------------------------------------
Slot-4 X450a-48t Operational 50
Slot-6 X450a-24xDC Operational 26
Slot-7 Empty 0
Slot-8 Empty 0
Slot-1 StackA.4 #

9 Configure the new slots in VLANs, IP subnetworks, and so forth as required.
Adding a Node to a Stack

This section describes how to add a node to a stack. Assume the original stack is connected as follows:
Slot-1 Stack.9 # show stacking stack-ports
---- ---- ----------------- ----------- -----
*1 1 00:04:96:26:6a:f1 Operational C-
*1 2 00:04:96:26:6a:f1 Operational C-
2 1 00:04:96:26:6c:93 Operational C-
2 2 00:04:96:26:6c:93 Operational C-
3 1 00:04:96:26:5f:4f Operational C-
3 2 00:04:96:26:5f:4f Operational CB
4 1 00:04:96:1f:a5:43 Operational CB
4 2 00:04:96:1f:a5:43 Operational C-
5 1 00:04:96:28:01:8f Operational C-
5 2 00:04:96:28:01:8f Operational C-
6 1 00:04:96:20:b2:5c Operational C-
6 2 00:04:96:20:b2:5c Operational C-
1 Connect the new switch to the stack using a stacking cable to join the stacking ports and form a
physical ring. The connections should be made such that the node appears in the natural position in
the stack and in the slot. The example below adds a new node that becomes slot 7.
■ The connection broken should be the one between node 00:04:96:20:b2:5c port 2 and node
00:04:96:26:6a:f1 port 1.
■ The new node 00:04:96:26:6c:92 port 1 should be connected to node 00:04:96:20:b2:5c port 2
■ The new node 00:04:96:26:6c:92 port 2 should be connected to node 00:04:96:26:6a:f1 port 1.
2 Power on the new node, if you have not already powered it on, and wait for it to be recognized by
the stack.
Assume the original stack is connected as follows:
Slot-1 Stack.9 # show stacking stack-ports
---- ---- ----------------- ----------- -----
*1 1 00:04:96:26:6a:f1 Operational C-
*1 2 00:04:96:26:6a:f1 Operational C-
2 1 00:04:96:26:6c:93 Operational C-
2 2 00:04:96:26:6c:93 Operational C-
3 1 00:04:96:26:5f:4f Operational C-
3 2 00:04:96:26:5f:4f Operational CB
4 1 00:04:96:1f:a5:43 Operational CB
4 2 00:04:96:1f:a5:43 Operational C-
5 1 00:04:96:28:01:8f Operational C-
5 2 00:04:96:28:01:8f Operational C-

6 1 00:04:96:20:b2:5c Operational C-
6 2 00:04:96:20:b2:5c Operational C-
The show stacking command can also be used to verify whether the node is present in the stack.
Do not be concerned with the state of the node at this time.
3 On the master node, run the show switch command to verify the image selected for the stack. Make
sure the same image on the new node contains the same EXOS release that is being run by the stack.
4 Verify the stack configuration using the show stacking configuration command.
5 Use the synchronize stacking node-address 00:04:96:26:6c:92 command.
6 Configure a unique slot number in the active topology.
Slot-1 stack.11 # configure stacking node-address 00:04:96:26:6c:92 slot-number 7
7 Configure the license-level restriction of the new node to be same as the other members in the stack.
Slot-1 stack.12 # configure stacking node-address 00:04:96:26:6c:92 license-level edge
8 (Optional) Configure the priority, master-capability, and the alternate-ip-address of the new node.
See Configuring a new stack for details.
9 Reboot the node alone using the command:
Slot-1 stacK.13 # reboot node-address 00:04:96:26:6c:92
Are you sure you want to reboot this stack node? (y/N) Yes
After the node restarts, it will be in the active topology as a standby node.
10 Verify the new stack and the slot using the commands show stacking and show slot.
Slot-1 stacK.18 # sh stack
------------------ ---- ----------- ------- ---
*00:04:96:26:6a:f1 1 Active Master CA-
00:04:96:26:6c:93 2 Active Standby CA-
00:04:96:26:5f:4f 3 Active Backup CA-
00:04:96:1f:a5:43 4 Active Standby CA-
00:04:96:28:01:8f 5 Active Standby CA-
00:04:96:20:b2:5c 6 Active Standby CA-
00:04:96:26:6c:92 7 Active Standby CA-
Slot-1 stacK.18 # sh slot

--------------------------------------------------------------------
Slot-4 SummitX450-24x SummitX450-24x Operational 26
Slot-5 SummitX450-24t SummitX450-24t Operational 26
Slot-6 SummitX450-24t SummitX450-24t Operational 26
Slot-8 Empty 0

You have joined a new node to the stack.
Configuring an Alternate IP Address

This configuration example shows you how to configure an Alternate IP address for a stack:
1 Verify the alternate IP address configuration using the show stacking configuration command:
Slot-1 stacK.13 # show stacking configuration
Stack MAC in use: 02:04:96:26:6c:92
------------------ --- --- ---- ------------------ --------------- ------- ---
*00:04:96:26:6a:f1 1 1 11 <none> <none> CcEeMm- Ee
00:04:96:26:6c:93 2 2 Auto <none> <none> CcEeMm- Ee
00:04:96:27:c8:c7 3 3 Auto <none> <none> CcEeMm- Ee
00:04:96:26:5f:4f 4 4 4 <none> <none> CcEeMm- Ee
00:04:96:1f:a5:43 5 5 Auto <none> <none> CcEeMm- Ee
00:04:96:28:01:8f 6 6 6 <none> <none> CcEeMm- Ee
00:04:96:20:b2:5c 7 7 Auto <none> <none> CcEeMm- Ee
00:04:96:26:6c:92 8 8 Auto <none> <none> CcEeMm- Ee
If you have continuous block of available IP addresses for all slots in a stack, you can use the automatic
option to configure the alternate IP address for the node. This will configure the alternate IP address for
all the slots in the stack.
2 Configure the allocated block of addresses as alternate IP addresses:

Slot-1 stacK.14 # configure stacking alternate-ip-address 10.127.4.131/24 10.127.4.254
automatic
Slot-1 stacKgopu.15 # show stacking configuration

------------------ --- --- ---- ------------------ --------------- ------- ---
*00:04:96:26:6a:f1 1 1 11 10.127.4.131/24 10.127.4.254 CcEeMm- Ee
00:04:96:26:6c:93 2 2 Auto 10.127.4.132/24 10.127.4.254 CcEeMm- Ee
00:04:96:27:c8:c7 3 3 Auto 10.127.4.133/24 10.127.4.254 CcEeMm- Ee
00:04:96:26:5f:4f 4 4 4 10.127.4.134/24 10.127.4.254 CcEeMm- Ee
00:04:96:1f:a5:43 5 5 Auto 10.127.4.135/24 10.127.4.254 CcEeMm- Ee
00:04:96:28:01:8f 6 6 6 10.127.4.136/24 10.127.4.254 CcEeMm- Ee
00:04:96:20:b2:5c 7 7 Auto 10.127.4.137/24 10.127.4.254 CcEeMm- Ee
00:04:96:26:6c:92 8 8 Auto 10.127.4.138/24 10.127.4.254 CcEeMm- Ee

Now if you want to change the alternate IP address of slot 4, use the command configure stacking
slot 4 alternate-ip-address.
Slot-1 stacK.16 # configure stacking slot 4 alternate-ip-address 10.127.4.139/24

10.127.4.254
Error: An alternate IP address must not be already configured for this node.
Since you have already configured the slot with an alternate IP address, you need to unconfigure the
alternate IP address before you configure a new alternate IP address for the slot.
3 Unconfigure the alternate IP address for slot 4 and configure a new alternate IP address for slot 4:
Slot-1 stacK.17 # unconfigure stacking slot 4 alternate-ip-address
Slot-1 stacK.18 # configure stacking slot 4 alternate-ip-address 10.127.4.139/24
10.127.4.254
Slot-1 stacK.19 # show stacking configuration
------------------ --- --- ---- ------------------ --------------- ------- ---
*00:04:96:26:6a:f1 1 1 11 10.127.4.131/24 10.127.4.254 CcEeMm- Ee
00:04:96:26:6c:93 2 2 Auto 10.127.4.132/24 10.127.4.254 CcEeMm- Ee
00:04:96:27:c8:c7 3 3 Auto 10.127.4.133/24 10.127.4.254 CcEeMm- Ee
00:04:96:26:5f:4f 4 4 4 10.127.4.139/24 10.127.4.254 CcEeMm- Ee
00:04:96:1f:a5:43 5 5 Auto 10.127.4.135/24 10.127.4.254 CcEeMm- Ee
00:04:96:28:01:8f 6 6 6 10.127.4.136/24 10.127.4.254 CcEeMm- Ee
00:04:96:20:b2:5c 7 7 Auto 10.127.4.137/24 10.127.4.254 CcEeMm- Ee
00:04:96:26:6c:92 8 8 Auto 10.127.4.138/24 10.127.4.254 CcEeMm- Ee
You have configured the alternate IP address.
Configuring a MAC address

This configuration example explains how to configure a MAC address for the stack. The following two
scenarios are described:
● Configuring Stack MAC address to use the current master's factory default MAC address.
● Configuring stack MAC address to use the MAC address of any node other than the master.

The following message appears in the show log after the stack restarts when the stack MAC is not
configured or configured inconsistently:
02/24/2007 15:48:58.78 <Erro:HAL.Stacking.CfgStkMACAddrInv> Slot-1: The stack MAC

address is not correctly configured on this node. The stack can not operate properly
in this condition. Please correct and reboot.
1 Use show stacking configuration to determine whether the MAC address of the stack is not
configured (note the capital M, lower-case m, and lower-case i flags are not set).
Slot-1 stack.3 # sh stack configuration
------------------ --- --- ---- ------------------ --------------- ------- ---
*00:04:96:26:6a:f1 1 1 11 10.127.4.131/24 10.127.4.254 CcEe--- Ee
00:04:96:26:6c:93 2 2 Auto 10.127.4.132/24 10.127.4.254 CcEe--- Ee
00:04:96:27:c8:c7 3 3 Auto 10.127.4.133/24 10.127.4.254 CcEe--- Ee
00:04:96:26:5f:4f 4 4 4 10.127.4.139/24 10.127.4.254 CcEe--- Ee
00:04:96:1f:a5:43 5 5 Auto 10.127.4.135/24 10.127.4.254 CcEe--- Ee
00:04:96:28:01:8f 6 6 6 10.127.4.136/24 10.127.4.254 CcEe--- Ee
00:04:96:20:b2:5c 7 7 Auto 10.127.4.137/24 10.127.4.254 CcEe--- Ee
00:04:96:26:6c:92 8 8 Auto 10.127.4.138/24 10.127.4.254 CcEe--- Ee
2 From the console of the master, enter the configure stacking mac-address command:
Slot-1 stack.43 # configure stacking mac-address
(debug) Slot-1 stack.44 # sh stack configuration
------------------ --- --- ---- ------------------ --------------- ------- ---
*00:04:96:26:6a:f1 1 1 11 10.127.4.131/24 10.127.4.254 CcEe--i Ee
00:04:96:26:6c:93 2 2 Auto 10.127.4.132/24 10.127.4.254 CcEe--i Ee
00:04:96:27:c8:c7 3 3 Auto 10.127.4.133/24 10.127.4.254 CcEe--i Ee
00:04:96:26:5f:4f 4 4 4 10.127.4.139/24 10.127.4.254 CcEe--i Ee
00:04:96:1f:a5:43 5 5 Auto 10.127.4.135/24 10.127.4.254 CcEe--i Ee
00:04:96:28:01:8f 6 6 6 10.127.4.136/24 10.127.4.254 CcEe--i Ee
00:04:96:20:b2:5c 7 7 Auto 10.127.4.137/24 10.127.4.254 CcEe--i Ee
00:04:96:26:6c:92 8 8 Auto 10.127.4.138/24 10.127.4.254 CcEe--i Ee

The Flags indicates in this case that the Stack MAC is configured and is not in use. After you restart
the stack, the “i” disappears from the Flags. To see if the stack MAC is consistently configured, use
the show stacking detail command and compare all configured stack MAC addresses of all
nodes for equality. In this case, they should be equal.
3 Reboot the stack.
4 Verify the new stack mac address using the show stacking configuration command.
Slot-1 stack.3 # show stacking configuration
Stack MAC in use: 02:04:96:26:6a:f1
------------------ --- --- ---- ------------------ --------------- ------- ---
*00:04:96:26:6a:f1 1 1 11 10.127.4.131/24 10.127.4.254 CcEeMm- Ee
00:04:96:26:6c:93 2 2 Auto 10.127.4.132/24 10.127.4.254 CcEeMm- Ee
00:04:96:27:c8:c7 3 3 Auto 10.127.4.133/24 10.127.4.254 CcEeMm- Ee
00:04:96:26:5f:4f 4 4 4 10.127.4.139/24 10.127.4.254 CcEeMm- Ee
00:04:96:1f:a5:43 5 5 Auto 10.127.4.135/24 10.127.4.254 CcEeMm- Ee
00:04:96:28:01:8f 6 6 6 10.127.4.136/24 10.127.4.254 CcEeMm- Ee
00:04:96:20:b2:5c 7 7 Auto 10.127.4.137/24 10.127.4.254 CcEeMm- Ee
00:04:96:26:6c:92 8 8 Auto 10.127.4.138/24 10.127.4.254 CcEeMm- Ee
You have successfully configured the stack MAC address to use the master’s factory default MAC
address. Now we will change it to use the factory default MAC address of slot 2.
5 From the master console, enter configure stacking slot 2 mac-address:

Slot-1 stack.4 # conf stacking slot 2 mac-address
6 Reboot the stack.
7 Use show stacking configuration to verify the new stack MAC address.
Slot-1 stack.2 # show stacking configuration
------------------ --- --- ---- ------------------ --------------- ------- ---
*00:04:96:26:6a:f1 1 1 11 10.127.4.131/24 10.127.4.254 CcEeMm- Ee
00:04:96:26:6c:93 2 2 Auto 10.127.4.132/24 10.127.4.254 CcEeMm- Ee
00:04:96:27:c8:c7 3 3 Auto 10.127.4.133/24 10.127.4.254 CcEeMm- Ee
00:04:96:26:5f:4f 4 4 4 10.127.4.139/24 10.127.4.254 CcEeMm- Ee
00:04:96:1f:a5:43 5 5 Auto 10.127.4.135/24 10.127.4.254 CcEeMm- Ee
00:04:96:28:01:8f 6 6 6 10.127.4.136/24 10.127.4.254 CcEeMm- Ee
00:04:96:20:b2:5c 7 7 Auto 10.127.4.137/24 10.127.4.254 CcEeMm- Ee
00:04:96:26:6c:92 8 8 Auto 10.127.4.138/24 10.127.4.254 CcEeMm- Ee


Now the stack has been configured to use the MAC address of slot 2 as the stack MAC address. To see
the difference between the two results, use show stacking slot 1 detail and look at the stack MAC
address that is configured.
Configuring the License Level Restriction

To configure the license level restriction:
1 Configure a license restriction of Edge so that all nodes can come up in the stack.
* X450a-24x.7 # configure stacking license-level edge
2 Restart the stack.
The stack restarts in the license-level Edge.
Bringing All Nodes Up to the Same License Level

1 Log into the master node.
2 Issue the show licenses command:
Slot-1 Stack.1 # show licenses
Enabled License Level:
Advanced Edge
Enabled Feature Packs:
None
Effective License Level:
Advanced Edge
Slot-1 Stack.2 #
3 For each active node, telnet into the node using the command telnet slot <slot-number>, run
the command show licenses, and collect the data before logging out.
4 Compare the Enabled License Level of all nodes. It is recommended that you purchase an upgrade
for all nodes that do not have the greatest of these license levels.

5 For any node that must be upgraded, telnet into the node using the command telnet slot <slot-
number>, run the enable license <key> command supplying the license key given to you by
Extreme Networks when you purchased the upgrade, run the show licenses command to insure
the upgrade was successful, and log out of the node.
6 If you had previously configured a license level restriction, from the master run the command
unconfigure stacking license-level and then run the command reboot to reboot the stack and
remove the restriction.
VLAN Management in Stacking

There is no additional restriction on VLAN ID usage or total number of VLANs due to the addition of
the SummitStack to EXOS.
The show vlan mgmt command will show the alternate management IP address as applied to the
VLAN on the local unit. This will allow you to see how the configured alternate management IP
address has been applied.
The show ipconfig Mgmt command will show the configured alternate management IP address as
applied to the VLAN on the local unit. This will allow you to see how the configured alternate
management IP address has been applied.
The multinetted VLAN indication will always appear as no. The alternate IP address is restricted to the
same subnet as the primary subnet configured for the Mgmt IP interface. As a result, only a single
subnet is used with the possibility of multiple station addresses. Further, you cannot configure a
secondary IP address on the Mgmt VLAN.
The show vlan mgmt command displays the following information:
Slot-1 Stack.35 # show vlan "Mgmt"

VLAN Interface with name Mgmt created by user
Admin State: Enabled Tagging: 802.1Q Tag 4095
Virtual router: VR-Mgmt
Primary IP: 10.1.4.1/24
Alternate IP: 10.1.4.2/24
IPv6: None
STPD: None
Protocol: Match all unfiltered protocols
Loopback: Disabled
NetLogin: Disabled
QosProfile: None configured
Ports: 1. (Number of active ports=1)
Untag: Mgmt-port on Mgmt-? is active
For the Mgmt VLAN, a secondary address cannot be configured and so the Secondary IP line does not
appear. The Alternate IP line shows either:
● The configured alternate management IP address if it has been activated

● <none> if it has not been configured
● Mismatch if it has been configured but does not exactly match the Primary IP subnet.

The show ip config mgmt command displays the following information:
Slot1 Stack.36 # show ipconfig Mgmt

Router Interface on VLAN Mgmt is enabled and up.
inet 10.66.4.74/24 broadcast 10.66.4.255 Mtu 1500
Alternate IP Address: 10.66.4.75/24
Flags:
AddrMaskRly NO BOOTP Host NO DirBcstHwFwd NO Fwd Bcast NO
IgnoreBcast NO IP Fwding NO IPmc Fwd NO Multinetted VLAN NO
IRDP Advert NO SendParam YES SendPortUn YES Send Redir YES
SendTimxceed YES SendUnreach YES TimeStampRly NO VRRP NO
For the Mgmt VLAN, a secondary address cannot be configured and so the Secondary IP line does not
appear. The Alternate IP Address line shows either:
● The configured alternate management IP address if it has been activated

● <none> if it has not been configured
● Mismatch if it has been configured but does not exactly match the Primary IP subnet.
Troubleshooting a Stack
Use this section to diagnose and troubleshoot common configuration errors for stacking. The most
common errors are:
● The stack did not come up as expected—Use show stacking, show stacking configuration, and
show stacking stack-ports commands to diagnose the problem. There could be incorrect
stacking cabling, a configuration error, or powered down nodes. Also check the log using the show
log command.
● The switch with the highest priority was not elected manager—nodes might have been powered up
at different times. Reboot all nodes in the stack simultaneously.
● A node appears in the stack as expected but does not appear to be operating as configured—Use the
show slot <slot> command to see if there is a license mismatch. For more information, see
“Managing Licenses on a SummitStack” on page 604.
A node can fail to become an active node for the following reasons:
● The stacking feature is not enabled on the node.

● Another node in the stack has the same slot number as this node.
● The node is isolated from the active topology of the stack by another node on which the stacking is
not enabled.
● The node is isolated from the active topology of the stack by another failed node.
To find the node failures and isolated nodes, use the show stacking command:
● If the node is not enabled for stacking, the node will show as Disabled.
● If the slot number of the node duplicates that of another node in the stack, either the node will show
as Failed, or some link between the node and the active topology of the stack will show as Inhibited.

● If the node is isolated by another node that is not enabled for stacking, some node between the node
and the active topology of the stack will show as Disabled.
● If the node is isolated by another failed node, the show stacking command will show the node as
Active but it will have the “O” flag set.
This section describes the following troubleshooting topics:
● Managing a Dual Master Situation on page 645

● Setting Traps for Stacking on page 647
● Rescuing a Stack That Has No Master-Capable Node on page 648
● Connecting to a SummitStack with No Master on page 648
Managing a Dual Master Situation

If a daisy chain is broken, or if a ring is broken in two places, it is possible to form two separate Active
Topologies. This results in a dual master situation.
Figure 65: Example of a split stack that results in Dual Master Situation
M1
B8
M2
S7 B3
P6 S4
S5
BD_161
P6 Node 6 is powered off

M Master node
B Backup node
S Standby node
X Indicates the broken link
For example, in Figure 65, a link is broken while a node in the ring was powered off. The link that is
now broken formerly connected the original master (M1) and backup (M2) nodes of a single active
topology.
All nodes in the stack except the powered off node are in the active topology and all nodes are
configured to be master-capable. Nodes 1, 7 and 8 forms an active topology and nodes 2, 3, 4, and 5

forms another active topology. Node M2 immediately transitions from backup to master node role.
Nodes B8 and B3 are elected in their respective active topologies as backup nodes.
If the backup node is on one stack and the master node is on the other, the backup node becomes a
master node because the situation is similar to that of master failure. Since both the stacks are
configured to operate as a single stack, there is confusion in your networks. For example, all of the
switch’s configured IP addresses appear to be duplicated. The management IP address will also be
duplicated since that address applies to the entire original stack.
To help mitigate the dual master problem, you can configure the master-capability and thus prevent
some nodes in the stack from operating in backup or master node roles. In addition, you can force all
nodes in the (broken) stack topology to restart and come up as not master-capable for the life of that
restart. The save command will save the configuration on all nodes in the active topology that have
been configured to be master-capable.
Standby nodes that exist in a severed stack segment that does not contain either the original master or
backup node will not attempt to become the master node. Instead, these nodes will reboot. After
rebooting, however, a master election process will occur among the nodes on this broken segment,
resulting in a dual master situation.
Dual master conditions are also possible when two non-adjacent nodes in a ring or a single (middle)
node in a daisy chain reboot. For a period of time, a rebooting node does not advertise itself to its
neighbors, resulting in temporary stacking link failures. This could cause node isolation, and the nodes
that are isolated perform as a severed stack segment depending on the circumstances of the severance:
● if the backup node is on the broken portion, it becomes a (dual) master;

● if the backup node is on the same portion as the master, all nodes on the (other) broken portion will
reboot.
When the rebooting nodes have sufficiently recovered, or when a severed stack is rejoined, the dual
master condition is resolved, resulting in the reboot of one of the master nodes. All standby and backup
nodes that had been acquired by the losing master node will also reboot.
You can avoid a dual master possibility during configuration by:
● Configuring the stack in a ring topology.

● Avoiding too many master-capable nodes while configuring larger stacks.
● Placing the nodes that provide stack redundancy (i.e., those nodes that are master-capable) such that
stacking link severances are unlikely.
Eliminating a Dual Master Situation Manually

To eliminate the dual master situation, you need to know all the nodes that are supposed to be in the
stack. You might lose the management connectivity to the master node because the other master node
duplicates the stack’s primary management IP address and stack MAC address.
1 If you lose the management connectivity, log into the master node using its alternate management IP
address.
2 Use the show stacking command to determine the nodes that have been lost from the stack. You
should already know all the nodes that are expected to be part of the stack.
3 Log into any node in the severed segment you wish to deactivate, either through its console port or
through the management interface using the alternate management IP address. Issue show
stacking to find whether the broken segment has indeed elected a new master node.

4 Reboot the broken segment forcing all nodes in the segment to come up as standby nodes using the
reboot stack-topology as-standby command.
If you have unsaved configuration changes, take care when selecting the stack segment to be
rebooted.
You should reboot the segment that has the smaller System UpTime.
If you know the node that was master of the unbroken stack, you can reboot the stack segment that
does not contain this master node. Otherwise, determine the System UpTime shown by each master
node.
If the System UpTimes of both masters are the same, you can reboot either segment without loss of
unsaved configuration changes. If the System UpTimes of both masters differ, you must reboot the
segment with the smaller System UpTime.
The above procedure is necessary only if you cannot reconnect the severed link in a timely manner. If
you can reconnect, the dual master condition will resolve itself. The formerly broken portion of the
stack will reboot and come up as standby nodes.
Automatic Resolution of the Dual Master Situation

When two stack segments are connected together and no slot number is duplicated on either segment, it
is assumed that this is a severed stack rejoin. It is possible that each stack segment has its own master.
Resolution of the dual master situation should generally be in favor of the original stack segment’s
master node. This is because the original stack segment may still retain the unsaved configuration. If the
severed segment was restarted before electing a new master node, the unsaved configuration is lost on
that segment.
The master election is done using the System UpTime. The master election process collects the System
UpTime information of the nodes. If a failover occurs, the System UpTime will be inherited by the new
master node, the new master node will continue to increase it as time passes. Thus the System UpTime
is the time since a master was first elected on a segment. When the stack is broken and both master and
backup nodes are on the same segment, the severed segment will always have the smaller System
UpTime.
If a stack severance results in the master and backup nodes being on different segments, both will have
the same System UpTime. In this case, the master is elected using the normal node role election method.
Setting Traps for Stacking

The stack generates traps that provide status information about the switches in the stack and also
stacking port status. Traps generated by the stack include:
● extremeStackMemberStatusChanged
extremeStackMemberSlotId—Indicates the slot ID
extremeStackMemberOperStatus—Indicates the slot state of the switch
The stack generates this trap when an overheat condition is detected on an active node:
● extremeStackMemberOverheat
This trap is generated when the node reaches a steady state. Whenever a member is added or deleted
from the stack, the change is indicated through this trap:
● extremeStackingPortStatusChanged
IfIndex—Interface Index of the port

extremeStackingPortRemoteMac—MAC Address of the remote switch attached to this port

extremeStackingPortLinkSpeed—Indicates 10/100/1000 Mbps
extremeStackingPortLinkStatus—Status of the link
The trap is generated whenever the status of a stacking port changes.
Connecting to a SummitStack with No Master

If an entire stack has no master node because the stack has been rebooted in standby only mode, you
can log in to a node by using a failsafe account.
If a new node has been added to the stack since the stack failsafe account was configured, logging in to
that node requires knowledge of the failsafe account information that is already configured into that
node's NVRAM. If you don’t know what the failsafe account and password is, try the default values
described in the Accessing the Switch chapter.
If you do not know the failsafe account and you still want to log in to the stack, you have to:
● Join the stack to another segment that has a master node to which the you have access
● Manually restart the stack to clear the as-standby condition if the reboot stack-topology as-
standby command was previously used
● Use the procedure described in “Rescuing a Stack That Has No Master-Capable Node” on page 648
Rescuing a Stack That Has No Master-Capable Node
NOTE
If a node becomes unbootable, refer to the Troubleshooting appendix for information.
You can have a stack with nodes that are all configured with the master-capability set to off.
For example, if a stack was operating with no redundancy (for example, with one master-capable node)
and the master node failed, all other nodes in the stack restart as standby nodes and there is no master
node.
In this situation, the only security information available is the failsafe account. If you know the failsafe
user name and password, you can log into any node and reconfigure master-capability or redundancy.
However, if you do not know the failsafe account information, there is another way you can change the
configuration.
At the login prompt, enter the following special login ID exactly as displayed here:
REBOOT AS MASTER-CAPABLE
You will get a prompt to enter a password. Do not enter any password, just press the enter key. The
following message appears:
Node reboot initiated with master-capability turned on.
This node then sets an internal indicator that is preserved across the reboot. While restarting, the node
ignores the node master-capability configuration and becomes the master node.

Since the save command saves the configuration file to all nodes, the node that just rebooted as master-
capable should have access to the security information that was configured for the stack. If a RADIUS
server is needed, the selected node requires a network connection for authentication.
The special login ID described above is available only if all the following conditions are met:
● The node supports SummitStack.

● Stacking mode is active on the node.
● All nodes in the active topology have master-capability turned off.
● There is no master node in the active topology.
If the above conditions are met, five minutes after starting the node and every five minutes after that,
the following message appears on the console:
Warning: the stack has no Master node and all active nodes are operating with
master-capability turned off. If you wish to reconfigure, you may log in
using the failsafe account. Alternatively, you may use the special login
REBOOT AS MASTER-CAPABLE
with no password to force a reboot of a node with master-capability
temporarily turned on.
Using the special login ID does not alter the master-capability configuration permanently. If you restart
a node that has been restarted with the special login ID, that node restarts using its configured master-
capability; unless you again use the special login ID to restart.
The procedure described here is generally not needed if another node that is master-capable is expected
to rejoin the stack. If this procedure is used, it is possible that the new master duplicates the master that
is expected to rejoin later.
Stacking Link Failure

A stacking link is said to be failed when one of the following happens:
● The stacking link is physically disconnected.
● The neighbor on an In Stack link stops transmitting topology information.
● The link stops transmitting topology information while a node restarts.
Based on the stacking topology, the stack behavior changes.
Ring Topology . All traffic paths that were directed through the failed link are redirected. All nodes
converge on the new (daisy chain) topology that results from the link break. The Topology Protocol that
determines the stack topology immediately informs other nodes that a link has failed. Each node will
start the process of redirecting traffic paths.
Daisy chain . A stacking link severance means a severed stack. The Topology Protocol will report the
loss of all nodes in the severed portion. Depending on master capability configuration and the original
location of the backup node, the severed portion may or may not elect a new master node. If it does, the
dual master condition may be in effect.
The show slot command display the slots that contains active nodes that are in the severed portion as
Empty.

FAQs on SummitStack
● How can I find the slot number of the master slot in a stack?
To find the slot number of the master slot, log in to any of the stack node and run the command
show stacking.
● How would I know whether there is a dual master situation in a stack?
A main symptom is loss of IP connectivity. Run the show stacking command to see whether all
expected nodes are still in the stack.
● How would I find the current topology of the stack?
Run show stacking command.
● Can I enable EAPS on stacking?
Yes. You can enable the EAPS on a stack. EAPS will operate in your networks even if an EAPS path
crosses through the stacking links. EAPS is not used as a redundancy protocol for the stacking ring.
● Why should I configure an Alternate IP address?
To login to the individual nodes using the management port of the node and to configure the nodes
individually. It is most beneficial in manually resolving a dual master situation since connectivity
using the alternate IP address is not affected by the dual master situation.

2 Using Switching and Routing Protocols
25 Ethernet Automatic Protection Switching

● Overview of the EAPS Protocol on page 654
● Fault Detection and Recovery on page 656
● Multiple EAPS Domains on page 659
● Configuring EAPS on a Switch on page 662
● Configuring EAPS Shared Ports on page 675
● EAPS Shared Port Configuration Rules on page 684
● EAPS Shared Port Configuration Examples on page 685
Licensing
You need a Core or an Advanced Core license to configure and use all of the Ethernet Automatic
Protection Switching (EAPS) features described in this chapter.
A subset of EAPS, called EAPS Edgemode, is available with an Edge or an Advanced Edge license. The
following features are available with EAPS Edgemode:
● Switches can belong to only one EAPS ring.
● You can have up to four EAPS domains using two matching ring ports.
EAPS shared ports is not supported with an Edge or an Advanced Edge license.
an Advanced Core license, respectively.
license. The BlackDiamond 12802 switch ships with an Advanced Edge license.
series switch, and Summit X450a series switch ship with an Advanced Edge license. To use the
complete EAPS functionality, including running two or more EAPS rings, having a switch belong to
multiple EAPS rings, or configuring shared-ports that allow multiple EAPS domains to share a common
link, you must upgrade to a Core software license. A SummitStack must also be operating at the Core
license level in order to use the complete EAPS functionality.
The Summit X450e series switch ships with an Edge license. You can upgrade to an Advanced Edge
license; however, both the Edge and the Advanced Edge licenses only support EAPS Edgemode. You
cannot upgrade to a Core software license.
NOTE
The Summit X450e series switch only supports EAPS Edgemode.

Ethernet Automatic Protection Switching
Overview of the EAPS Protocol

The EAPS protocol provides fast protection switching to Layer 2 switches interconnected in an Ethernet
ring topology, such as a Metropolitan Area Network (MAN) or large campuses (see Figure 66).
EAPS protection switching is similar to what can be achieved with the Spanning Tree Protocol (STP),
but EAPS offers the advantage of converging in less than 1 second when a link in the ring breaks.
An Ethernet ring built using EAPS can have resilience comparable to that provided by SONET rings, at
a lower cost and with fewer restraints (such as ring size). The EAPS technology developed by Extreme
Networks to increase the availability and robustness of Ethernet rings is described in RFC 3619: Extreme
Networks’ Ethernet Automatic Protection Switching (EAPS) Version 1.
EAPS operates by declaring an EAPS domain on a single ring. Any virtual LAN (VLAN) that warrants
fault protection is configured on all ring ports in the ring, and is then assigned to an EAPS domain. On
that ring domain, one switch, or node, is designated the master node (see Figure 67), while all other
nodes are designated as transit nodes.
Figure 66: Gigabit Ethernet fiber EAPS MAN ring
Transit
node
Transit
node
Transit
Gigabit Ethernet Fiber node
EAPS MAN ring
Transit
node
Master
node EW_070
One port of the master node is designated the master node’s primary port (P) to the ring; another port is
designated as the master node’s secondary port (S) to the ring. In normal operation, the master node
blocks the secondary port for all non-control traffic belonging to this EAPS domain, thereby avoiding a
loop in the ring, like STP. Layer 2 switching and learning mechanisms operate per existing standards on
this ring.

Overview of the EAPS Protocol
NOTE
Like the master node, each transit node is also configured with a primary port and a secondary port on the ring, but
the primary/secondary port distinction is ignored as long as the node is configured as a transit node.
Figure 67: EAPS operation
S4
S3 S5
S2 S6
P S
S1 Secondary port
Direction of is logically blocked
health-check
message Master
node EW_071
If the ring is complete, the master node logically blocks all data traffic in the transmit and receive
directions on the secondary port to prevent a loop. If the master node detects a break in the ring, it
unblocks its secondary port and allows data traffic to be transmitted and received through it.
Fast Convergence
The Fast Convergence mode allows EAPS to converge more rapidly. In EAPS Fast Convergence mode,
the link filters on EAPS ring ports are turned off. In this case, an instant notification is sent to the EAPS
process if a port’s state transitions from up to down or vice-versa.
You configure Fast Convergence for the entire switch, not by EAPS domain.
EAPS and Hitless Failover—Modular Switches and SummitStack

Only
chassis or use redundancy in a SummitStack, one node assumes the role of primary and another node
assumes the role of backup. The primary executes the switch’s management functions, and the backup
acts in a standby role. Hitless failover transfers switch management control from the primary to the
backup and maintains the state of EAPS. EAPS supports hitless failover. You do not explicitly configure
hitless failover support; rather, if you have two MSMs installed in a chassis or you are operating with
redundancy in a SummitStack, hitless failover is available.

NOTE
To support hitless failover, the primary node replicates all EAPS protocol data units (PDUs) to the
backup, which allows the backup to be aware of the state of the EAPS domain. Since both nodes receive
EAPS PDUs, each node maintains equivalent EAPS states.
By knowing the state of the EAPS domain, the EAPS process running on the backup node can quickly
recover after a primary node failover. Although both nodes receive EAPS PDUs, only the primary
transmits EAPS PDUs to neighboring switches and actively participates in EAPS.
NOTE
page 1020 to confirm that your switch and both installed MSMs are running software that supports the
synchronize command.
To initiate hitless failover on a network that uses EAPS:

1 Confirm that the primary and backup nodes are synchronized and have identical software and
switch configurations using the show switch {detail} command. The output displays the status
of the nodes, with the primary node showing MASTER and the backup node showing BACKUP
(InSync).
● If the primary and backup nodes are not synchronized and both nodes are running a version of
● If the primary and backup nodes are synchronized, proceed to step 3.
2 If the primary and backup nodes are not synchronized, use the synchronize command to replicate
3 If the primary and backup nodes are synchronized, use the run failover (formerly run msm-
failover) command to initiate failover.
For more detailed information about verifying the status of the MSMs and system redundancy, see
Fault Detection and Recovery

EAPS fault detection on a ring is based on a single control VLAN per EAPS domain. This EAPS domain
provides protection to one or more data-carrying VLANs called protected VLANs.
The control VLAN is used only to send and receive EAPS messages; the protected VLANs carry the
actual data traffic. As long as the ring is complete, the EAPS master node blocks the protected VLANs
from accessing its secondary port.

Fault Detection and Recovery
NOTE
The control VLAN is not blocked. Messages sent on the control VLAN must be allowed into the switch for the master
node to determine whether the ring is complete.
To avoid loops in the network, the control VLAN must be NOT be configured with an IP address, and ONLY ring
ports may be added to this VLAN.
A master node detects a ring fault in one of three ways:

● Link down message sent by a transit node
● Ring port down event sent by hardware layers
● Polling response
The rest of this section describes the fault detection methods and the applicable restoration options.
Link Down Message Sent by a Transit Node

When any transit node detects a loss of link connectivity on any of its ring ports, it immediately sends a
“link down” message on the control VLAN using its good link to the master node.
When the master node receives the “link down” message (see Figure 68), it immediately declares a
“failed” state and opens its logically blocked secondary port on all the protected VLANs. Now, traffic
can flow through the master’s secondary port. The master node also flushes its FDB and sends a
message on the control VLAN to all of its associated transit nodes to flush their forwarding databases as
well, so that all of the switches can learn the new paths to Layer 2 endstations on the reconfigured ring
topology.
Figure 68: EAPS fault detection and protection switching

Break
in ring S4 sends "link down"
S4 message to master node
S3 S5
S3 sends "link down"

message to
master node
S2 S6
P S
S1 Master node opens secondary port
to allow traffic to pass
Master
node EW_072

Ring Port Down Event Sent by Hardware Layer

When a ring port goes down on a master node switch, it is notified by the lower hardware layer and
immediately goes into a “failed” state.
If the ring port that goes down on the master node is the primary port, the secondary port is opened.
The normal operation of flushing the master node’s FDB and sending a “flush FDB” message to all
transit nodes is performed.
Polling
The master node transmits a health check packet on the control VLAN at a user-configurable interval
(see Figure 67). If the ring is complete, the master node receives the health-check packet on its
secondary port (the control VLAN is not blocked on the secondary port). When the master node
receives the health-check packet, it resets its failtimer and continues normal operation.
If the master node does not receive the health check packet before the failtimer interval expires and the
failtime expiry action is set to open-secondary-port, it declares a “failed” state and performs the same
steps described above:
● Unblocks its secondary port for access by the protected VLANs.
● Flushes its forwarding database (FDB).
● Sends a “flush FDB” message to its associated transit nodes.
Restoration Operations
The master node continues sending health check packets out its primary port even when the master
node is operating in the failed state. As long as there is a break in the ring, the fail period timer of the
master node continues to expire, and the master node remains in the failed state.
When the broken link is restored, the master receives its health check packet back on its secondary port
and again declares the ring to be complete. Again, the master node logically:
● Blocks the protected VLANs on its secondary port.
● Flushes its FDB.
● Sends a “flush FDB” message to its associated transit nodes.
During the time between when the transit node detects that the link is operable again and when the
master node detects that the ring is complete, the secondary port on the master node is still open and
data could start traversing the transit node port that just came up.
To prevent the possibility of a such a temporary loop, when the transit node detects that its failed link is
up, it performs these steps:
1 For the port that just came up, put all the protected VLANs traversing that port into a temporary
blocked state.
2 Remember which port has been temporarily blocked.
3 Set the state to Preforwarding.
When the master node receives its health check packet back on its secondary port and detects that the
ring is again complete, it sends a message to all its associated transit nodes to flush their forwarding
databases.

Multiple EAPS Domains
When the transit nodes receive the message to flush their forwarding databases, they perform these
steps:
1 Flush their forwarding databases on the protected VLANs.
2 If the port state is set to Preforwarding, unblock all the previously blocked protected VLANs for the
port.

This section illustrates how you can work with more than one EAPS domain. The scenarios described in
this section include the following:
● EAPS Data VLAN Spanning Two Rings Connected by One Switch on page 659
● Multiple EAPS Domains per Ring—Spatial Reuse on page 660
● Multiple EAPS Rings Sharing a Common Link on page 661
EAPS Data VLAN Spanning Two Rings Connected by One Switch

Figure 69 shows how a data VLAN could span two rings interconnected by a common switch—a
“figure eight” topology. In this example, there is an EAPS domain with its own control VLAN running
on ring 1 and another EAPS domain with its own control VLAN running on ring 2. A data VLAN that
spans both rings will be added as a protected VLAN to both EAPS domains. Switch S5 has two
instances of EAPS domains running on it: one for each ring.
Figure 69: EAPS data VLAN spanning two rings interconnected by one switch
S4 S6
S3 S7
Ring 1 S5 Ring 2
S
S2
P S S 8 Master
node
P
S1
S9
Master
node EW_073

Multiple EAPS Domains per Ring—Spatial Reuse

To take advantage of the spatial reuse technology and broaden the use of the ring’s bandwidth, EAPS
supports multiple EAPS domains running on the ring at the same time (Figure 70).
Figure 70: Multiple EAPS domains per ring
Master EAPS 1
Transit EAPS 2
S
EAPS 1
EAPS 1
Transit EAPS 1 Transit EAPS 1
Transit EAPS 2 Transit EAPS 2
EAPS 2
EAPS 2
S
Master EAPS 2
Transit EAPS 1
EX_100
So, a single ring might have two EAPS domains running on it. Each EAPS domain would have a
different EAPS master node. Each EAPS domain will protect its own set of protected VLANS.
In a spatial reuse configuration, do not add the same protected VLAN to both EAPS domains.
BlackDiamond 8800 Series Switch, SummitStack, and Summit Family of Switches Only. If you configure
EAPS spatial reuse and disable the EAPS domain on a transit node, this may cause the master node’s
fail timer to expire and EAPS PDUs to no longer be forwarded through out the domain. If this occurs,
make sure the EAPS domain is enabled on all of the transit nodes. To confirm the current state of EAPS
on the switch, use the show eaps command.

Spatial Reuse with EAPS Shared Ports

You can also use spatial reuse with EAPS shared ports as shown in Figure 71.
Figure 71: EAPS shared ports configuration with spatial reuse
EX_105
The advantage of this topology is that you can configure common links between multiple rings and
configure multiple EAPS domains on each ring. To utilize this topology and configure EAPS shared
ports, you must have a core or an advanced core license.
For information about configuring common links and EAPS shared ports, see “Configuring EAPS
Shared Ports” on page 675.
Multiple EAPS Rings Sharing a Common Link

When you configure EAPS on multiple rings with a common link, you may experience a loop situation
across both rings. To solve this problem you can configure EAPS shared ports.
NOTE
You must have a core or an advanced core license to use the EAPS shared port feature.
In the example shown in Figure 69, switch S5 could be a single point of failure. If switch S5 were to go
down, users on Ring 1 would not be able to communicate with users on Ring 2. To make the network
more resilient, you can add another switch, S10. The link connecting S5 to S10 is knows as the common
link, as shown in Figure 72.

Figure 72: Multiple EAPS domains sharing a common link with EAPS shared ports
S4 S5 S6
S3 Controller S7
EAPS1 EAPS2
link ID=1 Common
link
S2 Partner S8
P S S P
S1 S 10 S9
Master Master
node node EW_095
The switches on either end of the common link must be configured as controller and a partner. For
information about configuring common links, see “Configuring EAPS Shared Ports” on page 675.
NOTE
If the shared port is not configured and the common link goes down a superloop between the multiple EAPS
domains will occur.
NOTE
To take advantage of the Spatial Reuse technology in a shared-port environment in this software release, you can use
the existing solution of configuring EAPS plus STP.
Configuring EAPS on a Switch

To configure and enable an EAPS domain:
1 Create EAPS domain and assign the name.
2 Configure the control VLAN.
3 Configure the protected VLAN(s).
4 Add the control VLAN to EAPS domain.
5 Add the protected VLAN(s) to EAPS domain.
6 Configure EAPS mode, master or transit.
7 Configure EAPS port, secondary and primary.
8 If desired, configure timeout and action for failtimer expiration.*
9 If desired, configure the hello time for the health-check packets.*
10 Enable EAPS for the entire switch.

11 If desired, enable Fast Convergence.*

12 Enable EAPS for the specified domain.
Although you can enable EAPS before configuring these steps, the EAPS domain(s) will not run until
you configure these parameters. (The steps with * can be configured at any time, even after the EAPS
domains are running.)
NOTE
If you configure a VMAN on a switch running EAPS, make sure you configure the VMAN attributes on all of the
switches that participate in the EAPS domain. For more information about VMANs, see the section Chapter
7, “vMAN and Tunneling.”
This section further describes the following topics:

● Creating and Deleting an EAPS Domain on page 663
● Defining the EAPS Mode of the Switch on page 664
● Configuring EAPS Polling Timers on page 665
● Configuring the Primary and Secondary Ports on page 666
● Configuring the EAPS Control VLAN on page 666
● Adding the EAPS Protected VLANs on page 667
● Enabling and Disabling Fast Convergence on page 668
● Enabling and Disabling an EAPS Domain on page 668
● Enabling and Disabling EAPS on the Switch on page 668
● Unconfiguring an EAPS Ring Port on page 669
● Disabling EAPS Loop Protection Warning Messages on page 670
● Displaying EAPS Status and Counter Information on page 670
Creating and Deleting an EAPS Domain

Each EAPS domain is identified by a unique domain name.
To create an EAPS domain, use the following command:
create eaps <name>
The name parameter is a character string of up to 32 characters that identifies the EAPS domain to be
created.
NOTE
If you use the same name across categories (for example, STPD and EAPS names), Extreme Networks recommends
The following command example creates an EAPS domain named “eaps_1”:
create eaps eaps_1

To delete an EAPS domain, use the following command:
delete eaps <name>
The following command example deletes the EAPS domain “eaps_1”:
delete eaps eaps_1
Defining the EAPS Mode of the Switch

To configure the EAPS node type of the switch, use the following command:
configure eaps <name> mode [master | transit]
One node (or switch) on the ring must be configured as the master node for the specified domain; all
other nodes (or switches) on the ring are configured as transit nodes for the same domain.
Defining the Master Node

The following command example identifies this switch as the master node for the EAPS domain named
eaps_1.
configure eaps eaps_1 mode master
Defining the Transit Node

If you configure a switch to be a transit node for an EAPS domain, the switch displays by default
messages to:
● Remind you to configure a master node in the EAPS domain.
● Notify you that changing a master node to a transit node might cause a loop in the network. If you
have not assigned a new master node before changing the current master node to a transit node, you
might cause a loop in the network.
When prompted, do one of the following:

● Enter y to identify the switch as a transit node.
● Enter n or press [Return] to cancel this action.
The following command example identifies this switch as a transit node for the EAPS domain named
eaps_1.
configure eaps eaps_1 mode transit
The switch displays the following warning message and prompts you to confirm this action:
WARNING: Make sure this specific EAPS domain has a Master node in the ring. If
you change this node from EAPS master to EAPS transit, you could cause a
loop in the network.
Are you sure you want to change mode to transit? (y/n)
Extreme Networks recommends that you keep the loop protection warning messages enabled. If you
have considerable knowledge and experience with EAPS, you might find the EAPS loop protection
warning messages unnecessary. For more information see, “Disabling EAPS Loop Protection Warning
Messages” on page 670.

Configuring EAPS Polling Timers

To set the values of the polling timers the master node uses for the EAPS health check packet that is
circulated around the ring for an EAPS domain, use the following commands:
configure eaps <name> hellotime <seconds>
configure eaps <name> failtime <seconds>
NOTE
These commands apply only to the master node. If you configure the polling timers for a transit node, they will be
ignored. If you later reconfigure that transit node as the master node, the polling timer values will be used as the
current values.
Use the hellotime keyword and its associated seconds parameter to specify the amount of time the
master node waits between transmissions of health check packets on the control VLAN. The value for
seconds must be greater than 0 when you are configuring a master node. The default value is 1 second.
NOTE
Increasing the hellotime value keeps the processor from sending and processing too many health check packets.
Use the failtime keyword and seconds parameters to specify the amount of time the master node
waits before the failtimer expires.
The seconds parameter must be greater than the configured value for hellotime. The default value is 3
seconds.
To configure the action taken if there is a break in the ring, use the following command:
configure eaps <name> failtime expiry-action [open-secondary-port | send-alert]
You can configure the action taken when the failtimer expires by using the configure eaps failtime
expiry-action command. Use the send-alert parameter to send an alert when the failtimer expires.
Instead of going into a “failed” state, the master node remains in a “Complete” or “Init” state, maintains
the secondary port blocking, and writes a critical error message to syslog warning the user that there is
a fault in the ring. An SNMP trap is also sent.
Use the open-secondary-port parameter to open the secondary port when the failtimer expires.
NOTE
Increasing the failtime value provides more protection by waiting longer to receive a health check packet when the
network is congested.
The following command examples configure the hellotime value for the EAPS domain “eaps_1” to
2 seconds, the failtimer value to 15 seconds, and the failtimer expiry-action to open the secondary port
if the failtimer expires:
configure eaps eaps_1 hellotime 2

configure eaps eaps_1 failtime 15
configure eaps eaps_1 failtimer expiry-action open-secondary-port

Configuring the Primary and Secondary Ports

Each node on the ring connects to the ring through two ring ports. As part of the protection switching
scheme, one port must be configured as the primary port, and the other must be configured as the
secondary port.
If the ring is complete, the master node prevents a loop by logically blocking all data traffic in the
transmit and receive directions on its secondary port. If the master node subsequently detects a break in
the ring, it unblocks its secondary port and allows data traffic to be transmitted and received through it.
To configure a node port as primary or secondary, use the following command:
configure eaps <name> [primary | secondary] port <ports>
The following command example adds port 1 of the module installed in slot 8 of the switch to the EAPS
domain “eaps_1” as the primary port.
configure eaps eaps_1 primary port 8:1
Messages Displayed when Adding EAPS Ring Ports to a VLAN

If you attempt to add EAPS ring ports to a VLAN that is not protected by EAPS, the switch prompts
you by default to confirm this action.
For example, if you use the configure vlan <vlan_name> add ports <port_list> command, and
the ports that you are attempting to add to the VLAN are currently used by EAPS as either primary or
secondary ring ports, the switch displays the following message:
Make sure <vlan_name> is protected by EAPS. Adding EAPS ring ports to a VLAN could
cause a loop in the network.
Do you really want to add these ports (y/n)
Enter y to add the ports to the VLAN. Enter n or press [Return] to cancel this action.
If you see this message, either configure the VLAN as an EAPS protected VLAN by using the
configure eaps add protected vlan command or add ports that the EAPS domain does not use as
primary or secondary ring ports.
Configuring the EAPS Control VLAN

You must configure one control VLAN for each EAPS domain. The control VLAN is used only to send
and receive EAPS messages.
NOTE
A control VLAN cannot belong to more than one EAPS domain. If the domain is active, you cannot delete the
domain or modify the configuration of the control VLAN.

To configure the EAPS control VLAN for the domain, use the following command:
configure eaps <name> add control vlan <vlan_name>
NOTE
The control VLAN must NOT be configured with an IP address. In addition, only ring ports may be added to this
control VLAN. No other ports can be members of this VLAN. Failure to observe these restrictions can result in a loop
in the network.
NOTE
The ring ports of the control VLAN must be tagged.
By default, EAPS PDUs are automatically assigned an 802.1p priority of seven. With a priority of seven,
EAPS PDUs take precedence within the EAPS topology thereby giving priority to EAPS traffic,
including EAPS control VLAN traffic and control VLAN messages. This ensures that the control VLAN
messages reach their intended destinations. You do not need to configure a QoS profile for the control
VLAN.
However, if you have a protected (data) VLAN with a QoS profile of QP8, you may need to create a
QoS profile of QP8 for the control VLAN. Whether you need to configure the control VLAN with a QoS
profile of QP8 depends entirely on your configuration.
NOTE
On the BlackDiamond 8800 original modules and the Summit X450 series switch, assigning a QoS profile to the
control VLAN affects the amount of available ACL masks. For more information see “Conserving ACL Masks and
Rules—BlackDiamond 8800 Original Series Modules and the Summit X450 Series Switches Only” on page 408.
The following command example adds the control VLAN “keys” to the EAPS domain “eaps_1”.
configure eaps eaps_1 add control vlan keys
Adding the EAPS Protected VLANs

You must add one or more protected VLANs for each EAPS domain. The protected VLANs are the data-
carrying VLANs.
NOTE
When you configure the VLAN that will act as a protected VLAN, the ring ports of the protected VLAN must be
tagged (except in the case of the default VLAN).
To add an EAPS protected VLAN, use the following command:
configure eaps <name> add protected vlan <vlan_name>
NOTE
As long as the ring is complete, the master node blocks the protected VLANs on its secondary port.

If the protected VLAN has a QoS profile of QP8, make sure the control VLAN has the appropriate QoS
profile setting. For more information, see the previous section “Configuring the EAPS Control VLAN”
on page 666.
The following command example adds the protected VLAN “orchid” to the EAPS domain “eaps_1.”
configure eaps eaps_1 add protected vlan orchid
Enabling and Disabling Fast Convergence

You enable Fast Convergence on the entire switch; this feature ensures convergence in less than 50
milliseconds.
To enable or disable Fast Convergence on the switch, use the following command:
configure eaps fast-convergence [off | on]
Enabling and Disabling an EAPS Domain

To enable a specific EAPS domain, use the following command:
enable eaps {<name>}
To disable a specific EAPS domain, use the following command:
disable eaps {<name>}
To prevent loops in the network, the switch displays by default the following warning message and
prompts you to disable EAPS for the specified domain:
WARNING: Disabling specific EAPS domain could cause a loop in the network!
Are you sure you want to disable this specific EAPS domain? (y/n)

● Enter y to disable EAPS for the specified domain.
Enabling and Disabling EAPS on the Switch

To enable the EAPS function for the entire switch, use the following command:
enable eaps
To disable the EAPS function for the entire switch, use the following command:
disable eaps

To prevent loops in the network, the switch displays by default the following warning message and
prompts you to disable EAPS for the entire switch:
WARNING: Disabling EAPS on the switch could cause a loop in the network!
Are you sure you want to disable EAPS? (y/n)

● Enter y to disable EAPS for the entire switch.
Unconfiguring an EAPS Ring Port

Unconfiguring an EAPS port sets its internal configuration state to INVALID, which causes the port to
appear in the Idle state with a port status of Unknown when you use the show eaps {<eapsDomain>}
{detail} command to display the status information about the port.
To unconfigure an EAPS primary or secondary ring port for an EAPS domain, use the following
command:
unconfigure eaps <eapsDomain> [primary | secondary] port
To prevent loops in the network, the switch displays by default a warning message and prompts you to
unconfigure the specified EAPS primary or secondary ring port. When prompted, do one of the
following:
● Enter y to unconfigure the specified port.
The following command example unconfigures this node’s EAPS primary ring port on the domain
“eaps_1”:
unconfigure eaps eaps_1 primary port
The switch displays the following warning message:
WARNING: Unconfiguring the Primary port from the EAPS domain could cause a loop in
the network!
Are you sure you want to unconfigure the Primary EAPS Port? (y/n)
Enter y to continue and unconfigure the EAPS primary ring port. Enter n to cancel this action.
The switch displays a similar warning message if you unconfigure the secondary EAPS port.

Disabling EAPS Loop Protection Warning Messages

The switch displays by default loop protection messages when configuring the following EAPS
parameters:
● Adding EAPS primary or secondary ring ports to a VLAN
● Deleting a protected VLAN
● Disabling the global EAPS setting on the switch
● Disabling an EAPS domain
● Configuring an EAPS domain as a transit node
● Unconfiguring EAPS primary or secondary ring ports from an EAPS domain
warning messages unnecessary. For example, if you use a script to configure your EAPS settings,
disabling the warning messages allows you to configure EAPS without replying to each interactive yes/
no question.
To disable loop protection messages, use the following command:
configure eaps config-warnings off
To re-enable loop protection messages, use the following command:
configure eaps config-warnings on
Displaying EAPS Status and Counter Information

You can display EAPS status, configuration, and counter information.
To display EAPS status and configuration information, use the following command:
show eaps {<eapsDomain>} {detail}
NOTE
You may see a slightly different display, depending on whether you display the master node or the transit node.
If you enter the show eaps command without a keyword, the command displays less than with the
detail keyword.
If you specify the optional domain eapsDomain parameter, the command displays status information for
a specific EAPS domain.
The display from the show eaps detail command shows all the information shown in the show eaps
<eapsDomain> command, but displays information for all configured EAPS domains.
Table 83 explains the fields on the EAPS display.

Table 83: show eaps display fields

Field Description
EAPS Enabled Current state of EAPS on this switch:
• Yes—EAPS is enabled on the switch.
• No—EAPS is not enabled.
EAPS Fast Convergence Displays only when Fast Convergence is on.
EAPS Display Config Displays the setting for loop protection messages:
Warnings
On—Loop protection messages are displayed (this is the default behavior).
Off—Loop protection messages are not displayed.
Number of EAPS instances Number of EAPS domains created. The maximum number of EAPS domains per
switch is 128.
Name The configured name for this EAPS domain.
State On a transit node, the command displays one of the following states:
• Idle—The EAPS domain has been enabled, but the configuration is not
complete.
• Links-Up—This EAPS domain is running, and both its ports are up and in the
forwarding state.
• Links-Down—This EAPS domain is running, but one or both of its ports are
down.
• Preforwarding—This EAPS domain is running, and both of its ports are up,
but one of them is in a temporary blocked state.
On a master node, the command displays one of the following states:
• Idle—The EAPS domain has been enabled, but the configuration is not
complete.
• Init—The EAPS domain has started but has not yet determined the status of
the ring. The secondary port is in a blocked state.
• Complete—The ring is in the complete state for this EAPS domain.
• Failed—There is a break in the ring for this EAPS domain.
• Pre-Init—The EAPS domain has started operation for Init state and has sent a
request to lower hardware layers to block the secondary port. It is in transient
state waiting for acknowledgement from hardware layer indicating the
operation is completed.
• Pre-Complete—The EAPS domain has started operation for Complete state
and has sent a request to lower hardware layers to block the secondary port.
It is in transient state waiting for acknowledgement from the hardware layer
indicating the operation is completed.
• [Failtimer Expired]—When the failtimer expires and its action is set to send-
alert, this flag is set. This flag indicates there is a misconfiguration or
hardware problem in the EAPS ring. The EAPS master node will continue to
remain in COMPLETE or INIT state with it’s secondary port blocking.
[Running: …] • Yes—This EAPS domain is running.
• No—This EAPS domain is not running.
Enabled Indicates whether EAPS is enabled on this domain:
• Y—EAPS is enabled on this domain.
• N—EAPS is not enabled.
Mode The configured EAPS mode for this switch: transit (T) or master (M).
Primary/Secondary port The port numbers assigned as the EAPS primary and secondary ports. On the
master node, the port distinction indicates which port is blocked to avoid a loop.

Table 83: show eaps display fields (Continued)

Field Description
Port status Indicates port status as one of the following states:
• Unknown—This EAPS domain is not running, so the port status has not yet
been determined.
• Up—The port is up and is forwarding data.
• Down—The port is down.
• Blocked—The port is up, but data is blocked from being forwarded.
Tag status Tagged status of the control VLAN:
• Tagged—The control VLAN has this port assigned to it, and the port is tagged
in the VLAN.
• Untagged—The control VLAN has this port assigned to it, but the port is
untagged in the control VLAN.
• Undetermined—Either a VLAN has not been added as the control VLAN to
this EAPS domain or this port has not been added to the control VLAN.
Hello Timer interval The configured value of the timer in seconds, specifying the time that the master
node waits between transmissions of health check packets.
Fail Timer interval The configured value of the timer in seconds, specifying the time that the master
node waits before the failtimer expires.
Failtimer expiry action Displays the action taken when the failtimer expires:
• Send-alert—Sends a critical message to the syslog when the failtimer expires.
• Open-secondary-port—Opens the secondary port when the failtimer expires.
Preforwarding Timer intervala The configured value of the timer. This value is set internally by the EAPS
software. The set value is 15 seconds.
Note: If two links in an EAPS domain go down at the same time and one link
comes back up, it will take 15 seconds for the reconnected link to start
receiving traffic again.
Last update Indicates the last time a hello packet is received from the master node.
EAPS Domain has … Lists the assigned name and ID of the control VLAN.
Controller Vlans
EAPS Domain has … Lists the assigned names and VLAN IDs of all the protected VLANs configured
Protected Vlansb on this EAPS domain.
Number of Protected Vlans The count of protected VLANs configured on this EAPS domain.
a. This field applies only to master nodes; it does not display for a transit mode.
b. These fields apply only to transit nodes; they are not displayed for a master node.
To display EAPS counter information, use the following command:
show eaps counters [<eapsDomain> | global]
If you specify the name of an EAPS domain, the switch displays counter information related to only
that domain.
If you specify the global keyword, the switch displays EAPS counter information when the events
counted are not applicable to any specific EAPS domain. The output displayed is calculated for all
configured EAPS domains, not just one specific EAPS domain.

Table 84 describes the significant fields and values in the display output of the show eaps counters
<eapsDomain> command:
Table 84: show eaps counters <eapsDomain> display

Field Description
Rx-Health Indicates the EAPS domain received EAPS Health PDUs.
Rx-RingUp-FlushFdb Indicates the EAPS ring is up, and the EAPS domain received EAPS
RingUp-FlushFdb PDUs to flush the FDB.
Rx-RingDown-FlushFdb Indicates the EAPS ring is down, and the EAPS domain received EAPS
RingDown-FlushFdb PDUs to flush the FDB.
Rx-Link-Down Indicates the EAPS domain received EAPS Link-Down PDUs and took
down the link.
Rx-Flush-Fdb Indicates the EAPS domain received EAPS Flush-Fdb PDUs and
flushed the FDB.
Rx-Suspend-Prefwd-Timer Indicates the EAPS domain received EAPS Suspend-Preforward-Timer
PDUs.
NOTE: Switches running ExtremeWare send this PDU during an MSM
failover. Switches running ExtremeXOS 10.1 or later do not send or
receive this PDU.
Rx-Query-Link-Status Indicates the EAPS domain received EAPS Query-Link-Status PDUs.
Rx-Link-Up Indicates the EAPS domain received EAPS Link-Up PDUs and brought
the link back up.
Rx-Unknown Indicates the EAPS domain dropped unknown EAPS PDUs.
Rx-Another-Master Indicates the EAPS domain dropped EAPS PDUs because there is
another Master switch in the same EAPS domain.
Rx-Unconfigured-Port Indicates the EAPS domain dropped EAPS PDUs because the ingress
port is not configured to be a ring port for the EAPS domain and the
corresponding control VLAN.
Rx-Health-Pdu-Pri-Port Indicates the EAPS domain dropped EAPS Health PDUs because the
primary port received them instead of the secondary port.
NOTE: The secondary port of the Master switch must receive EAPS
Health PDUs, not the primary port.
Tx-Health Indicates the EAPS domain sent EAPS Health PDUs.
Tx-RingUp-FlushFdb Indicates the EAPS ring is up, and the EAPS domain sent EAPS
RingUp-FlushFdb PDUs to flush the FDB.
Tx-RingDown-FlushFdb Indicates the EAPS ring is down, and the EAPS domain sent EAPS
RingDown-FlushFdb PDUs to flush the FDB.
Tx-Link-Down Indicates the EAPS domain sent EAPS Link-Down PDUs because the
link went down.
Tx-Flush-Fdb Indicates the EAPS domain sent EAPS Flush-Fdb PDUs because the
FDB needs to be flushed.
Tx-Suspend-Prefwd-Timer Indicates the EAPS domain sent EAPS Suspend-Preforward-Timer
PDUs.
NOTE: Switches running ExtremeWare send this PDU during an MSM
failover. Switches running ExtremeXOS 10.1 or later do not send or
receive this PDU. This counter should remain at 0.
Tx-Query-Link-Status Indicates the EAPS domain sent EAPS Query-Link-Status PDUs.
Tx-Link-Up Indicates the EAPS domain sent EAPS Link-Up PDUs and the link is
up.

Table 84: show eaps counters <eapsDomain> display (Continued)

Field Description
Tx-Unknown Indicates the number of unknown EAPS PDUs sent by the EAPS
domain.
NOTE: Unknown EAPS PDUs can be a new type of PDU that the
switch does not track in the sending routine.
Tx-Transmit-Err Indicates the number of EAPS PDUs the EAPS domain was unable to
send because of an error.
Fw-Link-Down Indicates the number of EAPS Link-Down PDUs received by the EAPS
domain and forwarded in slow path.
Fw-Flush-Fdb Indicates the number of EAPS Flush-Fdb PDUs received by the EAPS
domain and forwarded in slow path.
FW-Query-Link-Status Indicates the number of EAPS Query-Link-Status PDUs received by the
EAPS domain and forwarded in slow path.
Fw-Unknown Indicates the number of unknown EAPS PDUs forwarded in slow path.
switch does not track in the forwarding routine.
Fw-Transmit-Er Indicates the number of EAPS PDUs the EAPS domain was unable to
forward in slow path because of an error.
NOTE
Rx and Fw counters—If a PDU is received, processed, and consumed, only the Rx counter will increment. If a PDU
is forwarded in slow path, both the Rx counter and Fw counter will increment.
global command:
Table 85: show eaps counters global display

Field Description
Rx-Failed Indicates an error occurred when receiving packets from the Layer 2
forwarding engine.
Rx-Invalid-Vlan-Intf Indicates that the VLAN interface for the incoming VLAN cannot be
found.
Rx-Undersize-Pkt Indicates the length of the packet is less than the length of the
header.
Rx-Invalid-8021Q-Tag Indicates the VlanTypeLength field in the Ethernet header does not
match the default Ethernet value for the 802.1Q tag.
Rx-Invalid-SNAP-Type Indicates an invalid Subnetwork Access Protocol (SNAP) value in the
Ethernet header.
Rx-Invalid-OUI Indicates the Organizational Unique Identifier (OUI) value in the
Ethernet header does not match 00:E0:2B.
Rx-EEP-Unsupported-Version Indicates an unsupported Extreme Encapsulation Protocol (EEP)
version. The EEP version should be 1.
Rx-EEP-Invalid-Length Indicates the length of the EEP header is greater than the length of
the packet.
Rx-EEP-Checksum-Invalid Indicates the EEP checksum is invalid.

Configuring EAPS Shared Ports
Table 85: show eaps counters global display (Continued)

Field Description
Rx-Domain-Invalid Indicates the control VLAN’s incoming PDU is not associated with an
EAPS domain.
Rx-Lif-Invalid Indicates that EAPS is unable to determine the logical interface (LIF)
for the ingress port.
Rx-Lif-Down Indicates the LIF for the ingress port is in the Down state.
Tx-Failed Indicates an error occurred when sending packets to the Layer 2
forwarding engine.
Clearing the EAPS Counters

The counters continue to increment until you explicitly clear the information. By clearing the counters,
you can see fresh statistics for the time period you are monitoring. To clear, reset all of the counters
gathered by EAPS, use one of the following commands:
● clear counters
● clear eaps counters
NOTE
You must have a core or an advanced core license to use the EAPS shared port feature.
When multiple EAPS rings share a physical link, this link is referred to as the common link. Each node
on either end of this common link should be configured with an EAPS shared port instance. You must
configure one node of the common link as the controller and the node on the other end as the partner.
Each common link in the EAPS network must have a unique link ID. The controller and partner shared
ports belonging to the same common link must have matching link IDs. These controller and partner
nodes with matching link IDs must be directly connected to each other. No other instance in the
network should have that link ID. During normal operation, both the controller and the partner send
segment health check messages on their EAPS domain every second.
In the situation described above, if there is no EAPS shared ports configuration, and the common link
fails, this might result in a superloop.
Steady State
In steady state when the common link is up, both the controller and partner are said to be in the
“ready” state. After EAPS has converged and the EAPS master node has blocked its own secondary
ports, the controller puts all its ports into “forwarding,” and goes back to “ready” state.

Figure 73: Multiple EAPS domain steady state
S3 P3 S9
P2
S1 P4
Controller S6 EAPS3
EAPS1
P1
Common S10
S4 link ID=10 link EAPS2 S7
P5
S2 S8
P8
Partner
Master
S5 P6 S11
P7
Master Master EX_104
Figure 73 shows a multiple EAPS domain steady state, where:

● EAPS1 is the EAPS domain for ring S1, S3, S4, S5, and S2
● P1, P2, P3, and P4 are the ports on switch S1
● P5, P6, P7, and P8 are the ports on switch S2
● S5, S8, and S11 are the master nodes of their respective EAPS domains
● S3, S4, S6, S7, S9, and S10 are the transit nodes of their respective EAPS domains
● S1 and S2 are running EAPSv2
● S1 is the controller
● S2 is the partner
● P1 is the EAPS shared port on switch S1
● P5 is the EAPS shared port on switch S2
● Link ID 10 is the unique identifier for the common link
Common Link Failures

If a single common link fails, the configured controller (S1) and partner (S2) take steps to prevent a
superloop.
Assuming there is a single data VLAN configured on all three EAPS domains, the controller (S1) keeps
one port open (called “Active-Open”). The remaining segment ports are “blocked” to prevent a
superloop.
In Figure 74, P2 is the “Active-Open” port on S1. Ports P3 and P4 are “blocked.” The master nodes (S5,
S8, and S11) open their secondary ports.

Figure 74: EAPS domain common link failure
S3 P3
EAPS3
S9
P2
Active-Open
S1
S6
Controller P4
P1
S4 x EAPS2
S7 S10
EAPS1
S2
S8
Partner
Master
S5 S11
Master Master
EW_102b
When the common link is restored, the controller goes into Preforwarding state. After the controller
receives notification from the master nodes that they have converged and blocked their secondary ports,
the controller opens all ports.
If you have an EAPS configuration with multiple common links and a second common link fails, the
controllers continue to take steps to prevent a superloop. In addition to having one controller with an
“Active-Open” port, the controller with the smallest link ID becomes the “root blocker.” There can be
only one “root blocker” in the network.
Port Considerations
In an EAPS shared port configuration, the segment ports are sorted in ascending order based on their
port number, not the order you add an EAPS domain to EAPS shared ports. This is particularly useful
when planning your EAPS configuration.
The benefit of sorting ports in ascending order is evident if a common link fails. The port with the
lowest port number among the segment ports in the “UP” state becomes the “Active Open” port. When
configuring your EAPS domain, keep the port numbers in mind. For high bandwidth links, utilize
lower port numbers, and for low bandwidth links, utilize higher port numbers. This way, if a common
link fails, the high bandwidth link is still available.
NOTE
The order you add EAPS domains to EAPS shared ports is relevant if the EAPS domains have matching ring ports
and participate in spatial reuse. In this case, the show eaps shared-port {<port>} {detail} command
displays the newly added EAPS domain after all other existing EAPS domains with the same matching ring port.
To display EAPS shared port status information, use the following command:
show eaps shared-port {<port>} {detail}

Flushing the FDBs

When a controller goes into or out of the “blocking” state, the controller sends a “flush fdb” message to
flush all of the FDBs of the switches in its segments. Each switch in the path of the “flush fdb” message
flushes its FDB.
In a network with multiple EAPS ports in the blocking state, the “flush fdb” message gets propagated
across the boundaries of the EAPS domains.
Creating and Deleting a Shared Port

To configure a common link, you must create a shared port on each switch belonging to the common
link. To create a shared port, use the following command:
create eaps shared-port <ports>
Where ports is the common link port.
NOTE
A switch can have a maximum of two shared ports.
To delete a shared port on the switch, use the following command:
delete eaps shared-port <ports>
Defining the Mode of the Shared Port

The shared port on one end of the common link must be configured to be the controller. This is the end
responsible for blocking ports when the common link fails thereby preventing the superloop.
The shared port on the other end of the common link must be configured to be the partner. This end
does not participate in any form of blocking. It is responsible for only sending and receiving health-
check messages.
To configure the mode of the shared port, use the following command:
configure eaps shared-port <ports> mode <controller | partner>
Configuring the Link ID of the Shared Port

Each common link in the EAPS network must have a unique link ID. The controller and partner shared
ports belonging to the same common link must have matching link IDs. No other instance in the
network should have that link ID.
If you have multiple adjacent common links, Extreme Networks recommends that you configure the
link IDs in ascending order of adjacency. For example, if you have an EAPS configuration with three
adjacent common links, moving from left to right of the topology, configure the link IDs from the
lowest to the highest value.

To configure the link ID of the shared port, use the following command:
configure eaps shared-port <ports> link-id <id>
The link ID range is 1 to 65535.
Configuring the Shared Port Segment Timer

To configure the segment timer, use the following command:
configure eaps shared-port <ports> segment-timeout expiry-action [segment-down | send-
alert]

● segment-down—If the controller or partner switch’s segment timer expires, that segment is set to
“down,” and a query is not sent through the ring.
● send-alert—If the controller or partner switch’s segment timer expires, that switch keeps the
segment up, with the failed flag set, and sends a warning message to the log.
All segments, including the controller and partner shared ports belonging to the same common link,
must use the same segment timer expiry action. By default, the segment timer expiry action is send-alert
and the timer is set to 3 seconds.
Unconfiguring an EAPS Shared Port

To unconfigure a link ID on a shared port, use the following command:
unconfigure eaps shared-port <ports> link-id
To unconfigure the mode on a shared port, use the following command:
unconfigure eaps shared-port <ports> mode
To delete a shared port, use the following command:
delete eaps shared-port <ports>
Displaying EAPS Shared-Port Status and Counter Information

You can display EAPS shared port status, configuration, and counter information.
To display EAPS shared port status information, use the following command:
show eaps shared-port {<port>} {detail}
If you enter the show eaps shared-port command without an argument or keyword, the command
displays a summary of status information for all configured EAPS shared ports.
If you specify an EAPS shared-port, the command displays information about that specific port.
Otherwise, the command displays information about all of the shared-ports configured on the switch.
You can use the detail keyword to display more detailed status information about the segments and
VLANs associated with each shared port.

Table 86 describes the significant fields and values in the display output of the show eaps shared-
port {<port>} {detail} commands.
Table 86: show eaps shared-port display fields

Field Description
Shared Port Displays the port number of the shared port.
Mode Indicates whether the switch on either end of the common link is a
controller or partner. The mode is configured by the user.
Link ID The link ID is the unique common link identifier configured by the
user.
Up Displays one of the following states:
• Yes—Indicates that the link ID and the mode are configured.
• No—Indicates that the link ID or the mode is not configured.
State Displays one of the following states:
• Idle—Shared-port instance is not running.
• Ready—The EAPS shared-port instance is running, the neighbor
can be reached, and the common link is up.
• Blocking—The EAPS shared-port instance is running, the neighbor
cannot be reached, or the common link is down.
• Preforwarding—The EAPS shared-port instance is in a blocking
state, and the common link came up. To prevent a superloop, a
temporary blocking state is created before going into Ready state.
Domain Count Indicates the number of EAPS domains sharing the common link.
VLAN Count Indicates the total number of VLANs that are protected under the
EAPS domains sharing this common link.
Nbr Displays one of the following states:
• Yes—Indicates that the EAPS instance on the other end of the
common link is configured with matching link ID and opposite
modes. For example, if one end of the common link is configured
as a controller, the other end must be configured as a partner.
• Err—Indicates that the EAPS instance on the other end of the
common link is configured with a matching link ID, but the modes
are configured the same (for example, both modes are configured
as controller, or both modes are configured as partner.
• No—Indicates one or more of the following:
- The switch on the other end of the common link is not running.
- The shared port has not been created.
- The link IDs on each side of the common link do not match.
- The common link, and any other segment between the controller
and partner are not fully connected.
RB State Displays one of the following states:
• None—This EAPS shared-port is not the “root blocker.”
• Active—This EAPS shared-port is the “root blocker” and is
currently active.
• Inactive—This EAPS shared-port is the “root blocker” but is
currently inactive.
RB ID The ID of the root blocker. If the value is none, there are not two or
more common-link failures.

Table 86: show eaps shared-port display fields (Continued)

Field Description
Active Open (available with the • None—Indicates that there is no Active-Open port on the VLAN.
detail keyword)
• Port #—Indicates the port that is Active-Open and is in a
forwarding state.
Segment Timer expiry action • Segment down—Specifies that if the controller or partner switch
detect a down segment, that segment stays down and a query is
not sent through the ring. The switch marks the segment status as
"Down."
• Send alert—Specifies that if the controller or partner switch detect
a down segment, that switch keeps the segment up and sends a
warning message to the log (default). The switch sends a trap alert
and sets the failed flag [F].
Segment Port (available with the The segment port is the other ring port of an EAPS domain that is not
detail keyword or by specifying a the shared-port.
shared port)
Status (available with the detail • Up—There is connectivity to the neighboring EAPS shared-port via
keyword or by specifying a shared port) this port.
• Down—There is a break in the path to the neighboring EAPS
shared-port via this port.
• Blocking-Up—The path is Up, but due to the "root blocker" being
in the Active state, this port is blocked to prevent a loop.
• Blocking-Down—The “root blocker” is in the Active State; however,
the path is Down. Because the path is Down, there is no need to
block the “root blocker” port to prevent a loop.
• [F]—The segment timer has expired but has not received an
explicit link-down notification. The port remains in the Up state,
with the timer expired flag set to True.
EAPS Domain (available with the The EAPS domain having the segment port as one of its ring ports.
detail keyword or by specifying a
shared port)
Vlan-port count (available with the The total number of VLANs being protected under this segment port.
detail keyword or by specifying a
shared port)
Adjacent Blocking Id (available with • None—The neighbor on this port is not reporting a Controller in the
the detail keyword or by specifying a Blocking state.
shared port)
• <Link-Id>—The neighbor on this port is a controller in the Blocking
state with a link ID of <Link-Id>.
Segment RB Id (available with the • None—The neighbor on this port is not aware of a "root blocker" in
detail keyword or by specifying a the network.
shared port)
• <RB-Id>—The neighbor on this port has determined that there is a
"root blocker" in the network with a link ID of <RB-Id>.
Vlan (available with the detail Displays a list of VLANs protected by the segment port.
keyword or by specifying a shared port)

Table 86: show eaps shared-port display fields (Continued)

Field Description
Virtual-port Status (available with the This information appears for the Controller, when it is in either the
detail keyword or by specifying a Blocking or Preforwarding state.
shared port)
• Active-Open—This VLAN or port is in the Forwarding state and has
connectivity to the neighboring EAPS shared port via this port.
• Open—This VLAN or port is in the Forwarding state but does not
have connectivity to the neighboring EAPS shared port via this port.
• Blocked—This VLAN or port is in the Blocking state to prevent a
• Down—This port’s link is down.
• Active—At this moment, this VLAN or port is not being handled by
EAPS shared port. Rather, this VLAN or port is being handled by
the regular EAPS protocol.
To display EAPS shared port counter information, use the following command:
show eaps counters shared-port [global | <port> {segment-port <segport>

{<eapsDomain>}}]
If you specify the global keyword, the switch displays general counter information for all configured
EAPS shared port instances. The output displayed is calculated for all configured EAPS shared ports;
not just one specific shared port instance.
If you specify a particular EAPS shared port, the switch displays counter information related to only
that shared port.
If you specify a particular EAPS segment port, the switch displays counter information related to only
that segment port for the specified EAPS domain.
shared-port global command:
Table 87: show eaps counters shared-port global display

Field Description
Rx-Invalid-Instance Displays the number of dropped EAPS shared-port PDUs because
there is not a valid EAPS shared port instance for the incoming port.
Rx-Unknown Displays the number of unknown EAPS PDUs dropped by the shared
port instances.
Fw-Invalid-Instance Displays the number of EAPS shared-port PDUs that could not be
forwarded in slow path because the shared port instances could not
find a valid EAPS shared port instance for the outgoing port.

shared-port <port> command:
Table 88: show eaps counters shared-port <port> display

Field Description
Rx-Seg-Health Indicates the shared port instance received EAPS shared ports
Segment-Health-Check PDUs.
Rx-Path-Detect Indicates the shared port instance received EAPS shared ports Path-
Detect PDUs.
Rx-Flush-Notify Indicates the shared port instance received EAPS shared ports Flush-
Notify PDUs and flushed the FDB.
If this PDU reaches a port of the shared ports pair that initiated the
PDU, the shared port instance may terminate the PDU. Otherwise, the
shared port instance forwards the PDU.
Rx-Unknown Displays the number of unknown EAPS PDUs dropped by the shared
port instance.
Rx-Seg-Health-Dropped Displays the number of EAPS shared ports Segment-Health-Check
PDUs dropped by the shared port instance.
This counter increments if the Segment-Health-Check PDU returns to
the sending switch. If that occurs, the switch drops the Segment-
Health-Check PDU.
Rx-Path-Detect-Dropped Displays the number of EAPS shared ports Path-Detect PDUs dropped
by the shared port instance.
This counter increments in the following situations:
• If the packet’s Fwd-id matches the EAPS shared port’s Link-Id, the
port is not in the blocking state, and the incoming port is a
segment port.
• If the packet’s Link-Id matches the EAPS shared port’s Link-Id, the
port is not in the blocking state, and the incoming port is a
segment port.
Rx-Flush-Notify-Dropped Displays the number of EAPS shared ports Flush-Notify-Dropped PDUs
dropped by the shared port instance.
This counter increments in the following situations:
• If the Flush-Notify-Dropped PDU returns to the sending switch.
• If the packet’s Fwd-Id matches the EAPS shared port’s Link-Id and
the port is not in the blocking state.
Rx-Dropped-Invalid-Port Displays the number of EAPS shared ports PDUs dropped by the
shared port instance because it does not exist.
Tx-Seg-Health Indicates the shared port instance sent EAPS shared ports Segment-
Health-Check PDUs.
Tx-Path-Detect Indicates the shared port instance sent EAPS shared ports Path-Detect
PDUs.
NOTE: This counter appears under Common Link Port Stats and
should always be 0.
Tx-Flush-Notify Indicates the shared port instance sent EAPS shared ports Flush-
Notify PDUs to flush the FDB.
should always be 0.

Table 88: show eaps counters shared-port <port> display (Continued)

Field Description
Tx-Flush-Fdb Indicates the shared port instance sent EAPS Flush-Fdb PDUs
because the FDB needs to be flushed.
should always be 0.
Tx-Unknown Indicates the number of unknown EAPS PDUs sent by the shared port
instance.
switch does not track in the sending routine.
Tx-Transmit-Err Indicates the number of EAPS PDUs the shared port instance was
unable to send because of an error.
Fw-Seg-Health Indicates the number of EAPS shared ports Segment-Health-Check
PDUs received by the shared port instance and forwarded in slow
path.
Fw-Path-Detect Indicates the number of EAPS shared ports Path-Detect PDUs received
by the shared port instance and forwarded in slow path.
Fw-Flush-Notify Indicates the number of EAPS Flush-Notify PDUs received by the
shared port instance and forwarded in slow path to flush the FDB.
Fw-Flush-Fdb Indicates the number of EAPS Flush-Fdb PDUs received by the shared
port instance and forwarded in slow path.
Fw-Unknown Indicates the number of unknown EAPS PDUs forwarded in slow path.
switch does not track in the forwarding routine.
Fw-Transmit-Err Indicates the number of EAPS PDUs the shared port instance was
unable to forward in slow path because of an error.
Clearing the EAPS Counters

The counters continue to increment until you explicitly clear the information. By clearing the counters,
you can see fresh statistics for the time period you are monitoring. To clear, reset all of the counters
gathered by EAPS, use one of the following commands:
● clear counters
● clear eaps counters
EAPS Shared Port Configuration Rules

The following rules apply to EAPS shared port configurations:
● The controller and partner shared ports on either side of a common link must have the same
link ID.
● Each common link in the network must have a unique link ID.
● The modes on either side of a common link must be different from each other; one must be a
controller and one must be a partner.
● There can be only up to two shared ports per switch.

EAPS Shared Port Configuration Examples
● There cannot be more than one controller on a switch.

Valid combinations on any one switch are:
■ 1 controller
■ 1 partner
■ 1 controller and 1 partner
■ 2 partners
● A shared port cannot be configured on an EAPS master’s secondary port.
● You must have a core or an advanced core license to use the EAPS shared port feature.

This section provides examples of EAPS shared port configurations.
Basic Configuration
This example, shown in Figure 75, is the most basic configuration; two EAPS domains with a single
common link between them.
Figure 75: EAPS shared port basic configuration
S4 S5 S6
S3 Controller S7
EAPS1 EAPS2
link ID=1 Common
link
S2 Partner S8
P S S P
S1 S 10 S9
Master Master
node node EW_095

Basic Core Configuration

This configuration, shown in Figure 76, shows a core with access rings. In this topology, there are two
EAPS common links.
Figure 76: EAPS shared port basic core configuration
Master
S node
S4 S7 S 12 P
P1:2 P1:3
S5 S9
S3 Controller Controller EAPS3
EAPS1 EAPS2 S
P1:1
Common link link ID=1 Common link link ID=2 S 11
P Master
S2 S6 S 10 node
Partner Partner
P S
S1 S P EAPS4
S8 S 13
Master EW_096
node Master
node
Right Angle Configuration

In this topology, there are still two EAPS common links, but the common links are adjacent to each
other. To configure a right angle configuration, there must be two common links configured on one of
the switches. Figure 77 shows a Right Angle configuration.
Figure 77: EAPS shared port right angle configuration
Master
S
node
S5 S6
EAPS1
P
S7 S 10
link ID=2 Partner
S 4 Controller
Common Controller EAPS4
link
link ID=1 Master
EAPS2 Common S node
S3 link
EAPS3 S9
P
Partner P S
S2 S
S1
Master S8 P S 11
node EW_097
Master
node

Combined Basic Core and Right Angle Configuration

Figure 78 shows a combination Basic Core and Right Angle configuration.
Figure 78: Basic core and right angle configuration
Master
node S
S7 S8
S4 EAPS5
S5 P S9
Partner link ID=3 Controller
EAPS4 S 14
S3 Controller Common Partner
link EAPS3
EAPS1
Master
link ID=1 S node
P Common Common link ID=2
Master link link
node S 13
S2
P
Partner Controller
S EAPS2
S
S1 S6 P S 10 S 15
S 12 S S 11 P Master
node EW_098
Master
node
Large Core and Access Rings Configuration

Figure 79 shows a single large core ring with multiple access rings hanging off of it. This is an extension
of a basic core configuration.
Figure 79: Large core and access ring configuration

Master
node
P S
S4
S
S3 S5
EAPS3
S2 Controller Partner S6
link ID=40 Master
Partner Controller
S EAPS5
EAPS4 S
link ID=30 link ID=50
EAPS1
Master Master
node S1 S7 node
P
Controller Partner P
link ID=20
S 12 Controller Partner S8
EAPS2
S 11 S9
P
S S 10
Master EW_099
node

Advanced Configuration
Figure 80 shows an extension of the Basic Core and Right Angle configuration.
Figure 80: Advanced configuration
S3 S4 S S5
S Partner
Master Controller Master Controller
node
S2 link ID=1
link ID=2 EAPS2 EAPS3 S
link ID=4
Common
P Common
EAPS5 link Common
link
link S6
EAPS1
Common
S1 link
Partner Partner
Controller Partner
Master link ID=3 Controller
S 14 S 13 S 12 EAPS6 S7
S S9 S8
S
EAPS4
S 11 S 10
EW_101

26 Spanning Tree Protocol

● Overview of the Spanning Tree Protocol on page 689
● Spanning Tree Domains on page 693
● STP Configurations on page 703
● Per VLAN Spanning Tree on page 709
● Rapid Spanning Tree Protocol on page 709
● Multiple Spanning Tree Protocol on page 721
● STP Rules and Restrictions on page 732
● Configuring STP on the Switch on page 732
● Displaying STP Settings on page 739
Using the Spanning Tree Protocol (STP) functionality of the switch makes your network more fault
tolerant. The following sections explain more about STP and the STP features supported by
ExtremeXOS.
NOTE
STP is a part of the 802.1D bridge specification defined by the IEEE Computer Society. To explain STP in terms
used by the IEEE 802.1D specification, the switch will be referred to as a bridge.
ExtremeXOS 12.0 supports the new edition of the IEEE 802.1D standard (known as IEEE 802.1D-2004)
for STP, which incorporates enhancements from the IEEE 802.1t-2001, IEEE 802.1W, and IEEE 802.1y
standards. The IEEE 802.1D-2004 standard is backward compatible with the IEEE 802.1D-1998 standard.
For more information, see “Compatibility Between 802.1D-1998 and 802.1D-2004 STP Bridges” on
page 690.
Overview of the Spanning Tree Protocol

STP is a bridge-based mechanism for providing fault tolerance on networks. STP allows you to
implement parallel paths for network traffic and to ensure that redundant paths are:
● Disabled when the main paths are operational.
● Enabled if the main path fails.
NOTE
STP and Extreme Standby Router Protocol (ESRP) cannot be configured on the same Virtual LAN (VLAN)
simultaneously.

Spanning Tree Protocol
Compatibility Between 802.1D-1998 and 802.1D-2004 STP

Bridges
The IEEE 802.1D-2004 compliant bridges interoperate with the IEEE 802.1D-1998 compliant bridges. To
ensure seamless operation of your STP network, read this section before you configure STP on any
Extreme Networks device running ExtremeXOS 11.6 or later.
Differences in behavior between the two standards include the:

● Default port path cost
● Bridge priority
● Port priority
● Edge port behavior
This section describes the bridge behavior differences in more detail.
Default Port Path Cost

The 802.1D-2004 standard modified the default port path cost value to allow for higher link speeds. A
higher link speed can create a situation whereby an 802.1D-1998 compliant bridge could become the
more favorable transit path.
For example, in Figure 81, bridge A is the root bridge running the new 802.1D-2004 standard, bridges B
and C are running the old 802.1D-1998 standard, and bridges D, E, and F are running the new 802.1D-
2004 standard. In addition, all ports are 100 Mbps links. The ports on bridges B and C have a default
path cost of 19, and the ports on bridge A, D, E, and F have a default path cost of 200,000.
If you use the default port path costs, bridge D blocks its port to bridge E, and all traffic between
bridges D and E must traverse all of bridges in the network. Bridge D blocks its port to bridge E
because the path cost to the root bridge is less by going across bridges B and C (with a combined root
cost of 38) compared with going across bridge E (with a root cost of 200,000). In fact, if there were 100
bridges between bridges B, C, and D running the old 802.1D-1998 standard with the default port path
costs, bridge D would still use that path because the path cost is still higher going across bridge E.

Overview of the Spanning Tree Protocol
Figure 81: 802.1D-1998 and 802.1D-2004 mixed bridge topology
Switch A Root
bridge
Switch F
Switch B
Switch E
Switch C
Blocked
Switch D
IEEE 802.1D-1998
IEEE 802.1D-2004
EX_179
As a workaround and to prevent this situation, configure the port path cost to make links with the same
speed use the same path host value. In the example described above, configure the port path cost for
the 802.1D-2004 compliant bridges (bridges A, D, E, and F) to 19.
NOTE
You cannot configure the port path cost on bridges B and C to 200,000 because the path cost range setting for
802.1D-1998 compliant bridges is 1 to 65,535.
To configure the port path cost, use the following command:
configure stpd <stpd_name> ports cost [auto | <cost>] <port_list>
Bridge Priority
By configuring the STPD bridge priority, you make the bridge more or less likely to become the root
bridge. Unlike the 802.1D-1998 standard, the 802.1D-2004 standard restricts the bridge priority to a 16-
bit number that must be a multiple of 4,096. The new priority range is 0 to 61,440 and is subject to the
multiple of 4,096 restriction. The old priority range was 0 to 65,535 and was not subject to the multiple
of 4,096 restriction (except for MSTP configurations). The default bridge priority remains the same at
32,768.
If you have an ExtremeXOS 11.5 or earlier configuration that contains an STP or RSTP bridge priority
that is not a multiple of 4,096, the switch rejects the entry and the bridge priority returns to the default
value while loading the structure. The MSTP implementation in ExtremeXOS already uses multiples of
4,096 to determine the bridge priority.

To configure the bridge priority, use the following command:
configure stpd <stpd_name> priority <priority>
For example, to lower the numerical value of the priority (which gives the priority a higher precedence),
you subtract 4,096 from the default priority: 32,768 - 4,096 = 28,672. If you modify the priority by a
value other than 4,096, the switch automatically changes the priority to the lower priority value. For
example, if you configure a priority of 31,000, the switch automatically changes the priority to 28,672.
Port Priority
The port priority value is always paired with the port number to make up the 16-bit port identifier,
which is used in various STP operations and the STP state machines. Unlike the 802.1D-1998 standard,
the 802.1D-2004 standard uses only the four most significant bits for the port priority and it must be a
multiple of 16. The new priority range available is 0 to 240 and is subject to the multiple of 16
restriction. The 802.1D-1998 standard uses the eight most significant bits for the port priority. The old
priority range was 0 to 31 and was not subject to the multiple of 16 restriction.
To preserve backward compatibility and to use ExtremeXOS 11.5 or earlier configurations, the existing
configure stpd ports priority command is available in ExtremeXOS 11.6 and later. If you have an
ExtremeXOS 11.5 or earlier configuration, the switch interprets the port priority based on the 802.1D-
1998 standard. If the switch reads a value that is not supported in ExtremeXOS 11.6 or later, the switch
rejects the entry.
ExtremeXOS 11.6 introduced support for a new ports priority command: configure stpd ports
port-priority. When you save the port priority value in an ExtremeXOS 11.6 or later configuration,
the switch saves it as the new command configure stpd ports port-priority with the
corresponding change in value.
For example, if the switch reads the configure stpd ports priority 16 command from an
ExtremeXOS 11.5 or earlier configuration, (which is equivalent to the command configure stpd
ports priority 8 entered through CLI), the switch saves the value in an ExtremeXOS 11.6 or later
configuration as configure stpd ports port-priority 128.
Edge Port Behavior

In ExtremeXOS 11.5 or earlier, Extreme Networks had two edge port implementations: edge port and
edge port with safeguard. The 802.1D-2004 standard has a bridge detection state machine, which
introduced a third implementation of edge port behavior in ExtremeXOS 11.6. The following list
describes the behaviors of the different edge port implementations:
● Edge port (ExtremeXOS 11.5 and earlier):
■ The port does not send BPDUs
■ The port does not run a state machine
■ If BPDUs are received, the port discards the BPDU and enters the blocking state
■ If subsequent BPDUs are not received, the port remains in the forwarding state
● Edge port with safeguard configured (ExtremeXOS 11.5 and 11.4 only):
■ The port sends BPDUs
■ When configured for MSTP, the port runs a partial state machine
■ If BPDUs are received, the port enters the blocking state
■ If subsequent BPDUs are not received, the port attempts to enter the forwarding state

Spanning Tree Domains
● Edge port running 802.1D-2004 with edge-guard enabled (ExtremeXOS 11.6 or later):
■ The port sends BPDUs
■ The port runs a state machine
■ If BPDUs are received, the port behaves as a normal RSTP port by entering the forwarding state
and participating in RSTP
■ If subsequent BPDUs are not received, the port attempts to become the edge port again
Edge port with safeguard was introduced in ExtremeXOS 11.4 to prevent accidental or deliberate
misconfigurations (loops) by having edge ports enter the blocking state upon receiving a BPDU. The
802.1D-2004 standard implements a bridge detection mechanism that causes an edge port to transition
to a non-edge port upon receiving a BPDU; however, if the former edge port does not receive any
subsequent BPDUs during a pre-determined interval, the port attempts to become an edge port.
If an 802.1D-2004 compliant safeguard port (edge port) connects to an 802.1D-1998 compliant edge port
with safeguard configured, the old safeguard port enters the blocking state. Although the new
safeguard port becomes a designated port, the link is not complete (and thus no loop is formed)
because one side of the link is blocked.

The switch can be partitioned into multiple virtual bridges. Each virtual bridge can run an independent
Spanning Tree instance. Each Spanning Tree instance is called a Spanning Tree Domain (STPD). Each
STPD has its own root bridge and active path. After an STPD is created, one or more VLANs can be
assigned to it.
A physical port can belong to multiple STPDs. In addition, a VLAN can span multiple STPDs.
The key points to remember when configuring VLANs and STP are:
● Each VLAN forms an independent broadcast domain.
● STP blocks paths to create a loop-free environment.
● Within any given STPD, all VLANs belonging to it use the same spanning tree.
To create an STPD, use the following command:
create stpd <stpd_name>
To delete an STPD, use the following command:
delete stpd <stpd_name>
For detailed information about configuring STP and various STP parameters on the switch, see
“Configuring STP on the Switch” on page 732.

● Member VLANs on page 694
● STPD Modes on page 695
● Encapsulation Modes on page 696
● STP States on page 697
● Binding Ports on page 698

● Rapid Root Failover on page 700

● STPD BPDU Tunneling on page 700
● STP and Hitless Failover—Modular Switches Only on page 702
Member VLANs
When you add a VLAN to an STPD, that VLAN becomes a member of the STPD. The two types of
member VLANs in an STPD are:
● Carrier
● Protected
Carrier VLAN
A carrier VLAN defines the scope of the STPD, which includes the physical and logical ports that
belong to the STPD and if configured, the 802.1Q tag used to transport Extreme Multiple Instance
Spanning Tree Protocol (EMISTP) or Per VLAN Spanning Tree (PVST+) encapsulated bridge protocol
data units (BPDUs) (see “Encapsulation Modes” on page 696 for more information about encapsulating
STP BPDUs). Only one carrier VLAN can exist in a given STPD, although some of its ports can be
outside the control of any STPD at the same time.
If you configure EMISTP or PVST+, the StpdID must be identical to the VLANid of the carrier VLAN in
that STPD. See “Specifying the Carrier VLAN” on page 695 for an example.
If you have an 802.1D configuration, Extreme Networks recommends that you configure the StpdID to
be identical to the VLANid of the carrier VLAN in that STPD. See “Basic 802.1D Configuration
Example” on page 734 for an example.
If you configure Multiple Spanning Tree (MSTP—IEEE 802.1Q-2003, formerly IEEE 802.1s) you do not
need carrier VLANs for MSTP operation. With MSTP, you configure a Common and Internal Spanning
Tree (CIST) that controls the connectivity of interconnecting MSTP regions and sends BPDUs across the
regions to communicate the status of MSTP regions. All VLANs participating in the MSTP region have
the same privileges. For more information about MSTP, see “Multiple Spanning Tree Protocol” on
page 721.
Protected VLAN
Protected VLANs are all other VLANs that are members of the STPD. These VLANs “piggyback” on
the carrier VLAN. Protected VLANs do not transmit or receive STP BPDUs, but they are affected by
STP state changes and inherit the state of the carrier VLAN. Protected VLANs can participate in
multiple STPDs, but any particular port in the VLAN can belong to only one STPD. Also known as non-
carrier VLANs.
If you configure MSTP, all member VLANs in an MSTP region are protected VLANs. These VLANs do
not transmit or receive STP BPDUs, but they are affected by STP state changes communicated by the
CIST to the MSTP regions. Multiple spanning tree instances (MSTIs) cannot share the same protected
VLAN; however, any port in a protected VLAN can belong to multiple MSTIs. For more information
about MSTP, see “Multiple Spanning Tree Protocol” on page 721.

Specifying the Carrier VLAN

● Creates and enables an STPD named s8.
● Creates a carrier VLAN named v5.
● Assigns VLAN v5 to STPD s8.
● Creates the same tag ID for the VLAN and the STPD (the carrier VLAN’s VLANid must be identical
to the STPDs StpdID).
create vlan v5
configure vlan v5 tag 100
configure vlan v5 add ports 1:1-1:20 tagged
create stpd s8
configure stpd s8 add vlan v5 ports all emistp
configure stpd s8 tag 100
enable stpd s8
Notice how the tag number for the VLAN v5 and the STPD s8 is identical (the tag is 100). By using
identical tags, you have selected the carrier VLAN. The carrier VLAN’s VLANid is identical to the
STPD’s StpdID.
STPD Modes
An STPD has three modes of operation:
● 802.1D mode
Use this mode for backward compatibility with previous STP versions and for compatibility with
third-party switches using IEEE standard 802.1D. When configured in this mode, all rapid
configuration mechanisms are disabled.
● 802.1w mode
Use this mode for compatibility with Rapid Spanning Tree (RSTP). When configured in this mode,
all rapid configuration mechanisms are enabled. The benefit of this mode is available on point-to-
point links only and when the peer is likewise configured in 802.1w mode. If you do not select
point-to-point links and the peer is not configured for 802.1w mode, the STPD fails back to 802.1D
mode.
You enable or disable RSTP on a per STPD basis only. You do not enable RSTP on a per port basis.
For more information about RSTP and RSTP features, see “Rapid Spanning Tree Protocol” on
page 709.
● MSTP mode
Use this mode for compatibility with MSTP. MSTP is an extension of RSTP and offers the benefit of
better scaling with fast convergence. When configured in this mode, all rapid configuration
mechanisms are enabled. The benefit of MSTP is available only on point-to-point links and when
you configure the peer in MSTP or 802.1w mode. If you do not select point-to-point links and the
peer is not configured in 802.1w mode, the STPD fails back to 802.1D mode.
You must first configure a CIST before configuring any MSTIs in the region. You cannot delete or
disable a CIST if any of the MSTIs are active in the system.
You create only one MSTP region on the switch, and all switches that participate in the region must
have the same regional configurations. You enable or disable an MSTP on a per STPD basis only.
You do not enable MSTP on a per port basis.

If configured in MSTP mode, an STPD uses the 802.1D BPDU encapsulation mode by default. To
ensure correct operation of your MSTP STPDs, do not configure EMISTP or PVST+ encapsulation
mode for MSTP STPDs.
For more information about MSTP and MSTP features, see “Multiple Spanning Tree Protocol” on
page 721.
By default, the:
● STPD operates in 802.1D mode.
● Default device configuration contains a single STPD called s0.
● Default VLAN is a member of STPD s0 with autobind enabled.
To configure the mode of operation of an STPD, use the following command:
configure stpd <stpd_name> mode [dot1d | dot1w | mstp [cist | msti <instance>]]
All STP parameters default to the IEEE 802.1D values, as appropriate.
Encapsulation Modes
You can configure ports within an STPD to accept specific BPDU encapsulations. This STP port
encapsulation is separate from the STP mode of operation. For example, you can configure a port to
accept the PVST+ BPDU encapsulation while running in 802.1D mode.
An STP port has three possible encapsulation modes:

● 802.1D mode
Use this mode for backward compatibility with previous STP versions and for compatibility with
third-party switches using IEEE standard 802.1D. BPDUs are sent untagged in 802.1D mode. Because
of this, any given physical interface can have only one STPD running in 802.1D mode.
This encapsulation mode supports the following STPD modes of operation: 802.1D, 802.1w, and
MSTP.
● Extreme Multiple Instance Spanning Tree Protocol (EMISTP) mode
EMISTP mode is proprietary to Extreme Networks and is an extension of STP that allows a physical
port to belong to multiple STPDs by assigning the port to multiple VLANs. EMISTP adds significant
flexibility to STP network design. BPDUs are sent with an 802.1Q tag having an STPD instance
Identifier (StpdID) in the VLANid field.
This encapsulation mode supports the following STPD modes of operation: 802.1D and 802.1w.
● Per VLAN Spanning Tree (PVST+) mode
This mode implements PVST+ in compatibility with third-party switches running this version of
STP. The STPDs running in this mode have a one-to-one relationship with VLANs and send and
process packets in PVST+ format.
This encapsulation mode supports the following STPD modes of operation: 802.1D and 802.1w.
These encapsulation modes are for STP ports, not for physical ports. When a physical port belongs to
multiple STPDs, it is associated with multiple STP ports. It is possible for the physical port to run in
different modes for different domains to which it belongs.
If configured in MSTP mode, an STPD uses the 802.1D BPDU encapsulation mode by default. To ensure
correct operation of your MSTP STPDs, do not configure EMISTP or PVST+ encapsulation mode for
MSTP STPDs.

To configure the BPDU encapsulation mode for one or more STP ports, use the following command:
configure stpd <stpd_name> ports mode [dot1d | emistp | pvst-plus] <port_list>
To configure the default BPDU encapsulation mode on a per STPD basis, use the following command:
configure stpd <stpd_name> default-encapsulation [dot1d | emistp | pvst-plus]
Instead of accepting the default encapsulation modes of dot1d for the default STPD s0 and emistp for
all other STPDs, this command allows you to specify the type of BPDU encapsulation to use for all
ports added to the STPD (if not otherwise specified).
STPD Identifier
An StpdID is used to identify each STP domain. You assign the StpdID when configuring the domain,
and that carrier VLAN of that STPD cannot belong to another STPD. Unless all ports are running in
802.1D mode, an STPD with ports running in either EMISTP mode or PVST+ mode must be configured
with an StpdID.
An StpdID must be identical to the VLANid of the carrier VLAN in that STP domain. For an 802.1D
STPD, the VLANid can be either a user-defined ID or one automatically assigned by the switch.
NOTE
If an STPD contains at least one port not in 802.1D mode, you must configure the STPD with an StpdID.
MSTP uses two different methods to identify the STPDs that are part of the MSTP network. An instance
ID of 0 identifies the CIST. The switch assigns this ID automatically when you configure the CIST
STPD. An MSTI identifier (MSTI ID) identifies each STP domain that is part of an MSTP region. You
assign the MSTI ID when configuring the STPD that participates in the MSTP region. In an MSTP
region, MSTI IDs only have local significance. You can reuse MSTI IDs across MSTP regions. For more
information about MSTP and MSTP features, see “Multiple Spanning Tree Protocol” on page 721.
STP States
Each port that belongs to a member VLAN participating in STP exists in one of the following states:
● Blocking
A port in the blocking state does not accept ingress traffic, perform traffic forwarding, or learn MAC
source addresses. The port receives STP BPDUs. During STP initialization, the switch always enters
the blocking state.
● Listening
A port in the listening state does not accept ingress traffic, perform traffic forwarding, or learn MAC
source addresses. The port receives STP BPDUs. This is the first transitional state a port enters after
being in the blocking state. The bridge listens for BPDUs from neighboring bridge(s) to determine
whether the port should or should not be blocked.
● Learning
A port in the learning state does not accept ingress traffic or perform traffic forwarding, but it begins
to learn MAC source addresses. The port also receives and processes STP BPDUs. This is the second
transitional state after listening. From learning, the port will change to either blocking or forwarding.

● Forwarding
A port in the forwarding state accepts ingress traffic, learns new MAC source addresses, forwards
traffic, and receives and processes STP BPDUs.
● Disabled
A port in the disabled state does not participate in STP; however, it will forward traffic and learn
new MAC source addresses.
Binding Ports
The two ways to bind (add) ports to an STPD are: manually and automatically. By default, ports are
manually added to an STPD.
NOTE
The default VLAN and STPD S0 are already on the switch.
Manually Binding Ports

To manually bind ports, use one of the following commands:
● configure stpd <stpd_name> add vlan <vlan_name> ports [all | <port_list>] {[dot1d
| emistp | pvst-plus]}
● configure vlan <vlan_name> add ports [all | <port_list>] {tagged | untagged} stpd
<stpd_name> {[dot1d | emistp | pvst-plus]}
The first command adds all ports or a list of ports within the specified VLAN to an STPD. For EMISTP
and PVST+, the carrier VLAN must already exist on the same set of ports. The second command adds
all ports or a list of ports to the specified VLAN and STPD at the same time. If the ports are added to
the VLAN but not to the STPD, the ports remain in the VLAN.
For EMISTP and PVST+, if the specified VLAN is not the carrier VLAN and the specified ports are not
bound to the carrier VLAN, the system displays an error message. If you configure MSTP on your
switch, MSTP does not need carrier VLANs.
NOTE
The carrier VLAN’s VLANid must be identical to the StpdID of the STP domain.
If you add a protected VLAN or port, that addition inherits the carrier VLAN’s encapsulation mode
unless you specify the encapsulation mode when you execute the configure stpd add vlan or
configure vlan add ports stpd commands. If you specify an encapsulation mode (dot1d, emistp,
or pvst-plus), the STP port mode is changed to match; otherwise, the STP port inherits either the
carrier VLANs encapsulation mode on that port or the STPD’s default encapsulation mode.
For MSTP, you do not need carrier a VLAN. A CIST controls the connectivity of interconnecting MSTP
regions and sends BPDUs across the regions to communicate region status. You must use the dot1d
encapsulation mode in an MSTP environment. For more information about MSTP, see the section
“Multiple Spanning Tree Protocol” on page 721.

To remove ports, use the following command:
configure stpd <stpd_name> delete vlan <vlan_name> ports [all | <port_list>]
If you manually delete a protected VLAN or port, only that VLAN or port is removed. If you manually
delete a carrier VLAN or port, all VLANs on that port (both carrier and protected) are deleted from that
STPD.
To learn more about member VLANs, see “Member VLANs” on page 694. For more detailed
information about these command line interface (CLI) commands, see the ExtremeXOS Command
Reference Guide.
Automatically Binding Ports

To automatically bind ports to an STPD when the ports are added to a VLAN, use the following
command:
enable stpd <stpd_name> auto-bind vlan <vlan_name>
The autobind feature is disabled on user-created STPDs. The autobind feature is enabled on the default
VLAN that participates in the default STPD S0.
For EMISTP or PVST+, when you issue this command, any port or list of ports that you add to the
carrier VLAN are automatically added to the STPD with autobind enabled. In addition, any port or list
of ports that you remove from a carrier VLAN are automatically removed from the STPD. This feature
allows the STPD to increase or decrease its span as ports are added to or removed from a carrier VLAN.
NOTE
The carrier VLAN’s VLANid must be identical to the StpdID of the STP domain.
Enabling autobind on a protected VLAN does not expand the boundary of the STPD. If the same set of
ports are members of the protected VLAN and the carrier VLAN, protected VLANs are aware of STP
state changes. For example, assume you have the following scenario:
● Carrier VLAN named v1
● v1 contains ports 3:1-3:2
● Protected VLAN named v2
● v2 contains ports 3:1-3:4
Since v1 contains ports 3:1-3:2, v2 is aware only of the STP changes for ports 3:1 and 3:2, respectively.
Ports 3:3 and 3:4 are not part of the STPD, which is why v2 is not aware of any STP changes for those
ports.
In addition, enabling autobind on a protected VLAN causes ports to be automatically added or

removed as the carrier VLAN changes.
For MSTP, when you issue this command, any port or list of ports that gets automatically added to an
MSTI are automatically inherited by the CIST. In addition, any port or list of ports that you remove
from an MSTI protected VLAN are automatically removed from the CIST. For more information, see
“Automatically Inheriting Ports—MSTP Only.”

To remove ports, use the following command:
configure stpd <stpd_name> delete vlan <vlan_name> ports [all | <port_list>]
If you manually delete a port from the STPD on a VLAN that has been added by autobind,
ExtremeXOS records the deletion so that the port does not get automatically added to the STPD after a
system restart.
To learn more about the member VLANs, see “Member VLANs” on page 694. For more detailed
information about these CLI commands, see the ExtremeXOS Command Reference Guide.
Automatically Inheriting Ports—MSTP Only

In an MSTP environment, whether you manually or automatically bind a port to an MSTI in an MSTP
region, the switch automatically binds that port to the CIST. The CIST handles BPDU processing for
itself and all of the MSTIs; therefore, the CIST must inherit ports from the MSTIs in order to transmit
and receive BPDUs. You can only delete ports from the CIST if it is no longer a member of an MSTI.
For more information about MSTP, see “Multiple Spanning Tree Protocol” on page 721.
Rapid Root Failover

ExtremeXOS supports rapid root failover for faster STP failover recovery times in STP 802.1D mode. If
the active root port link goes down, ExtremeXOS recalculates STP and elects a new root port. The rapid
root failover feature allows the new root port to immediately begin forwarding, skipping the standard
listening and learning phases. Rapid root failover occurs only when the link goes down and not when
there is any other root port failure, such as missing BPDUs.
The default setting for this feature is disabled. To enable rapid root failover, use the following
command:
enable stpd <stpd_name> rapid-root-failover
To display the configuration, use the following command:
show stpd {<stpd_name> | detail}
STPD BPDU Tunneling

You can configure ExtremeXOS to allow a BPDU to traverse a VLAN without being processed by STP.
This is known as BPDU tunneling. There are differences in how to configure this behavior in
ExtremeWare and ExtremeXOS. The examples in this section show how you might have used this
feature in ExtremeWare and how to configure STPD BPDU tunneling with ExtremeXOS.
You may be more familiar configuring STPD BPDU tunneling on Extreme Networks devices running
ExtremeWare. In ExtremeWare, you specify the VLAN to add to the STPD and then you “ignore” the
STP BPDUs. By ignoring the STP BPDUs, the switch prevents the ports in the VLAN from becoming
part of the STPD. In ExtremeXOS, you specify the VLAN to add to the STPD and then you disable that
STPD. By disabling the STPD, the switch allows the ports associated with the VLAN to continue
passing traffic and the switch forwards the BPDUs without adding any STP information.

Figure 82: Sample network using STPD BPDU tunneling
STPD 3
VLAN 3
STPD 2
VLAN 2
STPD 1
VLAN 1
1:1 1:2 1:3 1:1 1:2 1:1 1:3
Switch A Switch B Switch C

2:1 2:1 2:1
Service provider L2
EX_180
The examples described below assume you have already done the following:
● Created the VLANs v1, v2, and v3
● Configured the VLAN tags
● Created the STPDs s1, s2, and s3
● Configured the STPD tags
● Configured the mode of operation for the STPD
● Configured the STP ports
● Enabled STPD
The following example shows how to configure STPD BPDU tunneling on devices running
ExtremeWare:
Switch A
enable ignore-bpdu vlan v1

configure vlan v1 add ports 1:1,2:1 tagged stpd s1

Switch B

Switch C

The following example shows how to configure STPD BPDU tunneling on devices running
ExtremeXOS:
Switch A
configure vlan v1 add ports 1:1,2:1 tagged

configure stpd s1 add vlan v1 ports 1:1
disable s1
disable s2
disable s3
Switch B

disable s1
disable s2
Switch C

disable s1
disable s3
STP and Hitless Failover—Modular Switches Only

chassis or you are using redundancy in a SummitStack, one node assumes the role of primary and the
other node assumes the role of backup. The primary executes the switch’s management functions, and
the backup acts in a standby role. Hitless failover transfers switch management control from the
primary to the backup and maintains the state of STP. STP supports hitless failover. You do not

STP Configurations
explicitly configure hitless failover support; rather, if you have two nodes installed, hitless failover is
available.
NOTE
To support hitless failover, the primary node replicates STP BPDUs to the backup, which allows the
nodes to run STP in parallel. Although both primary and backup node receive STP BPDUs, only the
primary transmits STP BPDUs to neighboring switches and participates in STP.
NOTE
page 1020 to confirm that both primary and backup nodes are running software that supports the synchronize
command.
To initiate hitless failover on a network that uses STP:

1 Confirm that the nodes are synchronized and have identical software and switch configurations
using the show switch {detail} command. The output displays the status of the primary and
backup nodes, with the primary node showing MASTER and the backup node showing BACKUP
(InSync).
3 If the nodes are synchronized, use the run failover (formerly run msm-failover) command to
initiate failover.
For more detailed information about verifying the status of the primary and backup nodes, and system
redundancy, see “Understanding System Redundancy—Modular Switches and SummitStack Only” on
page 77. For more information about hitless failover, see “Understanding Hitless Failover Support—
Modular Switches and SummitStack Only” on page 82.
STP Configurations
When you assign VLANs to an STPD, pay careful attention to the STP configuration and its effect on
the forwarding of VLAN traffic.
This section describes three types of STP configurations:

● Basic STP
● Multiple STPDs on a single port (which uses EMISTP)
● A VLAN that spans multiple STPDs

Basic STP Configuration

This section describes a basic, 802.1D STP configuration. Figure 83 illustrates a network that uses VLAN
tagging for trunk connections. The following four VLANs have been defined:
● Sales is defined on switch A, switch B, and switch M.
● Personnel is defined on switch A, switch B, and switch M.
● Manufacturing is defined on switch Y, switch Z, and switch M.
● Engineering is defined on switch Y, switch Z, and switch M.
● Marketing is defined on all switches (switch A, switch B, switch Y, switch Z, and switch M).
Two STPDs are defined:

● STPD1 contains VLANs Sales and Personnel.
● STPD2 contains VLANs Manufacturing and Engineering.
The carrier and protected VLANs are also defined:

● Sales is the carrier VLAN on STPD1.
● Personnel is a protected VLAN on STPD1.
● Manufacturing is a protected VLAN on STPD2.
● Engineering is the carrier VLAN on STPD2.
● Marketing is a member of both STPD1 and STPD2 and is a protected VLAN.
Figure 83: Multiple STPDs
Sales, Personnel, Marketing Manufacturing, Engineering, Marketing
Switch A Switch Y
Switch B Switch Z
Switch M
STPD 1 STPD 2
Sales, Personnel, Manufacturing, Engineering, Marketing EX_048
When the switches in this configuration boot-up, STP configures each STPD such that the topology
contains no active loops. STP could configure the topology in a number of ways to make it loop-free.

STP Configurations
In Figure 83, the connection between switch A and switch B is put into blocking state, and the
connection between switch Y and switch Z is put into blocking state. After STP converges, all the
VLANs can communicate, and all bridging loops are prevented.
The protected VLAN Marketing, which has been assigned to both STPD1 and STPD2, communicates
using all five switches. The topology has no loops, because STP has already blocked the port connection
between switch A and switch B and between switch Y and switch Z.
Within a single STPD, you must be extra careful when configuring your VLANs. Figure 84 illustrates a
network that has been incorrectly set up using a single STPD so that the STP configuration disables the
ability of the switches to forward VLAN traffic.
Figure 84: Incorrect tag-based STPD configuration

Marketing & Sales Marketing, Sales & Engineering
Switch 1 Switch 3
Switch 2
Sales & Engineering EX_049
The tag-based network in Figure 84 has the following configuration:

● Switch 1 contains VLAN Marketing and VLAN Sales.
● Switch 2 contains VLAN Engineering and VLAN Sales.
● Switch 3 contains VLAN Marketing, VLAN Engineering, and VLAN Sales.
● The tagged trunk connections for three switches form a triangular loop that is not permitted in an
STP topology.
● All VLANs in each switch are members of the same STPD.
STP can block traffic between switch 1 and switch 3 by disabling the trunk ports for that connection on
each switch.
Switch 2 has no ports assigned to VLAN Marketing. Therefore, if the trunk for VLAN Marketing on
switches 1 and 3 is blocked, the traffic for VLAN Marketing will not be able to traverse the switches.
NOTE
If an STPD contains multiple VLANs, all VLANs should be configured on all ports in that domain, except for ports
that connect to hosts (edge ports).

Multiple STPDs on a Port

Traditional 802.1D STP has some inherent limitations when addressing networks that have multiple
VLANs and multiple STPDs. For example, consider the sample depicted in Figure 85.
Figure 85: Limitations of traditional STPD
A B A B
S1 S2 S1 S2
A B A B
EX_050
The two switches are connected by a pair of parallel links. Both switches run two VLANs, A and B. To
achieve load-balancing between the two links using the traditional approach, you would have to
associate A and B with two different STPDs, called S1 and S2, respectively, and make the left link carry
VLAN A traffic while the right link carries VLAN B traffic (or vice versa). If the right link fails, S2 is
broken and VLAN B traffic is disrupted.
To optimize the solution, you can use the Extreme Multiple Instance Spanning (EMISTP) mode, which
allows a port to belong to multiple STPDs. EMISTP adds significant flexibility to STP network design.
Referring to Figure 85, using EMISTP, you can configure all four ports to belong to both VLANs.
Assuming that S1 and S2 still correspond to VLANs A and B, respectively, you can fine-tune STP
parameters to make the left link active in S1 and blocking in S2, while the right link is active in S2 and
blocking in S1. Again, if the right link fails, the left link is elected active by the STP algorithm for S2,
without affecting normal switching of data traffic.
Using EMISTP, an STPD becomes more of an abstract concept. The STPD does not necessarily
correspond to a physical domain; it is better regarded as a vehicle to carry VLANs that have STP
instances. Because VLANs can overlap, so do STPDs. However, even if the different STPDs share the
entire topology or part of the redundant topology, the STPDs react to topology change events in an
independent fashion.
VLANs Spanning Multiple STPDs

Traditionally, the mapping from VLANs to STP instances have been one-to-one or many-to-one. In both
cases, a VLAN is wholly contained in a single instance. In practical deployment there are cases in which
a one-to-many mapping is desirable. In a typical large enterprise network, for example, VLANs span
multiple sites and/or buildings. Each site represents a redundant looped area. However, between any
two sites the topology is usually very simple.

STP Configurations
Alternatively, the same VLAN may span multiple large geographical areas (because they belong to the
same enterprise) and may traverse a great many nodes. In this case, it is desirable to have multiple STP
domains operating in a single VLAN, one for each looped area. The justifications include the following:
● The complexity of the STP algorithm increases, and performance drops, with the size and complexity
of the network. The 802.1D standard specifies a maximum network diameter of seven hops. By
segregating a big VLAN into multiple STPDs, you reduce complexity and enhance performance.
● Local to each site, there may be other smaller VLANs that share the same redundant looped area
with the large VLAN. Some STPDs must be created to protect those VLAN. The ability to partition
VLANs allows the large VLAN to be “piggybacked” in those STPDs in a site-specific fashion.
Figure 86 has five domains. VLANs green, blue, brown, and yellow are local to each domain. VLAN red
spans all of the four domains. Using a VLAN that spans multiple STPDS, you do not have to create a
separate domain for VLAN red. Instead, VLAN red is “piggybacked” onto those domains local to other
VLANs.
Figure 86: VLANs spanning multiple STPDs
VLAN red
S1 VLAN green S2
VLAN yellow
VLAN red
VLAN red S4
VLAN brown
VLAN red
S3
VLAN blue
EX_051
In addition, the configuration in Figure 86 has these features:

● Each site can be administered by a different organization or department within the enterprise.
Having a site-specific STP implementation makes the administration more flexible and convenient.
● Between the sites the connections usually traverse distribution switches in ways that are known
beforehand to be “safe” with STP. In other words, the looped areas are already well-defined.
EMISTP Deployment Constraints

Although EMISTP greatly enhances STP capability, these features must deployed with care. This section
describes configuration issues that, if not followed, could lead to an improper deployment of EMISTP.
This section also provides the following restrictive principles to abide by in network design:
● Although a physical port can belong to multiple STPDs, any VLAN on that port can be in only one
domain. Put another way, a VLAN cannot belong to two STPDs on the same physical port.
● Although a VLAN can span multiple domains, any LAN segment in that VLAN must be in the same
STPD. VLANs traverse STPDs only inside switches, not across links. On a single switch, however,
bridge ports for the same VLAN can be assigned to different STPDs. This scenario is illustrated in
Figure 87.

Figure 87: VLANs traverse domains inside switches
S1
S1
S2
S2
Correct Wrong EX_052
● The VLAN partition feature is deployed under the premise that the overall inter-domain topology
for that VLAN is loop-free. Consider the case in Figure 88, VLAN red (the only VLAN in the figure)
spans STPDs 1, 2, and 3. Inside each domain, STP produces a loop-free topology. However, VLAN
red is still looped, because the three domains form a ring among themselves.
Figure 88: Looped VLAN topology
Domain 2
Domain 1
Domain 3
EX_053
● A necessary (but not sufficient) condition for a loop-free inter-domain topology is that every two
domains only meet at a single crossing point.
NOTE
You can use MSTP to overcome the EMISTP constraints described in this section. See “Multiple Spanning Tree
Protocol” on page 721 for information about MSTP.

Per VLAN Spanning Tree
Per VLAN Spanning Tree

Switching products that implement Per VLAN Spanning Tree (PVST) have been in existence for many
years and are widely deployed. To support STP configurations that use PVST, ExtremeXOS has an
operational mode called PVST+.
NOTE
In this document, PVST and PVST+ are used interchangeably. PVST+ is an enhanced version of PVST that is
interoperable with 802.1Q STP. The following discussions are in regard to PVST+, if not specifically mentioned.
STPD VLAN Mapping

Each VLAN participating in PVST+ must be in a separate STPD, and the VLAN number (VLANid)
must be the same as the STPD identifier (StpdID). As a result, PVST+ protected VLANs cannot be
partitioned.
This fact does not exclude other non-PVST+ protected VLANs from being grouped into the same STPD.
A protected PVST+ VLAN can be joined by multiple non-PVST+ protected VLANs to be in the same
STPD.
Native VLAN
In PVST+, the native VLAN must be peered with the default VLAN on Extreme Networks devices, as
both are the only VLAN allowed to send and receive untagged packets on the physical port.
Third-party PVST+ devices send VLAN 1 packets in a special manner. ExtremeXOS does not support
PVST+ for VLAN 1. Therefore, when the switch receives a packet for VLAN 1, the packet is dropped.
When a PVST+ instance is disabled, the fact that PVST+ uses a different packet format raises an issue. If
the STPD also contains ports not in PVST+ mode, the flooded packet has an incompatible format with
those ports. The packet is not recognized by the devices connected to those ports.
Rapid Spanning Tree Protocol

The Rapid Spanning Tree Protocol (RSTP), originally in the IEEE 802.1w standard and now part of the
IEEE 802.1D-2004 standard, provides an enhanced spanning tree algorithm that improves the
convergence speed of bridged networks. RSTP takes advantage of point-to-point links in the network
and actively confirms that a port can safely transition to the forwarding state without relying on any
timer configurations. If a network topology change or failure occurs, RSTP rapidly recovers network
connectivity by confirming the change locally before propagating that change to other devices across the
network. For broadcast links, there is no difference in convergence time between STP and RSTP.
RSTP supersedes legacy STP protocols, supports the existing STP parameters and configurations, and
allows for seamless interoperability with legacy STP.

RSTP Concepts
This section describes the following important RSTP concepts:
● Port Roles on page 710
● Link Types on page 711
Port Roles
RSTP uses information from BPDUs to assign port roles for each LAN segment. Port roles are not user-
configurable. Port role assignments are determined based on the following criteria:
● A unique bridge identifier (MAC address) associated with each bridge
● The path cost associated with each bridge port
● A port identifier associated with each bridge port
RSTP assigns one of the following port roles to bridge ports in the network, as described in Table 89.
Table 89: RSTP port roles

Port Role Description
Root Provides the shortest (lowest) path cost to the root bridge. Each bridge has only one root port; the
root bridge does not have a root port. If a bridge has two or more ports with the same path cost,
the port with the best port identifier becomes the root port.
Designated Provides the shortest path connection to the root bridge for the attached LAN segment. To
prevent loops in the network, there is only one designated port on each LAN segment. To select
the designated port, all bridges that are connected to a particular segment listen to each other’s
BPDUs and agree on the bridge sending the best BPDU. The corresponding port on that bridge
becomes the designated port. If there are two or more ports connected to the LAN, the port with
the best port identifier (lowest MAC address) becomes the designated port.
Alternate Provides an alternate path to the root bridge and the root port.
Backup Supports the designated port on the same attached LAN segment. Backup ports exist only when
the bridge is connected as a self-loop or to a shared-media segment.
Disabled A port in the disabled state does not participate in RSTP; however, it will forward traffic and
learn new MAC source addresses.
When RSTP stabilizes, all:

● Root ports and designated ports are in the forwarding state.
● Alternate ports and backup ports are in the blocking state.
RSTP makes the distinction between the alternate and backup port roles to describe the rapid transition
of the alternate port to the forwarding state if the root port fails.
In ExtremeXOS 11.5 and earlier, ports that connect to non-STP devices are edge ports. Edge ports do not
participate in RSTP, and their role is not confirmed. Edge ports immediately enter the forwarding state
unless the port receives a BPDU. In that case, edge ports enter the blocking state. The edge port remains
in the blocking state until it stops receiving BPDUs and the message age timer expires.
ExtremeXOS 11.6 and later support an enhanced bridge detection method, which is part of the 802.1D-
2004 standard. Ports that connect to non-STP devices are still considered edge ports. However, if you
have an 802.1D-2004 compliant edge port, the bridge detection mechanism causes the edge port to
transition to a non-edge port upon receiving a BPDU. If the former edge port does not receive a
subsequent BPDU during a pre-determined interval, the port attempts to become an edge port.

In summation, this is the behavior of an edge port running ExtremeXOS 11.6 or later:
● The port sends BPDUs
● The port runs a state machine
● If BPDUs are received, the port behaves as a normal RSTP port by entering the forwarding state and
participating in RSTP
● If subsequent BPDUs are not received, the port attempts to become the edge port again
Link Types
With RSTP, you can configure the link type of a port in an STPD. RSTP tries to rapidly move
designated point-to-point links into the forwarding state when a network topology change or failure
occurs. For rapid convergence to occur, the port must be configured as a point-to-point link.
Table 90 describes the link types.
Table 90: RSTP link types

Port Link Type Description
Auto Specifies the switch to automatically determine the port link type. An auto link behaves like a
point-to-point link if the link is in full-duplex mode or if link aggregation is enabled on the
port. Otherwise, the link behaves like a broadcast link used for 802.1w configurations.
Edge Specifies a port that does not have a bridge attached.
ExtremeXOS 11.6 or later—An edge port is held in the STP forwarding state unless a BPDU is
received by the port. In that case, the port behaves as a normal RSTP port. The port is no
longer considered an edge port. If the port does not receive subsequent BPDUs during a pre-
determined time, the port attempts to become an edge port.
ExtremeXOS 11.5 or earlier—An edge port is placed and held in the STP forwarding state
unless a BPDU is received by the port. In that case, an edge port enters and remains in the
blocking state until it stops receiving BPDUs and the message age timer expires.
Broadcast Specifies a port attached to a LAN segment with more than two bridges. A port with a
broadcast link type cannot participate in rapid reconfiguration using RSTP or MSTP. By
default, all ports are broadcast links.
Point-to-point Specifies a port attached to a LAN segment with only two bridges. A port with port-to-port link
type can participate in rapid reconfiguration. Used for 802.1w and MSTP configurations.
Configuring Link Types. By default, all ports are broadcast links. To configure the ports in an STPD, use
configure stpd <stpd_name> ports link-type [[auto | broadcast | point-to-point]

<port_list> | edge <port_list> {edge-safeguard [enable | disable]}]

● auto—Configures the ports as auto links. If the link is in full-duplex mode or if link aggregation is
enabled on the port, an auto link behaves like a point-to-point link.
● broadcast—Configures the ports as broadcast ports. By default, all ports are broadcast links.
● point-to-point—Configures the ports for rapid reconfiguration in an RSTP or MSTP environment.
● edge—Configures the ports as edge ports. For information about edge safeguard, see “Configuring
Edge Safeguard—ExtremeXOS 11.5 and 11.4 Only” next.

To change the existing configuration of a port in an STPD, and return the port to factory defaults, use
unconfigure stpd <stpd_name> ports link-type <port_list>
To display detailed information about the ports in an STPD, use the following command:
show stpd <stpd_name> ports {[detail | <port_list> {detail}]}
Configuring Edge Safeguard—ExtremeXOS 11.5 and 11.4 Only. Loop prevention and detection on an edge
port configured for RSTP is called “edge safeguard.” You configure edge safeguard on RSTP edge ports
to prevent accidental or deliberate misconfigurations (loops) resulting from connecting two edge ports
together or by connecting a hub or other non-STP switch to an edge port. Edge safeguard also limits the
impact of broadcast storms that might occur on edge ports. This advanced loop prevention mechanism
improves network resiliency but does not interfere with the rapid convergence of edge ports.
An edge port configured with edge safeguard immediately enters the forwarding state and transmits
BPDUs. If a loop is detected, STP blocks the port. By default, an edge port without edge safeguard
configured immediately enters the forwarding state but does not transmit BPDUs unless a BPDU is
received by that edge port.
You can also configure edge safeguard for loop prevention and detection on an MSTP edge port.
To configure an edge port and enable edge safeguard on that port, use the following command:
configure stpd <stpd_name> ports link-type [[auto | broadcast | point-to-point]

If you have already configured a port as an edge port and you want to enable edge safeguard on the
port, use the following command:
configure stpd <stpd_name> ports edge-safeguard enable <port_list>
To disable edge safeguard on an edge port, use one of the following commands:
● configure stpd <stpd_name> ports edge-safeguard disable <port_list>
● configure stpd <stpd_name> ports link-type [[auto | broadcast | point-to-point]
RSTP Timers
For RSTP to rapidly recover network connectivity, RSTP requires timer expiration. RSTP derives many
of the timer values from the existing configured STP timers to meet its rapid recovery requirements
rather than relying on additional timer configurations. Table 91 describes the user-configurable timers,
and Table 92 describes the timers that are derived from other timers and not user-configurable.
Table 91: User-configurable timers
Timer Description
Hello The root bridge uses the hello timer to send out configuration BPDUs through all of
its forwarding ports at a predetermined, regular time interval. The default value is 2
seconds. The range is 1 to 10 seconds.
Forward delay A port moving from the blocking state to the forwarding state uses the forward delay
timer to transition through the listening and learning states. In RSTP, this timer
complements the rapid configuration behavior. If none of the rapid rules are in
effect, the port uses legacy STP rules to move to the forwarding state. The default
is 15 seconds. The range is 4 to 30 seconds.

Table 92: Derived timers

Timer Description
TCN The root port uses the topology change notification (TCN) timer when it detects a
change in the network topology. The TCN timer stops when the topology change
timer expires or upon receipt of a topology change acknowledgement. The default
value is the same as the value for the bridge hello timer.
Topology change The topology change timer determines the total time it takes the forwarding ports to
send configuration BPDUs. The default value for the topology change timer depends
upon the mode of the port:
• 802.1D mode—The sum of the forward delay timer value (default value is
15 seconds; range of 4 to 30 seconds) and the maximum age timer value
(default value is 20 seconds; range of 6 to 40 seconds).
• 802.1w mode—Double the hello timer value (default value is 4 seconds).
Message age A port uses the message age timer to time out receiving BPDUs. When a port
receives a superior or equal BPDU, the timer restarts. When the timer expires, the
port becomes a designated port and a configuration update occurs. If the bridge
operates in 1w mode and receives an inferior BPDU, the timer expires early. The
default value is the same as the STPD bridge max age parameter.
Hold A port uses the hold timer to restrict the rate that successive BPDUs can be sent.
The default value is the same as the value for the bridge hello timer.
Recent backup The timer starts when a port leaves the backup role. When this timer is running, the
port cannot become a root port. The default value is double the hello time
(4 seconds).
Recent root The timer starts when a port leaves the root port role. When this timer is running,
another port cannot become a root port unless the associated port is put into the
blocking state. The default value is the same as the forward delay time.
The protocol migration timer is neither user-configurable nor derived; it has a set value of 3 seconds.
The timer starts when a port transitions from STP (802.1D) mode to RSTP (802.1w) mode and vice-
versa. This timer must expire before further mode transitions can occur.
RSTP Operation
In an RSTP environment, a point-to-point link LAN segment has two bridges. A switch that considers
itself the unique, designated bridge for the attached LAN segment sends a “propose” message to the
other bridge to request a confirmation of its role. The other bridge on that LAN segment replies with an
“agree” message if it agrees with the proposal. The receiving bridge immediately moves its designated
port into the forwarding state.
Before a bridge replies with an “agree” message, it reverts all of its designated ports into the blocking
state. This introduces a temporary partition into the network. The bridge then sends another “propose”
message on all of its designated ports for further confirmation. Because all of the connections are
blocked, the bridge immediately sends an “agree” message to unblock the proposing port without
having to wait for further confirmations to come back or without the worry of temporary loops.
Beginning with the root bridge, each bridge in the network engages in the exchange of “propose” and
“agree” messages until they reach the edge ports. Edge ports connect to non-STP devices and do not
participate in RSTP. Their role does not need to be confirmed. If you have an 802.1D-2004 compliant
edge port, the bridge detection mechanism causes the edge port to transition to a non-edge port upon
receiving a BPDU. If the former edge port does not receive a subsequent BPDU during a pre-
determined interval, the port attempts to become an edge port.

RSTP attempts to transition root ports and designated ports to the forwarding state and alternate ports
and backup ports to the blocking state as rapidly as possible.
A port transitions to the forwarding state if any of the following is true. The port:
● Has been in either a root or designated port role long enough that the spanning tree information
supporting this role assignment has reached all of the bridges in the network.
NOTE
RSTP is backward compatible with STP, so if a port does not move to the forwarding state with any of the RSTP
rapid transition rules, a forward delay timer starts and STP behavior takes over.
● Is now a root port and no other ports have a recent role assignment that contradicts with its root
port role.
● Is a designated port and attaches to another bridge by a point-to-point link and receives an “agree”
message from the other bridge port.
● Is an edge port.
An edge port is a port connected to a non-STP device and is in the forwarding state.
The following sections provide more information about RSTP behavior.
Root Port Rapid Behavior

In Figure 89, the diagram on the left displays the initial network topology with a single bridge having
the following:
● Two ports are connected to a shared LAN segment.
● One port is the designated port.
● One port is the backup port.
The diagram on the right displays a new bridge that:

● Is connected to the LAN segment
● Has a superior STP bridge priority
● Becomes the root bridge and sends a BPDU to the LAN that is received by both ports on the old
bridge

Figure 89: Example of root port rapid behavior
Inital topology New topology
Bridge Bridge
Backup Designated Backup Designated

port port port port
LAN segment
Superior STP
bridge priority
Root
bridge
EX_054
If the backup port receives the BPDU first, STP processes this packet and temporarily elects this port as
the new root port while the designated port’s role remains unchanged. If the new root port is
immediately put into the forwarding state, there is a loop between these two ports.
To prevent this type of loop from occurring, the recent backup timer starts. The root port transition rule
does not allow a new root port to be in the forwarding state until the recent backup timer expires.
Another situation may arise if you have more than one bridge and you lower the port cost for the
alternate port, which makes it the new root port. The previous root port is now an alternate port.
Depending on your STP implementation, STP may set the new root port to the forwarding state before
setting the alternate port to the blocking state. This may cause a loop.
To prevent this type of loop from occurring, the recent root timer starts when the port leaves the root
port role. The timer stops if the port enters the blocking state. RSTP requires that the recent root timer
stop on the previous root port before the new root port can enter the forwarding state.
Designated Port Rapid Behavior

When a port becomes a new designated port, or the STP priority changes on an existing designated
port, the port becomes an unsynced designated port. For an unsynced designated port to rapidly move
into the forwarding state, the port must propose a confirmation of its role on the attached LAN segment
(unless the port is an edge port). Upon receiving an “agree” message, the port immediately enters the
forwarding state.
If the receiving bridge does not agree and it has a superior STP priority, the receiving bridge replies
with its own BPDU. Otherwise, the receiving bridge keeps silent, and the proposing port enters the
forwarding state and starts the forward delay timer.
The link between the new designated port and the LAN segment must be a point-to-point link. If there
is a multi-access link, the “propose” message is sent to multiple recipients. If only one of the recipients
agrees with the proposal, the port can erroneously enter the forwarding state after receiving a single
“agree” message.

Receiving Bridge Behavior

The receiving bridge must decide whether or not to accept a proposal from a port. Upon receiving a
proposal for a root port, the receiving bridge:
● Processes the BPDU and computes the new STP topology.
● Synchronizes all of the designated ports if the receiving port is the root port of the new topology.
● Puts all unsynced, designated ports into the blocking state.
● Sends down further “propose” messages.
● Sends back an “agree” message through the root port.
If the receiving bridge receives a proposal for a designated port, the bridge replies with its own BPDU.
If the proposal is for an alternate or backup port, the bridge keeps silent.
Propagating Topology Change Information

When a change occurs in the topology of the network, such events are communicated through the
network.
In an RSTP environment, only non-edge ports entering the forwarding state cause a topology change. A
loss of network connectivity is not considered a topology change; however, a gain in network
connectivity must be communicated. When an RSTP bridge detects a topology change, that bridge starts
the topology change timer, sets the topology change flag on its BPDUs, floods all of the forwarding
ports in the network (including the root ports), and flushes the learned MAC address entries.
Rapid Reconvergence
This section describes the RSTP rapid behavior following a topology change. In this example, the bridge
priorities are assigned based on the order of their alphabetical letters; bridge A has a higher priority
than bridge F.

Suppose we have a network, as shown in Figure 90, with six bridges (bridge A through bridge F) where
the following is true:
● Bridge A is the root bridge.
● Bridge D contains an alternate port in the blocking state.
● All other ports in the network are in the forwarding state.
Figure 90: Initial network configuration
A B C
A,0 A,1 A,2
F E D
A,1 A,2 A,3
Designated Root Blocked

port port port EX_055a
The network reconverges in the following way:

1 If the link between bridge A and bridge F goes down, bridge F detects the root port is down. At this
point, bridge F:
● Immediately disables that port from the STP.
● Performs a configuration update.
As shown in Figure 91, after the configuration update, bridge F:
● Considers itself the new root bridge.
● Sends a BPDU message on its designated port to bridge E.
Figure 91: Down link detected
A B C
A,0 A,1 A,2
Down
link BPDU
F E D
F,0 A,2 A,3
Designated Root
port port EX_055b

2 Bridge E believes that bridge A is the root bridge. When bridge E receives the BPDU on its root port
from bridge F, bridge E:
● Determines that it received an inferior BPDU.
● Immediately begins the max age timer on its root port.
As shown in Figure 92, after the configuration update, bridge E:
● Regards itself as the new root bridge.
● Sends BPDU messages on both of its designated ports to bridges F and D, respectively.
Figure 92: New root bridge selected
A B C
A,0 A,1 A,2
Designated
port Root
port
F E D
F,0 E,0 A,3
BPDU
EX_055c
3 As shown in Figure 93, when bridge F receives the superior BPDU and configuration update from
bridge E, bridge F:
● Decides that the receiving port is the root port.
● Determines that bridge E is the root bridge.
Figure 93: Communicating new root bridge status to neighbors
A B C
A,0 A,1 A,2
Designated
port Root
port
F E D
E,1 E,0 A,3
EX_055d

4 Bridge D believes that bridge A is the root bridge. When bridge D receives the BPDU from bridge E
on its alternate port, bridge D:
● Immediately begins the max age timer on its alternate port.
As shown in Figure 94, after the configuration update, bridge D:
● Moves the alternate port to a designated port.
● Sends a “propose” message to bridge E to solicit confirmation of its designated role and to
rapidly move the port into the designated state.
Figure 94: Sending a propose message to confirm a port role
A B C
A,0 A,1 A,2
Designated
port Root
port
F E D
E,1 E,0 A,3
Propose BPDU
EX_055e
5 Upon receiving the proposal, bridge E (as shown in Figure 95):

● Changes its receiving port to a root port.
The existing designated port enters the blocking state.
Bridge E then sends:
● A “propose” message to bridge F.
● An “agree” message from its root port to bridge D.
Figure 95: Communicating port status to neighbors
A B C
A,0 A,1 A,2
Designated Root
port port
F E D
E,1 A,4 A,3
Agree BPDU
EX_055f

6 To complete the topology change (as shown in Figure 96):

● Bridge D moves the port that received the “agree” message into the forwarding state.
● Bridge F confirms that its receiving port (the port that received the “propose” message) is the
root port, and immediately replies with an “agree” message to bridge E to unblock the proposing
port.
Figure 96: Completing the topology change
A B C
A,0 A,1 A,2
Root Designated
port port
F E D
A,5 A,4 A,3
EX_055g
Figure 97 displays the new topology.
Figure 97: Final network configuration
A B C
A,0 A,1 A,2
Root Designated
port port
F E D
A,5 A,4 A,3
EX_055h
Compatibility With STP (802.1D)

RSTP interoperates with legacy STP protocols; however, the rapid convergence benefits are lost when
interacting with legacy STP bridges.
Each RSTP bridge contains a port protocol migration state machine to ensure that the ports in the STPD
operate in the correct, configured mode. The state machine is a protocol entity within each bridge
configured to run in 802.1w mode. For example, a compatibility issue occurs if you configure 802.1w
mode and the bridge receives an 802.1D BPDU on a port. The receiving port starts the protocol
migration timer and remains in 802.1D mode until the bridge stops receiving 802.1D BPDUs. Each time
the bridge receives an 802.1D BPDU, the timer restarts. When the port migration timer expires, no more
802.1D BPDUs have been received, and the bridge returns to its configured setting, which is 802.1w
mode.

Multiple Spanning Tree Protocol

The Multiple Spanning Tree Protocol (MSTP), based on IEEE 802.1Q-2003 (formerly known as IEEE
802.1s), allows the bundling of multiple VLANs into one spanning tree topology. This concept is not
new to Extreme Networks. Like MSTP, Extreme Networks proprietary EMISTP implementation can
achieve the same capabilities of sharing a virtual network topology among multiple VLANs; however,
MSTP overcomes some of the challenges facing EMISTP, including enhanced loop protection
mechanisms and new capabilities to achieve better scaling.
MSTP logically divides a Layer 2 network into regions. Each region has a unique identifier and contains
multiple spanning tree instances (MSTIs). An MSTI is a spanning tree domain that operates within and
is bounded by a region. MSTIs control the topology inside the regions. The Common and Internal
Spanning Tree (CIST) is a single spanning tree domain that interconnects MSTP regions. The CIST is
responsible for creating a loop-free topology by exchanging and propagating BPDUs across regions to
form a Common Spanning Tree (CST).
MSTP uses RSTP as its converging algorithm and is interoperable with the legacy STP protocols: STP
(802.1D) and RSTP (802.1w). MSTP has three major advantages over 802.1D, 802.1w, and other
proprietary implementations:
● To save control path bandwidth and provide improved scalability, MSTP uses regions to localize
BPDU traffic. BPDUs containing information about MSTIs contained within an MSTP region do not
cross that region’s boundary.
● A single BPDU transmitted from a port can contain information for up to 64 STPDs. MSTP BPDU
processing utilizes less resources compared to 802.1D or 802.1w where one BPDU corresponds to
one STPD.
● In a typical network, a group of VLANs usually share the same physical topology. Dedicating a
spanning tree per VLAN like PVST+ is CPU intensive and does not scale very well. MSTP makes it
possible for a single STPD to handle multiple VLANs.
MSTP Concepts
This section describes the following MSTP concepts:
● MSTP Regions on page 722
● Common and Internal Spanning Tree on page 724
● Multiple Spanning Tree Instances on page 725
● Boundary Ports on page 726
● MSTP Port Roles on page 727
● MSTP Port States on page 727
● MSTP Link Types on page 728
● MSTP Edge Safeguard on page 728
● MSTP Timers on page 728
● MSTP Hop Counts on page 728
● Configuring MSTP on the Switch on page 728

MSTP Regions
An MSTP network consists of either individual MSTP regions connected to the rest of the network with
802.1D and 802.1w bridges or as individual MSTP regions connected to each other. An MSTP region
defines the logical boundary of the network. With MSTP, you can divide a large network into smaller
areas similar to an OSPF area or a BGP Autonomous System, which contain a group of switches under
a single administration. Each MSTP region has a unique identifier, is bound together by one CIST that
spans the entire network. A bridge participates in only one MSTP region at a time.
An MSTP region can hide its internal STPDs and present itself as a virtual 802.1w bridge to other
interconnected regions or 802.1w bridges because the port roles are encoded in 802.1w and MSTP
BPDUs.
By default, the switch uses the MAC address of the switch to generate an MSTP region. Since each
MAC address is unique, every switch is in its own region by default. For multiple switches to be part of
an MSTP region, you must configure each switch in the region with the same MSTP region identifiers.
See “Configuring MSTP Region Identifiers” on page 723 for information.
In Figure 98, all bridges inside MSTP regions 1 and 2 are MSTP bridges; bridges outside of the regions
are either 802.1D or 802.1w bridges.
Figure 98: Sample MSTP topology with two MSTP regions
= boundary port
= master port
= MSTI root port (10) CIST root bridge
A C
B
(40) CIST
(20) CIST regional root,
regional MSTI
root regional root
MSTP MSTP
D Region 1 E I Region 2 J
F K
(50) MSTI (60) (80) (90)
regional
root
G H
(100) EX_167

Configuring MSTP Region Identifiers. For multiple switches to be part of an MSTP region, you must
configure each switch in the region with the same MSTP configuration attributes, also known as MSTP
region identifiers.
The following list describes the MSTP region identifiers:

● Region Name—This indicates the name of the MSTP region. In the Extreme Networks
implementation, the maximum length of the name is 32 characters and can be a combination of
alphanumeric characters and underscores ( _ ).
● Format Selector—This indicates a number to identify the format of MSTP BPDUs. The default is 0.
● Revision Level—This identifier is reserved for future use; however, the switch uses and displays a
default of 3.
The switches inside a region exchange BPDUs that contain information for MSTIs. The switches
connected outside of the region exchange CIST information. By having devices look at the region
identifiers, MSTP discovers the logical boundary of a region:
● Configuring the MSTP region name
To configure the MSTP region name, use the following command:
configure mstp region <regionName>
The maximum length of the region name is 32 characters and can be a combination of alphanumeric
characters and underscores ( _ ). You can configure only one MSTP region on the switch at any
given time.
If you have an active MSTP region, Extreme Networks recommends that you disable all active
STPDs in the region before renaming the region on all of the participating switches.
● Configuring the MSTP BPDU format identifier
To configure the number used to identify MSTP BPDUs, use the following command:
configure mstp format <format_identifier>
By default, the value used to identify the MSTP BPDUs is 0. The range is 0 to 255.
If you have an active MSTP region, Extreme Networks recommends that you disable all active
STPDs in the region before modifying the value used to identify MSTP BPDUs on all participating
switches.
● Configuring the MSTP revision level
Although the configure mstp revision <revision> command is available on the CLI, this
command is reserved for future use.
Unconfiguring an MSTP Region. Before you unconfigure an MSTP region, Extreme Networks
recommends that you disable all active STPDs in the region.
To unconfigure the MSTP region on the switch, use the following command:
unconfigure mstp region
After you issue this command, all of the MSTP settings return to their default values. See “Configuring
MSTP Region Identifiers” on page 723 for information about the default settings.

Common and Internal Spanning Tree

As previously described, MSTP logically divides a Layer 2 network into regions. The Common and
Internal Spanning Tree (CIST) is a single spanning tree domain that interconnects MSTP regions. The
CIST is responsible for creating a loop-free topology by exchanging and propagating BPDUs across
regions to form a Common Spanning Tree (CST). In essence, the CIST is similar to having a large
spanning tree across the entire network. The CIST has its own root bridge that is common to all MSTP
regions, and each MSTP region elects a CIST regional root that connects that region to the CIST thereby
forming a CST.
The switch assigns the CIST an instance ID of 0, which allows the CIST to send BPDUs for itself in
addition to all of the MSTIs within an MSTP region. Inside a region, the BPDUs contain CIST records
and piggybacked M-records. The CIST records contain information about the CIST, and the M-records
contain information about the MSTIs. Boundary ports exchange only CIST record BPDUs.
All MSTP configurations require a CIST domain. You must first configure the CIST domain before
configuring any MSTIs. By default, all MSTI ports in the region are inherited by the CIST. You cannot
delete or disable a CIST if any of the MSTIs are active in the system.
Configuring the CIST. To configure an STPD as the CIST, use the following command and specify the
mstp [cist] parameters:
You enable MSTP on a per STPD basis only. By specifying the mstp [cist] parameters, you configure
the mode of operation for the STPD as MSTP, and you identify the STPD to be the CIST.
CIST Root Bridge. In a Layer 2 network, the bridge with the lowest bridge ID becomes the CIST root
bridge. The parameters (vectors) that define the root bridge include the following:
● User-defined bridge priority (by default, the bridge priority is 32,768)
● MAC address
The CIST root bridge can be either inside or outside an MSTP region. The CIST root bridge is unique for
all regions and non-MSTP bridges, regardless of its location.
For more information about configuring the bridge ID, see the configure stpd priority command in
the ExtremeXOS Command Reference Guide.
CIST Regional Root Bridge. Within an MSTP region, the bridge with the lowest path cost to the CIST root
bridge is the CIST regional root bridge. The path cost, also known as the CIST external path cost, is a
function of the link speed and number of hops. If there is more than one bridge with the same path
cost, the bridge with the lowest bridge ID becomes the CIST regional root. If the CIST root is inside an
MSTP region, the same bridge is the CIST regional root for that region because it has the lowest path
cost to the CIST root. If the CIST root is outside an MSTP region, all regions connect to the CIST root
via their CIST regional roots.
The total path cost to the CIST root bridge from any bridge in an MSTP region consists of the CIST
internal path cost (the path cost of the bridge to the CIST regional root bridge) and the CIST external
path cost. To build a loop-free topology within a region, the CIST uses the external and internal path
costs, and the MSTI uses only the internal path cost.

Looking at MSTP region 1 in Figure 99, the total path cost for the bridge with ID 60 consists of an
external path cost of A and an internal path cost of E.
Figure 99: Close-up of MSTP Region 1
(10) CIST root bridge
= boundary port
= master port
(20) CIST
= MSTI root port
regional root
MSTP
D E
Region 1
F G
(50) MSTI (60)

regional root
EX_168
CIST Root Port. The port on the CIST regional root bridge that connects to the CIST root bridge is the
CIST root port (also known as the master port for MSTIs). The CIST root port is the master port for all
MSTIs in that region, and it is the only port that connects the entire region to the CIST root.
If a bridge is both the CIST root bridge and the CIST regional root bridge, there is no CIST root port on
that bridge.
Enabling the CIST. To enable the CIST, use the following command and specify the CIST domain as the
<stpd_name>:
enable stpd {<stpd_name>}
Multiple Spanning Tree Instances

As previously described, multiple spanning tree instances (MSTIs) control the topology inside an MSTP
region. An MSTI is a spanning tree domain that operates within and is bounded by a region; an MSTI
does not exchange BPDUs with or send notifications to other regions. You identify an MSTI on a per
region basis. The MSTI ID does not have any significance outside of its region so you can re-use IDs
across regions. An MSTI consists of a group of VLANs, which can share the same network topology.
Each MSTI has its own root bridge and a tree spanning its bridges and LAN segments.
You can map multiple VLANs to an MSTI; however, multiple MSTIs cannot share the same VLAN.
Configuring the MSTI and the MSTI ID. MSTP uses the MSTI ID, not an Stpd ID, to identify the spanning
tree contained within the region. As previously described, the MSTI ID only has significance within its
local region, so you can re-use IDs across regions.

To configure the MSTI that is inside an MSTP region and its associated MSTI ID, use the following
command and specify the mstp [msti <instance>] parameters:
The range of the MSTI instance ID is 1 to 4,094.
MSTP STPDs use 802.1D BPDU encapsulation mode by default. To ensure correct operation of your
MSTP STPDs, do not configure EMISTP or PVST+ encapsulation mode for MSTP STPDs. For more
information, see “Encapsulation Modes” on page 696.
MSTI Regional Root Bridge. Each MSTI independently chooses its own root bridge. For example, if two
MSTIs are bounded to a region, there is a maximum of two MSTI regional roots and one CIST regional
root.
The bridge with the lowest bridge ID becomes the MSTI regional root bridge. The parameters that
define the root bridge include the following:
● User-defined bridge priority (by default, the bridge priority is 32,768)
● MAC address
Within an MSTP region, the cost from a bridge to the MSTI regional root bridge is known as the MSTI
internal path cost. Looking at MSTP region 1 in Figure 99 on page 725, the bridge with ID 60 has a path
cost of F to the MSTI regional root bridge.
The MSTI regional root bridge can be the same as or different from the CIST regional root bridge of that
region. You achieve this by assigning different priorities to the STP instances configured as the MSTIs
and the CIST. For more information about configuring the bridge ID, see the configure stpd
priority command in the ExtremeXOS Command Reference Guide.
MSTI Root Port. The port on the bridge that has the lowest path cost to the MSTI regional root bridge is
the MSTI root port. If a bridge has two or more ports with the same path cost, the port with the best
port identifier becomes the root port.
Enabling the MSTI. To enable the MSTI, use the following command and specify the MSTI domain as
the <stpd_name>:
Boundary Ports
Boundary ports are bridge ports that are only connected to other MSTP regions or 802.1D or 802.1w
bridges. The ports that are not at a region boundary are called internal ports. The boundary ports
exchange only CIST BPDUs. A CIST BPDU originated from the CIST root enters a region through the
CIST root port and egresses through boundary ports. This behavior simulates a region similar to an
802.1w bridge, which receives BPDUs on its root ports and forwards updated BPDUs on designated
ports.

Figure 100 shows an MSTP network that consists of two MSTP regions. Each region has its own CIST
regional root and is connected to the CIST root through master ports. The CIST regional roots in each
region are the MSTP bridges having the lowest CIST external root path cost. The CIST root is the bridge
with the lowest bridge ID and is an 802.1w bridge outside of either MSTP region.
Figure 100: Sample MSTP topology with two MSTP regions
= boundary port
= master port
= MSTI root port (10) CIST root bridge
A C
B
(40) CIST
(20) CIST regional root,
regional MSTI
root regional root
MSTP MSTP
D Region 1 E I Region 2 J
F K
(50) MSTI (60) (80) (90)
regional
root
G H
(100) EX_167
MSTP Region 1 and MSTP Region 2 are connected to the CIST root through directly connected ports,
identified as master ports. The bridge with ID 100 connects to the CIST root through Region 1, Region 2,
or segment B. For this bridge, either Region 1 or Region 2 can be the designated region or segment B
can be the designated segment. The CIST BPDUs egressing from the boundary ports carry the CIST
regional root as the designated bridge. This positions the entire MSTP region as one virtual bridge.
The CIST controls the port roles and the state of the boundary ports. A master port is always
forwarding for all CIST and MSTI VLANs. If the CIST sets a boundary port to the discarding state, the
CIST blocks traffic for all VLANs mapped to it and the MSTIs within that region. Each MSTI blocks
traffic for their member VLANs and puts their internal ports into the forwarding or blocking state
depending on the MSTI port roles. For more information about port states, see “MSTP Port States” on
page 727.
MSTP Port Roles

MSTP uses the same port roles as RSTP (Root, Designated, Alternate, and Backup), as described in
Table 89 on page 710. In addition to those port roles, MSTP introduces a new port role: Master. A Master
port is the port that connects an MSTI to the CIST root.
MSTP Port States

MSTP uses the same port states as RSTP (Listening, Learning, Forwarding, and Blocking). In the
Extreme Networks MSTP implementation, the listening state is not truly implemented as FDB learning

cannot be done when the port is not in the forwarding state. Ports in the blocking state listen but do not
accept ingress traffic, perform traffic forwarding, or learn MAC source address; however, the port
receives and processes BPDUs. For more information about all of the STP port states, see “STP States”
on page 697.
MSTP Link Types

MSTP uses the same link types as STP and RSTP, respectively. In an MSTP environment, configure the
same link types for the CIST and all MSTIs. For more information about the link types, see “Link
Types” on page 711.
MSTP Edge Safeguard

You can configure edge safeguard for loop prevention and detection on an MSTP edge port. For more
information see “Configuring Edge Safeguard—ExtremeXOS 11.5 and 11.4 Only” on page 712.
MSTP Timers
MSTP uses the same timers as STP and RSTP, respectively. For more information, see “RSTP Timers” on
page 712.
MSTP Hop Counts

In an MSTP environment, the hop count has the same purpose as the maxage timer for 802.1D and
802.1w environments. The CIST hop count is used within and outside a region. The MSTI hop count is
used only inside of the region. In addition, if the other end is an 802.1D or 802.1w bridge, the maxage
timer is used for interoperability between the protocols.
The BPDUs use hop counts to age out information and to notify neighbors of a topology change. To
configure the hop count, use the following command:
configure stpd <stpd_name> max-hop-count <hopcount>
By default, the hop count of a BPDU is 20 hops. The range is 6 to 40 hops.
Configuring MSTP on the Switch

To configure and enable MSTP:
1 Create the MSTP region using the following command:
configure mstp region <regionName>
2 Create and configure the CIST, which forms the CST, using the following commands:
configure stpd <stpd_name> mode mstp cist
NOTE
You can configure the default STPD, S0 as the CIST.

NOTE
No VLAN can be bound to the CIST and no ports can be added to the CIST. Therefore, the VLAN should be
bound to the MSTI and the “show MSTI port” command will show the VLAN ports. The ports added to the MSTI
are bound automatically to the CIST even though they are not added to it.
3 Enable the CIST using the following command:

4 Create and configure MSTIs using the following commands:
configure stpd <stpd_name> mode mstp msti <instance>
5 Add VLANs to the MSTIs using one of the following commands:
● Manually binding ports
configure stpd <stpd_name> add vlan <vlan_name> ports [all | <port_list>] {[dot1d
configure vlan <vlan_name> add ports [all | <port_list>] {tagged | untagged} stpd
<stpd_name> {[dot1d | emistp | pvst-plus]}
● Automatically binding ports to an STPD when ports are added to a member VLAN
enable stpd <stpd_name> auto-bind vlan <vlan_name>
6 Enable the MSTIs.
For a more detailed configuration example, see “MSTP Configuration Example” on page 737.

MSTP Operation
To further illustrate how MSTP operates and converges, Figure 101 displays a network with two MSTP
regions. Each region contains three MSTP bridges and one MSTI. The overall network topology also
contains one CIST root bridge (Switch A, which has the lowest bridge ID), one interconnecting 802.1w
bridge (Switch D), and 10 full duplex, point-to-point segments. VLAN Default spans all of the bridges
and segments in the network, VLAN engineering is local to its respective region, and STPD S0 is
configured as the CIST on all bridges.
Figure 101: MSTP topology with the CIST root bridge contained within a region
= boundary port
= master port
= MSTI root port
1 Switch D 2
MSTP MSTP
Region 1 Region 2
3
CIST root CIST
Switch A CIST regional root
Switch E regional root
MSTI regional root
4 5 6 7
8 9 10
Switch B Switch C Switch F Switch G
MSTI regional root
STPD SI configured on switch A-C STPD SI configured on switch E-G

VLAN engineering assigned to SI VLAN finance assigned to SI
EX_165
MSTP Region 1 consists of the following:

● Three bridges named Switch A, Switch B, and Switch C
● One MSTI STPD named S1 with an MSTI ID of 1
● VLAN engineering mapped to the MSTI STPD, S1
● Switch A as the CIST root bridge (this is the CIST root bridge for all regions)
● Switch A as the CIST regional root bridge
● Switch A as the MSTI regional root bridge
● Three boundary ports that connect to MSTP region 2 and other 802.1D or 802.1w bridges
MSTP Region 2 consists of the following:

● Three bridges named Switch E, Switch F, and Switch G
● One MSTI STPD named S1 with an MSTI ID of 1
NOTE
The MSTI ID does not have any significance outside of its region so you can re-use IDs across regions.

● VLAN finance mapped to the MSTI STPD, S1

● Switch E as the CIST regional root bridge
● Switch F as the MSTI regional root bridge
● One master port that connects to the CIST
● Three boundary ports that connect to MSTP region 1 and other 802.1D or 802.1w bridges
The following sequence describes how the MSTP topology convergences:

1 Determining the CIST root bridge, MSTP regions, and region boundaries.
Each bridge believes that it is the root bridge, so each bridge initially sends root bridge BPDUs
throughout the network. As bridges receive BPDUs and compare vectors, the bridge with the lowest
Bridge ID is elected the CIST root bridge. In our example, Switch A has the lowest Bridge ID and is
the CIST root bridge.
The bridges in the MSTP regions (Switches A, B, C, E, F, and G) advertise their region information
along with their bridge vectors.
Segments 1, 3, and 9 receive BPDUs from other regions and are identified as boundary ports for
region 1. Similarly, segments 2, 3, and 9 are identified as boundary ports for region 2.
2 Controlling boundary ports.
The CIST regional root is advertised as the Bridge ID in the BPDUs exiting the region. By sending
CIST BPDUs across regional boundaries, the CIST views the MSTP regions as virtual 802.1w bridges.
The CIST takes control of the boundary ports and only CIST BPDUs enter or exit a region boundary.
Each MSTP region has a CIST regional root bridge that communicates to the CIST root bridge. The
bridge with the lowest path cost becomes the CIST regional root bridge. The port on the CIST
regional root bridge that connects to the CIST root bridge is the CIST root port.
For region 1, Switch A has the lowest cost (0 in this example) and becomes the CIST regional root.
Since the bridge is both the CIST root bridge and the CIST regional root bridge, there is no CIST root
port on the bridge.
For region 2, Switch E is the CIST regional root bridge and so a port on that bridge becomes the
CIST root port.
3 Identifying MSTI regional roots.
Each MSTI in a region has an MSTI regional root bridge. MSTI regional roots are selected
independently of the CIST root and CIST regional root. The MSTP BPDUs have M-records for each
MSTI. Bridges belonging to an MSTI compare vectors in their M-records to elect the MSTI regional
root.
4 Converging the CIST.
The CIST views every region as a virtual bridge and calculates the topology using the 802.1w
algorithm. The CIST calculates the topology both inside and outside of a region.
5 Converging MSTIs.
After the CIST identifies the boundary ports, each MSTI in a domain converge their own trees using
802.1w.
At this point, all CIST and MSTIs have assigned port roles of root, designated, alternate, and backup
to their respective spanning trees. All root and designated ports transition to the forwarding state
while the remaining ports remain in the discarding state.
Propagating topology change information is similar to that described for RSTP. For more information
see, “Propagating Topology Change Information” on page 716.
For a configuration example, see “MSTP Configuration Example” on page 737.

STP Rules and Restrictions

This section summarizes the rules and restrictions for configuring STP as follows:
● The carrier VLAN must span all ports of the STPD. (This is not applicable to MSTP.)
● The StpdID must be the VLANid of the carrier VLAN; the carrier VLAN cannot be partitioned. (This
is not applicable to MSTP.)
● A default VLAN cannot be partitioned. If a VLAN traverses multiple STPDs, the VLAN must be
tagged.
● An STPD can carry, at most, one VLAN running in PVST+ mode, and its StpdID must be identical
with that VLANid. In addition, the PVST+ VLAN cannot be partitioned.
● The default VLAN of a PVST+ port must be identical to the native VLAN on the PVST+ device
connected to that port.
● If an STPD contains both PVST+ and non-PVST+ ports, that STPD must be enabled. If that STPD is
disabled, the BPDUs are flooded in the format of the incoming STP port, which may be incompatible
with those of the connected devices.
● The 802.1D ports must be untagged; and the EMISTP/PVST+ ports must be tagged in the carrier
VLAN.
● An STPD with multiple VLANs must contain only VLANs that belong to the same virtual router
instance.
● Automatically adding ports to an STPD (known as STP autobind) cannot be configured on a
Netlogin VLAN.
● STP cannot be configured on the following ports:
■ A mirroring target port.
■ A software-controlled redundant port.
■ A Netlogin port.
● MSTP and 802.1D STPDs cannot share a physical port.
● Only one MSTP region can be configured on a switch.
● In an MSTP environment, A VLAN can belong to one of the MSTIs.
● A VLAN can belong to only one MSTP domain.
● MSTP is not interoperable with PVST+.
● No VLAN can be bound to the CIST.
Configuring STP on the Switch

To configure basic STP:
1 Create one or more STPDs using the following command:
2 Add one or more VLANs to the STPD using the following command:
configure stpd <stpd_name> add vlan <vlan_name> ports [all | <port_list>] {[dot1d

3 Define the carrier VLAN using the following command:

configure stpd <stpd_name> tag <stpd_tag>
NOTE
The carrier VLAN’s VLANid must be identical to the StpdID of the STPD.
4 Enable STP for one or more STPDs using the following command:
After you have created the STPD, you can optionally configure STP parameters for the STPD.
NOTE
You should not configure any STP parameters unless you have considerable knowledge and experience with STP. The
default STP parameters are adequate for most networks.
The following parameters can be configured on each STPD:

● Hello time (In an MSTP environment, configure this only on the CIST.)
● Forward delay
● Max age (In an MSTP environment, configure this only on the CIST.)
● Max hop count (MSTP only)
● Bridge priority
● StpdID (STP, RSTP, EMISTP, and PVST+ only)
● MSTI ID (MSTP only)
The following parameters can be configured on each port:

● Path cost
● Port priority
● Port mode
NOTE
The device supports the RFC 1493 Bridge MIB, RSTP-03, and Extreme Networks STP MIB. Parameters of the s0
default STPD support RFC 1493 and RSTP-03. Parameters of any other STPD support the Extreme Networks STP
MIB.
NOTE
If an STPD contains at least one port not in 802.1D (dot1D) mode, the STPD must be configured with an StpdID.
The following section provides more detailed STP configuration examples, including 802.1D, EMISTP,
RSTP, and MSTP.

STP Configuration Examples

This section provides four configuration examples:
● Basic 802.1D STP on page 734
● EMISTP on page 735
● RSTP 802.1w on page 736
● MSTP 802.1s on page 737
Basic 802.1D Configuration Example

● Removes ports from the VLAN Default that will be added to VLAN Engineering.
● Creates the VLAN Engineering.
● Assigns a VLANid to the VLAN Engineering.
NOTE
If you do not explicitly configure the VLANid in your 802.1D deployment, use the show vlan command to see
the internal VLANid automatically assigned by the switch.
● Adds ports to the VLAN Engineering.

● Creates an STPD named Backbone_st.
● Configures the default encapsulation mode of dot1d for all ports added to STPD Backbone_st.
● Enables autobind to automatically add or remove ports from the STPD.
● Assigns the Engineering VLAN to the STPD.
● Assigns the carrier VLAN.
NOTE
To assign the carrier VLAN, the StpdID must be identical to the VLANid of the carrier VLAN.
● Enables STP.
configure vlan default delete ports 2:5-2:10
configure vlan engineering tag 150
configure vlan engineering add ports 2:5-2:10 untagged
create stpd backbone_st

configure stpd backbone_st default-encapsulation dot1d
enable stpd backbone_st auto-bind vlan engineering
configure stpd backbone_st tag 150
enable stpd backbone_st
By default, the port encapsulation mode for user-defined STPDs is emistp. In this example, you set it to
dot1d.

EMISTP Configuration Example

Figure 102 is an example of EMISTP.
Figure 102: EMISTP configuration example
VLAN red
S1 VLAN green S2
VLAN yellow
VLAN red
VLAN red S4
VLAN brown
VLAN red
S3
VLAN blue
EX_051
NOTE
By default, all ports added to a user-defined STPD are in emistp mode, unless otherwise specified.
The following commands configure the switch located between S1 and S2:
create vlan red
configure red tag 100
configure red add ports 1:1-1:4 tagged
create vlan green

configure green tag 200
configure green add ports 1:1-1:2 tagged
create vlan yellow

configure yellow tag 300
configure yellow add ports 1:3-1:4 tagged
create stpd s1
configure stpd s1 add green ports all
configure stpd s1 add red ports 1:1-1:2 emistp
enable stpd s1
create stpd s2
configure stpd s2 add yellow ports all
configure stpd s2 add red ports 1:3-1:4 emistp
enable stpd s2

RSTP 802.1w Configuration Example

Figure 103 is an example of a network with multiple STPDs that can benefit from RSTP. For RSTP to
work:
● Create an STPD.
● Configure the mode of operation for the STPD.
● Create the VLANs and assign the VLANid and the VLAN ports.
● Assign the carrier VLAN.
● Add the protected VLANs to the STPD.
● Configure the port link types.
● Enable STP.
Figure 103: RSTP example
Sales, Personnel, Marketing Manufacturing, Engineering, Marketing
Switch A Switch Y
Switch B Switch Z
Switch M
STPD 1 STPD 2
Sales, Personnel, Manufacturing, Engineering, Marketing EX_048
In this example, the commands configure switch A in STPD1 for rapid reconvergence. Use the same
commands to configure each switch and STPD in the network.
create stpd stpd1

configure stpd stpd1 mode dot1w
create vlan sales

create vlan personnel
create vlan marketing
configure vlan sales tag 100
configure vlan personnel tag 200
configure vlan marketing tag 300
configure vlan sales add ports 1:1,2:1 tagged

configure vlan personnel add ports 1:1,2:1 tagged

configure vlan marketing add ports 1:1,2:1 tagged
configure stpd stpd1 add vlan sales ports all

configure stpd stpd1 add vlan personnel ports all
configure stpd stpd1 add vlan marketing ports all
configure stpd stpd1 ports link-type point-to-point 1:1,2:1
configure stpd stpd1 tag 100
enable stpd stpd1
MSTP Configuration Example

Figure 104 is an example with multiple STPDs that can benefit from MSTP. In this example, we have
two MSTP regions that connect to each other and one external 802.1w bridge.
Figure 104: MSTP configuration example
= boundary port
= master port
= MSTI root port
CIST root
1 Switch D 2
MSTP MSTP
Region 1 Region 2
3 CIST regional root

CIST regional root CIST root port
Switch A MSTI regional root
Switch E
4 5 6 7
8 9 10
Switch B Switch C Switch F Switch G
MSTI regional root
STPD SI configured on switch A-C STPD SI configured on switch E-G

VLAN engineering assigned to SI VLAN finance assigned to SI
EX_166
For MSTP to work, complete the following steps on all switches in Region 1 and Region 2:
● Remove ports from the VLAN Default that will be added to VLAN Engineering.
● Create the VLAN Engineering.
● Assign a VLANid to the VLAN Engineering.

NOTE
If you do not explicitly configure the VLANid in your MSTP deployment, use the show vlan command to see
the internal VLANid automatically assigned by the switch.
● Add ports to the VLAN Engineering.

● Create the MSTP region.
NOTE
You can configure only one MSTP region on the switch at any given time.
● Create the STPD to be used as the CIST, and configure the mode of operation for the STPD.
● Specify the priority for the CIST.
● Enable the CIST.
● Create the STPD to be used as an MSTI and configure the mode of operation for the STPD.
● Specify the priority for the MSTI.
● Assign the VLAN Engineering to the MSTI.
● Configure the port link type.
● Enable the MSTI.
On the external switch (the switch that is not in a region):

● Create an STPD that has the same name as the CIST, and configure the mode of operation for the
STPD.
● Specify the priority of the STPD.
● Enable the STPD.
NOTE
In the following sample configurations, any lines marked (Default) represent default settings and do not need to be
explicitly configured. STPD s0 already exists on the switch.
In the following example, the commands configure Switch A in Region 1 for MSTP. Use the same
commands to configure each switch in Region 1:

configure vlan engineering add port 2-3 tagged
configure mstp region region1

create stpd s0 (Default)
configure stpd s0 mode mstp cist
configure stpd s0 priority 32768 (Default)
enable stpd s0
create stpd s1
configure stpd s1 mode mstp msti 1
enable stpd s1 auto-bind vlan engineering
configure stpd s1 ports link-type point-to-point 2-3

Displaying STP Settings
enable stpd s1
In the following example, the commands configure Switch E in Region 2 for MSTP. Use the same
commands to configure each switch in Region 2:
create vlan finance

configure vlan finance tag 2
configure vlan finance add port 2-3 tagged
configure mstp region region2

configure stpd s0 mode mstp cist
enable stpd s0
create stpd s1
configure stpd s1 mode mstp msti 1
enable stpd s1 auto-bind vlan finance
enable stpd s1
In the following example, the commands configure switch D, the external switch. Switch D becomes the
CIST root bridge:

configure stpd s0 mode dot1w
configure stpd s0 priority 28672
enable stpd s0 auto-bind vlan Default
enable stpd s0

To display STPD settings, use the following command:
show stpd {<stpd_name> | detail}
To display more detailed information for one or more STPDs, specify the detail option.

● STPD name
● STPD state
● STPD mode of operation
● Rapid Root Failover
● Tag
● Ports
● Active VLANs
● Bridge Priority
● Bridge ID
● Designated root

● STPD configuration information
If you have MSTP configured on the switch, this command displays additional information:
● MSTP Region
● Format Identifier
● Revision Level
● Common and Internal Spanning Tree (CIST)
● Total number of MST Instances (MSTI)
To display the state of a port that participates in STP, use the following command:
show stpd <stpd_name> ports {[detail | <port_list> {detail}]}
To display more detailed information for one or more ports in the specified STPD, including
participating VLANs, specify the detail option.

● STPD port configuration
● STPD port mode of operation
● STPD path cost
● STPD priority
● STPD state (root bridge, and so on)
● Port role (root designated, alternate and so on)
● STPD port state (forwarding, blocking, and so on)
● Configured port link type
● Operational port link type
● Edge port settings (inconsistent behavior, edge safeguard setting)
● MSTP port role (internal or boundary)
If you have MSTP configured and specify the detail option, this command displays additional
information:
● MSTP internal path cost
● MSTP timers
● STPD VLAN Settings

If you have a VLAN that spans multiple STPDs, use the show vlan <vlan_name> stpd command to
display the STP configuration of the ports assigned to that specific VLAN.
The command displays the following:

● STPD port configuration
● STPD port mode of operation
● STPD path cost
● STPD priority
● STPD state (root bridge, and so on)
● Port role (root designated, alternate and so on)
● STPD port state (forwarding, blocking, and so on)
● Configured port link type
● Operational port link type


27 Extreme Standby Router Protocol

● Overview of ESRP on page 744
● ESRP Concepts on page 745
● Determining the ESRP Master on page 751
● Configuring an ESRP Domain on a Switch on page 756
● Advanced ESRP Features on page 759
● Displaying ESRP Information on page 767
● Using ELRP with ESRP on page 767
● ESRP Examples on page 770
● ESRP Cautions on page 775
Licensing
You need an Advanced Edge, Core, or Advanced Core license to configure and use all of the Extreme
Standby Router Protocol (ESRP) features described in this chapter.
If you have an Edge license, you can configure your Extreme Networks switch to be ESRP-aware. ESRP-
aware switches do not actively participate in ESRP; however, when connected to a network with ESRP-
enabled switches, the ESRP-aware switches reliably perform failover and failback scenarios in the
prescribed recovery times.
series switch, and Summit X450a series switch ship with an Advanced Edge license.
The Summit X450e series switch and the Summit X250e series switch ship with an Edge license, which
supports only ESRP-aware. You can, however, upgrade to an Advanced Edge license that supports
ESRP.
A SummitStack must be operating at the Advanced Edge license level to configure and use all ESRP
features described in this chapter.

Extreme Standby Router Protocol
Overview of ESRP
The Extreme Standby Router Protocol (ESRP) is a feature of ExtremeXOS that allows multiple switches
to provide redundant routing services to users. From the workstation’s perspective, there is only one
default router (that has one IP address and one MAC address), so address resolution protocol (ARP)
cache entries in client workstations do not need to be refreshed or aged out. ESRP is available only on
Extreme Networks switches.
In addition to providing Layer 3 routing redundancy for IP and IPX, ESRP also provides Layer 2
redundancy. You can use these “layered” redundancy features in combination or independently.
You do not have to configure the switch for routing to make valuable use of ESRP. The Layer 2
redundancy features of ESRP offer fast failure recovery and provide for dual-homed system design. In
some instances, depending on network system design, ESRP can provide better resiliency than using
Spanning Tree Protocol (STP) or Virtual Router Redundancy Protocol (VRRP).
Extreme Networks recommends that all switches participating in ESRP run the same version of
ExtremeXOS.

● ESRP Modes of Operation on page 744
● ESRP and ELRP on page 744
● Reasons to Use ESRP on page 745
ESRP Modes of Operation

ExtremeXOS has two modes of ESRP operation: standard and extended. Select standard ESRP if your
network contains some switches running ExtremeWare, others running ExtremeXOS, and a combination
of those switches participating in ESRP. Standard ESRP is backward compatible with and supports the
ESRP functionality of ExtremeWare.
Select extended ESRP if your network contains switches running only ExtremeXOS. Extended mode
ESRP supports and is compatible with switches running ExtremeXOS. By default, ExtremeXOS operates
in extended mode.
In addition to the modes of operation, ESRP has an auto toggle feature. Depending on the mode of
operation configured on the neighbor switch, the mode of operation at this end will toggle to the same
mode of operation as the neighbor.
For more detailed information about the ESRP modes of operation, see Standard and Extended ESRP on
page 748.
ESRP and ELRP

Support for the Extreme Loop Recovery Protocol (ELRP) was introduced in ExtremeXOS 11.1. ELRP
allows you to prevent, detect, and recover from Layer 2 loops in the network. For more information
about configuring ELRP in an ESRP environment, see Using ELRP with ESRP on page 767. For more
information about standalone ELRP, see Using Standalone ELRP to Perform Loop Tests on page 1041.

ESRP Concepts
Reasons to Use ESRP

You can use ESRP to achieve edge-level or aggregation-level redundancy. Deploying ESRP in this area
of the network allows you to simplify your network design, which is important in designing a stable
network. ESRP also works well in meshed networks where Layer 2 loop protection and Layer 3
redundancy are simultaneously required.
ESRP Concepts
You configure ESRP on a per domain basis on each switch. A maximum of two switches can participate
in providing redundant Layer 3 or Layer 2 services to a single Virtual LAN (VLAN). If you configure
and use ESRP groups, more than two switches can provide redundant Layer 2 or Layer 3 services to a
single VLAN. The switches exchange keep-alive packets for each VLAN independently. Only one
switch (the master) can actively provide Layer 3 routing and/or Layer 2 switching for each VLAN. This
switch handles the forwarding, ARP requests, and routing for this particular VLAN. Other participating
switches for the VLAN are in slave mode waiting for an ESRP state change.
For a VLAN within an ESRP domain, each participating switch uses the same MAC address and must
be configured with the same IP address or IPX NetID. It is possible for one switch to be a master switch
for one or more VLANs while being a slave switch for other VLANs, thus allowing the load to be split
across participating switches.
To participate in ESRP, the following must be true:

● A VLAN can belong to only one ESRP domain.
● The IP address for the VLANs participating in an ESRP domain must be identical.
● All switches in the ESRP network must use the same election algorithm, otherwise loss of
connectivity, broadcast storms, or other unpredictable behavior may occur.
● If you have an untagged master VLAN, you must specify an ESRP domain ID. The domain ID must
be identical on all switches participating in ESRP for that particular domain.
● If you have a tagged master VLAN, ESRP uses the 802.1Q tag (VLANid) of the master VLAN for the
ESRP domain ID. If you do not use the VLANid as the domain ID, you must specify a different
domain ID. As previously described, the domain ID must be identical on all switches participating in
ESRP for that particular domain.
NOTE
If you configure the Open Shortest Path First (OSPF) routing protocol and ESRP, you must manually configure an
OSPF router identifier (ID). Be sure that you configure a unique OSPF router ID on each switch running ESRP. For
more information on configuring OSPF, see Chapter 10, “OSPF.”

Figure 105 displays a basic ESRP topology.
Figure 105: Example of a basic ESRP topology
ESRP Core Switch #1 ESRP Core Switch #2

State Domain Group State Domain Group
Master corpnet1 0 Slave corpnet1 0
Master corpnet2 0 Slave corpnet2 0
Slave corpnet3 0 Master corpnet3 0
Corpnet1, Corpnet2 Corpnet3

advertised ESRP advertised virtual mac:
virtual mac: 00:E0:2B:00:00:80
00:E0:2B:00:00:80
ESRP-aware ESRP-aware ESRP-aware EX_099
ESRP-Aware Switches
Extreme Networks switches that are not actively participating in ESRP but are connected on a network
that has other Extreme Networks switches running ESRP are ESRP-aware. When ESRP-aware switches
are attached to ESRP-enabled switches, the ESRP-aware switches reliably perform failover and failback
scenarios in the prescribed recovery times.
If Extreme Networks switches running ESRP are connected to Layer 2 switches that are manufactured
by third-party vendors, the failover times for traffic local to that segment may appear longer, depending
on the application involved and the FDB timer used by the other vendor’s Layer 2 switch. ESRP can be
used with Layer 2 switches from other vendors, but the recovery times vary.
The VLANs associated with the ports connecting an ESRP-aware switch to an ESRP-enabled switch
must be configured using an 802.1Q tag on the connecting port; or, if only a single VLAN is involved,
as untagged using the protocol filter any. ESRP will not function correctly if the ESRP-aware switch
interconnection port is configured for a protocol-sensitive VLAN using untagged traffic. You can also
use port restart in this scenario. For more information about port restart, see ESRP Port Restart on
page 762.

ESRP Concepts
Configuring ESRP-Aware Switches

For an Extreme Networks switch to be ESRP-aware, you must create an ESRP domain on the aware
switch, add a master VLAN to that ESRP domain, add a member VLAN to that ESRP domain if
configured, and configure a domain ID if necessary.
To participate as an ESRP-aware switch, the following must be true:

● The ESRP domain name must identical on all switches (ESRP-enabled and ESRP-aware) participating
in ESRP for that particular domain.
● The master VLAN name and IP address must be identical on all switches (ESRP-enabled and ESRP-
aware) participating in ESRP for that particular domain.
● If configured, the member VLAN name and IP address must be identical on all switches (ESRP-
enabled and ESRP-aware) participating in ESRP for that particular domain.
● The domain ID must be identical on all switches (ESRP-enabled or ESRP-aware) participating in
ESRP for that particular domain.
■ If you have an untagged master VLAN, you must specify an ESRP domain ID.
■ If you have a tagged master VLAN, ESRP uses the 802.1Q tag (VLANid) of the master VLAN for
the ESRP domain ID. If you do not use the VLANid as the domain ID, you must specify a
different domain ID.
NOTE
Before you begin, make a note of the ESRP domain parameters on the ESRP-enabled switch. That way you can
easily refer to your notes while creating the ESRP domain on the ESRP-aware switch.
To configure an ESRP-aware switch:

1 Create an ESRP domain using the create esrp <esrpDomain> command.
If you have an Edge license, only ESRP-aware functionality is supported; you cannot enable the
ESRP domain.
If you have an Advanced Edge, a Core, or an Advanced Core license, do not enable the domain after
you create it.
2 Add a master VLAN to your ESRP domain using the configure esrp <esrpDomain> add master
<vlan_name> command.
3 If configured, add the member VLANs to your ESRP domain using the configure esrp
<esrpDomain> add member <vlan_name> command.
4 If necessary, configure a domain ID for the ESRP domain using the configure esrp <esrpDomain>
domain-id <number> command.
Displaying ESRP-Aware Information

To display ESRP-aware information, use the following command:
show esrp {<name>}
The display includes the group number and MAC address for the master of the group, as well as the
age of the information.

Standard and Extended ESRP

ESRP has two modes of operation: standard and extended. By default, ExtremeXOS operates in
extended mode. To configure a different mode of operation, use the following command:
configure esrp mode [extended | standard]
Standard mode is backward compatible with and supports the ESRP functionality of switches running
ExtremeWare. ESRP functionality available in extended mode is not applicable in standard mode. Use
standard mode if your network contains both switches running ExtremeWare and switches running
ExtremeXOS participating in ESRP.
Extended mode supports and is compatible with switches running ExtremeXOS while participating in
ESRP. Use extended mode if your network contains only switches running ExtremeXOS.
The following list describes the major differences in behavior between standard and extended mode:
● Handshaking
In standard mode, events such as link flapping cause the ESRP master switch to generate a large
number of packets and to increase processing time.
To prevent this, extended mode supports handshaking. Handshaking occurs when a switch requests
a state change, forces its neighbor to acknowledge the change, and the neighbor sends an
acknowledgement to the requesting switch. For example, if a slave switch wants to become the
master, it enters the pre-master state, notifies the neighbor switch, and forces the neighbor to
acknowledge the change. The neighbor then sends an acknowledgement back to the slave switch.
While the requesting switch waits for the acknowledgements, future updates are suppressed to make
sure the neighbor does not act on incorrect data.
● Stickiness
In standard mode, if an event causes the ESRP master switch to fail over to the slave, it becomes the
new master. If another event occurs, the new master switch returns to the slave and you have
experienced two network interruptions.
To prevent this, extended mode supports the sticky election metric. The default election algorithm
uses the sticky metric. For example, if an event causes the ESRP master switch to fail over to the
slave, it becomes the new master and has a higher sticky value. If another event occurs, for example
adding active ports to the slave, the new master does not fail back to the original master even if the
slave has more active ports. After sticky is set on the master, regardless of changes to its neighbor’s
election algorithm, the new master retains its position. Sticky algorithms provide for fewer network
interruptions than non-sticky algorithms. Sticky is set only on the master switch.
● Port weight
In standard mode, the port count calculation does not take into account the available bandwidth of
the ports. For example, a switch with a one Gigabit Ethernet uplink may be unable to become master
because another switch has a load-shared group of four fast Ethernet links. The active port count
only consider the number of active ports, not the bandwidth of those ports.
In extended mode, the active port count considers the number of active ports and the port weight
configuration also considers the bandwidth of those ports. You enable port weight only on the load-
shared master port.
● Domain ID
In standard mode, ESRP packets do not contain domain information; therefore, the only information
about the packet comes from the receiving port.
The concept of domain ID is applicable only to extended mode. A domain ID in the packet clearly
classifies the packet, associates a received ESRP PDU to a specific ESRP domain, and tells the
receiving port where the packet came from. In extended mode, you must have a domain ID for each

ESRP Concepts
ESRP domain. Each switch participating in ESRP for a particular domain must have the same
domain ID configured.
The ESRP domain ID is determined from one of the following user-configured parameters:
■ ESRP domain number created with the configure esrp <esrpDomain> domain-id <number>
command
■ 802.1Q tag (VLANid) of the tagged master VLAN
● Hello messages
In standard mode, both the master switch and slave switch send periodic ESRP hello messages. This
causes an increase in packet processing by both the master and slave.
In extended mode, the master switch sends periodic ESRP hello messages. This reduces the amount
of packet processing, increases the amount of available link bandwidth, and does not impact
communicating state changes between switches.
In addition to the modes of operation, ESRP has an auto toggle feature. Depending on the mode of
operation configured on the neighbor switch, the mode of operation at this end will toggle to the same
mode of operation as the neighbor. For example, if you use the default mode—extended—and your
ESRP domain contains a switch running ExtremeXOS that detects a neighbor switch running
ExtremeWare, the mode automatically changes to standard for that domain. This action causes the
switch to enter the neutral state and re-elect the ESRP master. Since you are using the default mode of
operation, and the switch running ExtremeXOS detected a neighbor switch running ExtremeWare, the
ExtremeXOS switch toggles to standard mode although the configured mode of operation remains as
extended.
ESRP Domains
ESRP domains allow you to configure multiple VLANs under the control of a single instance of the
ESRP protocol. By grouping multiple VLANs under one ESRP domain, the ESRP protocol can scale to
provide protection to large numbers of VLANs. All VLANs within an ESRP domain simultaneously
share the same active and standby router and failover router, as long as one port of each member
VLAN belongs to the domain master.
Depending on the election policy used, when a port in a member VLAN belongs to the domain master,
the member VLAN ports are considered when determining the ESRP master. You can configure a
maximum of 64 ESRP domains in a network.
If you disable an ESRP domain, the switch notifies its neighbor that the ESRP domain is going down,
and the neighbor clears its neighbor table. If the master switch receives this information, it enters the
neutral state to prevent a network loop. If the slave switch receives this information, it enters the
neutral state.
ESRP Domain IDs

ESRP packets do not identify themselves to which domain they belong; you either configure a domain
ID or the ESRP domain uses the 802.1Q tag (VLANid) of the master VLAN. A domain ID in the packet
clearly classifies the packet, associates a received ESRP PDU to a specific ESRP domain, and tells the
receiving port where the packet came from.

Linking ESRP Switches

When considering system design using ESRP, Extreme Networks recommends using a direct link.
Direct links between ESRP switches are useful under the following conditions:
● A direct link can provide a more direct routed path, if the ESRP switches are routing and supporting
multiple VLANs where the master/slave configuration is split such that one switch is master for
some VLANs and a second switch is master for other VLANs. The direct link can contain a unique
router-to-router VLAN/subnet, so that the most direct routed path between two VLANs with
different master switches uses a direct link, instead of forwarding traffic through another set of
connected routers.
● A direct link can be used as a highly reliable method to exchange ESRP hello messages, so that the
possibility of having multiple masters for the same VLAN is lessened if all downstream Layer 2
switches fail.
● A direct link is necessary for the ESRP host attach (HA) option. The direct link is used to provide
Layer 2 forwarding services through an ESRP slave switch.
Direct links may contain a router-to-router VLAN, along with other VLANs participating in an ESRP
domain. If multiple VLANs are used on the direct links, use 802.1Q tagging. The direct links may be
aggregated into a load-shared group, if desired. If multiple ESRP domains share a host port, each
VLAN must be in a different ESRP group.
ESRP and Hitless Failover—Modular Switches and SummitStack

Only
chassis, one node assumes the role of primary and another node assumes the role of backup node. The
primary node executes the switch’s management functions, and the backup node acts in a standby role.
Hitless failover transfers switch management control from the primary node to the backup node and
maintains the state of ESRP. The ESRP extended version supports hitless failover.
For hitless failover support, ESRP switches and the primary and backup nodes must run a version of
ExtremeXOS that supports hitless failover and operate in ESRP extended mode.
NOTE
The ESRP domain on the primary node is active and participates in the ESRP protocol. The ESRP
domain on the backup node is in the neutral state listening for configuration changes, tracking failures,
and checkpointing messages and link state events. When you initiate node failover, the master ESRP
switch notifies its neighbor ESRP switch about the failover. After the neighbor receives information
from the master switch, the neighbor remains in its current state and waits for the failover to occur.
After the failover from the primary node to the backup node is complete, the master ESRP switch
notifies the neighbor so the neighbor can relinquish its current state.

Determining the ESRP Master
NOTE
Before initiating failover, review the section Synchronizing Nodes—Modular Switches and SummitStack Only on
page 1020 to confirm that the primary and backup nodes are running software that supports the synchronize
command.
To initiate hitless failover on a network that uses ESRP:

(InSync).
After you confirm the MSMs are synchronized, proceed to step 3.
Understanding System Redundancy—Modular Switches and SummitStack Only on page 77. For more
information about hitless failover, see Understanding Hitless Failover Support—Modular Switches and
SummitStack Only on page 82.

The system determines the ESRP master switch (providing Layer 3 routing and/or Layer 2 switching
services for a VLAN) using the following default factors:
● Stickiness—The switch with the higher sticky value has higher priority. When an ESRP domain
claims master, its sticky value is set to 1 (available only in extended mode).
● Active ports—The switch that has the greatest number of active ports takes highest precedence.
● Tracking information—Various types of tracking are used to determine if the switch performing the
master ESRP function has connectivity to the outside world. ExtremeXOS supports the following
types of tracking:
■ VLAN—Tracks any active port connectivity to one designated VLANs. An ESRP domain can
track one VLAN, and the tracked VLAN should not be a member of any other ESRP domain in
the system.
■ IP route table entry—Tracks specific learned routes from the IP route table.
■ Ping—Tracks ICMP ping connectivity to specified devices.
■ Environment (health checks)—Tracks the environment of the switch, including power supply and
chassis temperature.

If any of the configured tracking mechanisms fail, the master ESRP switch relinquishes status as
master, and remains in slave mode for as long as the tracking mechanism continues to fail.
● ESRP priority—This is a user-defined field. The range of the priority value is 0 to 255; a higher
number has higher priority, except for 255. The default priority setting is 0. A priority setting of 255
makes an ESRP switch a standby switch that remains in slave mode until you change the priority
setting. Extreme Networks recommends this setting for system maintenance. A switch with a
priority setting of 255 will never become the master.
● System MAC address—The switch with the higher MAC address has higher priority.
● Active port weight—The switch that has the highest port weight takes precedence. The bandwidth of
the port automatically determines the port weight (available only in extended mode).
You can configure the precedence order of the factors used by the system to determine the master ESRP
switch. For more information about configuring the ESRP election metrics, see ESRP Election
Algorithms on page 754.
Master Switch Behavior

If a switch is master, it actively provides Layer 3 routing services to other VLANs, and Layer 2
switching between all the ports of that VLAN. Additionally, the switch exchanges ESRP packets with
other switches that are in slave mode.
Pre-Master Switch Behavior

A pre-master switch is ready to transition to master, but is going through possible loop detection before
changing to the master state. Upon entering the pre-master state, the switch sends ESRP packets to
other switches on that same VLAN. If the switch finds itself superior to its neighbor, and successfully
executes loop detection techniques, the switch transitions to master. This temporary state avoids the
possibility of having simultaneous masters.
Slave Switch Behavior

If a switch is in slave mode, it exchanges ESRP packets with other switches on that same VLAN. When
a switch is in slave mode, it does not perform Layer 3 routing or Layer 2 switching services for the
VLAN. From a Layer 3 routing protocol perspective (for example, RIP or OSPF), when in slave mode
for the VLAN, the switch marks the router interface associated with that VLAN as down. From a Layer
2 switching perspective, no forwarding occurs between the member ports of the VLAN; this prevents
loops and maintains redundancy.
If you configure the switch to use the optional ESRP HA configuration, the switch continues Layer 2
forwarding to the master. For more information, see ESRP Host Attach on page 762.
Neutral Switch Behavior

The neutral state is the initial state entered into by the switch. In a neutral state, the switch waits for
ESRP to initialize and run. A neutral switch does not participate in ESRP elections. If the switch leaves
the neutral state, it enters the slave state.

Electing the Master Switch

A new master can be elected in one of the following ways:
● A communicated parameter change
● Loss of communication between master and slave(s)
If a parameter determines the master changes (for example, link loss or priority change), the election of
the new master typically occurs within one second. A parameter change triggers a handshake between
the routers. As long as both routers agree upon the state transition, new master election is immediate.
If a switch in slave mode loses its connection with the master, a new election occurs (using the same
precedence order indicated on page 751 or using a configured precedence order described in ESRP
Election Algorithms on page 754). The new election typically takes place in three times the defined
timer cycle (8 seconds by default).
Before the switch transitions to the master state, it enters a temporary pre-master state. While in the pre-
master state, the switch sends ESRP PDUs until the pre-master state timeout expires. Depending upon
the election algorithm, the switch may then enter the master or slave state. Traffic is unaffected by the
pre-master state because the master continues to operate normally. The pre-master state avoids the
possibility of having simultaneous masters.
You can configure the pre-master state timeout using the following command:
configure esrp <esrpDomain> timer premaster <seconds>
CAUTION
Configure the pre-master state timeout only with guidance from Extreme Networks personnel. Misconfiguration can
severely degrade the performance of ESRP and your switch.
ESRP Failover Time

ESRP Failover time is largely determined by the following factors:
● ESRP hello timer setting.
● ESRP neighbor timer setting.
● The routing protocol being used for interrouter connectivity if Layer 3 redundancy is used; OSPF
failover time is faster than RIP failover time.
The failover time associated with the ESRP protocol depends on the timer setting and the nature of the
failure. The default hello timer setting is 2 seconds; the range is 2 to 1024 seconds. The default neighbor
timer setting is 8 seconds; the range is 3*hello to 1024 seconds. The failover time depends on the type of
event that caused ESRP to failover. In most cases, a non-hardware failover is less than 1 second, and a
hardware failover is 8 seconds.
If routing is configured, the failover of the particular routing protocol (such as RIP V1, RIP V2, or OSPF)
is added to the failover time associated with ESRP.
If you use OSPF, make your OSPF configuration passive. A passive configuration acts as a stub area
and helps decrease the time it takes for recalculating the network. A passive configuration also
maintains a stable OSPF core.
For more information about the ESRP timers and configuring the ESRP timers, see the ExtremeXOS

ESRP Election Algorithms

You configure the switch to use one of 15 different election algorithms to select the ESRP master. ESRP
uses the default election policy for extended mode. If you have an ESRP domain operating in standard
mode, the domain ignores the sticky and weight algorithms.
To change the election algorithm, you must first disable the ESRP domain and then configure the new
election algorithm. If you attempt to change the election algorithm without disabling the domain first,
an error message appears.
To disable the ESRP domain, use the following command:
disable esrp {<esrpDomain>}
To modify the election algorithm, use the following command:
configure esrp <esrpDomain> election-policy [ports > track > priority | ports > track
> priority > mac | priority > mac | priority > ports > track > mac | priority > track
> ports > mac | sticky > ports > track > priority | sticky > ports > track > priority
> mac | sticky > ports > weight > track > priority > mac | sticky > priority > mac |
sticky > priority > ports > track > mac | sticky > priority > track > ports > mac |
sticky > track > ports > priority | sticky > track > ports > priority > mac | track >
ports > priority | track > ports > priority > mac]configure esrp <esrpDomain>
election-policy [ports > track > priority | ports > track > priority > mac | priority
> mac | priority > ports > track > mac | priority > track > ports > mac | sticky >
ports > track > priority | sticky > ports > track > priority > mac | sticky > ports >
weight > track > priority > mac | sticky > priority > mac | sticky > priority > ports
> track > mac | sticky > priority > track > ports > mac | sticky > track > ports >
priority | sticky > track > ports > priority > mac | track > ports > priority | track
> ports > priority > mac]
If you attempt to use an election algorithm not supported by the switch, an error message similar to the
following appears:
ERROR: Specified election-policy is not supported!
Supported Policies:
1. sticky > ports > weight > track > priority > mac
2. ports > track > priority
3. sticky > ports > track > priority
4. ports > track > priority > mac
5. sticky > ports > track > priority > mac
6. priority > mac
7. sticky > priority > mac
8. priority > ports > track > mac
9. sticky > priority > ports > track > mac
10. priority > track > ports > mac
11. sticky > priority > track > ports > mac
12. track > ports > priority
13. sticky > track > ports > priority
14. track > ports > priority > mac
15. sticky > track > ports > priority > mac
Table 93 describes the ESRP election algorithms. Each algorithm considers the election factors in a
different order of precedence. The election algorithms that use sticky and weight are only available in
extended mode.

Table 93: ESRP election algorithms

Election Algorithm Description
ports > track > priority Specifies that this ESRP domain should consider election factors in
the following order: Active ports, tracking information, ESRP priority.
ports > track > priority > mac Specifies that this ESRP domain should consider election factors in
the following order: Active ports, tracking information, ESRP priority,
MAC address.
NOTE: This is the default election algorithm for standard mode.
priority > mac Specifies that this ESRP domain should consider election factors in
the following order: ESRP priority, MAC address.
priority > ports > track > mac Specifies that this ESRP domain should consider election factors in
the following order: ESRP priority, active ports, tracking information,
MAC address.
priority > track > ports > mac Specifies that this ESRP domain should consider election factors in
the following order: ESRP priority, tracking information, active ports,
MAC address.
sticky > ports > track > priority Specifies that this ESRP domain should consider election factors in
the following order: Stickiness, active ports, tracking information,
ESRP priority.
sticky > ports > track > priority > mac Specifies that this ESRP domain should consider election factors in
the following order: Stickiness, active ports, tracking information,
ESRP priority, MAC address.
sticky > ports > weight > track > Specifies that this ESRP domain should consider election factors in
priority > mac the following order: Stickiness, active ports, port weight, tracking
information, ESRP priority, MAC address.
NOTE: Beginning with ExtremeXOS 11.1 and later, this is the default
election algorithm for extended mode.
sticky > priority > ports > track > mac Specifies that this ESRP domain should consider election factors in
the following order: Stickiness, ESRP priority, active ports, tracking
information, MAC address.
sticky > priority > track > ports > mac Specifies that this ESRP domain should consider election factors in
the following order: Stickiness, ESRP priority, tracking information,
active ports, MAC address.
sticky > priority > mac Specifies that this ESRP domain should consider election factors in
the following order: Stickiness, ESRP priority, MAC address.
sticky > track > ports > priority Specifies that this ESRP domain should consider election factors in
the following order: Stickiness, tracking information, active ports,
ESRP priority.
sticky > track > ports > priority > mac Specifies that this ESRP domain should consider election factors in
the following order: Stickiness, tracking information, active ports,
ESRP priority, MAC address.
track > ports > priority Specifies that this ESRP domain should consider election factors in
the following order: Tracking information, active ports, ESRP priority.
track > ports > priority > mac Specifies that this ESRP domain should consider election factors in
the following order: Tracking information, active ports, ESRP priority,
MAC address.
CAUTION
All switches in the ESRP network must use the same election algorithm, otherwise loss of connectivity, broadcast
storms, or other unpredictable behavior may occur.

NOTE
If you have a network that contains a combination of switches running ExtremeXOS and ExtremeWare, only the
ports-track-priority-mac election algorithm is compatible with ExtremeWare releases before version 6.0.
Configuring an ESRP Domain on a Switch

To create, configure, and enable a basic ESRP domain:
1 Create and configure the master VLAN.
2 Create and configure the member VLAN(s), if applicable.
3 Create the ESRP domain.
4 Configure the ESRP domain ID, if applicable.
5 Add the master VLAN to the ESRP domain.
6 Add the member VLAN(s) to the ESRP domain, if applicable.
7 Enable ESRP for the specified ESRP domain.
The instructions that follow assume that you have already created and configured the VLANs to use as
the ESRP domain master and member VLANs. For more information about creating VLANs, see
Chapter 5, Virtual LANs. For more information about ESRP master and member VLANs, see Adding
VLANs to an ESRP Domain on page 757.
You can also configure other ESRP domain parameters, including ESRP:
● Mode of operation as described in Standard and Extended ESRP on page 748.
● Timers as described in the ExtremeXOS Command Reference Guide.
● Election algorithms as described in ESRP Election Algorithms on page 754.
● Tracking as described in ESRP Tracking on page 759.
● Port restart as described in ESRP Port Restart on page 762.
● Host attach as described in ESRP Host Attach on page 762.
● Groups as described in ESRP Groups on page 764.
For more detailed information about all of the commands used to create, configure, enable, and disable
an ESRP domain, refer to the ExtremeXOS Command Reference Guide.
Creating and Deleting an ESRP Domain

You specify a unique ESRP domain name to identify each ESRP domain in your network.
To create an ESRP domain, use the following command:
create esrp <esrpDomain>
The esrpDomain parameter is a character string of up to 32 characters that identifies the ESRP domain
to be created.

Configuring an ESRP Domain on a Switch
NOTE
If you use the same name across categories (for example, STPD and ESRP names) Extreme Networks recommends
that you specify the appropriate keyword as well as the actual name. If you do not specify the keyword, the switch
may display an error message.
The following example creates an ESRP domain named esrp1:
create esrp esrp1
To delete an ESRP domain, use the following command:
delete esrp <esrpDomain>
The following example deletes the ESRP domain named esrp1:
delete esrp esrp1
Configuring the ESRP Domain ID

If you choose not use the 802.1Q tag (VLANid) of the master VLAN, or you have an untagged master
VLAN, you must create a domain ID before you can enable the ESRP domain. For more information
about ESRP domains and the ESRP domain ID, see ESRP Domains on page 749.
To configure an ESRP domain ID, use the following command:
configure esrp <esrpDomain> domain-id <number>
The number parameter specifies the number of the domain ID. The user-configured ID range is 4096
through 65,535.
The following example creates a domain ID of 4097 for ESRP domain esrp1:
configure esrp esrp1 domain-id 4097
Adding VLANs to an ESRP Domain

This section assumes that you have already created and configured the VLANs that you want to add to
the ESRP domain.
Adding and Deleting a Master VLAN

The master VLAN is the VLAN on the ESRP domain that exchanges ESRP PDUs and data between a
pair of ESRP-enabled devices. You must configure one master VLAN for each ESRP domain, and a
master VLAN can belong to only one ESRP domain.
To add a master VLAN to an ESRP domain, use the following command:
configure esrp <esrpDomain> add master <vlan_name>
The esrpDomain parameter specifies the name of the ESRP domain, and the vlan_name parameter
specifies the name of the master VLAN.

The following example adds the VLAN sales as the master VLAN to ESRP domain esrp1:
configure esrp esrp1 add master sales
To delete a master VLAN, you must first disable the ESRP domain before removing the master VLAN
using the disable esrp {<esrpDomain>} command.
To delete a master VLAN from an ESRP domain, use the following command:
configure esrp <esrpDomain> delete master <vlan_name>
The following examples removes the master VLAN sales from ESRP domain esrp1:
configure esrp esrp1 delete master sales
Adding and Deleting a Member VLAN

The member VLAN can belong to only one ESRP domain, and you configure zero or more member
VLANs for each ESRP domain. The state of the ESRP device determines whether the member VLAN is
in the forwarding or blocking state.
To add a member VLAN to an ESRP domain, use the following command:
configure esrp <esrpDomain> add member <vlan_name>
The esrpDomain parameter specifies the name of the ESRP domain, and the vlan_name parameter
specifies the name of the master VLAN.
The following example adds the VLAN purple as a member VLAN to ESRP domain esrp1:
configure esrp esrp1 add member purple
To delete a member VLAN from an ESRP domain, use the following command:
configure esrp <esrpDomain> delete member <vlan_name>
The following example removes the member VLAN purple from ESRP domain esrp1:
configure esrp esrp1 delete member purple
Enabling and Disabling an ESRP Domain

To enable a specific ESRP domain, use the following command:
enable esrp <esrpDomain>
To disable a specific ESRP domain, use the following command:
disable esrp {<esrpDomain>}

Advanced ESRP Features

This section describes the following advanced ESRP features:
● ESRP Tracking on page 759
● ESRP Port Restart on page 762
● ESRP Host Attach on page 762
● ESRP Port Weight and Don’t Count on page 763
● ESRP Groups on page 764
● Selective Forwarding on page 765
ESRP Tracking
Tracking information is used to track various forms of connectivity from the ESRP switch to the outside
world. This section describes the following ESRP tracking options:
● ESRP Environment Tracking on page 759
● ESRP VLAN Tracking on page 760
● ESRP Route Table Tracking on page 760
● ESRP Ping Tracking on page 760
● Displaying ESRP Tracking Information on page 761
ESRP Environment Tracking

You can configure ESRP to track hardware status. If a power supply fails, if the chassis is overheating,
or if a non-fully loaded power supply is detected, the priority for the ESRP domain will change to the
failover settings.
NOTE
ExtremeXOS determines the maximum available power required for the switch by calculating the number of power
supplies and the power required by the installed modules. Enabling environmental tracking on the switch without
enough power budget causes tracking to fail. In this case, the tracking failure occurs by design.
To configure the failover priority for an ESRP domain:

1 Set the failover priority, using the following command:
configure esrp <esrpDomain> add track-environment failover <priority>
2 Assign the priority flag precedence over the active ports count, using the following command:
configure esrp <esrpDomain> election-policy [ports > track > priority | ports >
track > priority > mac | priority > mac | priority > ports > track > mac |
priority > track > ports > mac | sticky > ports > track > priority | sticky >
ports > track > priority > mac | sticky > ports > weight > track > priority >
mac | sticky > priority > mac | sticky > priority > ports > track > mac | sticky
> priority > track > ports > mac | sticky > track > ports > priority | sticky >
track > ports > priority > mac | track > ports > priority | track > ports >
priority > mac]
Because the priority of both ESRP domains are set to the same value, ESRP will use the active ports
count to determine the master ESRP domain.

ESRP VLAN Tracking

You can configure an ESRP domain to track port connectivity to a specified VLAN as criteria for ESRP
failover. The number of VLAN active ports are tracked. If the switch is no longer connected to the
specified VLAN, the switch automatically relinquishes master status and remains in slave mode. You
can track a maximum of one VLAN.
Configuring VLAN Tracking. To add a tracked VLAN, use the following command:
configure esrp <esrpDomain> add track-vlan <vlan_name>
Disabling VLAN Tracking. To delete a tracked VLAN, use the following command:
configure esrp <esrpDomain> delete track-vlan <vlan_name>
ESRP Route Table Tracking

You can configure ESRP to track specified routes in the route table as criteria for ESRP failover. If all of
the configured routes are not available within the route table, the switch automatically relinquishes
master status and remains in slave mode. You can track a maximum of eight routes per route table.
Configuring Route Table Tracking. To add a tracked route, use the following command:
configure esrp <esrpDomain> add track-iproute <ipaddress>/<masklength>
Disabling Route Table Tracking. To delete a tracked route, use the following command:
configure esrp <esrpDomain> delete track-iproute <ipaddress>/<masklength>
ESRP Ping Tracking

You can configure ESRP to track connectivity using a simple ping to any device. This may represent the
default route of the switch, or any device meaningful to network connectivity of the master ESRP
switch. The switch automatically relinquishes master status and remains in slave mode if a ping
keepalive fails. You can configure a maximum of eight ping tracks.
NOTE
The ESRP ping tracking option cannot be configured to ping an IP address within an ESRP VLAN subnet. It should
be configured on some other normal VLAN across the router boundary.
Configuring Ping Tracking. To configure ping tracking, use the following command:
configure esrp <esrpDomain> add track-ping <ipaddress> frequency <seconds> miss
<misses>
The seconds parameter specifies the number of seconds between ping requests. The range is 1 to 600
seconds.
The misses parameter specifies the number of consecutive ping failures that will initiate failover to an
ESRP slave. The range is 1 to 256 pings.
Disabling Ping Tracking. To disable ping tracking, use the following command:
configure esrp <esrpDomain> delete track-ping <ipaddress>

Displaying ESRP Tracking Information

You can view the status of ESRP tracking on a per domain basis. The information displayed includes
the type of tracking used by the ESRP domain and how you configured the tracking option.
To view the status of tracked devices, use the following command:
show esrp <name>
ESRP Tracking Example

Figure 106 is an example of ESRP tracking.
Figure 106: ESRP tracking
ESRP master
200.1.1.1/24
vlan esrp1 (track-vlan)

vlan vlan1
Host 2: Router
200.1.1.14/24
Gateway:
200.1.1.1
L2 switch
10.10.10.121
Host 1:
200.1.1.13/24
Gateway: ESRP slave
200.1.1.1 200.1.1.2/24
EX_094
To configure VLAN tracking, use the following command:
configure esrp esrp1 add track-vlan vlan1
Using the tracking mechanism, if VLAN1 fails, the ESRP master realizes that there is no path to the
upstream router via the master switch and implements an ESRP failover to the slave switch.
To configure route table tracking, use the following command:
configure esrp esrp1 add track-iproute 10.10.10.0/24
The route specified in this command must exist in the IP routing table. When the route is no longer
available, the switch implements an ESRP failover to the slave switch.
To configure ping tracking, use the following command:
configure esrp esrp1 add track-ping 10.10.10.121 frequency 2 miss 2

The specified IP address is tracked. If the fail rate is exceeded, the switch implements an ESRP failover
to the slave switch.
ESRP Port Restart

You can configure ESRP to restart ports in the ESRP master domain when the downstream switch is
from a third-party vendor. This action takes down and restarts the port link to clear and refresh the
downstream ARP table.
To configure port restart, use the following command:
configure esrp ports <ports> restart
To disable port restart, use the following command:
configure esrp ports <ports> no-restart
If a switch becomes a slave, ESRP takes down (disconnects) the physical links of member ports that
have port restart enabled. The disconnection of these ports causes downstream devices to remove the
ports from their FDB tables. This feature allows you to use ESRP in networks that include equipment
from other vendors. After 2 seconds, the ports re-establish connection with the ESRP switch.
To remove a port from the restart configuration, delete the port from the VLAN and re-add it.
ESRP Host Attach

ESRP host attach (HA) is an optional ESRP configuration that allows you to connect active hosts directly
to an ESRP master or slave switch. Normally, the Layer 2 redundancy and loop prevention capabilities
of ESRP do not allow packet forwarding from the slave ESRP switch. ESRP HA allows configured ports
that do not represent loops to the network to continue Layer 2 operation independent of their ESRP
status.
ESRP HA is designed for redundancy for dual-homed server connections. HA allows the network to
continue Layer 2 forwarding regardless of the ESRP status. Do not use ESRP HA to interconnect devices
on the slave ESRP switch instead of connecting directly to the ESRP master switch.
The ESRP HA option is useful if you are using dual-homed network interface cards (NICs) for server
farms, as shown in Figure 107. The ESRP HA option is also useful where an unblocked Layer 2
environment is necessary to allow high-availability security.

Figure 107: ESRP host attach
OSPF/BGP-4
EX_095
ESRP VLANs that share ESRP HA ports must be members of different ESRP groups. Each port can have
a maximum of seven VLANs.
If you use load sharing with the ESRP HA feature, configure the load-sharing group first and then
enable HA on the group.
Other applications allow lower-cost redundant routing configurations because hosts can be directly
attached to the switch involved with ESRP. HA also requires at least one link between the master and
the slave ESRP switch for carrying traffic and to exchange ESRP hello packets.
ESRP domains that share ESRP HA ports must be members of different ESRP groups.
NOTE
Do not use the ESRP HA feature with the following protocols: STP, Ethernet Automatic Protection Switching (EAPS),
or VRRP. A broadcast storm may occur.
To configure a port to be a host port, use the following command:
configure esrp ports <ports> mode [host | normal]
ESRP Port Weight and Don’t Count

In an ESRP domain, the switch automatically calculates the port weight based on the bandwidth of the
port. ESRP uses the port weight to determine the master ESRP switch.
For load-shared ports, configure the master port in the load-share group with the port weight. A load-
shared port has an aggregate weight of all of its member ports. If you add or delete a load-shared port
(or trunk), the master load-shared port weight is updated.
If you do not want to count host ports and normal ports as active, configure the ESRP port weight on
those ports. Their weight becomes 0 and that allows the port to be part of the VLAN, but if a link
failure occurs, it will not trigger a reconvergence. With this configuration, ESRP experiences fewer state

changes due to frequent client activities like rebooting and unplugging laptops. This port is known as a
don’t-count port.
To configure the port weight on either a host attach port or a normal port, use the following command:
configure esrp ports <ports> weight [auto | <port-weight>]
ESRP Groups
ExtremeXOS supports running multiple instances of ESRP within the same VLAN or broadcast domain.
This functionality is called an ESRP group. Although other uses exist, the most typical application for
multiple ESRP groups is when two or more sets of ESRP switches are providing fast-failover protection
within a subnet.
A maximum of seven distinct ESRP groups can be supported on a single ESRP switch, and a maximum
of seven ESRP groups can be defined within the same network broadcast domain. You can configure a
maximum of 32 ESRP groups in a network.
For example, two ESRP switches provide Layer 2/Layer 3 connectivity and redundancy for the subnet,
while another two ESRP switches provide Layer 2 connectivity and redundancy for a portion of the
same subnet. Figure 108 shows ESRP groups.
Figure 108: ESRP groups
ESRP ESRP
Group1 Group1
Master Standby
ESRP ESRP Group2 Standby

Group2 (L2 only)
Master
(L2 only)
EX_096
An additional use for ESRP groups is ESRP HA, described on page 762.

Selective Forwarding
An ESRP-aware switch floods ESRP PDUs from all ports in an ESRP-aware VLAN. This flooding creates
unnecessary network traffic because some ports forward ESRP PDUs to switches that are not running
the same ESRP groups. You can select the ports that are appropriate for forwarding ESRP PDUs by
configuring selective forwarding on an ESRP-aware VLAN and thus reduce this excess traffic.
Configuring selective forwarding creates a port list of only those ports that forward to the ESRP groups
that are associated with an ESRP-aware VLAN. This ESRP-aware port list is then used for forwarding
ESRP PDUs.
NOTE
Extreme Networks recommends keeping the default settings unless you have considerable knowledge and experience
with ESRP.
To configure selective forwarding, use the following command:
configure esrp <domain> aware add selective-forward-ports <portlist> {group <group

number>}
where the following is true:
● domain—Specifies an ESRP domain name.

● portlist—Specifies the ports to be added to the ESRP-aware port list
● group number—Specifies the ESRP group within the given domain name.
When an ESRP-aware switch receives an ESRP PDU on a domain, the software looks up the group to
which the PDU belongs. If the group is found, the ESRP-aware switch processes the PDU then and
forwards it according to the group’s specified aware selective forwarding port list. If no selective
forwarding port list is configured, the switch forwards the PDU from all of the ports of the domain’s
master VLAN. If the group is not found, the PDU is discarded.
When a user adds one or more ports to the ESRP-aware port list (for example, 5:1 and 6:2) that are not
part of the master VLAN, the following message is displayed.
Warning: Port 5:1, 6:2 not currently a member of master vlan
The ports will still be added to the ESRP-aware port list; however, PDUs will not be forwarded out of
those ports until they are added to the master VLAN.
To disable selective forwarding, use the following command:
configure esrp <domain> aware delete selective-forward-ports <all| portlist> {group

<group number>}
where the following is true:
● domain—Specifies an ESRP domain name.

● all—Specifies all of the ports to be disabled.
● portlist—Specifies the selected ports to be disabled.
● group number—Specifies the ESRP group within the given domain name.

Displaying Selective Forwarding Information

To display all selective forwarding information for a given domain, use the following command:
show esrp ><domain> aware
For example, to display the ESRP-aware information for domain d1, enter:
show esrp d1 aware
The output is similar to the following two tables:
Domain: d1
Vlan: vesrp1
-----------------------------------------------------
Group Port Count Selective Forward Ports
-----------------------------------------------------
0 5 5:1, 5:2, 7:31, 7:32: 8:1
3 2 5:1, 8:1
-------------------------------------------------------------------------
Group Master MAC Rx Fwd FDB Flush Fwd Ports
-------------------------------------------------------------------------
0 00:12:00:33:44:55 10 10 1 selective
1 00:22:00:12:21:1F 77 77 3 default
3 00:02:00:13:11:11 99 99 3 selective
The first of the two tables displays the following description of the Selective Forward Ports.
● Group—The number of an ESRP group within the given domain.

● Port Count—The number of ports in the group that are selected to forward PDUs on the master
VLAN.
● Selective Forward Ports—The list of ports in the group that are selected to forward PDUs on the
master VLAN.
To specify this table only, use the following command:
show esrp <domain> aware selective-forward-ports
The second of the two tables displays the following statistical information about port activity.
● Group—The number of an ESRP group within the given domain.

● Master MAC—The MAC address for the master of the group.
● Rx—The number of PDUs received matching the domain/group pair.
● Fwd—The number of PDUs received and forwarded matching the domain/group pair.
● FDB Flush—The number of FDB Flush requests received from the ESRP Master for this domain/
group pair.
● Fwd Ports—Selective or Default.
Selective describes the group as having a configured aware port list for selective forwarding of PDUs
on the Master VLAN. The list of ports is displayed in the first table above.
Default describes the group as one where all the ports on the master VLAN forward the ESRP PDUs
that are received for the domain/group pair. Because there is no selective forwarding configuration
for this group, there is no entry in the first table.

Displaying ESRP Information
To specify this table only, use the following command:
show esrp <domain> aware statistics
Displaying ESRP Information

To view ESRP information, use the following command:
show esrp
Output from this command includes:

● The operational state of an ESRP domain and the state of its neighbor
● ESRP port configurations
To view more detailed information about an ESRP domain, use the following command and specify the
domain name:
show esrp {<name>}
Output from this command includes:

● The operational state of an ESRP domain
● ESRP election policy
● ESRP tracking information
● Timer statistics
● State change information
To view ESRP counter information for a specific domain, use the following command:
show esrp {<name>} counters
To view ESRP-aware information for a specific domain (including the group number, MAC address for
the master, and the age of information) use the following command:
show esrp {<name>}
For more information about any of the commands used to enable, disable, or configure ESRP, refer to
the ExtremeXOS Command Reference Guide.
Using ELRP with ESRP

Extreme Loop Recovery Protocol (ELRP) is a feature of ExtremeXOS that allows you to prevent, detect,
and recover from Layer 2 loops in the network. You can use ELRP with other protocols such as ESRP.
With ELRP, each switch, except for the sender, treats the ELRP protocol data unit (PDU) as a Layer 2
multicast packet. The sender uses the source and destination MAC addresses to identify the packet it
sends and receives. When the sender receives its original packet back, that triggers loop detection and
prevention. After a loop is detected, the loop recovery agent is notified of the event and takes the
necessary actions to recover from the loop. ELRP operates only on the sending switch; therefore, ELRP
operates transparently across the network.

How a loop recovers is dependent upon the protocol that uses the loop detection services provided by
ELRP. If you are using ELRP in an ESRP environment, ESRP may recover by transitioning the ESRP
domain from master to slave. This section describes how ESRP uses ELRP to recover from a loop and
the switch behavior.
Using ELRP with ESRP to Recover Loops

ELRP sends loop-detect packets to notify ESRP about loops in the network. In an ESRP environment,
when the current master goes down, one of the slaves becomes the master and continues to forward
Layer 2 and Layer 3 traffic for the ESRP domain. If a situation occurs when a slave incorrectly
concludes that the master is down, the slave incorrectly assumes the role of master. This introduces
more than one master on the ESRP domain which causes temporary loops and disruption in the
network.
ELRP on an ESRP Pre-Master Switch

A pre-master switch is an ESRP switch that is ready to transition to master but is going through
possible loop detection. A pre-master periodically sends out ELRP loop-detect packets (ELRP PDUs) for
a specified number of times and waits to make sure that none of the sent ELRP PDUs are received.
Transition to master occurs only after this additional check is completed. If any of the ELRP PDUs are
received, the switch transitions from pre-master to slave state. You configure pre-master ELRP loop
detection on a per ESRP domain basis.
ELRP on an ESRP Master Switch

A master switch is an ESRP switch that sends ELRP PDUs on its ESRP domain ports. If the master
switch receives an ELRP PDU that it sent, the master transitions to the slave. While in the slave state,
the switch transitions to the pre-master rate and periodically checks for loops before transitioning to the
master. The pre-master process is described in ELRP on an ESRP Pre-Master Switch on page 768. You
configure the master ELRP loop detection on a per ESRP domain basis.
Configuring ELRP
This section describes the commands used to configure ELRP for use with ESRP. By default, ELRP is
disabled.
Configuring Pre-Master Polling

If you enable the use of ELRP by ESRP in the pre-master state, ESRP requests ELRP packets sent to
ensure that there is no loop in the network before changing to the master state. If no packets are
received, there is no loop in the network. By default, the use of ELRP by ESRP in the pre-master state is
disabled.
To enable the use of ELRP by ESRP in the pre-master state on a per-ESRP domain basis, and to
configure how often and how many ELRP PDUs are sent in the pre-master state, use the following
command:
configure esrp <esrpDomain> elrp-premaster-poll enable {count <count> | interval

<interval>}

Using ELRP with ESRP

● esrpDomain—Specifies an ESRP domain name.
● count—Specifies the number of times the switch sends ELRP PDUs. The default is 3, and the range
is 1 to 32.
● interval—Specifies how often, in seconds, the ELRP PDUs are sent. The default is 1 seconds, and
the range is 1 to 32 seconds.
To disable the use of ELRP by ESRP in the pre-master state, use the following command:
configure esrp <esrpDomain> elrp-premaster-poll disable
Configuring Master Polling

If you enable the use of ELRP by ESRP in the master state, ESRP requests that ELRP packets are
periodically sent to ensure that there is no loop in the network while ESRP is in the master state. By
default, the use of ELRP by ESRP in the master state is disabled.
To enable the use of ELRP by ESRP in the master state on a per-ESRP domain basis, and to configure
how often the master checks for loops in the network, use the following command:
configure esrp <esrpDomain> elrp-master-poll enable {interval <interval>}

● esrpDomain—Specifies an ESRP domain name.
● interval—Specifies how often, in seconds, successive ELRP packets are sent. The default is 1
second, and the range is 1 to 64 seconds.
To disable the use of ELRP by ESRP in the master state, use the following command:
configure esrp <esrpDomain> elrp-master-poll disable
Configuring Ports
You can configure one or more ports of an ESRP domain where ELRP packet transmission is requested
by ESRP. This allows the ports in your network that might experience loops, such as ports that connect
to the master, slave, or ESRP-aware switches, to receive ELRP packets. You do not need to send ELRP
packets to host ports.
By default, all ports of the ESRP domain have ELRP transmission enabled on the ports.
If you change your network configuration, and a port no longer connects to a master, slave, or ESRP-
aware switch, you can disable ELRP transmission on that port. To disable ELRP transmission, use the
following command:
configure esrp <esrpDomain> delete elrp-poll ports [<ports> | all]
To enable ELRP transmission on a port, use the following command:
configure esrp <esrpDomain> add elrp-poll ports [<ports> | all]

Displaying ELRP Information

To display summary ELRP information, use the following command:
show elrp
In addition to displaying the enabled/disabled state of ELRP, the command displays the total number
of:
● Clients registered with ELRP
● ELRP packets transmitted
● ELRP packets received
For more information about the output associated with the show elrp command, see the ExtremeXOS
ESRP Examples
This section provides examples of ESRP configurations.
Single Domain Using Layer 2 and Layer 3 Redundancy

The example shown in Figure 109 uses a number of Extreme Networks devices as edge switches that
perform Layer 2 switching for ESRP domain esrp1 and VLAN Sales. The edge switches are dual-homed
to the BlackDiamond 10808 switches. The BlackDiamond 10808 switches perform Layer 2 switching
between the edge switches and Layer 3 routing to the outside world. Each edge switch is dual-homed
using active ports to two BlackDiamond 10808 switches. ESRP is enabled on each BlackDiamond 10808
switch for the ESRP domain esrp1 that interconnects to the edge switches. Each BlackDiamond 10808
switch has the VLAN Sales configured using the identical IP address. The BlackDiamond 10808 switches
then connect to the routed enterprise normally, using the desired routing protocol (for example, RIP or
OSPF).

ESRP Examples
Figure 109: Single ESRP domain using Layer 2 and Layer 3 redundancy
OSPF or RIP
Domain - esrp1, Domain - esrp1,

VLAN - Sales VLAN - Sales
(master) (standby)
EX_097
The BlackDiamond 10808 switch, acting as master for ESRP domain esrp1, performs both Layer 2
switching and Layer 3 routing services for VLAN Sales. The BlackDiamond 10808 switch in slave mode
for ESRP domain esrp1, performs neither for VLAN Sales, thus preventing bridging loops in the VLAN.
The BlackDiamond 10808 switch.
There are four paths between the BlackDiamond 10808 switches on VLAN Sales. All the paths are used
to send ESRP packets, allowing for four redundant paths for communication. The edge switches, being
ESRP-aware, allow traffic within the VLAN to failover quickly because these edge switches sense when
a master/slave transition occurs and flush FDB entries associated with the uplinks to the ESRP-enabled
BlackDiamond 10808 switches.

The following commands are used to configure both BlackDiamond 10808 switches. In this scenario, the
master is determined by the programmed MAC address of the switch because the number of active
links for the VLAN and the priority are identical to both switches.
This example assumes the following:

● ESRP election algorithm used is the default for standard mode (ports > track > priority >
mac).
● The Inter-router backbone is running OSPF, with other routed VLANs already properly configured.
Similar commands would be used to configure a switch on a network running RIP.
● Ports added to the VLAN have already been removed from VLAN default.
● IP address for the VLANs participating in ESRP must be identical.
NOTE
If your network has switches running ExtremeWare and ExtremeXOS participating in ESRP, Extreme Networks
recommends that the ExtremeXOS switches operate in ESRP standard mode. To change the mode of operation, use
the configure esrp mode [extended | standard] command.
Configuration commands for the BlackDiamond 10808 switches are as follows:
create vlan sales

configure vlan sales add ports 1:1-1:4
configure vlan sales ipaddress 10.1.2.3/24
enable ipforwarding
create esrp esrp1

enable esrp esrp1
configure ospf add vlan sales area 0.0.0.0 passive

configure ospf routerid 5.5.5.5
enable ospf
Multiple Domains Using Layer 2 and Layer 3 Redundancy

The example shown in Figure 110 illustrates an ESRP configuration that has multiple domains using
Layer 2 and Layer 3 redundancy.

ESRP Examples
Figure 110: Multiple ESRP domains using Layer 2 and Layer 3 redundancy
OSPF
or RIP
Sales master, Sales standby,
Engineering standby Engineering master
Sales Sales Sales + Engineering

Engineering
VLAN Sales, ESRP domain esrp1 - untagged link
VLAN Engineering, ESRP domain esrp2 - untagged link
VLANs Sales + Engineering, shared between ESRP domains
esrp1 + esrp2 - tagged link EX_098
This example builds on the previous example. It has the following features:
● An additional VLAN, Engineering, is added that uses Layer 2 redundancy.
● The VLAN Sales uses three active links to each BlackDiamond 10808 switch.
● The VLAN Engineering has two active links to each BlackDiamond 10808 switch.
● One of the edge devices carries traffic for both VLANs.
● The link between the third edge device and the first BlackDiamond 10808 switch uses 802.1Q tagging
to carry traffic from both VLANs traffic on one link. The BlackDiamond switch counts the link active
for each VLAN.
● The second BlackDiamond switch has a separate physical port for each VLAN connected to the third
edge switch.
In this example, the BlackDiamond switches are configured for ESRP such that the VLAN Sales
normally uses the first BlackDiamond switch and the VLAN Engineering normally uses the second
BlackDiamond switch. This is accomplished by manipulating the ESRP priority setting for each VLAN
for the particular BlackDiamond switch.

Configuration commands for the first BlackDiamond 10808 switch are as follows:
create vlan sales
configure vlan sales add ports 1:3, 1:5 tagged
configure vlan engineering add ports 1:4
configure vlan engineering add ports 1:3, 1:5 tagged
configure vlan engineering ipaddress 10.4.5.6/24
create esrp esrp1

configure esrp esrp1 priority 5
enable esrp esrp1
create esrp esrp2

configure esrp esrp2 add master engineering
enable esrp esrp2
Configuration commands for the second BlackDiamond 10808 switch are as follows:
create vlan sales
configure vlan sales add ports 1:5 tagged
configure vlan engineering add ports 1:4, 2:1
configure vlan engineering ipaddress 10.4.5.6/24
configure vlan engineering add ports 1:5 tagged
create esrp esrp1

configure esrp 1 add master sales
enable esrp esrp1
create esrp esrp2

configure esrp esrp2 add master engineering
configure esrp esrp2 priority 5
enable esrp esrp2

ESRP Cautions
ESRP Cautions
This section describes important details to be aware of when configuring ESRP.
Configuring ESRP and IP Multinetting

When configuring ESRP and IP multinetting on the same switch, the same set of IP addresses must be
configured for all involved VLANs.
ESRP and STP

A switch running ESRP should not simultaneously participate in STP for the same VLAN(s). Other
switches in the VLAN being protected by ESRP may run STP; the switch running ESRP forwards, but
does not filter, STP BPDUs. Therefore, you can combine ESRP and STP on a network and a VLAN, but
you must do so on separate devices. You should be careful to maintain ESRP connectivity between
ESRP master and slave switches when you design a network that uses ESRP and STP.
ESRP and VRRP

Do not configure ESRP and VRRP on the same VLAN or port. This configuration is not allowed or
supported.
ESRP Groups and Host Attach

ESRP domains that share ESRP HA ports must be members of different ESRP groups.
Port Configurations and ESRP

The following ports cannot be part of a VLAN that participates in an ESRP domain:
● A mirroring target port
● A software-controlled redundant port
● A Netlogin port
In addition, the following ESRP ports cannot be a mirroring, software-controlled redundant port, or
Netlogin port:
● Host Attach port.
● Don’t-count port. (This port has a port weight of 0.)
● Restart port.


28 Virtual Router Redundancy Protocol

● Overview of VRRP on page 778
● Determining the VRRP Master on page 779
● Additional VRRP Highlights on page 782
● VRRP Operation on page 783
● VRRP Configuration Parameters on page 785
● VRRP Examples on page 786
● VRRP Cautions on page 788
This chapter assumes that you are already familiar with the Virtual Router Redundancy Protocol
(VRRP). If not, refer to the following publications for additional information:
● RFC 2338—Virtual Router Redundancy Protocol (VRRP)
● RFC 2787—Definitions of Managed Objects for the Virtual Router Redundancy Protocol
● Draft IETF VRRP Specification v2.06
Licensing
You need an Advanced Edge, Core, or Advanced Core license to configure and use all of the VRRP
features described in this chapter.
VRRP functionality is not available with an Edge license.
series switch, and Summit X450a series switch ship with an Advanced Edge license.
The Summit X450e series switch ships with an Edge license, which does not support VRRP. You can,
however, upgrade to an Advanced Edge license that supports VRRP.
A SummitStack must be operating at the Advanced Edge license level to use all of the VRRP features
described in this chapter.

Virtual Router Redundancy Protocol
Overview of VRRP
Like the Extreme Standby Router Protocol (ESRP), VRRP allows multiple switches to provide redundant
routing services to users. VRRP is used to eliminate the single point of failure associated with manually
configuring a default gateway address on each host in a network. Without using VRRP, if the
configured default gateway fails, you must reconfigure each host on the network to use a different
router as the default gateway. VRRP provides a redundant path for the hosts. Using VRRP, if the
default gateway fails, the backup router assumes forwarding responsibilities.
VRRP and Hitless Failover—Modular Switches and SummitStack

Only
chassis, one node assumes the role of primary and another node assumes the role of backup. The
primary node executes the switch’s management functions, and the backup acts in a standby role.
Hitless failover transfers switch management control from the primary to the backup and maintains the
state of VRRP. VRRP supports hitless failover. You do not explicitly configure hitless failover support;
rather, if you have two nodes installed, hitless failover is available.
NOTE
protocol, platform, and support for hitless failover, see “Understanding Hitless Failover Support—Modular Switches
and SummitStack Only” in Chapter 4, “Managing the Switch.”
To support hitless failover, the primary node replicates VRRP protocol data units (PDUs) to the backup,
which allows the nodes to run VRRP in parallel. Although both nodes receive VRRP PDUs, only the
primary transmits VRRP PDUs to neighboring switches and participates in VRRP.
NOTE
page 1020 to confirm that the primary and backup nodes are running software that supports the synchronize
command.
To initiate hitless failover on a network that uses VRRP:

(InSync).
After you confirm the primary and backup nodes are synchronized, proceed to step 3.

Determining the VRRP Master

The VRRP master is determined by the following factors:
● VRRP priority—This is a user-defined field. The range of the priority value is 1 to 254; a higher
number has higher priority. The value of 255 is reserved for a router that is configured with the
virtual router IP address. A value of 0 is reserved for the master router, to indicate it is releasing
responsibility for the virtual router. The default value is 100.
● Higher IP address—If the routers have the same configured priority, the router with the higher IP
address becomes the master.
VRRP Tracking
Tracking information is used to track various forms of connectivity from the VRRP router to the outside
world. ExtremeXOS supports the use of the following VRRP tracking options:
● VRRP VLAN Tracking
● VRRP Route Table Tracking
● VRRP Ping Tracking
VRRP VLAN Tracking

You can configure VRRP to track connectivity of up to eight specified VLANs as criteria for failover. If
no active ports remain on the specified VLANs, the router automatically relinquishes master status and
remains in INIT mode.
Configuring VLAN Tracking. To add a tracked VLAN, use the following command:
configure vrrp vlan <vlan_name> vrid <vridval> add track-vlan <target_vlan_name>
Disabling VLAN Tracking. To delete a tracked VLAN, use the following command:
configure vrrp vlan <vlan_name> vrid <vridval> delete track-vlan <target_vlan_name>
VRRP Route Table Tracking

You can configure VRRP to track specified routes in the route table as criteria for VRRP failover. If any
of the configured routes are not available within the route table, the router automatically relinquishes
master status and remains in INIT mode.

Configuring Route Table Tracking. To add a tracked route, use the following command:
configure vrrp vlan <vlan_name> vrid <vridval> add track-iproute <ipaddress>/
<masklength>
Disabling Route Table Tracking. To delete a tracked route, use the following command:
configure vrrp vlan <vlan_name> vrid <vridval> delete track-iproute <ipaddress>/
<masklength>
VRRP Ping Tracking

You can configure VRRP to track connectivity using a simple ping to any outside responder. The
responder may represent the default route of the router, or any device meaningful to network
connectivity of the master VRRP router. If pinging the responder fails the specified number of times,
consecutively, the router automatically re-assumes the master role, as soon as the connection is
reestablished.
Configuring Ping Tracking. To add a tracked ping, use the following command:
configure vrrp vlan <vlan_name> vrid <vridval> add track-ping <ipaddress> frequency
<seconds> miss <misses>
The seconds parameter specifies the number of seconds between pings to the target IP address. The
range is 1 to 600 seconds.
The misses parameter specifies the number of misses allowed before this entry is considered to be
failing. The range is 1 to 255 pings.
Disabling Ping Tracking. To delete a tracked ping, use the following command:
configure vrrp vlan <vlan_name> vrid <vridval> delete track-ping <ipaddress>
Displaying VRRP Tracking Information

To view the status of tracked devices, use the following command:
show vrrp {detail}
VRRP Tracking Example

Figure 111 is an example of VRRP tracking.

Figure 111: VRRP tracking
VRRP master
200.1.1.1/24
(track-vlan)
vlan vlan1
Host 2:
200.1.1.14/24 Router
Gateway:
200.1.1.1
L2 switch
or hub
10.10.10.121
Host 1:
200.1.1.13/24
Gateway: VRRP backup
200.1.1.1 200.1.1.2/24
EX 067
To configure VLAN tracking, as shown in Figure 111, use the following command:
configure vrrp vlan vrrp1 vrid 2 add track-vlan vlan1
Using the tracking mechanism, if VLAN1 fails, the VRRP master realizes that there is no path to
upstream router via the master switch and implements a VRRP failover to the backup.
To configure route table tracking, as shown in Figure 111, use the following command:
configure vrrp vlan vrrp1 vrid 2 add track-iproute 10.10.10.0/24
The route specified in this command must exist in the IP routing table. When the route is no longer
available, the switch implements a VRRP failover to the backup.
To configure ping tracking, as shown in Figure 111, use the following command:
configure vrrp vlan vrrp1 vrid 2 add track-ping 10.10.10.121 frequency 2 miss 2
The specified IP address is tracked. If the fail rate is exceeded, the switch implements a VRRP failover
to the backup. A VRRP node with a priority of 255 may not recover from a ping-tracking failure if there
is a Layer 2 switch between it and another VRRP node. In cases where a Layer 2 switch is used to
connect VRRP nodes, Extreme Networks recommends that those nodes have priorities of less than 255.
Electing the Master Router

VRRP uses an election algorithm to dynamically assign responsibility for the master router to one of the
VRRP routers on the network. A VRRP router is elected master if the router has the highest priority (the
range is 1 to 254; 255 is a reserved number).
If the master router becomes unavailable, the election process provides dynamic failover and the
backup router that has the highest priority assumes the role of master.

A new master is elected when one of the following things happen:

● VRRP is disabled on the master router.
● Loss of communication occurs between master and backup router(s).
● Another VRRP router is attached to the VLAN, and the new router has the same priority as the
current master plus it has the higher IP address and is configured in the preempt mode as well.
When VRRP is disabled on the master interface, the master router sends an advertisement with the
priority set to 0 to all backup routers. This signals the backup routers that they do not need to wait for
the master down interval to expire, and the master election process for a new master can begin
immediately.
The master down interval is set as follows:
3 * advertisement interval + skew time
Where:
● The advertisement interval is a user-configurable option.
● The skew time is (256-priority/256).
NOTE
An extremely busy CPU can create a short dual master situation. To avoid this, increase the advertisement interval.
Additional VRRP Highlights

The following additional points pertain to VRRP:
● VRRP packets are encapsulated IP packets.
● The VRRP multicast address is 224.0.0.18.
● The virtual router MAC address is 00 00 5E 00 01 <vrid>
● Duplicate VRIDs are allowed on the router but not on the same interface.
● The maximum number of supported VRIDs per interface is seven.
● An interconnect link between VRRP routers should not be used, except when VRRP routers have
hosts directly attached.
● A maximum of 128 VRID instances are supported on the router.
● Up to seven unique VRIDs can be configured on the router. VRIDs can be re-used, but not on the
same interface.
● VRRP and the Spanning Tree Protocol (STP) can be simultaneously enabled on the same switch.
● Extreme Networks does not recommend simultaneously enabling VRRP and ESRP on the same
switch.

VRRP Operation
VRRP Operation
This section describes two VRRP network configurations:
● A simple VRRP network
● A fully redundant VRRP network
Simple VRRP Network Configuration

Figure 112 shows a simple VRRP network.
Figure 112: Simple VRRP network
Switch A Switch B
Switch A = Master Switch B = Backup
VRID = 1 VRID = 1
Virtual router IP address = 192.168.1.3 Virtual router IP address = 192.168.1.3
MAC address = 00-00-5E-00-01-01 MAC address = 00-00-5E-00-01-01
Priority = 255 Priority = 100
192.168.1.3 192.168.1.5
Default Gateway = 192.168.1.3

EX_068
In Figure 112, a virtual router is configured on Switch A and Switch B using these parameters:
● VRID is 1.
● MAC address is 00-00-5E-00-01-01.
● IP address is 192.168.1.3.
Switch A is configured with a priority of 255. This priority indicates that it is the master router. Switch
B is configured with a priority of 100. This indicates that it is a backup router.
The master router is responsible for forwarding packets sent to the virtual router. When the VRRP
network becomes active, the master router broadcasts an ARP request that contains the virtual router
MAC address (in this case, 00-00-5E-00-01-01) for each IP address associated with the virtual router.
Hosts on the network use the virtual router MAC address when they send traffic to the default
gateway.
The virtual router IP address is configured to be the real interface address of the IP address owner. The
IP address owner is usually the master router. The virtual router IP address is also configured on each
backup router. However, in the case of the backup router, this IP address is not associated with a
physical interface. Each physical interface on each backup router must have a unique IP address. The
virtual router IP address is also used as the default gateway address for each host on the network.
If the master router fails, the backup router assumes forwarding responsibility for traffic addressed to
the virtual router MAC address. However, because the IP address associated with the master router is

not physically located on the backup router, the backup router cannot reply to TCP/IP messages (such
as pings) sent to the virtual router.
Fully Redundant VRRP Network

You can use two or more VRRP-enabled switches to provide a fully redundant VRRP configuration on
your network. Figure 113 shows a fully redundant VRRP configuration.
Figure 113: Fully redundant VRRP configuration
Switch A Switch B
Master for virtual IP 192.168.1.3 Master for virtual IP 192.168.1.5
Master VRID = 1 Master VRID = 2
Backup for virtual IP 192.168.1.5 Backup for virtual IP 192.168.1.3
Backup VRID = 2 Backup VRID = 1
Default Route Backup Route EX_069
In Figure 113, switch A is configured as follows:

● IP address 192.168.1.3
● Master router for VRID 1
● Backup router for VRID 2
● MAC address 00-00-5E-00-01-01
Switch B is configured as follows:

● IP address 192.168.1.5
● Master router for VRID 2
● Backup router for VRID 1
● MAC address 00-00-5E-00-01-02
Both virtual routers are simultaneously operational. The traffic load from the four hosts is split between
them. Host 1 and host 2 are configured to use VRID 1 on switch A as their default gateway. Host 3 and
host 4 are configured to use VRID 2 on switch B as their default gateway. In the event that either switch
fails, the backup router configured is standing by to resume normal operation.

VRRP Configuration Parameters
VRRP Configuration Parameters

Table 94 lists the parameters that you configure on a VRRP router.
Table 94: VRRP configuration parameters
Parameter Description
vrid This is the virtual router identifier and is a configured item in the
range of 1 to 255. This parameter has no default value. For more
information, see the create vrrp vlan vrid command.
priority This priority value to be used by this VRRP router in the master
election process. A value of 255 is reserved for a router that is
configured with the virtual router IP address. A value of 0 is reserved
for the master router to indicate it is releasing responsibility for the
virtual router. The range is 1 to 254. The default value is 100. For
more information, see the configure vrrp vlan vrid
priority command.
ip_address This is the IP address associated with this virtual router. You can
associate one or more IP addresses to a virtual router. This parameter
has no default value. For more information, see the configure
vrrp vlan vrid add ipaddress command.
advertisement_interval Specifies the time interval between advertisements in seconds unless
otherwise specified as milliseconds. The default is 1 second.
The range is 1 through 255 seconds or 100 through 999
milliseconds.
NOTE: If you must configure the range in milliseconds, specify the
milliseconds keyword. If you enter a number from 100 through 255
and do not specify the milliseconds keyword, the interval defaults to
seconds. For more information, see the configure vrrp vlan
vrid advertisement-interval command.
skew_time This is the time to skew master_down_interval, in seconds. This value
is calculated as ((256-priority)/256).
master_down_interval This is the time interval for the backup router to declare master down,
in seconds. This value is calculated as
((3 * advertisement_interval) + skew_time).
preempt_mode This controls whether a higher priority backup router preempts a lower
priority master. A value of true allows preemption, and a value of false
prohibits preemption. The default setting is true.
NOTE: The router that owns the virtual router IP address always
preempts, independent of the setting of this parameter. For more
information, see the configure vrrp vlan vrid preempt or
the configure vrrp vlan vrid dont-preempt commands.

VRRP Examples
This section provides the configuration syntax for the two VRRP networks discussed in this chapter.
Configuring the Simple VRRP Network

Figure 114 shows the simple VRRP network described in “Simple VRRP Network Configuration” on
page 783.
Figure 114: Simple VRRP network
Switch A Switch B
Switch A = Master Switch B = Backup
VRID = 1 VRID = 1
Virtual router IP address = 192.168.1.3 Virtual router IP address = 192.168.1.3
Priority = 255 Priority = 100
192.168.1.3 192.168.1.5
Default Gateway = 192.168.1.3

EX_068
The following examples assume that you have already created the VLAN named vlan1 on the switch.
Configuration commands for switch A are as follows:

configure vlan vlan1 ipaddress 192.168.1.3/24
create vrrp vlan vlan1 vrid 1
configure vrrp vlan vlan1 vrid 1 prioirty 255
configure vrrp vlan vlan1 vrid 1 add 192.168.1.3
enable vrrp
Configuration commands for switch B are as follows:

enable vrrp

VRRP Examples
Configuring the Fully Redundant VRRP Network

Figure 115 shows the fully redundant VRRP network configuration described in “Fully Redundant
VRRP Network” on page 784.
Figure 115: Fully redundant VRRP configuration
Switch A Switch B
Master for virtual IP 192.168.1.3 Master for virtual IP 192.168.1.5
Master VRID = 1 Master VRID = 2
Backup for virtual IP 192.168.1.5 Backup for virtual IP 192.168.1.3
Backup VRID = 2 Backup VRID = 1
Default Route Backup Route EX_069
The following examples assume that you have already created the VLAN named vlan1 on the switch.
Configuration commands for switch A are as follows:

configure vrrp vlan vlan1 vrid 1 priority 255
enable vrrp
Configuration commands for switch B are as follows:

configure vrrp vlan vlan1 vrid 2 priority 255
enable vrrp

VRRP Cautions
This section describes important details to be aware of when configuring VRRP.
Assigning Multiple Virtual IP Addresses

It is possible to assign multiple virtual IP addresses to the same VRID for a VRRP VR. In this case, you
must meet the following conditions:
● Multiple virtual IP addresses must be on the same subnet.
● The switch cannot own any of the multiple IP addresses.
For example, if you have a VLAN named v1 configured with IP addresses 1.1.1.1/24 and 2.2.2.2/24, the
following configurations are allowed:
● VRRP VR on VLAN v1 with VRID 99 with virtual IP addresses 1.1.1.2 and 1.1.1.3
● VRRP VR on VLAN v1 with VRID 99 with virtual IP addresses 1.1.1.98 and 1.1.1.99
Using the VLAN v1 configuration described above, the following configurations are not allowed:
● VRRP VR on VLAN v1 with VRID 99 with virtual IP addresses 1.1.1.1 and 2.2.2.2 (the IP addresses
are not on the same subnet).
● VRRP VR on VLAN v1 with VRID 99 with virtual IP addresses 1.1.1.1 and 1.1.1.99 (the switch owns
IP address 1.1.1.1).
VRRP and ESRP

Do not configure VRRP and ESRP on the same VLAN or port. This configuration is not allowed or
supported.

29 MultiProtocol Label Switching (MPLS) Protocol
MPLS provides advanced IP services for the BlackDiamond 10808 and BlackDiamond 12000 chassis. The
Ethernet Modules shipped for these chassis contain advanced ASICs that support MPLS. Unlike older
BlackDiamond 6800 chassis, the newer chassis do not require an MPLS module to take advantage of
MPLS features. To configure MPLS on your switch, you need an enabled Feature Pack. Refer to Chapter
2, “ExtremeXOS Overview” for details.
This chapter covers the following topics:

● About MPLS on page 789
● MPLS Data Plane on page 791
● Configuring MPLS on page 799
● Configuration Example on page 807
● MPLS Configuration Constraints on page 809
About MPLS
NOTE
MPLS is only supported on the BD12000 “r” and the BD10808-1XL MSM series MSM modules.
This section describes MPLS and includes the following topics:
● How MPLS Works on page 789

● MPLS Terms and Acronyms on page 790
● Label Switched Paths on page 791
How MPLS Works

MPLS is a technology that allows routers to make protocol-independent forwarding decisions based on
fixed-length labels. The use of MPLS labels enables routers to avoid the processing overhead of delving
deeply into each packet and performing complex route lookup operations based upon destination IP
addresses.
In an MPLS environment, incoming packets are initially assigned “labels” by a Label Edge Router
(LER). The labels allow the packets to be more efficiently handled by MPLS-capable routers at each
point along the forwarding path.
An MPLS label essentially consists of a short fixed-length value carried within each packet header and
that identifies a Forwarding Equivalence Class (FEC). The FEC tells the router how to handle the
packet. An FEC is defined to be a group of packets that are forwarded in the same manner. Examples of
FECs include an IP prefix, a host address, or a VLAN ID. The label concept in MPLS is analogous to
other connection identifiers, such as an ATM VPI/VCI or a Frame Relay DLCI.

MultiProtocol Label Switching (MPLS) Protocol
By mapping to a specific FEC, the MPLS label efficiently provides the router with all of the local link
information needed for immediate forwarding to the next hop. MPLS creates a Label Switched Path
(LSP) along which each Label Switch Router (LSR) can make forwarding decisions based solely upon
the content of the labels. At each hop, the LSR simply strips off the existing label and applies a new one
that tells the next LSR how to forward the packet. This allows packets to be tunneled through an IP
network.
Figure 116 illustrates an MPLS network.
Figure 116: MPLS network
Egress
LER
Ingress
LER Destination IP
Source IP LSR network
network LSR
LSP
MPLS cloud MPLS_05
MPLS Terms and Acronyms

Table 95 defines common MPLS terms and acronyms.
Table 95: MPLS Terms and Acronyms

Term or Acronym Description
CSPF Constrained Shortest Path First. Route selection determined by an algorithm based on
available link bandwidth and path cost.
DoD Downstream-on-Demand. Distribution of labels as a result of explicit upstream label
requests.
DU Downstream Unsolicited. Distribution of labels downstream without an explicit label
request.
EXP bits A three-bit “experimental” field in an MPLS shim header.
FEC Forward Equivalence Class. A group of packets that are forwarded in the same manner
(for example, over the same Label Switched Path).
Label A short, fixed-length identifier used to forward packets from a given link.
Label stack A set of one or more MPLS labels used by MPLS to forward packets to the appropriate
destination.
Label swapping Lookup and replacement of an incoming label with the appropriate outgoing label.
LDP Label Distribution Protocol. A protocol defined by the IETF used to establish an MPLS
Label Switched Path (LSP).
LER Label Edge Router. A Label Switch Router that is at the beginning (ingress) or end
(egress) of a Label Switched Path.
LSP Label Switched Path. The unidirectional MPLS connection between two routers over
which packets are sent.
LSR Label Switch Router. A router that receives and transmits packets on an MPLS network.

MPLS Data Plane
Table 95: MPLS Terms and Acronyms (Continued)

Term or Acronym Description
MPLS MultiProtocol Label Switching. A set of protocols defined by the IETF used to transmit
information based on a label-switching forwarding algorithm.
NHLFE Next Hop Label Forwarding Entry. The NHLFE represents the MPLS router next hop
along the LSP.
PHP Penultimate Hop Popping. A label stack optimization used for conserving the number of
allocated labels.
PW Pseudo Wire. A logical point-to-point connection.
RSVP Resource ReSerVation Protocol. A resource setup protocol designed for an integrated
services network.
RSVP-TE The combination of RSVP and MPLS label signaling to provide traffic engineered LSPs
as specified in RFC 3209, RSVP-TE: Extensions to RSVP for LSP Tunnels.
Shim header MPLS-specific header information that is inserted between layer-2 and layer-3
information in the data packet.
SP Service Provider. An entity that provides network services for individuals or organizations.
TE Traffic Engineering. The provisioning of an autonomous flow along a specified network
path.
Transport LSP Any active LSP used to forward traffic through an MPLS network.
VPLS Virtual Private LAN Services. A multipoint Layer-2 VPN service that has the
property that all PW tunnels within a VPN are signaled with the same vcid,
where the vcid represents the VPN identifier.
VPN Virtual Private Network. A logical private network domain that spans a public or service
provider network infrastructure.
Label Switched Paths

Protocols that use label switching are connection-oriented. In MPLS, the connections are called Label
Switched Paths (LSPs) and are unidirectional in nature.
LSPs are established using LDP or RSVP-TE. Once established, an LSP can be used to carry IP traffic or
to tunnel other types of traffic, such as bridged MAC frames. The tunnel aspects of LSPs, which are
important in supporting virtual private networks (VPNs), result from the fact that forwarding is based
solely on labels and not on any other information carried within the packet.
MPLS Data Plane

This section describes how MPLS and IP routing work together to forward information on your
network.
This section covers the following topics:
● Label Switch Routers on page 792
● Supporting Quality of Service Features on page 793
● MPLS Layer on page 793
● Routing Using LSPs on page 795
● About MPLS Layer-2 VPNs on page 798

MPLS provides a great deal of flexibility for routing packets. Received IP unicast frames can be routed
normally or tunneled through LSPs. If a matching FEC exists for a received packet, the packet may be
transmitted using an LSP that is associated with the FEC. The packet is encapsulated using an MPLS
shim header before being transmitted.
Received MPLS packets can be label switched or routed normally toward the destination. Packets that
are in the middle of an LSP are label switched. The incoming label is swapped for a new outgoing label
and the packet is transmitted to the next LSR. For packets that have arrived at the end of an LSP (the
egress end of the LSP), the label is popped. If this label is the bottom of the stack, the shim header is
stripped and the packets are routed to the destination as normal IP packets.
NOTE
Multicast routing is not supported.
An MPLS domain is generally defined to be an OSPF or IS-IS autonomous system (AS). You can use
MPLS to reach destinations inside one of these AS types. You can also use MPLS to tunnel through all
or part of an AS in order to reach destinations outside of the AS.
Label Switch Routers

MPLS protocols are designed primarily for routed IP networks and are implemented by Label Switch
Routers (LSRs). The router where an LSP originates is called the ingress LSR, while the router where an
LSP terminates is called the egress LSR.
Ingress and egress LSRs are also referred to as Label Edge Routers (LERs). For any particular LSP, a
router is either an ingress LER, an intermediate LSR, or an egress LER. However, a router may function
as an LER for one LSP, while simultaneously functioning as an intermediate LSR for another LSP.
Figure 117 illustrates the three types of LSRs.
Figure 117: LSR types
LSR for
LSP A Destination
LSP A IP network
Ingress for LSP B
LSP B
LER LSR
Egress LER
Source IP for LSP B
network MPLS cloud
Destination
Egress LER for LSP A IP network
for LSP A
MPLS_12

MPLS Data Plane
The functions of the LSR types are described in Table 96.
Table 96: LSR Functions

LSR Function
Ingress LER Inserts one or more labels into packets transmitted onto an LSP.
Intermediate LSR Forwards packets via label swapping.
Egress LER Removes the last label(s) before forwarding packets received from an LSP.
Supporting Quality of Service Features

Quality of Service (QoS) LSP support is an important attribute of MPLS. MPLS supports the
Differentiated Services (DiffServ) model of QoS. The DiffServ QoS model is supported by mapping
different traffic classes to different LSPs, or by using the EXP bits in the MPLS shim header to identify
traffic classes with particular forwarding requirements.
MPLS Layer
This section describes the MPLS layer and includes the following topics:
● MPLS Shim Header on page 793
● MPLS Label Stack on page 794
● Penultimate Hop Popping on page 794
● Label Binding on page 795
● Label Space Partitioning on page 795
MPLS can be thought of as a shim-layer between layer 2 and layer 3 of the protocol stack. MPLS
provides connection services to layer-3 functions while making use of link-layer services from layer-2.
To achieve this, MPLS defines a shim header that is inserted between the link layer header and the
network layer header of transmitted frames. The format of a 32-bit MPLS shim header is illustrated in
Figure 118.
Figure 118: MPLS shim header
Label EXP bottom-of-stack TTL

(20 bits) (3 bits) (1 bits) (8 bits)
MPLS_01
MPLS Shim Header

The MPLS shim header contains the following fields:
● 20-bit label
● 3-bit experimental (EXP) field
The EXP field can be used to identify different traffic classes to support the DiffServ QoS model.

● 1-bit bottom-of-stack flag

The bottom-of-stack bit is set to 1 to indicate the last stack entry.
● 8-bit Time-To-Live (TTL) field.
The TTL field is used for loop mitigation, similar to the TTL field carried in IP headers.
MPLS Label Stack

The format of an MPLS label stack containing two MPLS shim header entries is shown in Figure 119.
Figure 119: MPLS label stack
bottom-of-stack bottom-of-stack
Label 1 EXP =0 TTL Label 2 EXP =1 TTL
MPLS_02
Figure 120 illustrates the format of a unicast MPLS frame on an Ethernet link. The MAC addresses are
those of the adjacent MPLS router interfaces. The x8847 Ethertype value indicates that the frame
contains a MPLS unicast packet. A different Ethertype value (x8848) is used to identify MPLS multicast
packets.
Figure 120: MPLS unicast frame on Ethernet
Ethertype MPLS remainder

MAC DA MAC SA
x8847 label stack of frame
MPLS_03
Figure 121 shows the format of a unicast MPLS frame that contains an 802.1Q VLAN tag. In both cases,
the Ethertype values no longer identify the network layer protocol type. This implies that, generally, the
protocol type must be inferable from the MPLS label value(s). For example, when only one type of
protocol is carried on a given LSP.
Figure 121: MPLS unicast frame on tagged Ethernet VLAN
Ethertype Ethertype MPLS remainder

MAC DA MAC SA VLAN tag
x8100 x8847 label stack of frame
MPLS_04
NOTE
For more detailed information on MPLS encapsulations, see RFC 3032, MPLS Label Stack Encoding.
Penultimate Hop Popping

Penultimate hop popping (PHP) is an LSR label stack processing optimization feature. When enabled,
the LSR can “pop” (or discard) the remaining label stack and forward the packet to the last router along
the LSP as a normal Ethernet packet.

MPLS Data Plane
By popping the label stack one hop prior to the LSP egress router, the egress router is spared having to
do two lookups. After the label stack has been popped by the penultimate hop LSR, the LSP egress
router must only perform an address lookup to forward the packet to the destination.
PHP label advertisements using implicit NULL labels can be optionally enabled. Support for receiving
implicit NULL label advertisements by neighbor LSRs is always enabled. For example, if an LSR
advertises implicit NULL labels for IP prefixes, the neighbor LSRs must support PHP.
Label Binding
Label binding is the process of, and the rules used to, associate labels with FECs. LSRs construct label
mappings and forwarding tables that comprise two types of labels: labels that are locally assigned and
labels that are remotely assigned.
Locally assigned labels are labels that are chosen and assigned locally by the LSR. For example, when
the LSR assigns a label for an advertised direct interface. This binding information is communicated to
neighboring LSRs. Neighbor LSRs view this binding information as remotely assigned.
Remotely assigned labels are labels that are assigned based on binding information received from
another LSR.
Label Space Partitioning

The Extreme MPLS implementation supports approximately 64 K locally-assigned labels. The Extreme
implementation partitions its label space as described in Table 97.
Table 97: MPLS Label Space Partitions

Label Range Label Partition Description
x00000-x0000f Defined/reserved by MPLS standards specified in RFC 3032.
x00010-x7fbff Dynamic LSR Partition—Used to identify intermediate LSR LSPs.
x7fc00-x803ff Static LSR Partition—Used to identify statically defined LSR LSPs.
x80400-xfffff Dynamic Egress Label Partition—Used for egress LER LSPs.
The data path uses the label partition bits in conjunction with the bottom-of-stack flag to efficiently
determine how a label should be processed.
No hard limits are imposed on the maximum size of the label stack, other than the constraint of not
exceeding the maximum frame size supported by the physical links comprising the LSP. You should
enable jumbo frame support on all ports. The jumbo frame size should be set to accommodate the
addition of a maximally-sized label stack. For example, a jumbo frame size of at least 1530 bytes is
needed to support a two-level label stack on a tagged Ethernet port and a jumbo frame size of at least
1548 bytes is needed to support a VPLS encapsulated MPLS frame.
Routing Using LSPs

● Routing Using Matching and Calculated LSP Next Hops on page 796
● Matching LSP Next Hops on page 797

● OSPF Calculated LSP Next Hops on page 797

● BGP Calculated LSP Next Hops on page 797
● LSP Precedence and Interaction on page 797
● Multivendor Support for Calculated LSPs on page 798
Routing Using Matching and Calculated LSP Next Hops

Normally, a route table prefix is associated with a gateway or next hop IP address. Using MPLS, a route
table prefix can also be associated with an LSP that can be used as the "next hop." There are two types
of LSP next hops that can be used to route a packet to its destination:
● Matching LSP next hop

An LSP is considered matching with respect to an FEC if it has been associated with that FEC via
LDP or RSVP-TE. An example of this is an IPv4 prefix for which a matching label mapping has been
received by LDP. Matching LSPs are supported for all route origin types.
● Calculated LSP next hop
An LSP is considered calculated with respect to an FEC if it has been associated with that FEC via a
routing protocol. Both OSPF and BGP can perform the calculations necessary to associate a route
table prefix with an LSP next hop.
Figure 122 illustrates the concept of matching and calculated LSPs.
Figure 122: Matching and calculated LSP next hops
LSR A LSR B LSR C

Router ID Router ID Router ID
10.3.1.1 Subnet 10.2.1.1 Subnet 10.1.1.1
10.0.1.0/24
30
31
32 10.0.2.0/24
Matching
33
LSPs
34
10.0.3.0/24
Bound labels
for LSR A MPLS_17
Table 98 describes the label bindings in the MPLS forwarding table for LSR A that are maintained for
FECs reachable via LSR A to LSR C, shown in Figure 122.
Table 98: Label Bindings for LSR A

Matching LSP Calculated LSP
Destination Next Hop Next Hop Label Next Hop Label
10.1.1.1/32 10.2.1.1 31 30
10.0.1.0/24 10.2.1.1 32 31
10.0.2.0/24 10.2.1.1 33 31
10.0.3.0/24 10.2.1.1 34 31

MPLS Data Plane
Matching LSP Next Hops

A matching LSP next hop is always preferred over a calculated LSP next hop. Matching LSP next hop
entries are added to the route table when an LSP becomes operational. They are deleted when an LSP is
no longer operational.
OSPF Calculated LSP Next Hops

Managing calculated LSP next hop entries is more involved. The OSPF Shortest Path First (SPF)
algorithm checks the availability of LSPs to remote OSPF routers during a calculation. The intra-area
SPF algorithm begins with the calculating router as the root of a graph. The graph is expanded by
examining the networks connected to the root and then examining the routers connected to those
networks. Continuing in this manner, the graph is built as a series of parent and child nodes. A check is
made for a matching LSP next hop as each entry is added. A check is also made for an LSP next hop
that can be inherited from the parent node. These inherited LSP next hops are referred to as "calculated"
LSP next hops. Thus, for each route table entry, the modified SPF algorithm determines whether a
matching LSP next hop is available and whether a calculated LSP next hop is available for use
whenever a matching LSP next hop is not present.
The modification to the SPF algorithm described above is important, because it enables the capabilities
provided by LDP or RVSP-TE LSPs to be fully utilized, while minimizing the resources devoted to label
management.
For example, in a network where all the LERs/LSRs implement this feature (such as an all-Extreme
MPLS network), labels only need to be advertised for the router IDs of the LERs/LSRs. Yet, LSPs can
still be used to route traffic destined for any OSPF route.
More specifically, LSPs can be used for all routes advertised by OSPF, with the possible exception of
LDP LSPs to routes summarized by OSPF area border routers (ABRs). The problem with using routes
summarized by OSPF ABRs is that route summarization can prevent label mappings from being
propagated for the links internal to the area being summarized, since an LSR will only propagate LDP
labels for FECs that exactly match a routing table entry.
BGP Calculated LSP Next Hops

BGP can also calculate how to use LSPs to reach BGP next hops. For example, an IBGP session is
established across the OSPF/MPLS backbone, and the communicating routers run both OSPF and IBGP.
When an IBGP route is installed, BGP determines whether a matching LSP next hop exists to the
destination. If not, it checks for an LSP to the BGP next hop. If an LSP exists to the BGP next hop, that
LSP is used as an LSP next hop for the IBGP route.
The recalculation requirements for BGP are similar to those for OSPF; when an LSP to a BGP next hop
router changes state, the BGP routing table entries must be checked to ensure their LSP next hop
information is still valid.
LSP Precedence and Interaction

A longest prefix match (LPM) is determined for all packets. If an LSP next hop is available, routed IP
traffic may be forwarded over an LSP using the LSP next hop. With respect to a given prefix, LSP next
hops can be either matching or calculated, and can be based on LDP or RSVP-TE LSPs. Matching LSP

next hops are preferred over calculated LSP next hops and RSVP-TE based LSPs are preferred over LDP
LSPs. Also, RSVP-TE LSPs can be configured to enable or disable their use as LSP next hops.
Therefore, if a more preferred LSP is established, routed IP traffic may begin to use a new LSP next
hop. Likewise, if a preferred LSP is torn down, routed traffic may begin to use the next best LSP next
hop. These changes can take place when there is an OSPF routing topology change, an LDP label
advertisement event, or a RSVP-TE signaling action.
Multivendor Support for Calculated LSPs

Unfortunately, some MPLS implementations do not support the ability to IP forward packets received
on an egress LSP to their OSPF router ID and/or BGP next hop address. If your MPLS network includes
equipment that does not support this type of IP forwarding, you can use configuration commands to
explicitly control the use of calculated LSP next hops.
The following configuration commands are available:
enable/disable iproute mpls-next-hop
This command enables and disables all use of LSP next hops. No IP traffic will be routed over an LSP
when mpls-next-hop IP routing capability is disabled.
enable/disable ospf mpls-next-hop
This command enables and disables the calculation of LSP next hops by OSPF.
enable/disable bgp mpls-next-hop
This command enables and disables the calculation of LSP next hops by BGP.
About MPLS Layer-2 VPNs

As networks grow and become more pervasive, the need to separate the physical network infrastructure
from the logical network or VLAN organization has become increasingly important. By logically
separating the network topology from the service provided by the physical network, services are more
easily managed, reliability through increased redundancy is improved, and you gain more efficient use
of the physical network infrastructure.
By mapping a VLAN to a specific set of MPLS pseudo wires (PWs), you can create virtual private
networks (VPNs). Within a VPN, all traffic is opaquely transported across the service provider network.
Each VPN can be managed and provisioned independently. See RFC 4447 and RFC 4448 for additional
information.
VPNs may have two or more customer points of presence (PoP). All PoPs are interconnected using
point-to-point connections called pseudo wires. If there are two PoPs in the VPN, the VPN is considered
to be point-to-point. If there are more than two PoPs in the VPN, the VPN is considered to be
multipoint. Multipoint VPNs must be fully-meshed. A multipoint VPN is also referred to as a Virtual
Private LAN Service (VPLS).
Layer-2 VPNs are constructed from a set of interconnected point-to-point MPLS pseudo wires. Pseudo
wire endpoint nodes operate as virtual VPN switches, bridging traffic between pseudo wires and the
local egress VLAN. Source MAC addresses within each VPN are associated with the local VLAN or the
pseudo wire from which the packet is received. Up to 224K MAC addresses can be cached. However,

Configuring MPLS
VPLS consumes 2 entries, reducing the maximum number of VPLS learned MAC addresses to 224K/2.
Within a VPN, once a MAC address has been learned, unicast traffic destined to the cached MAC
address is transmitted over a single pseudo wire. Integrated VPN MAC caching enhancement increases
network performance and improves VPN scalability.
For information about configuring MPLS Layer-2 VPNs, see Chapter 3, “Configuring MPLS VPLS
Layer-2 VPNs”.
Configuring MPLS
This section describes how to configure MPLS.
This section includes the following topics:

● Configuring MPLS LSR ID on page 799
● Configuring MPLS Interfaces on page 799
● Enabling MPLS on page 800
● Enabling MPLS Interfaces on page 800
● Propagation of IP TTL on page 800
● Configuring Penultimate Hop Popping on page 801
● Configuring QoS Mappings on page 801
● Mapping Dot1p to EXP Bits on page 802
● Resetting MPLS Configuration Parameter Values on page 802
● Displaying MPLS Configuration Information on page 803
Configuring MPLS LSR ID

The MPLS LSR ID must be configured before MPLS can be used. The address chosen must be a routable
IP address on the switch. It is suggested that this be set to the same IP address as a routing protocol ID
(for example, the OSPF Router ID). To configure the MPLS LSR ID, use the following command:
configure mpls lsr-id <ipaddress>
NOTE
The MPLS LSR ID must be configured before MPLS can be enabled. The LSR ID should be configured on a
loopback VLAN.
Configuring MPLS Interfaces

To use MPLS on a VLAN, MPLS must first be configured on that VLAN. To configure a specific VLAN
or all VLANs, use the following command:
configure mpls add [{vlan} <vlan_name> | vlan all]]
MPLS must be configured on a VLAN before it can be used to transmit or receive MPLS-encapsulated
frames. By default, MPLS is not configured on a newly created VLAN.

If all VLANs are selected, MPLS is configured on all VLANs. By default, neither LDP nor RSVP-TE are
configured on a VLAN. One of these protocols must be configured on a VLAN in order to establish
LSPs.
If you have enabled MPLS on an OSPF interface that is used to reach a particular destination, make sure
that you enable MPLS on all additional OSPF interfaces that can reach that same destination (for
example, enable MPLS on all VLANs that are connected to the backbone network).
Enabling MPLS
To enable MPLS on the switch, use the following command:
enable mpls
By default, MPLS is disabled on the switch. This command starts the MPLS process allowing MPLS
protocols to run and MPLS packets to be forwarded. Even though MPLS is enabled globally, individual
MPLS protocols must also be enabled and MPLS must be enabled on the MPLS configured VLANs.
To enable MPLS on specific VLANs, use the following command:
enable mpls [{vlan}<vlan_name>|vlan all]
Again, configuring a VLAN for MPLS does not enable MPLS on the VLAN. The VLAN must be
specifically MPLS enabled in order to send and receive MPLS packets over this interface.
MPLS must be enabled on all VLANs that transmit or receive MPLS-encapsulated frames. By default,
MPLS and MPLS label distribution protocols are disabled when MPLS is first configured on a VLAN.
Enabling MPLS Interfaces

To use MPLS on a VLAN, a MPLS label distribution protocol as well as MPLS itself must be enabled on
the VLAN. If all VLANs are selected, LDP is enabled on all VLANs for which MPLS has been
configured.
If you have enabled LDP and MPLS on an IGP interface that is used to reach a particular destination,
make sure that you enable LDP and MPLS on all additional IGP interfaces that can reach that same
destination (for example, enable LDP and MPLS on all OSPF VLANs that are connected to the backbone
network).
Propagation of IP TTL
The MPLS TTL is set to 255 on all packets originated by the switch. The MPLS TTL is also set to 255 on
all packets that enter a Pseudo Wire.
There are two modes of operation for routed IP packets: uniform TTL and pipe TTL mode. Currently,
switches that run Extreme OS support only the pipe TTL mode. In pipe TTL mode, the LSP is viewed as
a point-to-point link between the ingress LSR and the egress LSR; intermediate LSRs in the MPLS
network are not viewed as router hops from an IP TTL perspective. Thus, the IP TTL is decremented
once by the ingress LSR, and once by the egress LSR. In this mode, the MPLS TTL is set to 255 by the
ingress LSR, and is independent of the IP TTL.

Configuring MPLS
Configuring Penultimate Hop Popping

To enable or disable PHP, use the following command:
enable mpls php [{vlan} <vlan_name> | vlan all]
This command enables or disables whether PHP is requested by the egress LER. If vlan all is selected,
PHP is enabled on all VLANs on which MPLS has been configured. By default, PHP is disabled.
When PHP is enabled, PHP is requested on all LSPs for which the switch is the egress LER.
PHP is requested by assigning the Implicit Null Label in an advertised mapping. PHP is always
performed when requested by a peer (for example, when acting as an intermediate LSR).
Configuring QoS Mappings

ExtremeXOS provides examination and replacement services for the EXP field. These services behave
like the dot1p QoS commands. When EXP examination is enabled, the EXP value from the received
frame is used to assign the packet to a QoS profile. Once a packet is assigned to a QoS profile, the EXP
value may be overwritten in hardware before being transmitted. By default, the QoS profile EXP value
is equivalent to the QoS profile number one (QP1). To enable QoS for MPLS LSPs, use the following
command:
enable mpls exp examination
This command enables EXP examination for MPLS received packets. When EXP examination is enabled,
the EXP field in the outer or top label of the label stack is used to assign the received packet to an
internal switch qosprofile. If enabled, all MPLS packets are mapped to a qosprofile based on the
configured EXP bit value. That is, a packet with an EXP value of 0 is mapped to qosprofile QP1, a
packet with an EXP value of 1 is mapped to qosprofile QP2, etc. By default EXP examination is disabled
and all received MPLS packets are sent to QP1.
Each EXP value can be assigned a qosprofile. Multiple EXP values can be assigned to the same
qosprofile. The following command is used to configure the switch to route packets with a received
EXP value to a specific qosprofile.
configure mpls exp examination {value} <value> {qosprofile} <qosprofile>
The switch can overwrite the EXP value in the outer label of an MPLS packet. The following command
is used to enable the switch to replace the EXP value:
enable mpls exp replacement
This command enables EXP replacement for MPLS transmitted packets. When EXP replacement is
enabled, the EXP field in the outer or top label of the label stack is replaced with the EXP value
configured for the QoS profile for which the packet was assigned. By default, EXP replacement is
disabled and packets are propagated without modifying the EXP field value.
Each qosprofile can be assigned an EXP value used to overwrite the EXP bit field in the outer label of
the transmitted packet. All qosprofiles can be configured to overwrite the EXP bit field using the same
EXP value. The following command is used to configure the switch to overwrite MPLS packets
transmitted from a specific qosprofile with a new EXP value.
configure mpls exp replacement {qosprofile} <qosprofile> {value} <value>

Mapping Dot1p to EXP Bits

The priority of Ethernet tagged packets can be mapped into the MPLS network and vice versa using the
switch fabric qosprofile. Ethernet packets are assigned to a qosprofile by enabling dot1p examination.
MPLS packets are assigned to a qosprofile by enabling exp examination. When the packets egress the
switch, the dot1p and exp bit fields can be overwritten.
Enabling exp replacement instructs the switch to overwrite both the dot1p field and the exp field in the
outer most label.
Figure 123: Mapping Dot1p to EXP Bits
BlackDiamond 12K
MPLS Network
qosprofile 1-8
Pri=7
Received packets 1 Transmitted packets Pri=7
Pri=2 exp=7
2
Pri=1 3 Pri=2
exp=2
4
5 Pri=1
exp=1
6
7
8
By default, when dot1p examination and exp replacement are not enabled, all received packets be
routed through qosprofile qp1 and the packets will be transmitted with the dot1p and exp value of
zero. As shown in the figure above, the switch can be configured to route packets to a specific
qosprofile based on the dot1p value in the 802.1p bit field. In this example dot1p examination is
enabled. By default, when dot1p examination is enabled, packets are sent to the qosprofile that
corresponds to the dot1p value. A dot1p value of 0 maps to qosprofile 1, a dot1p value of 1 maps to
qosprofile 2, and so on. By default, mpls exp replacement is disabled. By enabling mpls exp
replacement, MPLS packets are transmitted with the configured qosprofile dot1p and exp value. By
default, these values correspond to the qosprofile. Qosprofile 1 overwrites the dot1p and exp fields
with 0, qosprofile 2 overwrites the dot1p and exp fields with 1, and so on.
To configure the reverse packet flow, mpls exp examination and dot1p replacement must be configured.
Enabling mpls exp examination instructs the switch to route received MPLS packets to a qosprofile
based on the EXP value. Enabling dot1p replacement instructs the switch to write the dot1p value in the
transmitted packet.
Resetting MPLS Configuration Parameter Values

To reset MPLS configuration parameters to their default values, first disable MPLS using the following
command:
disable mpls
and then, clear the MPLS configuration and reset the MPLS parameters to their default values using the
following command:
unconfigure mpls

Configuring MPLS
These commands affect the following configuration parameters:
● All VLANs are removed from MPLS.

● All EXP (examination and replacement) QOS mappings are reset.
● LSR-ID is reset.
● LDP and RSVP-TE are globally disabled.
● MPLS Traps are disabled.
● LDP and RSVP-TE timers are reset.
● LDP Advertisement settings are reset.
● LDP Loop Detection settings are reset.
● All RSVP-TE LSPs are deleted.
● All RSVP-TE Paths are deleted.
● All RSVP-TE Profiles are deleted.
● The default RSVP-TE Profile is reset.
Displaying MPLS Configuration Information

This section explains how to display the following MPLS configuration information:
● Displaying MPLS Basic Configuration Information on page 803
● Displaying MPLS Interface Information on page 804
● Displaying MPLS Label Information on page 804
● Displaying MPLS Label Mapping Information on page 806
● Displaying MPLS QoS Mapping Information on page 806
Displaying MPLS Basic Configuration Information

To display basic MPLS configuration information, use the following command:
show mpls
This command displays the general configuration of all the MPLS components and system wide
configuration. The output, shown below, displays the switch MPLS, RSVP-TE, and LDP configuration
status. It also shows the configuration status of SNMP trap, EXP examination, and EXP replacement
settings. The configured LSR ID and all configured MPLS VLANs are also shown.
* BD-10K.4 # show mpls

MPLS System
MPLS Admin Status : Enabled
MPLS Oper Status : Enabled
RSVP-TE Admin Status : Enabled
RSVP-TE Oper Status : Enabled
LDP Admin Status : Enabled
LDP Oper Status : Enabled
SNMP Traps : Disabled
EXP Examination : Disabled
EXP Replacement : Disabled
LSR ID : 192.99.1.5

MPLS VLANs : loopback

: blowingrock
: boone
: asheville
Displaying MPLS Interface Information

To display the MPLS interface information, use the following command:
show mpls interface {{vlan} <vlan_name>} {detail}
When the optional parameters are omitted, this command displays information for all the configured
MPLS VLAN interfaces. The summary MPLS interface information displayed includes the configured IP
address, approximate time RSVP-TE and LDP have been up, number of neighbors or adjacencies, and a
set of status flags.
* BD-10K.2 # show mpls interface
Local RSVP-TE LDP

VLAN Name IP Address MTU UpTm #Nbr UpTm #Adj Flags
---------------- --------------- ---- ---- ---- ---- ---- ------
loopback 192.99.1.5 1500 0m 0 16h 0 M-L-IU
blowingrock 192.60.40.5 1500 0m 0 14h 1 M-L-IU
boone 192.100.40.5 1500 0m 0 16h 1 M-L-IU
asheville 192.80.40.5 1500 0m 0 16h 1 M-L-IU
Flags: (M) MPLS Enabled, (R) RSVP-TE Enabled, (L) LDP Enabled,
(P) PHP Enabled, (I) IP Forwarding Enabled, (U) MPLS Operational
When the vlan parameter is specified, this command displays the current MPLS interface summary
configuration and status for only the specified vlan. If the optional detail keyword is specified, the
summary information is displayed in the detail format.
Displaying MPLS Label Information

This command displays the labels that have been received and advertised by LDP and RSVP-TE.
Advertised labels are labels that originate from the local switch and are associated with egress LSPs or
the incoming portion of a transit LSP. Received labels are labels that have originated from a peer or
neighboring LSR and are associated with ingress LSPs or the outgoing portion of a transit LSP.
To display a summary of how many MPLs labels have been advertised and received, use the following
command:
* BD-10K.13 # show mpls label summary

Total number of RSVP-TE labels advertised : 2
Total number of RSVP-TE labels received : 1
Total number of LDP labels advertised : 4
Total number of LDP labels received : 37
To display the MPLS labels advertised to other LSRs and MPLS labels received from other LSRs, use the
following command:
show mpls {ldp | rsvp-te} label {summary | {{advertised | received} {<label_number> |

implicit-null}}}

Configuring MPLS
Advertised labels represent egress LSPs or the incoming portion of a cross connect transit LSP. Looking
at the example below, advertised label 0x80403 has been sent to peer LSRs. Peer LSRs will insert this
label for packets trying to reach devices on the 192.80.40.0/24 subnet. Because this is an egress LSP, the
NHop Type field is vlan, indicating that the Next Hop specifies a local VLAN. The NextHop interface
for this label is identified by the VLAN asheville. There's no LSP name, since this label has been
advertised by LDP. RSVP-TE advertised labels are always associated with a named LSP. This is shown
in the example below for labels 0x80405 and 0x80406.
* BD-10K.14 # show mpls label advertised
Advertised Destination LSP Peer NHop

Label Mapping Flags Label Type NextHop LSPName
-----------------------------------------------------------------------------
0x80402 192.99.1.5/32 LE -- vlan loopback --
0x80403 192.80.40.0/24 LE -- vlan asheville --
0x80404 192.100.40.0/24 LE -- vlan boone --
0x80405 192.80.40.5/32 RE -- locl 192.80.40.5 Lsp3-5test2
0x80406 192.99.1.5/32 RE -- locl 192.99.1.5 Lsp3To5
0x8040b 192.60.40.0/24 LE -- vlan blowingrock --
(M)Multiple Next Hops (L)LDP (R)RSVP-TE

(T)Transit LSP (I)Ingress to LSP (E)Egress from LSP
Received labels represent ingress LSPs or the outgoing portion of a cross connect transit LSP. These are
labels that have been sent by other LSRs to the local switch. When forwarding traffic into an MPLS
network, the local switch will insert one of these labels. For example, label 0x80401 is used to forward
traffic destined to subnet 192.24.100.0/24. These packets are sent to the peer LSR using the next hop IP
address of 192.100.40.3.
* BD-10808.15 # show mpls label received

Received Destination LSP
Label Mapping Flags NextHop LSPName
----------------------------------------------------------------------------
0x80404 192.99.1.3/32 RI 192.100.40.3 Lsp5To3
0x80401 192.24.100.0/24 LI 192.100.40.3 --
0x80403 192.99.1.3/32 LI 192.100.40.3 --
0x80405 192.10.50.0/24 LI 192.100.40.3 --
0x00013 10.20.30.0/24 LI 192.100.40.3 --
0x00014 11.136.96.0/24 LI 192.100.40.3 --
0x00015 192.99.1.4/32 LI 192.100.40.3 --
0x00016 11.95.96.0/24 LI 192.100.40.3 --
0x00017 11.98.96.0/24 LI 192.100.40.3 --
0x00018 11.100.100.5/32 LI 192.100.40.3 --
0x00019 11.100.100.6/32 LI 192.100.40.3 --
0x0001a 11.100.100.7/32 LI 192.100.40.3 --
0x0001b 11.100.100.9/32 LI 192.100.40.3 --
0x0001c 11.100.100.48/32 LI 192.100.40.3 --
0x0001d 11.100.100.100/32 LI 192.100.40.3 --
0x0001e 11.100.100.203/32 LI 192.100.40.3 --
0x0001f 11.100.100.204/32 LI 192.100.40.3 --
0x00034 11.100.100.20/32 LI 192.100.40.3 --
(M)Multiple Next Hops (L)LDP (R)RSVP-TE

(T)Transit LSP (I)Ingress to LSP (E)Egress from LSP

Displaying MPLS Label Mapping Information

To display MPLS label mapping information, use the following command:
show mpls [ldp | rsvp-te] lsp {summary | {{advertised | received} {<label_number> | implicit-null}}}
This command displays information about how to forward packets that arrive labeled as MPLS packets.
As such, it shows how labels advertised to upstream peers are mapped to labels received from
downstream peers. This mapping is sometimes referred to as the Incoming Label Map (ILM).
When the label_number parameter is omitted, summary information is displayed for all incoming label
assignments that have been made by the switch. When the label_number is specified, summary
information is displayed for the specified label.
As can be seen below, the output display differs for label mappings signaled by LDP and RSVP-TE.
Please see the respective sections for additional information.
* BD-10K.5 # show mpls ldp lsp
Prefix Adv Label Peer Label Next Hop VLAN
192.168.0.4/32 0x80402 -- -- lpbk
192.168.0.2/32 0x00039 0x80400 11.0.2.2 vlan2
11.0.4.0/24 0x0003A 0x80401 11.0.2.2 vlan2
192.168.0.4/32 0x0003B 0x00013 11.0.2.2 vlan2
* BD-10K.6 # show mpls rsvp-te lsp

Ingress LSP Name Path Name Destination Transmit I/F UpTm
Flags
---------------- ---------------- --------------- ---------------- ----
--------
LSR1-LSR4 path1-2-4 192.168.0.4 vlan2 47m
UEP--OIV
Egress LSP Name Source IP Destination Receive I/F UpTm

---------------- --------------- --------------- ---------------- ----
LSR4-LSR1 192.168.0.4 192.168.0.1 vlan1 47m
Transit LSP Name Source IP Destination Receive I/F Transmit I/F UpTm
---------------- --------------- --------------- ------------- ------------- ----
LSR2-LSR3 192.168.0.2 192.168.0.3 vlan2 vlan1 47m
Flags: (U) Up, (E) Enabled, (P) Primary LSP, (S) Secondary LSP,
(R) Redundant Paths, (B) Bandwidth Requested, (O) ERO Specified,
(I) IP Traffic Allowed, (V) VPN Traffic Allowed,
(v) VPN Assigned Traffic Allowed
Displaying MPLS QoS Mapping Information

To display MPLS QoS mapping information, use the following command:
* BD-10K.10 # show mpls exp examination
EXP --> QoS Profile mapping:
00 --> QP1
01 --> QP2
02 --> QP3
03 --> QP4
04 --> QP5
05 --> QP6

Configuration Example
06 --> QP7
07 --> QP8
EXP Examination is disabled
* BD-10K.11 # show mpls exp replacement
QoS Profile --> EXP mapping:
QP1 --> 00
QP2 --> 01
QP3 --> 02
QP4 --> 03
QP5 --> 04
QP6 --> 05
QP7 --> 06
QP8 --> 07
EXP Replacement is disabled
* BD-10K.12 #
Configured mappings for both dot1p-to-exp and exp-to-dot1p are displayed.
The network configuration, shown in Figure 124, illustrates how to configure a BlackDiamond switch to
support a routed MPLS network.
Figure 124: MPLS configuration example
11
/24 .
0.
.0 LSR 3 vl 3.
.1 an 0/2
.0 n1 Router ID = 3 4
11 vla
192.168.0.3
9.9.9.0/24 12.12.12.0/24
unc OSPF backbone area duke
and
11 MPLS domain
LSR 1 .0 24 LSR 4
.2 0/
Router ID = vla .0 4. Router ID =
n2 /24 .0
.
n4
192.168.0.1 11 vla 192.168.0.4
LSR 2
Router ID = 192.168.0.2 MPLS_18
The four switches, labeled LSR 1, LSR 2, LSR 3, and LSR 4, have the same physical hardware
configuration. Each switch contains a GM-20XT, XM-2X and an MSM-5 module. The switches are all
interconnected via Gigabit Ethernet to form the OSPF backbone area and the MPLS domain. In this

example, two directly connected OSPF-disabled VLANs are shown: unc and duke. Traffic between unc
and duke follows routed paths over calculated LSPs established between LSR 1 and LSR 4.
The commands used to configure LSR 1 are described below. The remaining LSRs are configured
similarly.
The following commands configure the module types for the specific BlackDiamond slots:
configure slot 2 module xm-2x
configure slot 3 module gm-20xt
The following command sets the maximum jumbo frame size for the switch chassis to 1600:
configure jumbo-frame-size size 1600
The following commands create the VLANs:
create vlan lpbk

create vlan vlan1
create vlan vlan2
create vlan unc
The following commands configure the VLAN IP address and assign ports participating in each VLAN:
configure vlan lpbk ipaddress 192.168.0.1/32
enable ipf lpbk
enable loopback-mode lpbk
configure vlan default delete ports all
configure vlan vlan1 add port 3:2 untagged
configure vlan vlan2 add port 3:3 untagged
configure vlan unc ipaddress 9.9.9.1/24
configure vlan unc add port 3:4 untagged
The following commands enable IP forwarding on the configured VLANs. The MTU size is increased
on the MPLS VLANs to accommodate the MPLS shim header:
enable ipforwarding vlan vlan1
configure ip-mtu 1550 vlan vlan1
enable ipforwarding vlan vlan2
configure ip-mtu 1550 vlan vlan2
enable ipforwarding vlan unc
The following command configures the MPLS LSR ID:
configure mpls lsr-id 192.168.0.1
The following commands configure MPLS on VLANs vlan1 and vlan2:

configure mpls add vlan vlan1
configure mpls add vlan vlan2
The following commands enable LDP and MPLS on VLANs vlan1 and vlan2:
enable mpls ldp vlan1
enable mpls vlan1
enable mpls ldp vlan2
enable mpls vlan2

MPLS Configuration Constraints
The following command allows LDP to advertise a label mapping for the LSR ID:
configure mpls ldp advertise lsr-id
The following commands globally enable LDP and MPLS on the switch:
enable mpls protocol ldp
enable mpls
The following commands add vlan1 and vlan2 to the backbone area, each with a cost of 10. The 0.0.0.0
(backbone) area does not need to be created because it exists by default:
configure ospf add vlan vlan2 area 0.0.0.0
configure ospf vlan vlan2 cost 10
configure ospf add vlan vlan1 area 0.0.0.0
configure ospf vlan vlan1 cost 10
The following command enables distribution of local (direct) interfaces into the OSPF area:
enable ospf export direct cost 10 type ase-type-1
The following command configures the OSPF router ID on the switch and enables the distribution of a
route for the OSPF router ID in the router LSA. Originating the router ID as a host route allows other
routers in the same OSPF area to establish calculated LSPs for external routes to this router:
configure ospf routerid 11.0.1.11
It also allows this router to be a peer in a L2 VPN.
The following command enables OSPF:

enable ospf
MPLS Configuration Constraints

MPLS has the following configuration constraints:
● GVRP—GVRP is not supported over MPLS LSPs.

● Server Load Balancing (SLB)—SLB and MPLS are mutually exclusive functions. Both functions
cannot be simultaneously enabled.
● IP flow redirection—IP flow redirection commands and MPLS are mutually exclusive functions.
Both functions cannot be enabled simultaneously.
● IGMP snooping—OSPF and LDP session establishment require the MSM to receive and process IP
multicast frames. Therefore, IGMP snooping must be enabled to support MPLS.
● VPLS—VPLS requires that IGMP snooping be disabled on customer-facing VLANs.


30 Label Distribution Protocol
This chapter describes the Label Distribution Protocol (LDP), part of MPLS, and covers the following
topics:
● Overview of LDP on page 811

● MPLS LDP Signaling Modes on page 812
● Configuring LDP on page 814
The LDP features supported in this release include:
● Downstream Unsolicited label advertisement

● Liberal label retention
● Ordered control mode
● Advertisement of labels for direct interfaces, RIP routes, and static routes
● LDP Loop Detection
● Configurable LDP Timers
Overview of LDP
The Label Distribution Protocol (LDP) is a protocol defined by the IETF for the purpose of establishing
MPLS LSPs. Using LDP, peer LSRs exchange label binding information to create LSPs. This section
describes the Extreme Networks implementation of LDP, and includes the following topics:
● LDP Neighbor Discovery on page 811
● Advertising Labels on page 812
● Propagating Labels on page 812
LDP Neighbor Discovery

LDP includes a neighbor discovery protocol that runs over UDP. Using the basic discovery mechanism,
each LSR periodically multicasts a hello message to a well-known UDP port to which all LSRs listen.
These hello messages are transmitted to the all routers on this subnet multicast group. When a neighbor is
discovered, a hello-adjacency is formed and the LSR with the numerically greater IP address is denoted
as the active LSR.
Hello messages must continue to be received periodically for the hello-adjacency to be maintained. The
hold time that specifies the duration for which a hello message remains valid can be negotiated by the
peer LSRs as part of the HELLO exchange. During the HELLO exchange, each LSR proposes a value
and the lower of the two is used as the hold time.
Targeted LDP hello-adjacencies between potentially nondirectly connected LSRs are supported using an
extended discovery mechanism. In this case, targeted hello messages are periodically sent to a specific
IP address.

Label Distribution Protocol
After the hello-adjacency is formed, the active LSR initiates establishment of a TCP connection to the
peer LSR. At this point, an LDP session is initiated over the TCP connection. The LDP session consists
of an exchange of LDP messages that are used to setup, maintain, and release the session.
Advertising Labels
You can control whether label mappings are advertised for:
● Direct routes
● RIP routes
● Static routes
In these cases the switch is acting as the egress LER for these LSPs.
Propagating Labels
LDP propagates label mappings for FECs that exactly match a routing table entry. In the case of label
mappings received from an LDP peer, LDP checks for an exactly matching entry with a next hop IP
address that is associated with the LDP peer from which a label mapping was received.
MPLS LDP Signaling Modes

This section describes the LDP signaling modes, and includes the following topics:
● Label Advertisement Modes on page 812
● Label Retention Modes on page 813
● LSP Control Modes on page 813
Label Advertisement Modes

LDP provides two modes for advertising labels:
● Downstream-on-demand (DoD)
● Downstream unsolicited (DU)
Using DoD mode, label bindings are only distributed in response to explicit requests. A typical LSP
establishment flow begins when the ingress LER originates a label request message to request a label
binding for a particular FEC (for a particular IP address prefix or IP host address). The label request
message follows the normal routed path to the FEC. The egress LER responds with a label mapping
message that includes a label binding for the FEC. The label mapping message then follows the routed
path back to the ingress LSR, and a label binding is provided by each LSR along the path. LSP
establishment is complete when the ingress LER receives the label mapping message.
Conversely, using DU mode, an LSR may distribute label bindings to LSRs that have not specifically
requested them. These bindings are distributed using the label mapping message, as in downstream-on-
demand mode. From an LDP message perspective, the primary difference using DU mode is the lack of
a preceding label request message.

MPLS LDP Signaling Modes
Architecturally, the difference is more significant, because the DU mode is often associated with a
topology-driven strategy, where labels are routinely assigned to entries as they are inserted into the
routing database. In either case, an LSR only uses a label binding to switch traffic if the binding was
received from the current next hop for the associated FEC.
Both label advertisement modes can be concurrently deployed in the same network. However, for a
given adjacency, the two LSRs must agree on the discipline. Negotiation procedures specify that DU
mode be used when a conflict exists when using Ethernet links. Label request messages can still be used
when MPLS is operating in unsolicited mode.
The Extreme LDP implementation supports DU mode only.
Label Retention Modes

LDP provides two modes for label retention:
● Conservative
● Liberal
Using conservative label retention mode, an LSR retains only the label-to-FEC mappings that it
currently needs (mappings received from the current next hop for the FEC). Using liberal retention
mode, LSRs keep all the mappings that have been advertised to them. The trade-off is memory
resources saved by conservative mode versus the potential of quicker response to routing changes made
possible by liberal retention (for example, when the label binding for a new next hop is already resident
in memory).
The Extreme MPLS implementation supports liberal label retention, only.
LSP Control Modes

LDP provides two LSP control modes:
● Independent
● Ordered
Using independent LSP control, each LSR makes independent decisions to bind labels to FECs. By
contrast, using ordered LSP control, the initial label for an LSP is always assigned by the egress LSR for
the associated FEC (either in response to a label request message or by virtue of sending an unsolicited
label mapping message).
More specifically, using ordered LSP control, an LSR only binds a label to a particular FEC if it is the
egress LSR for the FEC, or if it has already received a label binding for the FEC from its next hop for the
FEC. True to its name, the mode provides a more controlled environment that yields benefits such as
preventing loops and ensuring use of consistent FECs throughout the network.
The Extreme MPLS implementation supports ordered LSP control, only.

Configuring LDP
This section describes the following tasks:
● Enabling LDP on the Switch on page 814
● Enabling LDP on a VLAN on page 814
● Enabling LDP Loop Detection on page 815
● Configuring an LDP Label Advertisement Filter on page 815
● Configuring LDP Session Timers on page 815
● Displaying LDP Information on page 816
● Displaying LDP Interface Information on page 817
● Displaying LDP Peer Session Information on page 818
● Displaying LDP Protocol Counters on page 818
● Displaying LDP Labels on page 819
● Displaying LDP LSP Forwarding Database on page 821
Enabling LDP on the Switch

To enable LDP on the switch, use the following command:
enable mpls protocol ldp
By default, LDP is disabled on the switch. This command enables MPLS to process LDP control packets
and to advertise and receive LDP labels.
Globally enabling the LDP protocol does not enable LDP on MPLS configured VLANs. The VLAN must
be specifically LDP-enabled in order to set up LDP adjacencies over the interface. See the next section
for enabling LDP on a VLAN.
Enabling LDP on a VLAN

Before LDP can be enabled on a VLAN, the VLAN must be configured for MPLS. Use the following
command to configure MPLS on a VLAN:
configure mpls add [{vlan} <vlan_name> | vlan all]
LDP is disabled by default when a VLAN is configured for MPLS.
To enable LDP on a VLAN, use the following command:

enable mpls ldp [{vlan} <vlan_name> | vlan all]
This command enables LDP on one or all VLANs for which MPLS has been configured.
To disable LDP on a VLAN, use the following command:
disable mpls ldp [{vlan} <vlan_name> | vlan all]
This command disables LDP on one or all VLANs for which MPLS has been configured. This command
terminates all LDP Hello Adjacencies and all established LDP LSPs that use the specified interface(s).

Configuring LDP
Enabling LDP Loop Detection

There are two types of LDP loop detection. By default both are disabled. Enabling loop detection allows
the switch to detect if a label advertisement loop exists within the network. Both hop count and path
vector loop detection methods are used if loop detection is enabled. To enable LDP loop detection, use
enable mpls ldp loop-detection
To disable LDP loop detection, use the following command:
disable mpls ldp loop-detection
These commands affect the LDP loop detection behavior for all LDP enabled VLANs.
Configuring an LDP Label Advertisement Filter

To configure an LDP label advertisement filter, use the following command:
configure mpls ldp advertise [{direct [all | lsr-id | none]} | {rip [all | none] |
{static [all | none]}
This command configures a filter to be used by LDP when originating unsolicited label mapping
advertisements to LDP neighbors.
You can configure how the advertisement filter is applied, as follows:
● direct—The advertisement filter is applied to the FECs associated with direct routes.
● rip—The advertisement filter is applied to the FECs associated with RIP routes.
● static—The advertisement filter is applied to the FECs associated with static routes.
You can configure the advertisement filter, as follows:
● all— Unsolicited label mappings are originated for all routes of the specified type (direct, RIP, or
static).
● lsr-id—An unsolicited label mapping is originated if a /32 direct route exists that matches the MPLS
LSR ID. This filter is the default setting for direct routes and is only available for direct routes.
● none—No unsolicited label mappings are originated for all routes of the specified type. This is the
default setting for RIP and static routes.
You can control the number of labels advertised using the configure mpls ldp advertise
command. Advertising labels for a large number of routes may increase the required number of labels
that must be allocated by LSRs. Care should be used to insure that the number of labels advertised by
LERs does not overwhelm the label capacity of the LSRs.
Configuring LDP Session Timers

To configure LDP session timers, use the following command:
configure mpls ldp timers [targeted | link] [{hello-time <hello_hold_seconds>} {keep-

alive-time <keep_alive_hold_seconds>}]

This command configures the LDP peer session timers for the switch. The LDP peer session timers are
separately configurable for link and targeted LDP hello adjacencies.
The hello-time <hello_hold_seconds> parameter specifies the amount of time (in seconds) that a
hello message received from a neighboring LSR remains valid. The rate at which hello messages are sent
is one third the configured hello-time. If a hello message is not received from a particular neighboring
LSR within the specified hello-time <hello_hold_seconds> then the hello-adjacency is not
maintained with that neighboring LSR. Should two peers have different configured hello-time values,
they will negotiate to use the lower value.
The session keep-alive time <keep_alive_hold_seconds> parameter specifies the time (in seconds)
during which an LDP message must be received for the LDP session to be maintained. The rate at
which keep alive messages are sent, provided there are no LDP messages transmitted, is one sixth the
configured keep-alive-time. If an LDP PDU is not received within the specified session keep-alive
time <keep_alive_hold_seconds> interval, the corresponding LDP session is torn down. Should two
peers have different configured keep-alive-time values, they will negotiate to use the lower value.
In the event that two peers have both a link and a targeted hello adjacency, the hello-time values for the
two hello adjacencies are negotiated separately. The keep-alive-time value is established based on
negotiations occurring when the LDP session is established following the first hello adjacency to be
established.
The minimum and maximum values for both the hello-time <hello_hold_seconds> and keep-
alive time <keep_alive_hold_seconds> are 6 and 65,534, respectively. Changes to targeted timers
only affect newly created targeted peers. Disabling and then enabling all VPLSs will cause all current
targeted peers to be re-created.
The default values are as follows:
● link ldp hello-time <hello_hold_seconds> – 15

● targeted-ldp hello-time <hello_hold_seconds> – 45
● link ldp hello-time <interval_time> – auto set to 1/3 the configured hello-time
● targeted-ldp hello-time <interval_time> – auto set to 1/3 the configured hello-time
● link ldp keep-alive <keep-alive hold-seconds> – 40
● targeted-ldp keep-alive <keep-alive hold-seconds> – 60
● link ldp keep-alive <interval_time> – auto set to 1/6 the configured keep-alive time
● targeted-ldp keep-alive <interval_time> – auto set to 1/6 the configured keep-alive time
Restoring LDP Session Timers

To restore the default values for LDP session timers, use the following command:
unconfigure mpls
This command can only be executed when MPLS is disabled and will restore all MPLS configuration
settings.
Displaying LDP Information

To display basic LDP configuration information, use the following command:
show mpls ldp

Configuring LDP
This command displays the general configuration of LDP. Some settings are not configurable. These
fields are identified with an asterisk (*). The remaining fields can be modified using LDP configuration
commands to change the behavior of LDP. A list of VLANs that have LDP enabled is shown at the
bottom of the display output.
* BD-10k.6 # show mpls ldp

LDP Status : Enabled
Protocol Version : v1*
Label Retention Mode : Liberal*
Label Distribution Method : Downstream Unsolicited*
LDP Loop Detection

Status : Disabled
Hop-Count Limit : 255
Path-Vector Limit : 255
LDP Targeted Timers

Hello Hold : 45 seconds
Keep Alive Hold : 60 seconds
LDP Link Timers

Hello Hold : 15 seconds
Keep Alive Hold : 40 seconds
Label Advertisement
Direct : All
Rip : None
Static : None
LDP VLANs : loopback

: blowingrock
: boone
: asheville
* Indicates parameters that cannot be modified
Displaying LDP Interface Information

To display LDP interface information, use the following command:
show mpls ldp interface {{vlan} <vlan_name>} {detail | counters}

This command displays the operational LDP interface information. All the VLANs that have LDP
enabled are listed. The negotiated LDP hello hold time is displayed. This is not the configured LDP
hello hold time but represents the value that is negotiated between the local switch and the connected
LDP peer. Hello messages are transmitted every 1/3 the negotiated hello hold time. In this example,
that would be every 5 seconds. The hello timer status information is displayed in milliseconds. The
approximate LDP uptime and status flags are also shown.
* BD-10K.7 # show mpls ldp interface
VLAN Name #Adj NegHHldTm NxtHello UpTm Flags
---------------- ---- --------- -------- ---- -----
loopback 0 15000 2230 16h MLU
blowingrock 1 15000 2200 14h MLU
boone 1 15000 2210 16h MLU

asheville 1 15000 2210 16h MLU
Flags: (M) MPLS Enabled, (L) LDP Enabled,

(U) LDP Operational
Displaying LDP Peer Session Information

To display MPLS LDP peer session information, use the following command:
show mpls ldp peer {<ipaddress>} {detail}
This command displays information about the status of LDP peer sessions. Summary information is
displayed for all known LDP peers and LDP peer sessions. If you specify the <ipaddress> of an LDP
peer, information for the single LDP peer is displayed. If you specify the detail keyword, additional
information is displayed in a comprehensive detailed format.
By default the information displayed includes:
● Peer sessions
● Peer state
● Uptime
● Number of hello adjacencies
If you specify the detail keyword, the following additional information is displayed:
● Discontinuity time
● Negotiated label distribution
● Next hop address
● Keep-Alive hold timer
● Hello adjacency details
Displaying LDP Protocol Counters

LDP control protocol packet error counters are maintained per interface. To view these counters, use the
following command:
show mpls ldp interface {{vlan} <vlan_name>} {detail | counters}
These counters may be useful in determining LDP issues between peers. Error counters that continually
increment should be investigated.
* BD-10K.20 # show mpls ldp interface vlan blowingrock counters
VLAN: blowingrock (192.60.40.5)

Link Targeted
Counter Adjacencies Adjacencies
------------------------------ ----------- -----------
Shutdown Notifications (Rcvd) 0 0
Shutdown Notifications (Sent) 0 0
Failed Session Attempts (NAKs) 0 0
Hello Errors 0 0
Parameters Advertised Errors 0 0

Configuring LDP
Max PDU Length Errors 0 0

Label Range Errors 0 0
Bad LDP ID Errors 0 0
Bad PDU Length Errors 0 0
Bad Msg Length Errors 0 0
Bad TLV Length Errors 0 0
Bad TLV Value Errors 0 0
Keep-Alive Timeout Errors 0 0
Omitting the optional vlan parameter will display counters for all LDP enabled
interfaces.
Clearing LDP Protocol Counters

To clear the LDP control protocol error counters, use the following command
clear counters mpls ldp {{{vlan} <vlan_name>} | lsp all}
Omitting the optional vlan parameter will clear counters for all LDP enabled interfaces.
Displaying LDP Labels

To display MPLS LDP labels, use the following command:
show mpls ldp label {advertised | received | retained | summary | <label_number>}
Both advertised and received LDP labels are displayed. The format is the same as the show mpls label
display format except that only LDP labels are displayed.
* BD-10K.11 # show mpls ldp label
Advertised Destination LSP Peer NHop

Label Mapping Flags Label Type NextHop Name
-------------------------------------------------------------------------------
0x80402 192.99.1.5/32 LE -- vlan loopback --
0x80403 192.80.40.0/24 LE -- vlan asheville --
0x80404 192.100.40.0/24 LE -- vlan boone --
0x8040b 192.60.40.0/24 LE -- vlan blowingrock --
0x80410 11.100.100.2/32 LE -- VPLS -- ncsu-vpn
0x80411 11.100.100.3/32 LE -- VPLS -- ncsu-vpn
0x80412 11.100.100.4/32 LE -- VPLS -- ncsu-vpn
Received Destination LSP

Label Mapping Flags NextHop Name
------------------------------------------------------------------------------
0x80401 192.24.100.0/24 LI 192.100.40.3 --
0x80403 192.99.1.3/32 LI 192.100.40.3 --
0x80405 192.10.50.0/24 LI 192.100.40.3 --
0x00013 10.20.30.0/24 LI 192.100.40.3 --
0x00014 11.136.96.0/24 LI 192.100.40.3 --
0x00015 192.99.1.4/32 LI 192.100.40.3 --
0x00016 11.95.96.0/24 LI 192.100.40.3 --
0x00017 11.98.96.0/24 LI 192.100.40.3 --

0x00018 11.100.100.5/32 LI 192.100.40.3 --

0x00019 11.100.100.6/32 LI 192.100.40.3 --
0x0001a 11.100.100.7/32 LI 192.100.40.3 --
0x0001b 11.100.100.9/32 LI 192.100.40.3 --
0x0001c 11.100.100.48/32 LI 192.100.40.3 --
0x0001d 11.100.100.100/32 LI 192.100.40.3 --
0x0001e 11.100.100.203/32 LI 192.100.40.3 --
0x0001f 11.100.100.204/32 LI 192.100.40.3 --
0x00034 11.100.100.20/32 LI 192.100.40.3 --
0x80441 11.100.100.2/32 LI 11.100.100.2 ncsu-vpn
0x80441 11.100.100.3/32 LI 11.100.100.3 ncsu-vpn
0x80444 11.100.100.4/32 LI 11.100.100.4 ncsu-vpn
Optionally specifying the advertised or received keyword limits the display to only the type of label
specified.
When the label for a VPLS is displayed, the Destination Mapping field contains the IP address of the
VPLS peer to/from which the label is being sent/received. The name field contains the local name of
the VPLS. For received labels, the NextHop field also contains the IP address of the VPLS peer.
Specifying the retained keyword allows the display of LDP label mappings received from LDP peers
that are being liberally retained. These can include label mappings for IPv4 prefixes for which no
corresponding route table entry exists. They can also include label mappings for VPLSs for which no
corresponding VPLS has been configured locally.
* BD-10K.21 # show mpls ldp label retained lsp
Peer Next
Prefix Label Hop Vlan
--------------------------------------------------------------------
14.14.14.0/24 0xcb008 192.100.40.3 boone
11.132.96.0/24 0x00082 192.100.40.3 boone
11.131.96.0/24 0x00084 192.100.40.3 boone
* BD-10K.21 # show mpls ldp label retained pseudo-wire
Peer Cntl Requested

Label Peer VPN-ID Word MTU
---------------------------------------------------
0x8c003 11.100.100.4 1000 No 1500
When displaying retained VPLS labels, the Cntl Word field indicates if the peer has requested the
inclusion of a control word when encapsulating the VPLS traffic as described in RFC 4385. Extreme
switches do not support the inclusion of a control word.
The Requested MTU field indicates the MTU value signaled by the peer. This value must match the
locally configured value for the VPLS pseudo wire to become operational. See the Configuring VPLS
MTU section for additional information.

Configuring LDP
Displaying LDP LSP Forwarding Database

To display information about LDP LSPs, use the following command:
show mpls ldp lsp {prefix <ipNetmask>} {egress | ingress | transit} {detail}
This command displays the LDP LSPs established to, from, and through this switch. By default, ingress,
egress, and transit LSPs are all displayed. By optionally specifying the LSP type, the output display is
filtered to show only the type of LSPs specified.
When all LSP types are being displayed, LSPs that show only an advertised label represent egress LSPs
from the network perspective or incoming LSPs from the switch perspective. LSPs that show only a
received label represent ingress LSPs from the network perspective or outgoing LSPs from the switch
perspective. LSPs that show both an incoming and an outgoing label represent transit LSPs. As Extreme
switches are merge-capable, all LDP transit LSPs can also be used as ingress LSPs.
The significance of the VLAN information shown depends on the LSP type. For ingress and transit
LSPs, the indicated VLAN is the MPLS interface used to reach the next hop peer. For egress LSPs, there
is no associated MPLS next hop interface. When the prefix being advertised is associated with a local
(direct) VLAN, that VLAN name is displayed. When the advertised prefix is associated with a static or
an RIP route, the VLAN field is empty.
Advertised labels have switch-wide significance and are generally advertised out multiple interfaces.
* BD-10K.15 # show mpls ldp lsp

Prefix Adv Label Peer Label Next Hop VLAN
192.99.1.5/32 0x80402 -- -- loopback
192.80.40.0/24 0x80403 -- -- asheville
192.100.40.0/24 0x80404 -- -- boone
192.60.40.0/24 0x8040b -- -- blowingrock
192.24.100.0/24 0x00015 0x80401 192.100.40.3 boone
192.99.1.3/32 0x00016 0x80403 192.100.40.3 boone
192.10.50.0/24 0x00018 0x80405 192.100.40.3 boone
10.20.30.0/24 0x0001c 0x00013 192.100.40.3 boone
11.136.96.0/24 0x0001d 0x00014 192.100.40.3 boone
Specifying the optional detail keyword displays each LSP in detail format. Additionally, received
packets and bytes are maintained for transit and egress (or incoming) LSPs. Specifying the keyword
prefix and a matching ipNetmask will restrict the display to a single entry.
*BD-10808.17 # show mpls ldp lsp prefix 11.108.96.0/24 detail

FEC IP/Prefix: 11.108.96.0/24:
Advertised Label: 0 (0)
Received Label : 0x18 (24)
Next Hop IP : 192.100.40.3
VLAN Name : boone
Packets received : 489
Bytes received : 46944


31 Configuring MPLS VPLS Layer-2 VPNs
This chapter describes VPLS Layer-2 VPN services and the following topics:
● Overview of MPLS Layer-2 VPNs on page 823

● VPLS Layer 2 VPN Characteristics on page 828
● Configuring MPLS Layer-2 VPNs on page 828
● VPLS VPN Configuration Examples on page 833
Overview of MPLS Layer-2 VPNs

The idea behind VPLS L2 VPNs over MPLS is to enable Layer-2 virtual private networking (VPN)
service offerings in a simple manner that is easy to deploy and operate. Layer-2 VPN services, based on
a combination of Ethernet and MPLS/IP technologies, are designed to enable service providers to offer
Ethernet business private line services. These services are also referred to as Virtual Private LAN
Services (VPLS). Layer-2 VPN services use a simple Layer-2 interface at the customer edge combined
with the resilience and scalability of an MPLS/IP core to provide VPN connectivity.
This section describes Layer-2 VPNs and includes the following topics:
● VPLS Support on page 823

● Layer-2 VPN Service Deliminators on page 824
● MPLS Pseudo Wires on page 824
● Layer-2 VPN Domains on page 826
● MAC Learning on page 827
● Spanning Tree Protocols on page 827
● IP Protocol Considerations on page 827
VPLS Support
The LDP L2 VPN VPLS implementation includes support for:
● LDP signaling support for PWid FEC for pseudo wire establishment.
● Use of LDP or RSVP-TE to establish transport LSPs.
● Tunnel endpoints are identified via configured IP addresses.
● Different VLAN IDs at each end of a pseudo wire (PW), with the VLAN ID set by the egress switch
to match that of the locally configured VLAN.
● Operations as PE VPLS node.
● VLAN, VMAN, and port edge services (no port-qualified VLAN service).
● Flooding Layer-2 packets to multiple PWs.
● VPLS peer scaling up to 32 peers per VPLS instance.

Configuring MPLS VPLS Layer-2 VPNs
NOTE
The implementation does not include support for Pseudo Wire participation in running the Spanning Tree Protocol.
Layer-2 VPN Service Deliminators

Service deliminators are used to define how the customer is identified to the VPLS. There are multiple
types of service deliminators. ExtremeXOS currently supports three types. The first is a VLAN service.
This service transparently interconnects two or more VLAN segments together over an MPLS network.
The configured VLAN IDs for the customer switch interfaces are not required to match, as long as the
egress LSR overwrites the VLAN tag with the locally defined VLAN ID. The second service is a VMAN
service. This service interconnects two or more VMAN segments. The service operates like the VLAN
service but uses the VMAN tag instead of the VLAN tag to identify the service. As with the VLAN
service, the interconnected VMAN segments do not need to have matching VMAN IDs. The third
service is a port service. This service transparently interconnects two or more ports together over an
MPLS network. Traffic is transported unmodified between ports.
The VLAN and VMAN service are configured by adding the service VLAN or a VMAN to the VPLS.
The port service is not explicitly configured but is emulated using a combination of VPLS capabilities.
First a VMAN must be configured and the port added to the VMAN. The service VMAN is then added
to the VPLS. At this point all traffic received on the port is VMAN encapsulated for transmission across
the VPLS. To transmit the traffic across the VPLS as it was received on the port, the VPLS is configured
to exclude the service tag. By excluding the service tag, the VMAN tag is stripped prior to being
transmitted from the switch. This configuration provides port mode service and allows one or multiple
ports to be associated with a VPLS.
MPLS Pseudo Wires

MPLS Pseudo Wire (PW) tunnels are logical connections between two LERs over an LSP. Pseudo Wires
are signaled based on the configured PW identifier (pwid). The signaled PW label is used to create a
two-label-stack shim header on PW encapsulated packets. The outer label is the transport LSP label
obtained from LDP or RSVP-TE and the inner label is the signaled PW label. LERs also signal the PW
type when attempting to establish a PW. ExtremeXOS supports only the PWid type FEC. The
Generalized ID FEC type is currently not supported.
Transporting 802.1Q Tagged Frames

When an 802.1Q Ethernet frame is encapsulated for transport over a VC tunnel, the entire frame is
included, except for the preamble and FCS.
There is a configuration option that determines whether the 4-byte VLAN tag field is included in the
transmitted packet. By default, the tag field is not included. If the tag field is not included, the egress
LER may add one. If it is included, the tag service identifier may be overwritten by the egress LER. The
ability to add a tag field or to overwrite the service identifier at the egress node allows two (possibly
independently administered) VLAN segments with different VLAN IDs to be treated as a single VLAN.
The following command can be used to include the VLAN tag field:
configure vpls <vpls_name> {dot1q [ethertype <hex_number> | tag [include | exclude]]}

{mtu <number>}

This command can also be used to control the overwriting of the 802.1Q ethertype when the VLAN tag
field is included. In this case, the ingress node prior to transmitting the encapsulated packet overwrites
the ethertype. This allows devices with configurable VLAN tag Ethertypes to interoperate.
Establishing LDP LSPs to PW Endpoints

Establishing a PW requires both an LSP and an LDP session between the two endpoints. The local PW
endpoint is the MPLS LSR ID. The remote PW endpoint is identified using an IP address configuration
parameter.
When using LDP to establish the LSPs, each endpoint needs to advertise a label mapping for an LSP to
its local endpoint address. To ensure that its LDP peers will use the label mapping, a corresponding IGP
route should also be advertised for the address. For example, when using OSPF, an OSPF route with
prefix length 32 should be advertised for the configured IP address.
It is highly recommended that you configure a loopback VLAN using the IP address of the local
endpoint (the MPLS LSR ID). Use prefix length 32 for the IP address configured for the loopback
VLAN. When you configure a loopback VLAN, the IP address used to identify the endpoint will remain
active, even when one or more of the LSR VLAN interfaces go down. Should a remote peer normally
use one of the "down" interfaces, the normal IGP and LDP recovery procedures will allow the PW to
use one of the remaining "up" interfaces to minimize the network outage.
You should also configure the loopback VLAN for MPLS using the configure mpls add vlan
<vlan_name> command. The addition of the loopback VLAN to MPLS causes LDP to include the IP
address in LDP Address Messages. Some implementations (including EXOS) require this information to
determine the correct LDP session over which to advertise label mappings for VC FECs (see “Using
LDP to Signal PW Label Mappings” on page 826).
NOTE
Neither MPLS nor LDP have to be enabled on the loopback VLAN.
There are two options to initiate the LDP advertisement of an LSP to the local MPLS LSR ID when a
loopback VLAN has been configured for that IP address:
● Configure MPLS LDP to advertise a direct interface whose IP address matches the LSR ID and has
prefix length 32. Use the configure mpls ldp advertise direct lsr-id command to do this.
● Configure MPLS LDP to advertise direct interfaces using the configure mpls ldp advertise
direct all command.
NOTE
This will cause LDP to advertise label mappings for all VLANs that have an IP address configured and have IP
forwarding enabled.
While both of the above methods will initiate the advertisement of a label mapping for an LSP to the
local endpoint, the first method is the preferred method.

Using LDP to Signal PW Label Mappings

Just as LDP advertises label mappings for LSPs, it can also be used to advertise label mappings for L2
VPNs. In this case, the signaled FEC information describes a particular L2 VPN. This FEC is often called
a Virtual Circuit FEC or VC FEC. The VC FEC information includes a PW ID that is a 32-bit numeric
field. Unlike LSP label advertisements that are usually sent to all possible upstream peers, the VC FEC
information is sent only to the configured remote endpoint.
When the first L2 VPN is configured to a remote peer, MPLS automatically creates a targeted Hello
Adjacency entity for establishing an LDP session. Once the session is established, LDP passes the VC
FEC label mapping associated with the L2 VPN. Once VC FECs for the same PW ID have been
exchanged in each direction, MPLS is ready to associate the PW with an LSP to the remote endpoint as
described in “LSP Selection” on page 826.
To determine the correct LDP session over which to send a VC FEC, MPLS checks the IP addresses
learned from its LDP peers via LDP Address Messages. EXOS MPLS expects to find the IP address of a
remote PW peer among the addresses received over the LDP session to that peer.
To insure that the local endpoint IP address is included in LDP Address Messages, it is highly
recommended to configure MPLS on a loopback VLAN as described in “Establishing LDP LSPs to PW
Endpoints” on page 825. Use the command configure mpls add vlan <vlan_name> to configure
MPLS on the loopback VLAN. It is not required that LDP or MPLS be enabled on the VLAN for the
associated IP address to advertised in LDP Address Messages.
LSP Selection
A Pseudo Wire can be configured to use any available LSP to the peer endpoint IP address or the PW
can be configured to use one or more specific "named" LSPs. In either case, the LSP has to egress
(terminate) at the remote endpoint. In the case of an LDP LSP, the LSP's FEC has to be a /32 prefix
length to the endpoint IP address. In the case of a RSVP-TE LSP, the destination address has to be that
of the remote endpoint. When configured to use any available LSP, MPLS gives preference to RSVP-TE
LSPs if any exist to the remote endpoint. As a single LSP will be chosen to carry the PW traffic, if
multiple LSPs of the chosen type exist, the decision of which LSP of this type to use is non-
deterministic.
The command configure vpls <vpls_name> peer <ipaddress> mpls add lsp <lsp_name> forces
the PW to use the specified named LSP. If multiple named LSPs are configured, only one is used to
carry the PW. The decision of which of the multiple configured LSPs to use is non-deterministic.
RSVP-TE can be configured to allow specific types of traffic on an LSP. By default, LSPs are used to
transport all traffic. Optionally, named LSPs can be configured to allow only IP traffic or only VPN
traffic. This can be used to control the LSP selection for specific types of packets. For example, if both
LDP and RSVP-TE LSPs exist and the RSVP-TE LSPs are configured to transport only VPN traffic, all IP
traffic will be forwarded using LDP LSPs. Since RSVP-TE LSPs are preferred over LDP LSPs, VPN
traffic will flow over the RSVP-TE LSPs. The command configure mpls lsp <lsp_name> transport
ip-traffic deny configures this behavior for the specified RSVP-TE LSP.
Layer-2 VPN Domains

Layer-2 VPN domains are created by adding PWs to each peer LSR to build a fully-meshed
interconnected VPLS. For each peer added, a PW is signaled that is used to carry traffic from the local
LSR to the remote peer LSR. Flood traffic from the local service (broadcast, multicast, and unknown
unicast packets) is replicated and forwarded across all PWs in the VPLS. Each peer will receive one
copy of the packet for delivery to its locally attached service. As MAC learning occurs on PWs, unicast

packets to a known destination MAC address are forwarded to the peer over the PW from which the
MAC address was learned.
MAC Learning
Learned MAC addresses are associated with PWs from which the packets are received. The learned
MAC address is always inserted into the forwarding database (FDB) as though it was learned on the
local service VLAN (and not the VLAN identified in the dot1q tag in the received PW packet). MAC
addresses learned from PWs use a different FDB aging timer than those MAC addresses learned on
Ethernet interfaces. Different FDB aging timers are maintained for Ethernet and pseudo wire VPLS FDB
entries. By default both aging timers are set to 300 seconds. However, the aging timers for each type of
FDB entry can be configured to different values. Note that PW FDB entries are not refreshed; they age
out based on the configured aging timer setting, unless you have disabled the aging timer. Ethernet
FDB entries automatically refresh with use, and do not age out unless they are not used for the length
of time configured for the aging timer. Any MAC address associated with a PW is automatically cleared
from the FDB when the PW label is withdrawn.
Spanning Tree Protocols

There is some debate as to the benefit of supporting Spanning Tree Protocols (STP) within a Layer-2
VPN. The idea is that STPs could be used to provide redundant VPN data paths that could be
unblocked if the STP detects a spanning tree topology failure. In general, it is believed that introducing
STP to VPLS increases network complexity with very little real benefit. Because each PW is carried over
an LSP, MPLS already provides a sufficient level of redundancy. For example, if a PW is using an LDP
established LSP, provided there are parallel routed paths to the PW endpoint, the PW will automatically
shift from a withdrawn or failed LSP to the next best available LSP. For transport LSPs established
using RSVP-TE, secondary LSPs can be configured that can be hot-swapped in the event of a primary
LSP failure. Thus, even though the underlying transport LSP may have changed, the Layer-2 VPN data
plane remains unaffected.
Currently, ExtremeXOS layer 2 protocols cannot be configured on VPLS domains. Likewise, the
following protocols cannot be enabled on a VPLS service VLAN:
● STP
● VRRP
● ESRP
● EAPS control VLAN
As such, for example, the BPDU functional address is not inserted into the FDB for this VLAN and all
received BPDU packets are flooded across the VPLS. In this scenario, a single large spanning tree
topology spanning all interconnected VPLS service sites is constructed. Note that this is not a
recommended configuration for VPLS. Depending on the packet latency within the backbone network,
STP timers may need to be tuned to build and maintain reliable topologies.
IP Protocol Considerations
ExtremeXOS allows an IP address to be configured for a VPLS service VLAN. This is permitted to allow
the switch to use IP ping and traceroute functions to/from other nodes on the VLAN. It is envisioned
that any such IP address will only be configured temporarily to assist in network verification.

As such, ExtremeXOS does not allow IP forwarding to be enabled on a VPLS service VLAN for either
IPv4 or IPv6. Therefore, any higher-level protocol that requires IP forwarding to be enabled cannot be
enabled on a VPLS service VLAN. For example, OSPF cannot be enabled on a VPLS service VLAN. In
addition, IGMP snooping cannot be enabled on a VPLS service VLAN.
VPLS Layer 2 VPN Characteristics

Characteristics of VPLS include:
● Use of LDP or RSVP-TE to establish the underlying pseudo wire transport LSPs.
● Pseudo wire endpoints are identified via configured VPLS peer IP addresses.
● Configuration of VPLS pseudo wire ID doubles as VPN ID.
● Customer packet VLAN tags may be overwritten on egress from the pseudo wire allowing local
service VLANs or VMANs interconnected over the same VPLS to have different IDs.
● Customer packet VLAN tags may be included or excluded from packets transmitted on the pseudo
wire.
● Customer packet VLAN tag Ethertype value may be modified before packet is transmitted on the
pseudo wire allowing local service VLANs or VMANs interconnected over the same VPLS to have
different 802.1Q tag Ethertype values.
● Support for full-mesh VPN architectures.
● Support for up to 32 VPLS peers per L2 VPN.
● Support for VLAN, VMAN, and Port VPLS services.
● Separately configurable VPLS FDB aging timer.
● Support for enabling and disabling both the VPLS and/or the VPLS service.
Configuring MPLS Layer-2 VPNs

This section describes how to configure MPLS Layer-2 VPNs, and includes the following topics:
● Configuring MPLS for Establishing VPLSs on page 829
● Creating a VPLS Domain on page 829
● Deleting a VPLS Domain on page 829
● Enabling a VPLS Domain on page 829
● Disabling a VPLS Domain on page 830
● Adding a VPLS Peer on page 830
● Deleting a VPLS Peer on page 830
● Adding a VPLS Service on page 830
● Deleting a VPLS Service on page 831
● Enabling a VPLS Service on page 831
● Disabling a VPLS Service on page 831
● Configuring VPLS Options on page 831
● Configuring VPLS MTU on page 832
● Configuring VPLS FDB Aging Timer on page 832

● Unconfiguring VPLS Options on page 832

● Displaying VPLS VPN Status on page 832
● Displaying VPLS VPN Peers on page 833
Configuring MPLS for Establishing VPLSs

As described in “Using LDP to Signal PW Label Mappings” on page 826, MPLS and LDP must be
properly configured in order to establish the PWs that make up a VPLS. MPLS should be enabled using
the enable mpls command and LDP should be enabled using the enable mpls protocol ldp
command.
Since the MPLS LSR ID is used as the local endpoint for PWs, it is highly desirable to create a loopback
VLAN whose associated IP address is that of the MPLS LSR ID. The configured prefix length should be
32. As described in “Establishing LDP LSPs to PW Endpoints” on page 825, configuring this loopback
VLAN for MPLS causes the address to be included in LDP Address Messages. Use the configure mpls
add vlan <vlan_name> command for this. It is not required that LDP or MPLS be enabled on the
VLAN for the address to be advertised by LDP. Use the configure mpls ldp advertise direct
lsr-id command to initiate an LDP label mapping for an LDP LSP to the local endpoint.
Creating a VPLS Domain

To create a VPLS L2 VPN, use the following command:
create vpls <vpls_name> fec-id-type pseudo-wire <pwid>
This command creates a named VPLS instance. Multiple VPLS domains can be created each
representing a different L2 VPN. The pwid is a number that is used to signal and identify which VPLS
is associated with each pseudo wire. All of the pseudo wires carrying traffic for a specific L2 VPN will
be signaled with the same pwid. No L2 VPN traffic is forwarded over the L2 VPN until at least one
VPLS peer is added to the VPLS and a service is associated with the VPLS. The configured pwid also
doubles as the VPLS VPN ID.
Deleting a VPLS Domain

To delete a VPLS L2 VPN, use the following command:
delete vpls [<vpls_name> | all]
This command deletes the named VPLS instance or all VPLS instances, depending on the keyword. All
VPLS peers and services associated with the deleted VPLS instance(s) are also deleted.
Enabling a VPLS Domain

To enable a VPLS L2 VPN, use the following command:
enable vpls [<vpls_name> | all]
This command enables a named VPLS instance. By default, a newly created VPLS is enabled.

Disabling a VPLS Domain

To disable a VPLS L2 VPN, use the following command:
disable vpls [<vpls_name> | all]
This command disables the named VPLS instance. When a VPLS is disabled, no traffic will flow across
the L2 VPN. The pseudo wires connecting this peer to all other configured peers are also terminated, so
the remote peers will no longer see this LSR as an active peer.
Adding a VPLS Peer

To add a VPLS peer to the L2 VPN, use the following command:
configure vpls <vpls_name> add peer <ipaddress>
This command adds a VPLS peer to the named VPLS. For each new peer added to the VPLS, a pseudo
wire is signaled to carry L2 VPN traffic for this VPLS. Up to 32 peers can be added to a VPLS. For each
peer added, that remote peer must also configure this local LSR as a peer for the VPLS. This insures
that the VPLS core is configured as a full mesh of VPLS peers. The VPLS names on each peer do not
have to match since the pseudo wire ID is used to correlate which VPLS each pseudo wire is associated.
Deleting a VPLS Peer

To delete a VPLS peer from the L2 VPN, use the following command:
configure vpls <vpls_name> delete peer [<ipaddress> | all]
This command deletes a VPLS peer from the named VPLS. Once the peer is deleted, that specified peer
is no longer a member of the VPLS. The peer must also be removed from all other VPLS peers to insure
a proper full mesh and to prevent connectivity issues.
Adding a VPLS Service

To add a VPLS service to the L2 VPN, use the following command:
configure vpls <vpls_name> add service [{vlan} <vlan_name> | {vman} <vman_name>]
This command adds a VPLS service to the named VPLS. Only one service can be added to each VPLS.
Traffic associated with the service is transported over the VPLS L2 VPN. Three basic types of services
are supported: vlan, vman, and port. Both the vlan and vman service are specified by adding the VLAN
or VMAN name to the VPLS. The port service is configured by adding a VMAN name to the VPLS and
configuring the VPLS to strip the VMAN tag. This allows incoming service traffic to be transported
across the VPLS L2 VPN exactly as it was received. See “Configuring VPLS Options” on page 831 for
information about configuring a VPLS to strip the VMAN tag.

Deleting a VPLS Service

To delete a VPLS service from the L2 VPN, use the following command:
configure vpls <vpls_name> delete service [{vlan} <vlan_name> | {vman} <vman_name>]
This command deletes a VPLS service from the named VPLS. Since there is no local service that needs
to be connected to the VPLS, the pseudo wires to each of the configured peers for this VPLS are
terminated.
Enabling a VPLS Service

To enable a VPLS service, use the following command:
enable vpls [<vpls_name> | all] service
This command enables the VPLS service configured for the named VPLS. By default, the any configured
VPLS service is enabled.
Disabling a VPLS Service

To delete a VPLS service from the L2 VPN, use the following command:
disable vpls [<vpls_name> | all] service
This command disables the VPLS service configured for the named VPLS. When the VPLS service is
disabled, the service is disconnected from the VPLS and disabled such that no packets are sent to or
received from the service. The pseudo wires to each of the configured peers for this VPLS are
terminated.
Configuring VPLS Options

To configure VPLS packet forwarding options, use the following command:
configure vpls <vpls_name> {dot1q [ethertype <hex_number> | tag [include | exclude]]}

{mtu <number>}
This command configures the service packet encapsulation options for the named VPLS. The options
should be configured the same for every LSR for this VPLS in order to prevent connectivity issues.
Specifying the dot1q ethertype forces the switch to overwrite the dot1q ethertype value in the service
packet. This can be used to interconnect to customer segments over the VPLS that are using different
configured ethertype values. By default, the dot1q tag in the service packet is not included. The switch
can be configured to strip or exclude the dot1q tag. This may be used to emulate port services or for
interoperability with equipment that may require tags.

Configuring VPLS MTU

To configure VPLS MTU size, use the following command:
configure vpls <vpls_name> mtu <number>
This command configures the MTU packet size for the named VPLS. By default, the VPLS MTU size is
set to 1500. This value must match the value signaled in the VC FEC information of the PW peer. If the
two values do not match, the PW will not be established.
Note that the actual value enforced by the hardware is determined by the jumbo frame settings on the
port(s) in the local VPLS service.
Configuring VPLS FDB Aging Timer

To configure VPLS aging timer, use the following command:
configure fdb vpls agingtime <seconds>
This command configures the VPLS aging time for the switch. By default the aging time is set to 300
seconds, which means that the FDB entry for a MAC address learned on a pseudo wire is removed after
300 seconds. Setting the aging time to 0 prevents learned FDB entries from ever being aged out.
Unconfiguring VPLS Options

To configure VPLS packet forwarding options, use the following command:
unconfigure vpls <vpls_name> dot1q ethertype
This command resets the dot1q ethertype for the specified VPLS to the default ethertype configured for
the switch.
Displaying VPLS VPN Status

To display the status of a configured VPLS, use the following command:
show vpls {<vpls_name>} {detail}
The show command displays all configured VPLS domains and the status of configured peers. When
the detail option is specified, additional VPLS information is displayed. Detailed information includes
transmit and receive pseudo wire labels, transport LSP label, receive packet and byte counts, peer state,
and next hop interface.
In the example below, the service name is the VLAN or VMAN that is configured for the VPLS and the
VPN ID is the configured pseudo wire ID. The VPLS flags provide additional summary status
information. Since each VPLS can have multiple configured peers, each peer is listed below the VPLS.
The peer state, which can be DOWN, SGNL, or UP, and peer flags are also show for each peer. Future
releases of ExtremeXOS will support spoke peers allowing hierarchical VPLS network architectures.

VPLS VPN Configuration Examples
* BD-10K.1 # show vpls

VPLS Name VPN ID Flags Services Name Peer IP State Flags
--------------- ------ ----- --------------- --------------- ----- -----
nc_mnt_region 1 EAX- blowingrock
192.99.1.3 UP C---
192.99.1.4 UP C---
VPLS Flags: (E) Admin Enabled, (A) Oper Active, (I) Include Tag,
(X) Exclude Tag, (T) Ethertype Configured
Peer Flags: (C) Core Peer, (S) Spoke Peer, (N) Named LSP Configured
Total number of configured VPLS: 1

Total number of active VPLS: 1
Total number of configured PWs: 2

Total number of active PWs: 2
Displaying VPLS VPN Peers

To display the status of a configured VPLS peer, use the following command:
show vpls peer <ipaddress> {detail}
This command lists all VPLS domains that the specified ipaddress is configured as a VPLS peer. The
status of each pseudo wire for each VPLS is also shown.

This section provides the following VPLS configuration examples:
● Basic Point-to-Point VPLS Configuration Example on page 833
● Multipoint Full Mesh VPLS Configuration Example on page 834
Basic Point-to-Point VPLS Configuration Example

This MPLS VPLS network configuration shown in Figure 125, builds upon the routed MPLS network
configuration example shown in Figure 124.

Figure 125: MPLS VPLS configuration example
11
4 .
0 /2 0.
3.
1. LSR 3 vl
.0
.
n1 an 0/2
Router ID = 3 4
11 vla 192.168.0.3
10.10.10.0/24 OSPF backbone area 10.10.10.0/24

unc-charlotte and unc-
Ps MPLS domain wilmington
eu
do
LSR 1 Wi LSR 4
re
n4 24
Router ID = Router ID =
11 vla
vla .0/
.0 n2
192.168.0.1 192.168.0.4
4
.2
0.
.0
.
11
/2
4
LSR 2
Router ID = 192.168.0.2 MPLS_28
Assuming that the routed MPLS network has already been configured as in the previous example, the
commands used to create the VPLS on LSR1 and LSR4 follow. The nc-university-vpn is a point-to-point
VPLS, since only two peers are participating.
The following command creates the VPLS with VPN ID 35. This command must be issued on both
LSR1 and LSR4.
create vpls nc-university-vpn fec-id-type pseudo-wire 35
On LSR1, the local VLAN service and the VPLS peer are configured.
configure vpls nc-university-vpn add service vlan unc-charlotte

configure vpls nc-university-vpn add peer 192.168.0.4
On LSR4, the local VLAN service and the VPLS peer are also configured.
configure vpls nc-university-vpn add service vlan unc-wilmington

Multipoint Full Mesh VPLS Configuration Example

The example shown in Figure 126 configures a four node full-mesh MPLS VPLS configuration by
adding two additional peers to the previous example.

Figure 126: Full mesh Configuration Example
10.10.10.0/24
unc-greensboro
11
/24 .
0.
.0 LSR 3 vl 3.
.1 an 0/2
.0 n1 Router ID = 3 4
11 vla
192.168.0.3
10.10.10.0/24 OSPF backbone area 10.10.10.0/24

unc-charlotte and unc-
MPLS domain wilmington
LSR 1 LSR 4
n4 24
Router ID = Router ID =
11 vla
vla .0/
.0 n2
192.168.0.1 10.10.10.0/24 192.168.0.4
4
.2
0.
.0
unc-asheville
.
11
/2
4
LSR 2
Router ID = 192.168.0.2 MPLS_29
LSR1
LSR2
configure vpls nc-university-vpn add service vlan unc-asheville
LSR3
configure vpls nc-university-vpn add service vlan unc-greensboro
LSR4


32 Configuring RSVP-TE
This chapter describes the Resource Reservation Protocol (RSVP), traffic engineering (TE) extensions to
RSVP, and how you configure RSVP-TE using Extreme XOS.
This chapter covers the following topics:
● RSVP Elements on page 838

● Traffic Engineering on page 842
● Establishing RSVP-TE LSPs on page 844
● RSVP-TE Implementation on page 844
● Configuring RSVP-TE on page 851
RSVP is a protocol that defines procedures for signaling QoS requirements and reserving the necessary
resources for a router to provide a requested service to all nodes along a data path.
RSVP is not a routing protocol. It works in conjunction with unicast and multicast routing protocols. An
RSVP process consults a local routing database to obtain routing information. Routing protocols
determine where packets get forwarded; RSVP is concerned with the QoS of those packets that are
forwarded in accordance with the routing protocol.
Reservation requests for a flow follow the same path through the network as the data comprising the
flow. RSVP reservations are unidirectional in nature, and the source initiates the reservation procedure
by transmitting a path message containing a traffic specification (Tspec) object. The Tspec describes the
source traffic characteristics in terms of peak data rate, average data rate, burst size, and minimum/
maximum packet sizes.
RSVP-TE is a set of traffic engineering extensions to RSVP. RSVP-TE extensions enable RSVP to be used
for traffic engineering in MPLS environments. The primary extensions add support for assigning MPLS
labels and specifying explicit paths as a sequence of loose and strict routes. These extensions are
supported by including label request and explicit route objects in the path message. A destination
responds to a label request by including a label object in its reserve message. Labels are then
subsequently assigned at each node the reserve message traverses. Thus, RSVP-TE operates in
downstream-on-demand label advertisement mode with ordered LSP control.
ExtremeXOS implementation of RSVP-TE complies with RFC3209 and includes support for:
● Configuration on a per VLAN interface

● Operation as either edge or core MPLS router
● Support for specifying explicitly routed paths
● Support for both loose and strict route objects
● Recording the route of an established path
● Bandwidth reservation and policy per LSP

Configuring RSVP-TE
● Signaling QoS along the RSVP path using the Tspec and Adspec objects
● Fixed Filter (FF) and Shared Explicit (SE) reservation styles
● Specifying RSVP-TE session attributes
● Scaling enhancements using Refresh Overhead Reduction extensions
● Improved link failure detection using the RSVP-TE Hello Message
● Ability to reroute traffic over pre-configured backup LSPs
RSVP Elements
This section describes the following elements of the RSVP protocol:
● Message Types on page 838
● Reservation Styles on page 841
Message Types
This section describes the RSVP message types and includes the following topics:
● Path Message on page 839
● Reserve Message on page 839
● Path Tear Message on page 840
● Reserve Tear Message on page 840
● Path Error Message on page 840
● Reserve Error Message on page 840
● Reserve Confirm Message on page 840
RSVP messages are passed between RSVP capable routers to establish, remove, and confirm resource
reservations along specified paths. RSVP messages are sent as raw IP datagrams with protocol number
46. Each LSR along the path must process RSVP control messages so that it can maintain RSVP session
state information. Therefore, most RSVP messages are transmitted with the IP Router Alert Option.
Including the IP Router Alert provides a convenient mechanism allowing the IP routing hardware to
intercept IP packets destined to a different IP address and deliver them to the RSVP control plane for
processing. This is needed to setup and refresh RSVP-TE LSPs that follow an explicitly specified
network path and thus may not use the normal routed next hop IP address. RSVP has two basic
message types, path message and reserve message, as shown in Figure 127.

RSVP Elements
Figure 127: RSVP Messages
Previous Incoming Outgoing Next

hops interfaces interfaces hops
Data Data
A a c C
Path Path
Resv Resv
Router
B D
Data Data
b d
Path Path
Resv Resv
B' D'
MPLS_27
In addition to the path and reserve messages, RSVP has the following additional message types:
● Path tear message

● Reserve tear message
● Path error message
● Reserve error message
● Reserve confirm message
Path Message
The RSVP path message is used to store state information about each node in the path. Each RSVP
sender transmits path messages downstream along routed1 paths to setup and maintain RSVP sessions.
Path messages follow the exact same path as the data flow, creating path states in each LSR along the
path. The IP source address of the path message must be an address of the sender it describes and the
IP destination address must be the endpoint address for the session. The path message is transmitted
with the IP Router Alert option2 since each router along the path must process the path message. Each
LSR is responsible for refreshing its path status by periodically transmitting a path message to the
downstream LSR.
In addition to the previous hop address, the path message contains the sender Tspec and Adspec. The
reservation message carries the flowspec.
Reserve Message
Each receiver host transmits an RSVP reservation request to its upstream neighbor. Reserve messages
carry reservation requests hop-by-hop along the reverse path. The IP destination address of a reserve
message is the unicast address of the previous-hop LSR, obtained from the session's path state. The IP
source address is the address of the node that originated the message. The reserve message creates and
maintains a reserve state in each node on the path. Each LSR is responsible for refreshing its reserve
status by periodically transmitting a reserve message to the upstream LSR.
1. The routed path may be the best routed path or an explicitly specified routed path using EROs.
2. IP Router Alert option is described in RFC 2113.

Configuring RSVP-TE
Reserve messages are eventually delivered to the sender, so that the sender can configure appropriate
traffic control parameters for the first hop node.
Path Tear Message

Path tear messages delete path state information reserved along the path. The message is initiated by
the path sender or by any LSR in which a path state timeout occurs or an LSP is preempted (due to
bandwidth reservations), and is sent downstream to the session's path endpoint. Path tear messages are
transmitted with the IP Router Alert option and are routed exactly the same as path messages. The IP
destination address must be the path endpoint and the source IP address must be the sender address
obtained from the session's path state for the path that is being torn down.
When a path state is deleted as the result of the path tear message, the related reservation state must
also be adjusted to maintain consistency in the node. The adjustment depends on the reservation style.
Reserve Tear Message

Reserve tear messages delete reservation state information. The message is initiated by the path
endpoint or any node along the path in which a reservation state has timed out or an LSP is preempted
(due to bandwidth reservations), and is sent upstream to the session's path sender. Reserve tear
messages are routed exactly the same as reserve messages. The IP destination address of a reserve
message is the unicast address of the previous-hop node, obtained from the session's reservation state.
The IP source address is the address of the node that originated the message.
If no reservation state matches the reserve tear message, the message is discarded. The reserve tear
message can delete any subset of the filter specification in FF-style or SE-style reservation state.
Reservation styles are described in Table 99.
Path Error Message

Path error messages are used to report processing errors for path messages. These messages are sent
upstream to the sender that issued the path message. The message is routed hop-by-hop using the path
state information maintained in each node. Path error messages are informational and do not modify
the path state within any node.
Reserve Error Message

Reserve error messages are used to report processing errors for reserve messages. In addition, reserve
error messages are used to report the spontaneous disruption of a reservation. Reserve error messages
travel downstream to the endpoint of the session. The message is forwarded hop-by-hop using the
reservation state information maintained in each node. Reserve error messages are informational and do
not modify the reservation state within any node.
Reserve Confirm Message

Reserve confirm messages are optionally transmitted to acknowledge a reservation request. These
messages are transmitted from the sender to the endpoint. The destination IP address is the IP address
of the endpoint and the source IP address is the address of the sender. Since none of the intermediate
path nodes need to process a reserve confirm message, the message is transmitted without the IP Router
Alert option.

RSVP Elements
Reservation Styles
A reservation style is a set of options that is included in the reservation request.
One reservation style concerns how reservations requested by different senders within the same session
are handled. This type of reservation style is handled in one of two ways: either create a distinct
reservation for each sender in the session, or use a single reservation that is shared among all packets of
the selected senders.
Another reservation style concerns how senders are selected. Again, there are two choices: an explicit list
of all selected senders or a wildcard that implies all senders in the session.
Table 99 describes the relationship between reservation attributes and styles.

.
Table 99: Reservation Attributes and Styles

Distinct Reservation Shared Reservation
Sender Selection Style Style
Explicit Fixed filter (FF) Shared explicit (SE)
Wildcard Not defined Wildcard filter (WF)
The following sections describe the three reservation styles:

● Fixed Filter on page 841
● Shared Explicit on page 841
● Wildcard on page 841
Fixed Filter
The fixed filter (FF) reservation style uses a distinct reservation and an explicit sender selection. This
means that each resource reservation is for a specific sender. The session resources are not shared with
other sender's packets. Because each reservation is identified with a single sender, a unique label is
assigned by the endpoint to each sender (i.e., point-to-point LSP reservation).
Shared Explicit
The shared explicit (SE) reservation style uses a shared reservation and an explicit sender selection. This
means that a single resource reservation is created that is shared by multiple senders. The endpoint may
specify which senders are to be included for the reservation. Because different senders are explicitly
listed in the RESV message, different labels may be assigned to each sender. Thus, multiple shared-
resource LSPs to the same endpoint can be created (i.e., multipoint-to-point LSP reservation). The
Extreme MPLS implementation requests SE reservation style when signaling RSVP-TE LSPs.
Wildcard
The wildcard (WF) reservation style uses the shared reservation and wildcard sender options. A
wildcard reservation creates a single reservation that is shared by data flows from all upstream senders.
The Extreme MPLS implementation does not support WF reservation style.

Configuring RSVP-TE
Traffic Engineering
MPLS Traffic Engineering (TE) extends RSVP to support several unique capabilities. By coupling RSVP
and MPLS, LSPs can be signaled along explicit paths with specific resource reservations. Additional
RSVP objects have been defined to provide TE extensions. These objects include the Label Request,
Label, Explicit Route, Record Route, and Session Attribute. Extreme's RSVP-TE implementation
supports all of these TE objects.
The following sections discusses the implementation of RSVP-TE for ExtremeXOS:
● RSVP Tunneling on page 842

● RSVP Objects on page 842
RSVP Tunneling
An RSVP tunnel sends traffic from an ingress node through an LSP. The traffic that flows through the
LSP is opaque (or tunneled) to the intermediate nodes along the path. Traffic flowing through the
tunnel to an intermediate node along the path is identified by the previous hop and is forwarded, based
on the label value(s), to the downstream node.
RSVP-TE can:
● Establish tunnels with or without QoS requirements.

● Dynamically reroute an established tunnel.
● Observe the actual route traversed by a tunnel.
● Identify and diagnose tunnels.
● Use administrative policy control to preempt an established tunnel.
● Perform downstream-on-demand label allocation, distribution, and binding.
Extended Tunnel ID
Some LSRs require their neighboring LSRs to include their Router ID in the Extended Tunnel ID field
when sending RSVP-TE messages. The Extended Tunnel ID is a globally unique identifier present in the
RSVP common header Session object (see RFC3209). To provide maximum compatibility with other
vendors’ implementations, the ExtremeXOS MPLS implementation accepts RSVP-TE messages
regardless of the Extended Tunnel ID value and always inserts the local Router ID into the Extended
Tunnel ID field prior to transmission of an RSVP-TE message.
RSVP Objects
This section describes the RSVP objects that are used to establish RSVP-TE LSPs:
● Label on page 843
● Label Request on page 843
● Explicit Route on page 843
● Record Route on page 843
● Session Attribute on page 844

Traffic Engineering
Label
The label object is carried in the reserve message and is used to communicate a next hop label for the
requested tunnel endpoint IP address upstream towards the sender.
Label Request
To create an RSVP-TE LSP, the sender on the MPLS path creates an RSVP path message and inserts the
label request object into the path message.
A label request object specifies that a label binding for the tunneled path is requested. It also provides
information about the network layer protocol that is carried by the tunnel. The network layer protocol
sent through a tunnel is not assumed to be IP and cannot be deduced from the Layer 2 protocol header,
which simply identifies the higher layer protocol as MPLS. Therefore, the Layer 3 Protocol ID (PID)
value must be set in the Label Request Object, so that the egress node can properly handle the tunneled
data.
NOTE
The ExtremeXOS RSVP-TE implementation supports only Label Request objects with no Label Range. Label Ranges
are used to signal ATM VPI/VCI or Frame Relay DLCI information for the LSP. These types of Label Requests are not
supported. In the ExtremeXOS RSVP-TE implementation the L3 PID value, which identifies the Layer 3 protocol of
the encapsulated traffic, is always set to 0x0800 (IP).
Explicit Route
The explicit route object specifies the route of the traffic as a sequence of nodes. Nodes may be loosely
or strictly specified.
The explicit route object is used by the MPLS sender if the sender knows about a route that:
● Has a high likelihood of meeting the QoS requirements of the tunnel

● Uses the network resources efficiently
● Satisfies policy criteria
If any of the above criteria are met, the sender can decide to use the explicit route for some or all of its
sessions. To do this, the sender node adds an explicit route object to the path message.
After the session has been established, the sender node can dynamically reroute the session (if, for
example, if discovers a better route) by changing the explicit route object.
Record Route
The record route object is used by the sender to receive information about the actual route traversed by
the RSVP-TE LSP. It is also used by the sender to request notification if there are changes to the routing
path. Intermediate or transit nodes can optionally use the RRO to provide loop detection.
To use the object, the sender adds the record route object to the path message.

Configuring RSVP-TE
Session Attribute
The session attribute object can also be added to the path message. It is used for identifying and
diagnosing the session. The session attribute includes the following information:
● Setup and hold priorities

● Resource affinities
● Local protection
Establishing RSVP-TE LSPs

Establishing LSPs requires every LSR along the path to support RSVP and the Traffic Engineering (TE)
extensions defined in RFC3209. The LSP endpoints attempt to detect non-RSVP capable LSRs by
comparing the time-to-live (TTL) value maintained in the RSVP common header with that of the IP
TTL. If these values are different, it is assumed that a non-RSVP capable LSR exists along the path. By
including the Label Request object in the path message, RSVP capable routers that do not support the
TE extensions can be detected. RSVP routers that do not support TE extensions will reply with the
"Unknown object class" error.
RSVP-TE LSPs are referred to as named LSPs. These LSPs have configurable names that are used to
identify the LSP within the CLI. The command create mpls rsvp-te lsp <lsp_name> destination
<ipaddress> allocates the internal resources for the LSP. The newly created LSP is not signaled until
the LSP path has been configured. The LSP can be configured to take a specific path through the
network or the administrator can let the switch choose the best path by specifying the path any. Up to
three paths may be configured for an LSP to provide redundancy. The command configure mpls
rsvp-te lsp <lsp_name> add path configures an LSP path for the LSP. Optionally, RSVP-TE profiles
may be applied to an LSP to change its properties. An RSVP-TE profile is a specific CLI container used
to hold configuration parameters associated with timers, bandwidth reservation, limits, and other
miscellaneous properties.
Once the RSVP-TE LSP is configured, the LSP is immediately signaled. If signaled successfully, the LSP
becomes active. The commands disable mpls rsvp-te lsp <lsp_name> and enable mpls rsvp-te
lsp <lsp_name> are used to tear down and re-signal the LSP. Disabling the LSP causes the LER to
send a path tear message to the destination, forcing the LSP down and all resources along the path to be
freed. Enabling the LSP instructs the LER to send a path message to the destination re-establishing the
LSP. The configuration of the LSP is not modified by the enable or disable LSP commands.
RSVP-TE Implementation
This section covers the following features of RSVP:
● Explicit Route Path LSPs on page 845
● Route Recording on page 846
● LSP Session Attributes on page 846
● Bandwidth Reservation on page 846
● Bandwidth Management for RSVP-TE LSPs on page 847
● Redundant LSPs on page 848
● Improving LSP Scaling on page 849

Explicit Route Path LSPs

An explicit route is a specified path through a routed network topology. The path may be strictly or
loosely specified. If strictly specified, each node or group of nodes along the path must be configured.
Thus, no deviation from the specified path is allowed.
Loosely specified paths allow for local flexibility in fulfilling the requested path to the destination. This
feature allows for significant leeway by the LSR in choosing the next hop when incomplete information
about the details of the path is generated by the LER. Each node along the path may use other metrics
to pick the next hop along the path, such as bandwidth available, class of service, or link cost. The
command configure mpls rsvp-te path <path_name> add ero is used to add an Explicit Route
Object to a path container.
An explicit routed path is encoded using the explicit route object (ERO) and is transmitted in the path
message. The ERO consists of a list of subobjects, each of which describes an abstract node. By
definition, an abstract node can be an IPv4 prefix, IPv6 prefix, or an autonomous system (AS) number.
The ExtremeXOS RSVP-TE implementation supports only IPv4 abstract nodes. The ExtremeXOS RSVP-
TE implementation supports both strict and loose IPv4 abstract nodes. Received path messages with
EROs that contain any other subobject type will result in the transmittal of an "Unknown object class"
error message. All LSRs along the specified path must support the inclusion of the ERO in the path
message for an explicitly routed path to be successfully set up.
An LSR receiving a path message containing an ERO must determine the next hop for this path. The
steps for selection of the next hop are as follows:
1 The receiving LSR evaluates the first subobject. If the subobject type is not supported or there is no
subobject, a "Bad ERO" error is returned. The abstract node is evaluated to ensure that this LSR was
the valid next hop for the path message. If the subobject is a strict abstract node, the abstract node
definition must match the local interface address. If it does, then this LSR is considered to be a
member of the abstract node. Additionally, if the /32 address matches a local interface address, the
path message must have been received on the direct interface corresponding to the /32 address. If
the abstract node is an IP prefix, the subnet configured for the interface from which the path
message was received must match the abstract node definition. In the event that this LSR is not part
of the strict abstract node definition, a "Bad initial subobject" error is returned. If the subobject is a
loose abstract node, the LSR determines if the abstract node definition corresponds to this LSR. If it
doesn't, the path message is transmitted along the best-routed or constrained optimized path to the
endpoint and the ERO is not modified. If it is, then processing of the ERO continues.
2 If there is no second subobject, the ERO is removed from the path message. If this LSR is not the end
of the path, the next hop is determined by the constrained optimized path (via CSPF) to the path
message endpoint.
3 If there is a second subobject, a check is made to determine if this LSR is a member of the abstract
node. If it is, the first subobject is deleted and the second subobject becomes the first subobject. This
process is repeated until either there is only one subobject or this LSR is not a member of the
abstract node as defined by the second subobject. Processing of the ERO is then repeated with step 2.
By repeating steps 2 and 3, any redundant subobjects that are part of this LSRs abstract node can be
removed from the ERO. If this operation were not performed, the next hop LSR might reject the path
message.
4 The LSR uses its CSPF to determine the next hop to the second subobject. If the first object is a /32
address, the first subobject is removed, since it would not be part of the next hop's abstract node.
The path message is then sent along the explicit path to the path message endpoint. No
determination is made to verify that the abstract node defined in the subobject is topologically
adjacent to this LSR. The next hop should verify this as part of its processing as defined in step 1.

Configuring RSVP-TE
If CSPF determines that a specific path needs to be taken through the network, additional EROs are
inserted into the path message.
Route Recording
The route a path takes can be recorded. Recording the path allows the ingress LER to know, on a hop-
by-hop basis, which LSRs the path traverses. Knowing the actual path of an LSP can be especially
useful for diagnosing various network issues.
Network path recording is configurable per LSP. This feature is configured by enabling route recording
for a specific RSVP-TE profile using the command configure mpls rsvp-te lsp profile
<lsp_profile_name> record enabled and associating the profile to an LSP. ExtremeXOS sets the label
recording desired flag in the path message if route recording has been enabled for the LSP.
If route recording is enabled, the record route object (RRO) is inserted into the path message using a
single RRO subobject, representing the ingress LER. When a path message that contains an RRO is
received by an Extreme LSR, an RRO IPv4 subobject representing the /32 address of the outgoing
interface of the path message is pushed onto the top1 of the first RRO. The updated RRO is returned in
the reserve message.
The label recording flag is only supported by ExtremeXOS for egress and transit LSPs. If an Extreme LSR
receives a path message with the label recording flag set in the RRO, the LSR will encode the LSP's label
into a label subobject and push it onto the RRO.
If a path message is received that contains an RRO, the Extreme LSR uses the RRO to perform loop
detection. The RRO is scanned to verify that the path message has not already traversed this LSR. If the
RRO contains an IPv4 subobject that represents a local LSR interface, the path message is dropped and
a “Routing Problem” error message is sent to the originating LER with an error value of “Loop
detected.”
LSP Session Attributes

Session attributes are signaled for configured RSVP-TE LSPs using the session attribute object without
resource affinities (i.e., LSP_TUNNEL Type). ExtremeXOS will use the setup and hold priority values to
preempt established LSPs in order to satisfy bandwidth requests. Lower hold priority LSPs are
preempted in order to satisfy the bandwidth request in a path message with a higher setup priority.
LSP attributes are configured by setting the priorities for a specific RSVP-TE profile using the command
configure mpls rsvp-te lsp profile <lsp_profile_name> setup-priority <priority> hold-
priority <priority> and associating the profile to the configured LSP.
Bandwidth Reservation
As mentioned previously, RSVP reservations are unidirectional in nature. The source initiates the
reservation procedure by transmitting a path message containing a sender Tspec object. The Tspec
describes the source traffic characteristics in terms of peak data rate, average data rate, burst size, and
minimum/maximum packet sizes. The path message can also contain an optional AdSpec object that is
updated by network elements along the path to indicate information such as the availability of
1. RRO is organized as a LIFO stack.

particular QoS services, the maximum bandwidth available along the path, the minimum path latency,
and the path maximum transmission unit (MTU).
ExtremeXOS supports LSR bandwidth reservation requests per LSP. Only the Int-Serv Controlled-Load
service request is supported. Bandwidth is always reserved on the physical ports that the LSP traverses.
Depending on the platform, the bandwidth reservation may also be policed. The network administrator
can verify that the requested bandwidth was actually reserved. In those cases when the bandwidth
reserved is less than the requested bandwidth, the LSP can be manually torn down, re-signaled using a
different path, or accepted. The LSR automatically attempts to find a path that best satisfies the
bandwidth request. Constrained path selections are supported using OSPF-TE or, in a future
ExtremeXOS release, ISIS-TE. Best effort LSPs are provisioned by specifying a reserved bandwidth as
best-effort. The reserved LSP bandwidth is configured by setting the bps rate for a specific RSVP-TE
profile, using the configure mpls rsvp-te lsp profile <lsp_profile_name> bandwidth
command and associating the profile to an LSP.
Accounting of bandwidth reserved through an Extreme LSR RSVP-TE enabled VLAN is supported. The
maximum available bandwidth per physical port or trunk group is enforced. Thus, the available
bandwidth specified in the Adspec object is not modified as the path message is forwarded to the LSP
endpoint. As reserve messages are processed, the reserved bandwidth specified in the Flowspec is
added to the total reserved bandwidth allocated for the physical ports.
Because LSP bandwidth is dynamically allocated, a configuration command is provided to reserve port
bandwidth for use by MPLS. The command configure mpls rsvp-te bandwidth committed-rate
<committed_bps> [Kbps | Mbps | Gbps] [{vlan} <vlan_name> | vlan all] {receive |
transmit | both} pre-reserves bandwidth from the specified MPLS enabled VLAN for RSVP-TE
traffic only. This pre-allocation of bandwidth is useful since other applications may compete with MPLS
for available bandwidth. By pre-reserving a portion of the MPLS interface's bandwidth capacity, MPLS
is guaranteed to have that amount of the MPLS interface's bandwidth to meet RSVP-TE LSP reservation
requests.
Bandwidth Management for RSVP-TE LSPs

If an RSVP-TE LSP is signaled through a switch with bandwidth parameters, the LSP bandwidth
request is granted or rejected based on the availability of bandwidth resources on the physical ports
that the LSP traverses. Data traffic through these switches is not policed and there are no guarantees
that the packet using the LSP will not be dropped.
NOTE
Per LSP rate limiting is not supported in this release.
The available bandwidth for each OSPF interface is continually updated within the OSPF area. As
RSVP-TE LSPs are established and torn down, the reserved bandwidth associated with these LSPs is
used to update the total bandwidth available through each OSPF interface. RSVP-TE and CSPF can use
the bandwidth information to determine the appropriate path that each LSP should take through the
network based on the LSP's profile parameters. LSP parameters that can affect CSPF's TE path
calculation include, the LSP setup priority and bandwidth configuration.
Available bandwidth is calculated for eight CoS levels. Each CoS uniquely maps to an LSP hold
priority. Thus, when an LSP is setup through the switch, the reserved bandwidth consumed is
associated with a CoS based on the signaled LSP hold priority. The available bandwidth is recalculated
and is advertised to its OSPF neighbors. Advertised bandwidth is calculated using graduated

Configuring RSVP-TE
bandwidth reporting methodology. Using this scheme, higher CoS levels will advertise available
bandwidth that includes allocated bandwidth for lower CoS levels. The reasoning for doing this is that
higher priority LSPs can preempt lower priority LSP. Thus, even though the bandwidth has been
allocated to a lower priority LSP, it is still available for use by higher priority LSPs
Example:
An interface is configured to reserve 250 Mbps for MPLS traffic. The following LSPs are established
through this interface. Remember, hold priority value of 0 is the highest priority and 7 is the lowest.
● LSP A, hold priority = 7, reserved = 50 Mbps

● LSP B, hold priority = 5, reserved = 100 Mbps
● LSP C, hold priority = 2, reserved = 25 Mbps
● LSP D, hold priority = 1, reserved = 25 Mbps
OSPF will advertise the following available bandwidth for each CoS. CoS 0 is the highest and CoS 7 is
the lowest:
● CoS 0 (hold = 0): 250 Mbps (No LSPs; all bandwidth available)
● CoS 1 (hold = 1): 225 Mbps (LSP D)
● CoS 2 (hold = 2): 200 Mbps (LSP C & D)
● CoS 5 (hold = 5): 100 Mbps (LSP B, C & D)
● CoS 6 (hold = 6): 100 Mbps (LSP B, C & D)
● CoS 7 (hold = 7): 50 Mbps (LSP A, B, C & D)
CSPF calculations only use the available bandwidth for the desired CoS, as specified by the LSP hold
priority. Thus in this example, if LSP "E", with a configured setup priority of 6, requires 150Mbps, CSPF
will calculate a path to the destination that does not go through the above interface, since only 100Mbps
worth of bandwidth is available.
Redundant LSPs
This section describes RSVP-TE redundancy provisioning, and includes the following topics:
● Primary LSP with Secondary LSP Paths on page 848
● Multipath LSPs on page 849
There are two methods for provisioning redundant RSVP-TE LSPs at the ingress LER, also referred to as
head-end LSP protection. The first mechanism uses the concept of configured secondary (or backup)
LSPs and the other configuration strategy relies on multipath LSPs.
Primary LSP with Secondary LSP Paths

Secondary RSVP-TE LSPs may be configured to provide backup LSPs in the event that the primary LSP
fails. Secondary LSPs are fully provisioned pre-established RSVP-TE LSPs that are maintained as
inactive until needed. If the primary LSP is torn down, the associated LSP next hop is removed from the
route table, and a new LSP next hop representing one of the active secondary LSPs is installed as the
preferred LSP. If there are multiple secondary LSPs available, the secondary LSP is randomly selected.

If the primary LSP is re-established, the primary LSP next hop information is re-installed and the
secondary LSP returns to inactive state.
Operation with L2 VPNs is similar. If a primary path fails, and a secondary LSP is available, VPLSs will
use the secondary LSP. When the primary LSP is re-established, VPLSs will again use the primary LSP.
Specifying redundant LSPs is accomplished by assigning secondary paths to an lsp. The configure
mpls rsvp-te lsp <lsp_name> add path <path_name> secondary command configures the
specified path as a backup LSP. A path different from the primary path must be specified. It is
recommended that defined paths be configured using EROs to specify different paths through the
network. Relying on the routing topology, by configuring the path to any, can create two LSPs that take
the same path. It is important to understand that the configured LSP signals multiple LSPs, up to three
(one primary and two secondary), but only one LSP can be used to forward traffic at any one time.
Multipath LSPs
Multiple RSVP-TE LSPs may exist or be configured to the same destination. The paths do not need to be
equal cost; all that is required is that all the LSPs to the same destination must have IP transport
enabled. In this scenario, LSP next hop information is communicated to the route table for up to eight
different named RSVP-TE LSPs. Locally originated traffic is distributed across each LSP based on
standard IP address hash algorithms. If one of the LSPs fails, the traffic is redistributed across the
remaining active named LSPs. Unlike backup LSP mechanism, all of the redundant multipath LSPs are
unique named LSPs and in general will all have primary configured paths.
Improving LSP Scaling

RSVP maintains path and reserve state by periodically sending refresh messages. Refresh messages
allow each LSR along the path to properly maintain reservation state information and to recover from
network failures. Because refresh messages are periodically sent for each path reservation, scaling the
number of RSVP-TE LSPs is an issue. Additionally, network requirements for faster failure detection
and improved LSP recovery times further exacerbate the scaling issue.
Several techniques are described in RFC 2961 RSVP Refresh Overhead Reduction to improve the scalability
of RSVP. These techniques include the bundle message, message ID extension, and summary refresh
extension. Support for these extensions is signaled between RSVP peers via the refresh-reduction-capable
bit in the flags field of the common RSVP header. Additionally, the hello extension, described in RFC3209,
provides a fourth scaling mechanism for RSVP. The hello extension is designed so that either peer can
use the mechanism regardless of how the other peer is configured. Therefore, support for the hello
extension is not signaled between RSVP peers. ExtremeXOS supports and is compliant with the RSVP-TE
scaling features described in RFC2961.
These features are summarized in the following sections:
● Bundle Message on page 850

● Summary Refresh Extension on page 850
● Message ID Extension on page 850
● Hello Extension on page 850
● Refresh Time on page 851
● Summary Refresh Time on page 851
● Bundle Time on page 851

Configuring RSVP-TE
Bundle Message
RSVP bundle messages aggregate multiple RSVP messages within a single PDU. The messages are
addressed directly to peer LSRs. Therefore, bundle messages are not sent with the IP Router Alert
option. Bundling multiple RSVP messages into a single PDU reduces the per packet overhead
associated with local delivery and local origination. Each bundle message must contain at least one
RSVP message. Transmission of RSVP messages may be delayed up to the number of seconds
configured for bundle time. The size of the bundle message is limited to the RSVP-TE interface MTU size.
Bundle messaging is enabled using the enable mpls rsvp-te bundle-message command.
Summary Refresh Extension

A summary refresh message is used to refresh RSVP states along an LSP without having to explicitly
send path and reserve refresh messages. This can substantially reduce the RSVP control bandwidth
overhead. Summary refresh messages contain a list of message_ID objects. Each message_ID object
identifies a path and reserve state to be refreshed. When summary refresh support is enabled, path and
reserve refresh messages are suppressed. If the message identifier value indicates that the RSVP state
has changed, the receiving LSR notifies the sender by transmitting a message_ID_NACK message. The
summary refresh rate is enabled using the enable mpls rsvp-te summary-refresh command.
Message ID Extension
The message ID extension provides reliable delivery of RSVP messages. It also provides a simple
mechanism for identifying refresh messages, which can greatly reduce refresh message processing on
the receiving LSR. The message ID extension defines three new objects: message_ID, message_ID_ACK,
and message_ID_NACK. The message_ID object contains a unique message identifier based on the
sender's IP address. Only one message_ID object is inserted into an RSVP message. The receiving LSR
can use the message_ID object to quickly refresh path and reserve states. If the message identifier value
in the message_ID object is greater than the locally saved message identifier value, then the RSVP
message represents a new or modified state. The receiving LSR must acknowledge an RSVP message
using the message_ID_ACK object if the sender set the ACK_desired flag in the message_ID object,
otherwise the message_ID acknowledgement is optional. The message_ID_ACK object may be included
in any unrelated RSVP message or in an RSVP ACK message. Message ID extension is required for both
bundle message and summary refresh, so this capability is automatically enabled if either of the other
capabilities is enabled.
Hello Extension
The RSVP hello message provides a quick and simple mechanism for detecting the loss of a peer RSVP-
TE LSR. The hello protocol is implemented using the RSVP soft-state model. RSVP hello messages may
be enabled independently of each LSR's peer. The hello protocol consists of two new objects:
hello_request and hello_ACK. If configured, an LSR will send a hello_request every hello interval. If a
hello_ACK is not received within a specified amount of time, the sending LSR assumes that its peer LSR
is no longer active. Once a peer LSR is deemed inactive, all reservation states associated with LSPs
established to or through the peer LSR must be freed and the LSPs torn down. The hello interval is
configurable using the command configure mpls rsvp-te timers session hello-time.

Configuring RSVP-TE
You can improve LSP scaling by configuring the following RSVP-TE parameters:
Refresh Time
The refresh time specifies the interval for sending refresh path messages. RSVP refresh messages
provide soft state link-level keep-alive information for previously established paths and enable the
switch to detect when an LSP is no longer active. RSVP sessions are torn down if an RSVP refresh
message is not received from a neighbor within [(keep-multiplier + 0.5) * 1.5 * refresh-time] seconds. The
valid refresh time may be set to any value between 1 and 36000 seconds. The default setting is 30
seconds. Configuring a longer refresh time reduces both switch and network overhead.
Summary Refresh Time

The summary refresh time, specified in tenths of a second, indicates the time interval for sending
summary refresh RSVP messages. The summary refresh time must be less than the configured refresh
time. The default summary refresh time is zero, indicating that no summary refresh RSVP messages are
sent. The summary refresh time value may be set to any value between zero to 100 (or 10 seconds). If
configured, the bundled and summary refresh RSVP messages are only sent to RSVP-TE peers
supporting RSVP refresh reduction.
Bundle Time
The bundle time, specified in tenths of a second, indicates the maximum amount of time a transmit
buffer is held so that multiple RSVP messages can be bundled into a single PDU. The default bundle
time is zero, indicating that RSVP message bundling is not enabled. The bundle time value can be set to
any value between zero and 30 (or 3 seconds).
Configuring RSVP-TE
This section describes the following tasks:
● Configuring RSVP-TE on the Switch on page 852
● Configuring RSVP-TE on a VLAN on page 852
● Configuring RSVP-TE Protocol Parameters on page 852
● Create an RSVP-TE Path on page 853
● Configuring an Explicit Route on page 853
● Reserve Bandwidth for MPLS on page 854
● Create an RSVP-TE Profile on page 855
● Configuring an RSVP-TE Profile on page 855
● Configuring an RSVP-TE LSP on page 856
● Adding a Path to an RSVP-TE LSP on page 856
● Displaying RSVP-TE LSP Configuration Information on page 857
● Displaying the RSVP-TE Paths on page 857
● Displaying the RSVP-TE Path Profile on page 857
● Displaying the RSVP-TE LSP on page 858

Configuring RSVP-TE
Configuring RSVP-TE on the Switch

To enable RSVP-TE on a switch, use the following command:
enable mpls protocol rsvp-te
To disable RSVP-TE on a switch, use the following command:
disable mpls protocol rsvp-te
Note that MPLS must be globally enabled using the enable mpls command before RSVP-TE can
become operational.
Configuring RSVP-TE on a VLAN

To configure RSVP-TE on one or all VLANs, use the following command:
enable mpls rsvp-te [{vlan} <vlan_name> | vlan all]
To disable RSVP-TE on a VLAN, use the following command:
disable mpls rsvp-te te [{vlan} <vlan_name> | vlan all]
This command disables RSVP-TE on one or all VLANs. Deleting RSVP-TE on a VLAN causes all TE
LSPs using that interface to be released, and prevents TE LSPs from being established or accepted on
the specified VLAN.
Configuring RSVP-TE Protocol Parameters

To configure RSVP-TE protocol parameters, use the following command:
configure mpls rsvp-te timers session[{bundle-message-time
<bundle_message_milliseconds>} {hello-keep-multiplier <hello_keep_number>} {hello-time
<hello_interval_seconds>}{refresh-keep-multiplier <refresh_keep_number>} {refresh-time
<refresh_seconds>}{summary-refresh-time <summary_refresh_milliseconds>}] [{vlan}
<vlan_name> | vlan all]
This command configures the RSVP-TE protocol parameters for the specified VLAN. The RSVP-TE
keyword all indicates that the configuration changes apply to all RSVP-TE enabled VLANs.
The hello-interval time specifies the RSVP hello packet transmission interval. The RSVP hello packet
is used by the switch to detect when a RSVP-TE peer is no longer reachable. If an RSVP hello packet is
not received from a peer within [hello-interval * keep-multiplier] seconds, the peer is declared down
and all RSVP sessions to and from that peer are torn down. The default hello-interval time is zero,
indicating that RSVP hellos are not enabled. The hello-interval may be set to any value between zero
and 60 seconds.
The refresh-time parameter specifies the interval for sending refresh path messages. RSVP refresh
messages provide “soft state” link-level keep-alive information for previously established paths and
enables the switch to detect when an LSP is no longer active. RSVP sessions are torn down if an RSVP
refresh message is not received from a neighbor within [(keep-multiplier + 0.5) * 1.5 * refresh-time]
seconds. The default refresh-time is 30 seconds and the default keep-multiplier value is three. The
minimum and maximum refresh-time values are one and 36,000 seconds (or ten hours) respectively.
The minimum and maximum keep-multiplier values are one and 255 respectively.

Configuring RSVP-TE
The bundle-time, specified in tenths of a second, indicates the maximum amount of time a transmit
buffer is held so that multiple RSVP messages can be bundled into a single PDU. The default bundle-
time is zero, indicating that RSVP message bundling is not enabled. The bundle-time value may be set
to any value between zero and 30 (or 3 seconds).
The summary-refresh-time, specified in tenths of a second, indicates the time interval for sending
summary refresh RSVP messages. The summary-refresh-time must be less than the configured
refresh-time. The default summary-refresh-time is zero, indicating that no summary refresh RSVP
messages are sent. The summary-refresh-time value may be set to any value between zero to 100 (or
10 seconds).
If configured, the bundled and summary refresh RSVP messages are only sent to RSVP-TE peers
supporting RSVP refresh reduction.
Create an RSVP-TE Path

To create an RSVP-TE routed path resource, use the following command:
create mpls rsvp-te path <path_name>
The <path_name> parameter is a character string that is to used to identify the path within the switch.
The <path_name> string must begin with an alphabetic character, and may contain up to 31 additional
alphanumeric characters. The RSVP-TE LSP is not signaled along the path until an LSP adds the
specified path name. The maximum number of configurable paths is 255.
To delete an RSVP-TE path, use the following command:
delete mpls rsvp-te path [<path_name> | all]
This command deletes a configured MPLS RSVP-TE routed path with the specified <path_name>. All
associated configuration information for <path_name> is deleted. A path cannot be deleted as long as
the <path_name> is associated with an LSP. If the all keyword is specified, all paths not associated
with an LSP are deleted.
Configuring an Explicit Route

To add an RSVP-TE explicit route, use the following command:
configure mpls rsvp-te path <path_name> add ero <ipNetmask> [strict | loose] {order
<number>}
This command adds an IP address to the explicit route object (ERO) for the specified path name. The
RSVP-TE routed path may be described by a configured sequence of the LSRs and/or subnets traversed
by the path. Each defined LSR or subnet represents an ERO subobject. Up to 64 subobjects can be added
to each path name.
The ipaddress keyword identifies an LSR using either a /32 address, which may represent an LSR
router ID, loopback address, or direct router interface, or an IP prefix, which represents a directly
connected subnet. Each IP address or prefix is included in the ERO as an IPv4 subobject.

Configuring RSVP-TE
If the IP address is specified as strict, the strict subobject must be topologically1 adjacent to the
previous subobject as listed in the ERO. If the IP address is specified as loose, the loose subobject is not
required to be topologically adjacent to the previous subobject as listed in the ERO. If omitted, the
default subobject attribute is loose. Each IP address or prefix is included in the ERO as an IPv4
subobject.
If the subobject matches a direct router interface or a directly attached subnet, the switch verifies that
the path message is received on the matching router interface.
The LSR path order is optionally specified using the order keyword. The order number parameter is
an integer value from 1 to 65535. IP prefixes with a lower order number are sequenced before IP
prefixes with a higher number. You can specify multiple addresses and assign them an order number.
The order number determines the path that the LSP follows. Thus, the LSP path follows the configured
path of the IP prefix with the order value from low to high. If the order keyword is not specified, the
number value for the LSR defaults to a value 100 higher than the current highest number value.
If the list of IP prefixes, added to the path, does not reflect an actual path through the network
topology, the path message is returned with an error from a downstream LSR and the LSP is not
established.
The order of a configured subobject can not be changed. The ERO subobject must be deleted and re-
added using a different order. If a subobject is added to or deleted from the ERO while the associated
LSP is established, the path is torn down and is resignaled using the new ERO.
Duplicate ERO subobjects are not allowed. Defining an ERO for the path is optional. If you do not
configure an ERO, the path is signaled along the best CSPF calculated path and the ERO is not included
in the path message. When the last subobject in the ERO of the path message is reached and the egress
IP node of the path has not been reached, the remaining path to the egress node is signaled along the
best CSPF calculated path. Specification of an ERO could lead to undesirable routed paths, so care
should be taken when terminating the ERO routed-path definition prior to the configured path egress
node.
To delete an RSVP-TE explicit route, use the following command:

configure mpls rsvp-te path <path_name> delete ero [all | <ipNetmask> | order
<number>]
This command deletes an LSR or subnet from the ERO for the specified path name. The LSR is specified
using the ipaddress, or order parameter. If an LSR is deleted from an ERO while the associated LSP is
established, the path is torn down and is resignaled using a new ERO. Use the all keyword to delete
the entire ERO from the path name. When there is no configured ERO, the path is no longer required to
take an explicit routed path. The path is then signaled along the best CSPF calculated path and no ERO
is included in the path message.
Reserve Bandwidth for MPLS

In order to manage the bandwidth used by MPLS RSVP-TE LSPs, a configured amount of bandwidth
must be reserved for MPLS. This bandwidth must be reserved on all MPLS interfaces over which RSVP-
TE LSPs requesting bandwidth are to be established. The following command is used to reserve
bandwidth for MPLS:
configure mpls rsvp-te bandwidth committed-rate <committed_rate_bps>
<committed_rate_unit> [{vlan} <vlan_name> | vlan all]
1. The LSP next hop matches the interface IP address of the immediate neighbor LSR.

Configuring RSVP-TE
The commited_rate_unit must be Kbps, Mbps, or Gbps. Choosing the all option will reserve the
specified amount of bandwidth on all MPLS interfaces.
The default-reserved value is zero. Therefore, no LSPs requesting bandwidth can be established until
bandwidth has been configured.
Create an RSVP-TE Profile

To create a traffic engineered LSP profile, use the following command:
create mpls rsvp-te profile <profile_name>
This command creates a configured RSVP-TE profile with the specified profile name. The default profile
cannot be deleted. If a profile is associated with a configured LSP, the profile cannot be deleted.
To delete a traffic engineered LSP profile, use the following command:
delete mpls rsvp-te profile [<profile_name> | all]
Configuring an RSVP-TE Profile

To configure an RSVP-TE profile, use the following command:
configure mpls rsvp-te profile <profile_name> {bandwidth [best-effort | [{committed-
rate <committed_bps> [Kbps | Mbps | Gbps]} {max-burst-size <burst_size> [Kb | Mb]}
{peak-rate <peak_bps> [Kbps | Mbps | Gbps]}]} {hold-priority <hold_priority>} {mtu
[<number> | use-local-interface]} {record [enabled | disabled]} {setup-priority
<setup_priority>}
A profile is a set of attributes that are applied to the LSP when the LSP is configured using the create
mpls rsvp-te lsp <lsp_name> destination <ipaddress> command. A default profile is provided
which cannot be deleted, but can be applied to any configured LSP. The profile name for the default
profile is default. The default profile parameter values are initially set to their respective default values.
The maximum number of configurable profiles is 255 (one of which is reserved for the default profile).
The bandwidth parameter specifies the desired reserved bandwidth for the LSP. Any positive integer
value is valid. You must append the characters, k for kilobits, m for megabits, or g for gigabits, to the
bps value to specify the unit of measure. The default bandwidth bps value is zero, which indicates that
the QoS for the LSP is best effort.
The max-burst-size and peak-rate parameters are signaled in the sender Tspec object and add
further definition of the expected traffic. The mtu parameter is also signaled in the sender Tspec object
and defines the expected maximum packet size that will sent over the LSP.
The setup-priority and hold-priority are optional parameters indicating the LSP priority. During
path set up, if the requested bandwidth cannot be reserved through the LSR, the setup-priority
parameter is compared to the hold-priority of existing LSPs to determine if any of the existing LSPs
need to be preempted to allow a higher priority LSP to be established. Lower numerical values
represent higher priorities. The setup-priority range is 0 to 7 and the default value is 7. The hold-
priority range is also 0 to 7 and the default value is 0.
The record keyword is used to enable hop-by-hop path recording. The enabled keyword causes the
record route object (RRO) to be inserted into the path message. The RRO is returned in the reserve

Configuring RSVP-TE
message and contains a list of IPv4 subobjects that describe the RSVP-TE path. Path recording by
default is disabled. When disabled, no RRO is inserted into the path message.
To delete an RSVP-TE path profile, use the following command:
delete mpls rsvp-te profile [<profile_name> | all]
This command deletes a configured RSVP-TE profile with the specified profile name. The default profile
cannot be deleted. If a profile is associated with a configured LSP, the profile cannot be deleted. If you
specify the all keyword, all profiles not associated with an LSP are deleted (except for the default
profile).
Configuring an RSVP-TE LSP

To add an RSVP-TE LSP, use the following command:
create mpls rsvp-te lsp <lsp_name> destination <ipaddress>
The <lsp_name> parameter is a character string that is to be used to identify the LSP within the switch.
The <lsp_name> string must begin with an alphabetic character and can contain up to 31 additional
alphanumeric characters. The LSP is not signaled until at least one LSP path is added. See “Adding a
Path to an RSVP-TE LSP” on page 856
To delete an RSVP-TE LSP, use the following command:
delete mpls rsvp-te lsp [<lsp_name> | all]
Deleting an LSP name disassociates all configured paths with this LSP and all configuration information
for the LSP name is deleted. If the LSP has been specified for use by a static route or a VPLS, that
configuration information is also deleted. If you specify the all keyword, all LSPs are deleted.
Adding a Path to an RSVP-TE LSP

To add a path to an RSVP-TE LSP, use the following command:
configure mpls rsvp-te lsp <lsp_name> add path [<path_name> | any] {profile
<profile_name>} {primary | secondary}
This command adds a configured path to the specified RSVP-TE LSP. The LSP name parameter is a
character string that is to be used to identify the LSP within the switch and must have been created
previously. The LSP is not signaled until a path is added to the LSP. Up to three paths can be defined
for the LSP: one primary and two secondary. All paths are signaled, but only one path is used to
forward traffic at any one time. The switch chooses the local MPLS VLAN interface from which to
signal the LSP. To force an LSP to use a specific local MPLS interface, configure the peer’s interface IP
address as the first ERO in the associated path. The profile name is optional. If omitted, the default
profile is applied to the LSP. The path name may be specified or the LSP can be configured to take any
path. For a given LSP, only one path can be configured to take any path through the MPLS network.
The specified path defaults to primary when no primary path has been configured for the LSP, and
defaults to secondary if the primary path has been previously configured for the LSP.
Each <path_name> added to an <lsp_name> must be unique, but a <path_name> can be associated with
multiple LSP names.

Configuring RSVP-TE
All configured primary and secondary paths for the <lsp_name> must have the same endpoint IP
address. For example, three paths can be configured for the <lsp_name>, but all paths should represent
different topological paths through the network to the same LSP endpoint.
Adding a secondary <path_name> designates a path as a hot-standby redundant path, used in the event
that the primary or the other secondary path cannot be established or fails. Provided the <path_name>
has not already been established, all paths are signaled as soon as they are associated with an
<lsp_name>. If the primary <path_name> fails, is not configured, or cannot be established after the
specified LSP retry-timeout, one of the configured secondary paths may become the active path for
<lsp_name>. All of the secondary paths have equal preference; the first one available is chosen. If at any
time the primary path is established, <lsp_name> immediately switches to using the primary path. If a
secondary path fails while in use, the remaining configured secondary path can become the active path
for <lsp_name>.
To delete a path from an RSVP-TE LSP, use the following command:
configure mpls rsvp-te lsp <lsp_name> delete path [<path_name> | any | all]
When you issue this command, the LSP associated with the path is immediately torn down. If the
deleted path represents the in-use LSP for <lsp_name> and another secondary path is configured, the
LSP immediately fails over to an alternate LSP. Because at least one path must be defined for each LSP,
the last configured path cannot be deleted from the LSP.
Displaying RSVP-TE LSP Configuration Information

To display RSVP-TE LSP configuration information, use the following command:
show mpls rsvp-te
This command displays summary configuration and status information for RSVP-TE. Global status of
RSVP-TE and the configured standard and rapid-retry LSP timer values are included in the display
output.
Displaying the RSVP-TE Paths

To display RSVP-TE paths, use the following command:
show mpls rsvp-te path {<path_name>} {detail}
This command displays the configuration and status information for MPLS RSVP-TE paths. Information
is listed in tabular format and includes the path name, number of configured ERO objects, number of
LSPs configured to use this path, the list of EROs and their type. Optionally specifying the detail
keyword displays the path information in verbose format, including all LSPs that are configured to use
the path.
Displaying the RSVP-TE Path Profile

To display RSVP-TE path profiles, use the following command:
show mpls rsvp-te profile {<profile_name>} {detail}
By default, this command displays all configured profile parameters for the specified profile. If the
profile name is omitted, the profile parameter values for all configured LSP profiles are displayed.

Configuring RSVP-TE
Optionally specifying the keyword detail displays the profile information is verbose format. When the
detail keyword is specified, all LSPs that are configured to use this profile are also displayed.
Displaying the RSVP-TE LSP

To displays the RSVP-TE LSP, use the following command:
show mpls rsvp-te lsp {{[destination | origin] <ipaddress>} {detail}
This command displays the LSP information associated with RSVP-TE that is used to forward packets
within the MPLS network. If no options are specified, summary information for all RSVP-TE LSPs is
displayed. This information includes the LSP name, LSP direction relative to the MPLS network (i.e.,
ingress, egress, or transit), incoming and or outgoing interface, configured destination, and uptime. If
the optional LSP name parameter is specified, only the LSP information for the specified ingress LSP
name is displayed. Optionally, the LSPs displayed can be further qualified by the keywords ingress,
egress, and transit. These keywords qualify the LSPs displayed from the perspective of the switch.
Ingress LSPs identify LSPs that originate from the switch into the MPLS network. Egress LSPs identify
LSPs that terminate at the switch from the MPLS network. Transit LSPs represent LSPs that traverse the
switch.
The optional destination keyword limits the display to only those LSPs that terminate at the specified
IP address. The optional origin keyword limits the display to only those LSPs that originate at the
specified IP address. Optionally specifying the detail keyword displays the LSP information is in
verbose format. Additional information displayed includes the configured path and profile, transmit
and/or receive labels, failure and retry counts, packet and byte counters, and recorded routes.
RSVP-TE LSPs comprise profiles, paths, and the actual LSP. This section describes how to configure an
RSVP-TE LSP.
Configuring RSVP LSPs is a multi-step process with some optional steps, depending on the specific
requirements of the LSP. Conceptually, a number of mandatory elements must be configured to create
an RSVP-TE LSP. In addition, you can also configure optional elements. In certain configurations, there
are also order dependencies.
The profile contains constraints that you wish to apply to the LSP. These constraints may affect the path
selected across the MPLS domain in order to meet those constraints. Examples of profile parameters
include bandwidth, setup, and hold priority relative to other configured LSPs.
The path can be used to specify the explicit path across the MPLS domain that the LSP should follow.
This is done using EROs. An ERO is an object, sent as part of the LSP setup request (path message) that
explicitly specifies the part of the path across the MPLS domain the setup request should follow. You
can configure both loose and strict EROs in a path.
Certain elements of configuration are order dependent. For example if you specify a profile or path
when creating an LSP, those path or profile definitions must already exist. Similarly a path must exist
before an ERO is created, as the ERO is added explicitly to the path.

The typical steps used to configure and verify an RSVP-TE LSP are as follows:
1 Create and configure a path (optional).

2 Reserve bandwidth for the LSP (optional).
3 Create and configure a profile (optional).
4 Create an LSP (mandatory).
5 Add a primary/secondary path to the LSP (mandatory).
6 Add a secondary path to the LSP (optional).
7 Verify LSP status (recommended).
Figure 128: RSVP-TE Configuration Example
London
Router ID 1.0.0.0
Pr
SP
im
yL
17
0
ary
0/3
2.2
ar
LS
3.2
im
5.2
P
Pr
5.2
3.3
2.2
6/3
17
172.25.23.8/30
17
Oxford Oxford
0
/3
Glasgow Birmingham
2.
28
25
University University
3.
Router ID Router ID
.2
.2
3.
25
32
2.0.0.0 4.0.0.0
2.
/3
0
17
Se
P
LS
co
nd
y
ar
ar
nd
y
LS
co
P
Se
Liverpool
Router ID 5.0.0.0 MPLS_24
The configuration example, shown in Figure 128, creates primary and secondary LSPs between the node
Glasgow and the node Birmingham. The steps specifically create an LSP between Glasgow and
Birmingham based on an explicitly routed path via London with bandwidth, and setup and hold
priority profile requirements. A secondary path is also created which, in the event of failure of a link or
node on the primary path, activates the secondary path for the LSP. This path is Glasgow, Birmingham
via Liverpool.

Configuring RSVP-TE
NOTE
Before configuring RSVP-TE LSPs, you need to enable the protocol on the switch, and an initial step of adding
RSVP-TE to a VLAN must be carried out for all VLANs over which the user wishes RSVP-TE LSPs to be signaled.
This is a one-time operation.
NOTE
A loopback VLAN with the LSR-ID should be added to MPLS to allow RSVP-TE LSPs to be established to the LSR-
ID.
The following commands configure RSVP-TE for the switch and add RSVP signaling capabilities to the
specified VLANs:
enable mpls
enable mpls protocol rsvp-te
configure mpls add vlan loopback
configure mpls add vlan gla-lon
enable mpls rsvp-te vlan gla-lon
enable mpls vlan gla-lon
configure mpls add vlan gla-liv
enable mpls rsvp-te vlan gla-liv
enable mpls vlan gla-liv
The following commands reserve bandwidth for RSVP-TE LSPs on these MPLS interfaces:
configure mpls rsvp-te bandwidth committed-rate 20 Mbps gla-lon

configure mpls rsvp-te bandwidth committed-rate 20 Mbps gla-liv
The following commands create and configure an LSP profile named Glasgow-Birmingham-pro. LSPs
that use the Glasgow-Birmingham-pro profile are signaled with a reserved bandwidth of 10 Mbps and
an LSP setup and hold priority of 5.
create mpls rsvp-te profile Glasgow-Birmingham-pro

configure mpls rsvp-te profile Glasgow-Birmingham-pro bandwidth committed-rate 10 m
configure mpls rsvp-te profile Glasgow-Birmingham-pro setup-priority 5 hold-priority 5
The following commands define the primary and secondary paths between Glasgow and Birmingham.
create mpls rsvp-te path Glasgow-Birmingham-pri-path

create mpls rsvp-te path Glasgow-Birmingham-sec-path
The following commands pin each path to an LSR, such that each path takes a different route to the
endpoint 4.0.0.0. Path Glasgow-Birmingham-pri-path is routed through LSR 1.0.0.0 and path Glasgow-
Birmingham-sec-path is routed through LSR 5.0.0.0.
configure mpls rsvp-te path Glasgow-Birmingham-pri-path add ero 1.0.0.0/32 loose

configure mpls rsvp-te path Glasgow-Birmingham-sec-path add ero 5.0.0.0/32 loose
The following commands create one RSVP-TE LSP with one primary and one secondary or backup
path. Each path uses the same profile.
create mpls resv-te lsp Glasgow-Birmingham-lsp destiation 4.0.0.0

configure mpls rsvp lsp Glasgow-Birmingham-lsp add path Glasgow-Birmingham-pri-path

profile Glasgow-Birmingham-pro primary
configure mpls rsvp lsp Glasgow-Birmingham-lsp add path Glasgow-Birmingham-sec-path
profile Glasgow-Birmingham-pro secondary
NOTE
The secondary LSP is signaled, however it remains in a standby state unless the primary path becomes unavailable.
By default, a VPLS pseudo wire flows over any available LSP. However, a VPLS pseudo wire can be
specifically directed to use a configured RSVP-TE based LSP. Configuration is no different from
configuring an LDP-based VPLS pseudo wire, except that the RSVP-TE LSP is explicitly specified. The
following command specifically directs a VPLS pseudo wire to use a previously configured RSVP-TE
LSP:
configure vpls Glasgow-Birmingham-cust1 peer 4.0.0.0 add mpls lsp Glasgow-Birmingham-

lsp

Configuring RSVP-TE

33 Operational Administration and Management
Service providers incur huge expenses related to managing their networks. They also assume some
liability with respect to contracted service level agreements (SLAs). In short, easily managed high
availability networks are key to service provider profitability. Service providers require the network
management tools to:
● Monitor the health of the network.

● Alert network administrators to specific network problems.
● Provide diagnostic information to help quickly isolate problems.
● Suggest corrective action to rectify the problem.
These tools must be integrated into a Network Management System (NMS) solution that provides a
simple, easy-to-use interface.
This chapter describes the ExtremeXOS OAM tools and includes the following sections:
● Failure Detection and Alerts on page 863

● Diagnostic Information on page 863
Failure Detection and Alerts

In general, failures are detected by control plane signaling protocol mechanisms. Control plane
protocols can detect both link and node failures. LDP uses a keep-alive protocol to maintain adjacencies.
LDP can detect failures within seconds. RSVP-TE, on the other hand is a stateless protocol. LSRs
constantly refresh LSPs by resending the path messages. The RSVP-TE standard allows for hello timer
configuration settings that can detect failures in less than a second. Detected failures generate an SNMP
trap and or log an error message. At times, the protocol failure detection methods are insufficient for
diagnosing and determining a network problem and additional tools are required.
Diagnostic Information
To assist with problem determination in an MPLS network, LSP ping support is included as a CLI ping
option in ExtremeXOS. The LSP ping support is based on draft-ietf-mpls-lsp-ping-13.txt. This draft
includes support for both connectivity verification and fault isolation for transport LSPs. Connectivity
verification is supported using a modified "ping" packet that is sent over the specified transport LSP.
Transport LSP fault isolation is supported using the LSP trace feature.
LSP ping is designed to catch failures where the control plane believes that a transport LSP is
operational but it is in fact not functioning correctly. LSP data plane corruption is far less likely to occur
than an LSP control plane failure, but the LSP ping is also useful for detecting possible latency issues.
The command ping mpls lsp [<lsp_name> | any <host> | prefix <ipNetmask>] {reply-mode
[ip | mpls]} is used to send MPLS ping packets over an LSP.

Operational Administration and Management
MPLS pings are sent to the well-known UDP port number 3503 with an IP in the 127.0.0.0/8 IP subnet.
The source IP address is set to the sender.
The time stamp field is supported for calculating round trip times and is accurate to 1/100 of a second.
When replying to a ping, the LSP ping response (i.e., MPLS echo reply) sequence number and time-
stamp fields are set to the LSP ping request (i.e., MPLS echo request) values. One MPLS echo response
will be sent for each MPLS echo request received. MPLS echo replies are sent out-of-band as a natively
IP routed IPv4 UDP packet.
Sending an MPLS echo reply via the MPLS control plane is not supported.
When the control plane detects a transport LSP failure, the LSR can switch to another LSP if one is
available. When the failure is detected by LSP ping, LSP trace can be used to assist in failure isolation.
Use the command traceroute mpls lsp [<lsp_name> | any <host> | prefix <ipNetmask>]
{reply-mode [ip | mpls]} {{from <from>} {ttl <ttl>} {next-hop <hopaddress>}} to
perform an MPLS trace.

34 IPv4 Unicast Routing
This chapter describes the following IPv4 topics:

● Overview of IPv4 Unicast Routing on page 865
● Proxy ARP on page 870
● Relative Route Priorities on page 869
● IP Route Sharing on page 869
● Configuring IPv4 Unicast Routing on page 871
● Verifying the IPv4 Unicast Routing Configuration on page 871
● Routing Configuration Example on page 872
● IPv4 Multinetting on page 873
● Configuring DHCP/BOOTP Relay on page 879
● UDP Forwarding on page 881
● IP Broadcast Handling on page 883
● IP Route Compression on page 885
This chapter assumes that you are already familiar with IP unicast routing. If not, refer to the following
publications for additional information:
● RFC 1256—ICMP Router Discovery Messages
● RFC 1812—Requirements for IP Version 4 Routers
NOTE
For more information on interior gateway protocols, see Chapter 8, “RIP,” or Chapter 10, “OSPF.” For information on
exterior gateway protocols, see Chapter 12, “Border Gateway Protocol.” For more information on switch support for
IPv6, see Chapter 7, “IPv6 Unicast Routing.”
Overview of IPv4 Unicast Routing

The switch provides full Layer 3, IPv4 unicast routing. It exchanges routing information with other
routers on the network using either the Routing Information Protocol (RIP) or the Open Shortest Path
First (OSPF) protocol. The switch dynamically builds and maintains a routing table and determines the
best path for each of its routes.
Each host using the IP unicast routing functionality of the switch must have a unique IP address
assigned. In addition, the default gateway assigned to the host must be the IP address of the router
interface.

IPv4 Unicast Routing
Beginning in release 11.2, ExtremeXOS can provide both IPv4 and IPv6 routing at the same time.
Separate routing tables are maintained for the two versions. Most commands that require you to specify
an IP address can now accept either an IPv4 or IPv6 address and act accordingly. Additionally, many of
the IP configuration, enabling, and display commands have added tokens for IPv4 and IPv6 to clarify
the version required. For simplicity, existing commands affect IPv4 by default and require you to
specify IPv6, so configurations from an earlier release will still correctly configure an IPv4 network.
Beginning with ExtremeXOS version 12.0, support for route optimization via route compression has
been added, to allow you to reduce the number of routes that are being installed in hardware. This
helps in route optimization and scaling. This feature allows you to install only less specific routes in the
table, when overlapping routes with same nexthop exists. This feature can be enabled for IPv4 routes
with the following command:
enable iproute compression
This feature can be disabled for IPv4 routes with the following command:
disable iproute compression
Router Interfaces
The routing software and hardware routes IP traffic between router interfaces. A router interface is
simply a virtual LAN (VLAN) that has an IP address assigned to it.
As you create VLANs with IP addresses belonging to different IP subnets, you can also choose to route
between the VLANs. Both the VLAN switching and IP routing function occur within the switch.
NOTE
Each IP address and mask assigned to a VLAN must represent a unique IP subnet. You cannot configure the same
IP address and subnet on different VLANs.
In Figure 129, a BlackDiamond switch is depicted with two VLANs defined; Finance and Personnel. All
ports on slots 1 and 3 are assigned to Finance; all ports on slots 2 and 4 are assigned to Personnel. Finance
belongs to the IP network 192.207.35.0; the router interface for Finance is assigned the IP address
192.207.35.1. Personnel belongs to the IP network 192.207.36.0; its router interface is assigned IP address
192.207.36.1. Traffic within each VLAN is switched using the Ethernet MAC addresses. Traffic between
the two VLANs is routed using the IP addresses.

Figure 129: Routing between VLANs
1 2 3 4 A B 5 6 7 8
192.207.35.1 192.207.36.1
192.207.35.0 192.207.36.0
Finance Personnel
1 2 3 4
192.207.35.11 192.207.35.13
192.207.36.14
192.207.36.12 EX_070
Populating the Routing Table

The switch maintains an IP routing table for both network routes and host routes. The table is
populated from the following sources:
● Dynamically, by way of routing protocol packets or by Internet Control Message Protocol (ICMP)
redirects exchanged with other routers
● Statically, by way of routes entered by the administrator:
■ Default routes, configured by the administrator
■ Locally, by way of interface addresses assigned to the system
■ By other static routes, as configured by the administrator
NOTE
If you define a default route and subsequently delete the VLAN on the subnet associated with the default route, the
invalid default route entry remains. You must manually delete the configured default route.
Dynamic Routes
Dynamic routes are typically learned by way of RIP or OSPF. Routers that use RIP or OSPF exchange
information in their routing tables in the form of advertisements. Using dynamic routes, the routing
table contains only networks that are reachable.
Dynamic routes are aged out of the table when an update for the network is not received for a period of
time, as determined by the routing protocol.

Static Routes
Static routes are manually entered into the routing table. Static routes are used to reach networks not
advertised by routers.
Static routes can also be used for security reasons, to control which routes you want advertised by the
router. You configure, if you want all static routes to be advertised, using one of the following
commands:
● enable rip export [bgp | direct | e-bgp | i-bgp | ospf | ospf-extern1 | ospf-
extern2 | ospf-inter | ospf-intra | static] [cost <number> {tag <number>} |
policy <policy-name>]
or
disable rip export [bgp | direct | e-bgp | i-bgp | ospf | ospf-extern1 | ospf-
extern2 | ospf-inter | ospf-intra | static]
● enable ospf export [bgp | direct | e-bgp | i-bgp | rip | static] [cost <cost>
type [ase-type-1 | ase-type-2] {tag <number>} | <policy-map>]
or
disable ospf export [bgp | direct | e-bgp | i-bgp | rip | static]
The default setting is disabled. Static routes are never aged out of the routing table.
A static route’s nexthop (gateway) must be associated with a valid IP subnet. An IP subnet is associated
with a single VLAN by its IP address and subnet mask. If the VLAN is subsequently deleted, the static
route entries using that subnet must be deleted manually.
Multiple Routes
When there are multiple, conflicting choices of a route to a particular destination, the router picks the
route with the longest matching network mask. If these are still equal, the router picks the route using
the following default criteria (in the order specified):
● Directly attached network interfaces
● Static routes
● ICMP redirects
● Dynamic routes
● Directly attached network interfaces that are not active.
NOTE
If you define multiple default routes, the route that has the lowest metric is used. If multiple default routes have the
same lowest metric, the system picks one of the routes.
You can also configure blackhole routes—traffic to these destinations is silently dropped.
The criteria for choosing from multiple routes with the longest matching network mask is set by
choosing the relative route priorities.

Relative Route Priorities

Table 100 lists the relative priorities assigned to routes depending on the learned source of the route.
NOTE
Although these priorities can be changed, do not attempt any manipulation unless you are expertly familiar with the
possible consequences.
Table 100: Relative route priorities

Route Origin Priority
Direct 10
BlackHole 50
Static 1100
ICMP 1200
EBGP 1700
IBGP 1900
OSPFIntra 2200
OSPFInter 2300
RIP 2400
OSPFExtern1 3200
OSPFExtern2 3300
BOOTP 5000
To change the relative route priority, use the following command:
configure iproute {ipv4} priority [rip | blackhole | bootp | ebgp | ibgp | icmp |
static | ospf-intra | ospf-inter | ospf-as-external | ospf-extern1 | ospf-extern2]
<priority>
IP Route Sharing
IP route sharing allows multiple equal-cost routes to be used concurrently. IP route sharing can be used
with static routes or with OSPF routes. In OSPF, this capability is referred to as equal cost multipath
(ECMP) routing. To use IP route sharing, use the following commands:
BlackDiamond 8800 family, SummitStack and Summit family:

enable iproute sharing
BlackDiamond 10808 and BlackDiamond 12800 series:.

enable iproute sharing
enable sharing
configure sharing address-based L2_L3
NOTE
IP route sharing on the BlackDiamond 10808 and 12800 series switches requires that link aggregation (load
sharing) be enabled and that you configure the L2_L3 load sharing algorithm, as shown in the example.

Next, configure static routes and/or OSPF as you would normally. For better resiliency when sharing
static routes, it is advised to configure static ARP and FDB entries for each gateway used by the shared
static routes. To do so, use the following commands:
configure iparp add <ip_addr> {vr <vr_name>} <mac>

create fdbentry <mac_addr> vlan <vlan_name> ports <port_list>
BlackDiamond 10808 and BlackDiamond 12800 series. For the BlackDiamond 10808 and BlackDiamond
12800 series, ExtremeXOS supports route sharing across up to eight static routes and up to eight ECMP
routes for OSPF.
BlackDiamond 8800 family, SummitStack, and Summit family. For the BlackDiamond 8800 family,
SummitStack, and Summit family switches, ExtremeXOS supports route sharing across up to four static
routes and up to four ECMP routes for OSPF, by default, but can be configured to support eight. To
configure the maximum number of ECMP gateways, use the following command:
configure iproute sharing max-gateways <max_gateways>
Using route sharing makes router troubleshooting more difficult because of the complexity in predicting
the path over which the traffic will travel.
Proxy ARP
Proxy Address Resolution Protocol (ARP) was first invented so that ARP-capable devices could respond
to ARP request packets on behalf of ARP-incapable devices. Proxy ARP can also be used to achieve
router redundancy and to simplify IP client configuration. The switch supports proxy ARP for this type
of network configuration. The section describes some example of using proxy ARP with the switch.
ARP-Incapable Devices
To configure the switch to respond to ARP requests on behalf of devices that are incapable of doing so,
you must configure the IP address and MAC address of the ARP-incapable device using the following
command:
configure iparp add proxy [<ipNetmask> | <ip_addr> {<mask>}] {vr <vr_name>} {<mac>}
{always}
After it is configured, the system responds to ARP requests on behalf of the device as long as the
following conditions are satisfied:
● The valid IP ARP request is received on a router interface.
● The target IP address matches the IP address configured in the proxy ARP table.
● The proxy ARP table entry indicates that the system should answer this ARP request, based on the
ingress VLAN and whether the always parameter is set. When the always option is set, the switch
will ARP for the host even if the ARP requester is on the same subnet as the requested host. If the
always option is not set, the switch will only answer if the ARP request comes in from a VLAN that
is not on the same subnet as the requested host.
When all the proxy ARP conditions are met, the switch formulates an ARP response using the
configured MAC address in the packet.

Configuring IPv4 Unicast Routing
Proxy ARP Between Subnets

In some networks, it is desirable to configure the IP host with a wider subnet than the actual subnet
mask of the segment. You can use proxy ARP so that the router answers ARP requests for devices
outside of the subnet. As a result, the host communicates as if all devices are local. In reality,
communication with devices outside of the subnet are proxied by the router.
For example, an IP host is configured with a class B address of 100.101.102.103 and a mask of
255.255.0.0. The switch is configured with the IP address 100.101.102.1 and a mask of 255.255.255.0. The
switch is also configured with a proxy ARP entry of IP address 100.101.0.0 and mask 255.255.0.0, without
the always parameter.
When the IP host tries to communicate with the host at address 100.101.45.67, the IP host communicates
as if the two hosts are on the same subnet, and sends out an IP ARP request. The switch answers on
behalf of the device at address 100.101.45.67, using its own MAC address. All subsequent data packets
from 100.101.102.103 are sent to the switch, and the switch routes the packets to 100.101.45.67.
Configuring IPv4 Unicast Routing

To configure IP unicast routing on the switch:
1 Create and configure two or more VLANs.
2 Assign each VLAN that will be using routing an IP address using the following command:
Ensure that each VLAN has a unique IP address.
3 Configure a default route using the following command:
unicast-only}
Default routes are used when the router has no other dynamic or static route to the requested
destination.
4 Turn on IP routing for one or all VLANs using the following command:
enable ipforwarding {ipv4 | broadcast | ignore-broadcast | fast-direct-broadcast}
{vlan <vlan_name>}
5 Configure the routing protocol, if required. For a simple network using RIP, the default
configuration may be acceptable.
6 Turn on RIP or OSPF using one of the following commands:
enable rip
enable ospf
Verifying the IPv4 Unicast Routing Configuration

Use the show iproute command to display the current configuration of IP unicast routing for the
switch and for each VLAN. The show iproute command displays the currently configured routes and
includes how each route was learned.

Additional verification commands include:

● show iparp—Displays the IP ARP table of the system.
● show ipconfig—Displays configuration information for one or more VLANs.
Routing Configuration Example

Figure 130 illustrates a BlackDiamond switch that has three VLANs defined as follows:
● Finance
■ All ports on slots 1 and 3 have been assigned.
■ IP address 192.207.35.1.
● Personnel
■ Protocol-sensitive VLAN using the IP protocol.
■ IP address 192.207.36.1.
● MyCompany
■ Port-based VLAN.
■ All ports on slots 1 through 4 have been assigned.
Figure 130: Unicast routing configuration example
1 2 3 4 A B 5 6 7 8
192.207.35.1 192.207.36.1
MyCompany
192.207.35.0 192.207.36.0
Finance Personnel
1 2 3 4
IP
IP = IP traffic
NetBIOS
NetBIOS
IP = NetBIOS traffic
IP
NetBIOS
NetBIOS EX_047

IPv4 Multinetting
The stations connected to the system generate a combination of IP traffic and NetBIOS traffic. The IP
traffic is filtered by the protocol-sensitive VLANs. All other traffic is directed to the VLAN MyCompany.
In this configuration, all IP traffic from stations connected to slots 1 and 3 have access to the router by
way of the VLAN Finance. Ports on slots 2 and 4 reach the router by way of the VLAN Personnel. All
other traffic (NetBIOS) is part of the VLAN MyCompany.
The example in Figure 130 is configured as follows:
create vlan Finance

create vlan Personnel
create vlan MyCompany
configure Finance protocol ip

configure Personnel protocol ip
configure Finance add port 1:*,3:*

configure Personnel add port 2:*,4:*
configure MyCompany add port all
configure Finance ipaddress 192.207.35.1

configure Personnel ipaddress 192.207.36.1
configure rip add vlan Finance
configure rip add vlan Personnel
enable ipforwarding
enable rip
IPv4 Multinetting
IP multinetting refers to having multiple IP networks on the same bridging domain (or VLAN). The
hosts connected to the same physical segment can belong to any one of the networks, so multiple
subnets can overlap onto the same physical segment. Any routing between the hosts in different
networks is done through the interface of the router. Typically, different IP networks will be on
different physical segments, but IP multinetting does not require this.
Multinetting can be a critical element in a transition strategy, allowing a legacy assignment of IP

addresses to coexist with newly configured hosts. However, because of the additional constraints
introduced in troubleshooting and bandwidth, Extreme Networks recommends that you use
multinetting as a transitional tactic only, and not as a long-term network design strategy.
Multinetting was not supported in ExtremeXOS 10.1, but versions of ExtremeWare before that
supported a multinetting implementation that required separate VLANs for each IP network. The
implementation introduced in ExtremeXOS 11.0 is simpler to configure, does not require that you create
a dummy multinetting protocol, and does not require that you create VLANs for each IP network. This
implementation does not require you to explicitly enable IP multinetting. Multinetting is automatically
enabled when a secondary IP address is assigned to a VLAN.
The following sections discuss these multinetting topics:

● Multinetting Topology on page 874
● How Multinetting Affects Other Features on page 875

● Configuring IPv4 Multinetting on page 878

● IP Multinetting Examples on page 879
Multinetting Topology
For an IP multinetted interface, one of the IP networks on the interface acts as the transit network for
the traffic that is routed by this interface. The transit network is the primary subnet for the interface.
The remaining multinetted subnets, called the secondary subnets, must be stub networks. This
restriction is required because it is not possible to associate the source of the incoming routed traffic to a
particular network. IP routing happens between the different subnets of the same VLAN (one arm
routing) and also between subnets of different VLANs.
Figure 131: Multinetted Network Topology
Transit
VLAN multi
network
Primary subnet
Secondary
subnet-1 Host
Secondary
subnet-2
BD10K
EX_102
Figure 131 shows a multinetted VLAN named multi. VLAN multi has three IP subnets so three IP
addresses have been configured for the VLAN. One of the subnets is the primary subnet and can be
connected to any transit network (for example, the Internet). The remaining two subnets are stub
networks, and multiple hosts such as management stations (such as user PCs and file servers) can be
connected to them. You should not put any additional routing or switching devices in the secondary
subnets to avoid routing loops. In Figure 131 the subnets are on separate physical segments, however,
multinetting can also support hosts from different IP subnets on the same physical segment.
When multinetting is configured on a VLAN, the switch can be reached using any of the subnet
addresses (primary or secondary) assigned to VLAN. This means that you can perform operations like
ping, Telnet, Trivial File Transfer Protocol (TFTP), Secure Shell 2 (SSH2), and others to the switch from
a host residing in either the primary or the secondary subnet of the VLAN. Other host functions (such
as traceroute) are also supported on the secondary interface of a VLAN.

IPv4 Multinetting
How Multinetting Affects Other Features

Multinetting will affect some other features in ExtremeXOS. The following sections explain how
multinetting affects both Layer 2 and Layer 3 features.
ARP
ARP operates on the interface and responds to every request coming from either the primary or
secondary subnet. When multiple subnets are configured on a VLAN and an ARP request is generated
by the switch over that VLAN, the source IP address of the ARP request must be a local IP address of
the subnet to which the destination IP address (which is being ARPed) belongs.
For example, if a switch multinets the subnets 10.0.0.0/24 and 20.0.0.0/24 (with VLAN IP addresses of
10.0.0.1 and 20.0.0.1), and generates an ARP request for the IP address 10.0.0.2, then the source IP
address in the ARP packet will be set to 10.0.0.1 and not to 20.0.0.1.
Route Manager
The Route Manager will install a route corresponding to each of the secondary interfaces. The route
origin will be direct, will be treated as a regular IP route, and can be used for IP data traffic forwarding.
These routes can also be redistributed into the various routing protocol domains if you configure route
redistribution.
IRDP
Some functional changes are required in Internet Router Discovery Protocol (IRDP) as result of IP
multinetting support. When IRDP is enabled on a Layer 3 VLAN, ExtremeXOS periodically sends ICMP
router advertisement messages through each subnet (primary and secondary) and responds to ICMP
router solicitation messages based on the source IP address of soliciting host.
Unicast Routing Protocols

Unicast routing protocols treat each IP network as an interface. The interface corresponding to the
primary subnet is the active interface, and the interfaces corresponding to the secondary subnet are
passive subnets.
For example, in the case of Open Shortest Path First (OSPF), the system treats each network as an
interface, and hello messages are not sent out or received over the non-primary interface. In this way,
the router link state advertisement (LSA) includes information to advertise that the primary network is
a transit network and the secondary networks are stub networks, thereby preventing any traffic from
being routed from a source in the secondary network.
Interface-based routing protocols (for example, OSPF) can be configured on per VLAN basis. A routing
protocol cannot be configured on an individual primary or secondary interface. Configuring a protocol
parameter on a VLAN automatically configures the parameter on all its associated primary and
secondary interfaces. The same logic applies to configuring IP forwarding, for example, on a VLAN.
Routing protocols in the multinetted environment advertise the secondary subnets to their peers in their
protocol exchange process. For example, for OSPF the secondary subnets are advertised as stub

networks in router LSAs. RIP also advertises secondary subnets to its peers residing on the primary
subnet.
OSPF. This section describes the behavior of OSPF in an IPv4 multinetting environment:
● Each network is treated as an interface, and hello messages are not sent out or received over the
non-primary interface. In this way, the router LSA includes information to advertise that the primary
network is a transit network and the secondary networks are stub networks, thereby preventing any
traffic from being routed from a source in the secondary network.
● Any inbound OSPF control packets from secondary interfaces are dropped.
● Direct routes corresponding to secondary interfaces can be exported into the OSPF domain (by
enabling export of direct routes), if OSPF is not enabled on the container VLAN.
● When you create an OSPF area address range for aggregation, you must consider the secondary
subnet addresses for any conflicts. That is, any secondary interface with the exact subnet address as
the range cannot be in another area.
● The automatic selection algorithm for the OSPF router ID considers the secondary interface
addresses also. The numerically highest interface address is selected as the OSPF router-id.
RIP. This section describes the behavior of the Routing Information Protocol (RIP) in an IP multinetting
environment:
● RIP does not send any routing information update on the secondary interfaces. However, RIP will
advertise networks corresponding to secondary interfaces in its routing information packet to the
primary interface.
● Any inbound RIP control packets from secondary interfaces are dropped.
● Direct routes corresponding to secondary interfaces can be exported into the RIP domain (by
enabling export of direct routes), if RIP is not enabled on the container VLAN.
BGP. There are no behavioral changes in the Border Gateway Protocol (BGP) in an IP multinetting
environment. This section describes a set of recommendations for using BGP with IP multinetting:
● Be careful of creating a BGP neighbor session with a BGP speaker residing in secondary subnet. This
situation may lead to routing loops.
● All secondary subnets are like stub networks, so you must configure BGP in such a way that the
BGP next hop becomes reachable using the primary subnet of a VLAN.
● When setting the BGP next hop using an inbound or outbound policy, ensure that the next hop is
reachable from the primary interface.
● A BGP static network's reachability can also be resolved from the secondary subnet.
● Secondary interface addresses can be used as the source interface for a BGP neighbor.
● Direct routes corresponding to secondary interfaces can be exported into the BGP domain (by
enabling export of direct routes).
IGMP Snooping and IGMP

Internet Group Management Protocol (IGMP) snooping and IGMP treat the VLAN as an interface.
Only control packets with a source address belonging to the IP networks configured on that interface
are accepted. IGMP accepts membership information that originates from hosts in both the primary and

IPv4 Multinetting
secondary subnets. The following describes the changes in behavior of IGMP in an IP multinetting
environment:
● A layer 3 VLAN will always use the primary IP address as the source address to send out an IGMP
query, and querier election is based on the primary IP address of interface. Because the RFC dictates
that there is only one querier per physical segment, the querier may be attached to any of configured
IP interfaces, including secondary interfaces, although this is not a recommended configuration.
● For a static IGMP group, the membership report is also sent out using the primary IP address.
● For local multicast groups such as 224.0.0.X, the membership report is sent out using the first IP
address configured on the interface, which is the primary IP address in ExtremeXOS.
● The source IP address check is disabled for any IGMP control packets (such as IGMP query and
IGMP membership report). Source IP address checking for IGMP control packet is disabled for all
VLANs, not just the multinetted VLANs.
Multicast Routing Protocols

For Protocol-Independent Multicast (PIM), the following behavior changes should be noted in a
multinetting environment:
● PIM does not peer with any other PIM router on a secondary subnet.
● PIM also processes data packets from the host on secondary subnets.
● PIM also accepts membership information from hosts on secondary subnets.
EAPS, ESRP, and STP

Control protocols like Ethernet Automatic Protection Switching (EAPS), Extreme Standby Router
Protocol (ESRP), and the Spanning Tree Protocol (STP) treat the VLAN as an interface. If the protocol
control packets are exchanged as Layer 3 packets, then the source address in the packet is validated
against the IP networks configured on that interface.
DHCP Server
The Dynamic Host Configuration Protocol (DHCP) server implementation in ExtremeXOS 11.0 will only
support address allocation on the primary IP interface of the configured VLAN. That is, all DHCP
clients residing on a bridging domain will have IP address belonging to the primary subnet. To add a
host on secondary subnet, you must manually configure the IP address information on that host.
DHCP Relay
When the switch is configured as a DHCP relay agent, it will forward the DHCP request received from
a client to the DHCP server. When doing so, the system sets the GIADDR field in the DHCP request
packet to the primary IP address of the ingress VLAN. This means that the DHCP server that resides on
a remote subnet will allocate an IP address for the client in the primary subnet range.
VRRP
The Virtual Router Redundancy Protocol (VRRP) protection can be provided for the primary as well as
for the secondary IP addresses of a VLAN. For multinetting, the IP address assigned to an VRRP virtual
router identifier (VRID) can be either the primary or the secondary IP addresses of the corresponding
VLAN.

For example, assume a VLAN v1 with two IP addresses: a primary IP address of 10.0.0.1/24, and a
secondary IP address of 20.0.0.1/24.
To provide VRRP protection to such a VLAN, you must configure one of the following:
● Configure VRRP in VLAN v1 with two VRRP VRIDs. One VRID will have the virtual IP address
10.0.0.1/24, and the other VRID will have the virtual IP address 20.0.0.1/24. The other VRRP router,
the one configured to act as backup, should be configured similarly.
—OR—
● Configure VRRP in VLAN v1 with two VRRP VRIDs. One VRID will have the virtual IP address as
10.0.0.1/24, and the other VRID will have the virtual IP address as 20.0.0.1/24.
It is possible for a VRRP VR to have additional virtual IP addresses assigned to it. In this case, the
following conditions must be met:
● Multiple virtual IP addresses for the same VRID must be on the same subnet.
● Multiple virtual IP addresses must all not be owned by the switch.
Assuming a VLAN v1 that has IP addresses 1.1.1.1/24 and 2.2.2.2/24, here are some more examples of
valid configurations:
● VRRP VR on v1 with VRID of 99 with virtual IP addresses of 1.1.1.2 and 1.1.1.3
Given the same VLAN v1 as above, here are some invalid configurations:
● VRRP VR on v1 with VRID of 99 with virtual IP addresses of 1.1.1.1 and 2.2.2.2 (the virtual IP
addresses are not on the same subnet)
● VRRP VR on v1 with VRID of 100 with virtual IP addresses of 2.2.2.2 and 1.1.1.1 (the virtual IP
addresses are not on the same subnet)
● VRRP VR on v1 with VRID of 99 with virtual IP addresses of 1.1.1.1 and 1.1.1.99 (one virtual IP
address is owned by the switch and one is not)
● VRRP VR on v1 with VRID of 100 with virtual IP addresses of 2.2.2.2 and 2.2.2.99 (one virtual IP
address is owned by the switch and one is not).
Configuring IPv4 Multinetting

You configure IP multinetting by adding a secondary IP address to a vlan. Use the following command
to add a secondary IP address:
configure vlan <vlan_name> add secondary-ipaddress [<ipaddress> {<netmask>} |
<ipNetmask>]
After you have added a secondary IP address, you cannot change the primary IP address of a VLAN
until you first delete all the secondary IP addresses. To delete secondary IP addresses, use the following
command:
configure vlan <vlan_name> delete secondary-ipaddress [<ipaddress> | all]

Configuring DHCP/BOOTP Relay
IP Multinetting Examples
The following example configures a switch to have one multinetted segment (port 5:5) that contains
three subnets (192.168.34.0/24, 192.168.35.0/24, and 192.168.37.0/24).
configure default delete port 5:5
create vlan multinet
configure multinet ipaddress 192.168.34.1/24
configure multinet add secondary-ipaddress 192.168.35.1/24
configure multinet add secondary-ipaddress 192.168.37.1/24
configure multinet add port 5:5
enable ipforwarding
The following example configures a switch to have one multinetted segment (port 5:5) that contains
three subnets (192.168.34.0, 192.168.35.0, and 192.168.37.0). It also configures a second multinetted
segment consisting of two subnets (192.168.36.0 and 172.16.45.0). The second multinetted segment spans
three ports (1:8, 2:9, and 3:10). RIP is enabled on both multinetted segments.
configure default delete port 5:5

create vlan multinet
configure multinet ipaddress 192.168.34.1
configure multinet add secondary-ipaddress 192.168.35.1
configure multinet add secondary-ipaddress 192.168.37.1
configure multinet add port 5:5
configure default delete port 1:8, 2:9, 3:10
create vlan multinet_2
configure multinet_2 ipaddress 192.168.36.1
configure multinet_2 add secondary-ipaddress 172.16.45.1
configure multinet_2 add port 1:8, 2:9, 3:10
configure rip add vlan multinet
configure rip add vlan multinet_2
enable rip
enable ipforwarding
Configuring DHCP/BOOTP Relay

After IP unicast routing has been configured, you can configure the switch to forward Dynamic Host
Configuration Protocol (DHCP) or BOOTP requests coming from clients on subnets being serviced by
the switch and going to hosts on different subnets. This feature can be used in various applications,
including DHCP services between Windows NT servers and clients running Windows 95. To configure
the relay function:
1 Configure VLANs and IP unicast routing.
2 Enable the DHCP or BOOTP relay function, using the following command:
enable bootprelay {vr <vrid>}
3 Configure the addresses to which DHCP or BOOTP requests should be directed, using the following
command:
configure bootprelay add <ip_address> {vr <vrid>}
To delete an entry, use the following command:
configure bootprelay delete [<ip_address> | all] {vr <vrid>}

Configuring the DHCP Relay Agent Option (Option 82)

After configuring and enabling the DHCP/BOOTP relay feature, you can enable the DHCP relay agent
option feature. This feature inserts a piece of information, called option 82, into any DHCP request
packet that is to be relayed by the switch. Similarly, if a DHCP reply received by the switch contains a
valid relay agent option, the option will be stripped from the packet before it is relayed to the client.
The DHCP relay agent option consists of two pieces of data, called sub-options. The first is the agent
circuit ID sub-option, and the second is the agent remote ID sub-option. When the DHCP relay agent
option is enabled on switches running ExtremeXOS, the value of these sub-options is set as follows:
● Agent circuit ID sub-option: Contains the VLAN information, module information, and ID of the
port on which the original DHCP request packet was received. This ID is encoded as ((slot_number *
1000) + port_number). For example, if the DHCP request were received on port 3:12, the agent circuit
ID value would be 3012. On non-slot-based switches, the agent circuit ID value is simply the port
number.
● Agent remote ID sub-option: Always contains the Ethernet MAC address of the relaying switch.
You can display the Ethernet MAC address of the switch by issuing the show switch command.
To enable the DHCP relay agent option, use the following command after configuring the DHCP/
BOOTP relay function:
configure bootprelay dhcp-agent information option
When DHCP option 82 is enabled, two types of packets need to be handled:
● DHCP Request: When the switch (relay agent) receives a DHCP request, option 82 is added at the
end of the packet. If the option has already been enabled, then the action taken depends on the
configured policy (drop packet, keep existing option 82 value, or replace the existing option). If the
incoming DHCP request is tagged, then that VLAN ID is added to the circuit ID sub option of
option 82; otherwise, the default VLAN ID is added.
● DHCP Reply: When the option 82 information check is enabled, the packets received from the
DHCP server are checked for option 82 information. If the remote ID sub-option is the switch's MAC
address, the packet is sent to the client; if not, the packet is dropped. If the check is not enabled. the
packets are forwarded as-is.
To disable the DHCP relay agent option, use the following command:
unconfigure bootprelay dhcp-agent information option
In some instances, a DHCP server may not properly handle a DHCP request packet containing a relay
agent option. To prevent DHCP reply packets with invalid or missing relay agent options from being
forwarded to the client, use the following command:
configure bootprelay dhcp-agent information check
To disable checking of DHCP replies, use this command:
unconfigure bootprelay dhcp-agent information check
A DHCP relay agent may receive a client DHCP packet that has been forwarded from another relay
agent. If this relayed packet already contains a relay agent option, then the switch will handle this
packet according to the configured DHCP relay agent option policy. The possible actions are to replace

UDP Forwarding
the option information, to keep the information, or to drop packets containing option 82 information. To
configure this policy, use the following command:
configure bootprelay dhcp-agent information policy [drop | keep | replace]
The default relay policy is replace. To configure the policy to the default, use this command:
unconfigure bootprelay dhcp-agent information policy
For more general information about the DHCP relay agent information option, refer to RFC 3046.
Verifying the DHCP/BOOTP Relay Configuration

To verify the DHCP/BOOTP relay configuration, use the following command:
show bootprelay
This command displays the configuration of the BOOTP relay service and the addresses that are
currently configured.
UDP Forwarding
UDP Forwarding is a flexible and generalized routing utility for handling the directed forwarding of
broadcast UDP packets. UDP Forwarding enables you to configure your switch so that inbound
broadcast UDP packets on a VLAN are forwarded to a particular destination IP address or VLAN.
UDP Forwarding allows applications, such as multiple DHCP relay services from differing sets of
VLANs, to be directed to different DHCP servers.
The following rules apply to UDP broadcast packets handled by this feature:
● If the UDP profile includes BOOTP or DHCP, it is handled according to guidelines specified in RFC
1542.
● If the UDP profile includes other types of traffic, these packets have the IP destination address
modified as configured, and changes are made to the IP and UDP checksums and decrements to the
TTL field, as appropriate.
If UDP Forwarding is used for BOOTP or DHCP forwarding purposes, do not configure or use the
existing bootprelay function. However, if the previous bootprelay functions are adequate, you may
continue to use them.
NOTE
UDP Forwarding only works across a layer 3 boundary and currently, UDP Forwarding can be applied to IPv4 packets
only, not to IPv6 packets.
Configuring UDP Forwarding

To configure UDP Forwarding, create an policy file for your UDP profile, and then associate the profile
with a VLAN using the following command:
configure vlan <vlan_name> udp-profile [<profilename> | none]

You can apply a UDP Forwarding policy only to an L3 VLAN (a VLAN having at least one IP address
configured on it). If no IP address is configured on the VLAN, then the command will be rejected.
UDP profiles are similar to ACL policy files. UDP profiles use a subset of the match conditions allowed
for ACLs. A UDP Forwarding policy must contain only the following attributes. Unrecognized
attributes will be ignored.
● Match attributes
■ Destination UDP Port Number (destination-port)
■ Source IP address (source-ipaddress)
● Action modified (set) attributes
■ Destination IP address (destination-ipaddress)
■ VLAN name (vlan)
Policy files used for UDP Forwarding are processed differently from standard policy files. Instead of
terminating when an entry’s match clause becomes true, each entry in the policy file will be processed
and the corresponding action will be taken for each true match clause.
For example, if the following policy file is used as a UDP Forwarding profile, any packets destined for
UDP port 67 will be sent to IP address 20.0.0.5 AND flooded to VLAN to7:
entry one {
if match all {
} then {
destination-ipaddress 20.0.0.5 ;
}
}
entry two {
if match all {
} then {
vlan "to7" ;
}
}
If you include more than one VLAN set attribute or more than one destination-ipaddress set attribute in
one policy entry, the last one will be accepted and the rest ignored. Therefore, you can have two valid
set statements in each entry of a UDP Forwarding policy; one a destination-ipaddress and one a VLAN.
ExtremeXOS currently allows a maximum of eight entries in a UDP Forwarding policy, so you can
define a maximum of 16 destinations for one inbound broadcast UDP packet: eight IP address and eight
VLANs.
NOTE
It is strongly advised not to have more than eight entries in a UDP Forwarding profile. The UDP Forwarding module
will process those entries even if the entries do not contain any attributes for UDP Forwarding. Having more than
eight entries will drastically reduce the performance of the system. If the inbound UDP traffic rate is very high,
having more than eight entries could cause the system to freeze or become locked.

IP Broadcast Handling
NOTE
If you rename a VLAN referred to in your UDP Forwarding profile, you must manually edit the policy to reflect the
new name, and refresh the policy.
You can also validate whether the UDP profile has been successfully associated with the VLAN by
using the command show policy {<policy-name> | detail}. UDP Forwarding is implemented as
part of the netTools process, so the command will display netTools as a user of the policy.
To remove a policy, use the none form of the following command:
configure vlan <vlan_name> udp-profile [<profilename> | none]
or use this command:
unconfigure vlan <vlan_name> udp-profile
For more information about creating and editing policy files, see Chapter 11, “Policy Manager.” For
more information about ACL policy files, see Chapter 12, “Access Lists (ACLs).”
UDP Echo Server

You can use UDP echo packets to measure the transit time for data between the transmitting and
receiving end.
To enable UDP echo server support, use the following command:
enable udp-echo-server {vr <vrid>}{udp-port <port>}
To disable UDP echo server support, use the following command:
disable udp-echo-server {vr <vrid>}
Beginning with version 12.0, the capability to IP subnet directed broadcast forwarding was added to
ExtremeXOS. In ExtremeXOS, IP subnet directed broadcast forwarding is done in the software by
default; if you want to perform forwarding in the hardware, see the command reference pages on IP
forwarding in the Extreme XOS 12.0 Command Reference Guide.
IP Broadcast Handling Details

To understand how IP broadcast handling functions in ExtremeXOS, consider the following two cases:
A system sends an IP packet (for example, via the ping command) to an IP subnet directed broadcast
address which is directly connected to that system. In this case, the IP packet goes out as an L2
broadcast with the destination media access control (DMAC) addresses all set to FF, while the source
media access control (SMAc) is set to the system MAC. This packet is sent out of all the ports of the
VLAN.

In the second case, a system sends a packet (for example, via the ping command) to an IP subnet
directed broadcast address which is remotely connected via a gateway. In this case, the IP packet goes
out as an L2 unicast with DMAC equal to the gateway's MAC address, while the SMAC is set to the
system MAC. At the gateway router, the existing IP packet forwarding mechanism is sufficient to send
the packet out of the correct interface if the router is not the final hop router.
When the packet reaches the final hop router, which is directly connected to the target IP subnet, IP
directed broadcast forwarding needs to be turned on. The IP broadcast handling feature is applicable
only at the final hop router directly attached to the target subnet. At the final hop router, when IP
subnet directed broadcast forwarding is enabled on an IP vlan via the command line, the following
happens:
● Some basic validity checks are performed (for example, checking to see if the VLAN has IP enabled)
● A subnet broadcast route entry for the subnet is installed. For example, consider a system with the
following configuration:
Vlan-A = 10.1.1.0/24, ports 1:1, 1:2, 1:3, 1:4
Vlan-B = 20.1.1.0/24, ports 1:5, 1:6, 1:7, 1:8
Vlan-C = 30.1.1.0/24, ports 1:9, 1:10, 1:11
If you enable IP directed broadcast forwarding on VLAN-A, you should install a route entry for
10.1.1.1.255 on this system.
● A packet arriving on port 1:5 VLAN-B with destination IP (DIP) set to 10.1.1.255, the source IP (SIP)
set to 20.1.1.3, the DMAC set to the router MAC, and the SMAC set to the originating system MAC,
arrives at the installed route entry and is sent out on all the ports of VLAN-A, with DMAC set to be
all FF and the SMAC set to the router's system MAC.
● An IP packet arriving on port 1:1 VLAN-A with the DIP set to 10.1.1.255, the SIP set to 10.1.1.3, the
DMAC set to all FF, and the SMAC set to the originator’s MAC, causes the L2 to be flooded out of
all the ports of VLAN-A.
When IP subnet directed broadcast is disabled on an IP VLAN, n each port of the VLAN, all IP subnet
directed broadcast entries are deleted.
NOTE
IP subnet directed broadcast forwarding is fast-path.
Command-line Support for IP Broadcast Handling

The enable ipforwarding and disable ipforwarding commands were enhanced to support the
enhanced IP broadcast handling functionality. Specifically, the following two keywords were added to
the commands:
● fast-direct-broadcast, which specifies to forward IP broadcast packets in the hardware and send a
copy to the CPU
● ignore-broadcast, which specifies to forward IP broadcast packets in the hardware but not send a
copy to the CPU
For further details, refer to the appropriate command reference pages in the Extreme XOS 12.0 Command
Reference Guide.

NOTE
These two keywords are available on the BlackDiamond 10808 switch only.
IP Route Compression
Route handling optimization and compression allow you to reduce the number of routes that are being
installed in hardware by installing less specific ones in cases of overlapping routes. The route pruning
technique is implemented as part of the Route Manager (RtMgr) process.
When a route is added or deleted or updated, the pruning algorithm is applied to see if the new route
and/or its immediate children can be compressed or uncompressed as follows:
● If the parent node (immediate less specific route) of the newly added IP prefix has the same gateway
as the new IP prefix, the newly added prefix is compressed.
● If the gateways of the newly added IP prefix and its immediate children are the same, the child
nodes are compressed.
● If the gateways of the newly added IP prefix and its immediate children are not the same, and the
child nodes had been previously compressed, the child nodes are uncompressed.
Enabling and Disabling Route Compression

The following command enables or disables IP route compression for a specified address family and
VR. The default family is IPv4. The default VR is the current CLI context VR.
[enable | disable] iproute {ipv4 | ipv6} compression {vr <vrname>}
The following example enables IP route compression for the IPv4 address family and the VR of the
current CLI context.
enable iproute compression
The following example disables IP route compression for the IPv4 address family and the VR of the
current CLI context.
disable iproute compression
Currently route compression is done in chunking when the user issues the enable/disable iproute
compression command. Because RtMgr processes a limited number of IP prefix nodes per second,
route compression does not cause any performance limitation. Also, incremental route addition or
deletion does not cause any performance change or delay when IP route compression is enabled.
Viewing the Route Compression

View a compressed route using the following command:
show iproute
Sample output:
Ori Destination Gateway Mtr Flags VLAN Duration

#be 3.0.0.0/8 111.222.0.5 7 UG-D---um--- feed 0d:19h:52m:49s
#be 4.0.0.0/8 111.222.0.5 5 UG-D---um--- feed 0d:19h:52m:49s

#be 4.0.0.0/9 111.222.0.5 5 UG-D---um--c feed 0d:19h:52m:49s

……………………………………………………
Origin(Ori): (b) BlackHole, (be) EBGP, (bg) BGP, (bi) IBGP, (bo) BOOTP
(ct) CBT, (d) Direct, (df) DownIF, (dv) DVMRP, (e1) ISISL1Ext
(e2) ISISL2Ext, (h) Hardcoded, (i) ICMP, (i1) ISISL1 (i2) ISISL2
(mb) MBGP, (mbe) MBGPExt, (mbi) MBGPInter
(mo) MOSPF (o) OSPF, (o1) OSPFExt1, (o2) OSPFExt2
(oa) OSPFIntra, (oe) OSPFAsExt, (or) OSPFInter, (pd) PIM-DM, (ps) PIM-SM
(r) RIP, (ra) RtAdvrt, (s) Static, (sv) SLB_VIP, (un) UnKnown
(*) Preferred unicast route (@) Preferred multicast route
(#) Preferred unicast and multicast route
Flags: (B) BlackHole, (D) Dynamic, (G) Gateway, (H) Host Route
(m) Multicast, (P) LPM-routing, (R) Modified, (S) Static
(u) Unicast, (U) Up (c) Compressed
Mask distribution:
19 routes at length 8 9 routes at length 9
Route Origin distribution:

7 routes from Direct 184816 routes from EBGP
Total number of routes = 184823

Total number of compressed routes = 93274
Display an iproute summary using the following command:
show iproute summary
Sample output:
=================ROUTE SUMMARY=================
Mask distribution:
1 routes at length 32
Route origin distribution:

6 Static 3 Direct

Display a Route Manager configuration using the following command:
show configuration rtmgr
Sample output:
#
# Module rtmgr configuration.
#
disable iproute sharing

configure iproute priority mpls 20

………
disable icmp timestamp vlan "to62"
enable ip-option loose-source-route
enable iproute compression ipv4 vr "VR-Default"
Event Log Messages

Event log messages are given in the following circumstances:
● When compression/uncompression start and end.
[ Severity level: Debug -Summary ]
● During each chunking start and end
[ Severity level: Debug -Verbose ]
● When a route is compressed/uncompressed.
[ Severity level: Debug -Verbose ]
Exceptional Scenarios
This section explains instances of exceptional route compression behavior.
● When a node does not have any best route
Consider the routing table shown in Table 101. When a node does not have any best route, children
are uncompressed, if they were already compressed. Also this node will be uncompressed, if it had
previously been compressed.
Table 101: Route Manager’s table when there is no best route for a node
Prefix Gateway Number of best paths Compressed?
192.0.0.0/8 10.203.174.68 1 No
192.168.0.0/16 10.203.174.68 0 No
192.168.224.0/24 10.203.174.68 1 No
192.168.225.0/24 10.203.174.68 1 No
● When a node contains only a multicast route

Route compression is applied to unicast routes only. If a node contains only a multicast route, the
compression algorithm is not applied to the node. Therefore multicast nodes are considered as nodes
with no best unicast routes as shown in Table 102.
Table 102: Route Manager’s table when a node contains only a multicast route
Prefix Gateway Unicast/Multicast Compressed?
192.0.0.0/8 10.203.174.68 Unicast Route No
192.168.0.0/16 10.203.174.68 Multicast Route No
192.168.224.0/24 10.203.174.68 Unicast Route No
192.168.225.0/24 10.203.174.68 Unicast Route No

ECMP Handling When IP Route Sharing Is Enabled

The nodes that have ECMP paths are compressed only if the following conditions are met; otherwise,
potential sub-optimal forwarding will occur:
● The number of ECMP paths for a given node must match the number of ECMP paths in its parent
node.
● A given node’s set of gateways must match its parent’s set of gateways.
Table 103 shows how compression is applied for the nodes with ECMP paths when IP route sharing is
enabled. Sample routes with ECMP paths are taken for illustration. The Reason field in the table
provides information about why the compression is applied or not applied for the node.
Table 103: Route Manager’s table when IP route sharing is enabled

Prefix Gateways Compressed? Reason
20.0.0.0/8 Gw1: 30.1.10.1, NO This is the top node.
Gw2: 50.1.10.1
20.1.10.0/24 Gw1: 30.1.10.1 NO Number of gateways did not match. This node has
only one gateway, while the parent node has two.
20.2.10.0/24 Gw1: 30.1.10.1, NO Number of gateways match. But one of the ECMP
Gw2: 60.1.10.1 paths (gateway 60.1.10.1) does not match with its
parent’s ECMP paths.
20.3.10.0/24 Gw1: 30.1.10.1, YES Number of gateways matches with its parent. Also all
Gw2: 50.1.10.1 the gateways match with parent.
20.4.10.0/24 Gw1: 30.1.10.1, NO Number of gateways does not match with its parent.
Gw2: 50.1.10.1
Gw3: 60.1.10.1
20.1.10.44/32 Gw1: 30.1.10.1 NO Number of gateways did not match. [This node has
Gw2: 50.1.10.1 ECMP paths. But the parent node 20.1.10.0 does
not have ECMP path.]
Table 104 shows only uncompressed routes.
Table 104: HAL(TCAM)/Kernel routing table when IP route sharing is enabled

Prefix Gateway
20.0.0.8/16 Gw1: 30.1.10.1,
Gw2: 50.1.10.1
20.1.10.0/24 Gw1: 30.1.10.1
20.2.10.0/24 Gw1: 30.1.10.1,
Gw2: 60.1.10.1
20.4.10.0/24 Gw1: 30.1.10.1,
Gw2: 50.1.10.1
Gw3: 60.1.10.1
20.1.10.44/32 Gw1: 30.1.10.1
Gw2: 50.1.10.1
Sample output is shown below:
* (debug) BD-12804.9 # enable iproute sharing

* (debug) BD-12804.10 # show iproute
#s Default Route 12.1.10.10 1 UG---S-um--- v1 0d:20h:1m:3s


#d 12.1.10.0/24 12.1.10.62 1 U------um--- v1 0d:20h:1m:3s
d 16.1.10.0/24 16.1.10.62 1 -------um--- v16 0d:20h:1m:4s
#d 22.1.10.0/24 22.1.10.62 1 U------um--- v2 0d:20h:1m:4s
#s 33.33.33.0/24 12.1.10.25 1 UG---S-um--- v1 0d:20h:1m:3s
#s 55.0.0.0/8 12.1.10.10 1 UG---S-um--- v1 0d:20h:1m:3s
#s 55.0.0.0/8 22.1.10.33 1 UG---S-um--- v2 0d:20h:1m:3s
#s 55.2.1.1/32 12.1.10.22 1 UG---S-um--- v1 0d:20h:1m:3s
#s 55.5.5.1/32 12.1.10.44 1 UG---S-um--- v1 0d:20h:1m:3s
#s 66.0.0.0/8 12.1.10.12 1 UG---S-um--- v1 0d:20h:1m:3s
#s 66.0.0.0/16 12.1.10.12 1 UG---S-um--c v1 0d:20h:1m:3s
#d 70.1.10.0/24 70.1.10.62 1 U------um--- v7 0d:20h:1m:4s
#s 78.0.0.0/8 12.1.10.10 1 UG---S-um--- v1 0d:20h:1m:3s
#s 79.0.0.0/8 12.1.10.10 1 UG---S-um--c v1 0d:20h:1m:3s
#s 79.0.0.0/8 12.1.10.12 1 UG---S-um--c v1 0d:20h:1m:3s
#s 80.0.0.0/8 12.1.10.10 1 UG---S-um--- v1 0d:20h:1m:3s
#d 80.1.10.0/24 80.1.10.62 1 U------um--- v8 0d:20h:1m:4s
#s 81.0.0.0/8 12.1.10.10 1 UG---S-um--- v1 0d:20h:1m:3s
#s 81.0.0.0/8 12.1.10.12 1 UG---S-um--- v1 0d:20h:1m:3s
#s 81.0.0.0/8 12.1.10.13 1 UG---S-um--- v1 0d:20h:1m:3s
#s 82.0.0.0/8 12.1.10.10 1 UG---S-um--- v1 0d:20h:1m:3s
#s 83.0.0.0/8 12.1.10.10 1 UG---S-um--- v1 0d:20h:1m:3s
#d 91.1.10.0/24 91.1.10.62 1 U------um--- v9 0d:20h:1m:4s
#d 92.1.10.0/24 92.1.10.62 1 U------um--- v10 0d:20h:1m:6s
#d 93.1.10.0/24 93.1.10.62 1 U------um--- v11 0d:20h:1m:6s
(mb) MBGP, (mbe) MBGPExt, (mbi) MBGPInter, (mp) MPLS Lsp
(L) Matching LDP LSP, (l) Calculated LDP LSP, (m) Multicast
(P) LPM-routing, (R) Modified, (S) Static, (s) Static LSP
(T) Matching RSVP-TE LSP, (t) Calculated RSVP-TE LSP, (u) Unicast, (U) Up
(c) Compressed Route
Mask distribution:
2 default routes 12 routes at length 8

8 routes from Direct 18 routes from Static


ECMP Handling When IP Route Sharing Is Disabled

If IP route sharing is disabled, the first best path is installed in the hardware table, if multiple best paths
are available. Hence the compression algorithm considers the first best path for ECMP cases. As shown
in the Route Manager Table in Table 105, when IP route sharing is disabled, all routes are compressed,
except the first one in this case.
Table 105: Route Manager’s table when IP route sharing is disabled

Prefix Gateways Compressed?
20.0.0.0/8 Gw1: 30.1.10.1, NO
Gw2: 50.1.10.1
20.1.10.0/24 Gw1: 30.1.10.1 YES
20.2.10.0/24 Gw1: 30.1.10.1, YES
Gw2: 60.1.10.1
20.3.10.0/24 Gw1: 30.1.10.1, YES
Gw2: 50.1.10.1
20.4.10.0/24 Gw1: 30.1.10.1, YES
Gw2: 50.1.10.1
Gw3: 60.1.10.1
20.1.10.44/32 Gw1: 30.1.10.1 YES
Gw2: 50.1.10.1
Table 106: HAL(TCAM)/Kernel routing table when IP route sharing is disabled

Prefix Gateway
20.0.0.8/16 Gw1: 30.1.10.1,
Sample output is shown below:
* (debug) BD-12804.7 # disable iproute sharing

* (debug) BD-12804.8 # show iproute
#d 12.1.10.0/24 12.1.10.62 1 U------um--- v1 0d:19h:58m:59s
d 16.1.10.0/24 16.1.10.62 1 -------um--- v16 0d:19h:58m:59s
#d 22.1.10.0/24 22.1.10.62 1 U------um--- v2 0d:19h:58m:59s
#s 33.33.33.0/24 12.1.10.25 1 UG---S-um--- v1 0d:19h:58m:58s
#s 55.0.0.0/8 12.1.10.10 1 UG---S-um--c v1 0d:19h:58m:58s
#s 55.0.0.0/8 22.1.10.33 1 UG---S-um--- v2 0d:19h:58m:58s
#s 55.2.1.1/32 12.1.10.22 1 UG---S-um--- v1 0d:19h:58m:58s
#s 55.5.5.1/32 12.1.10.44 1 UG---S-um--- v1 0d:19h:58m:58s
#s 66.0.0.0/8 12.1.10.12 1 UG---S-um--- v1 0d:19h:58m:58s
#s 66.0.0.0/16 12.1.10.12 1 UG---S-um--c v1 0d:19h:58m:58s
#d 70.1.10.0/24 70.1.10.62 1 U------um--- v7 0d:19h:58m:59s
#s 78.0.0.0/8 12.1.10.10 1 UG---S-um--c v1 0d:19h:58m:58s
#s 79.0.0.0/8 12.1.10.10 1 UG---S-um--c v1 0d:19h:58m:58s
#s 79.0.0.0/8 12.1.10.12 1 UG---S-um--- v1 0d:19h:58m:58s
#s 80.0.0.0/8 12.1.10.10 1 UG---S-um--c v1 0d:19h:58m:58s
#d 80.1.10.0/24 80.1.10.62 1 U------um--- v8 0d:19h:58m:59s
#s 81.0.0.0/8 12.1.10.10 1 UG---S-um--c v1 0d:19h:58m:58s
#s 81.0.0.0/8 12.1.10.12 1 UG---S-um--- v1 0d:19h:58m:58s
#s 81.0.0.0/8 12.1.10.13 1 UG---S-um--- v1 0d:19h:58m:58s

#s 82.0.0.0/8 12.1.10.10 1 UG---S-um--c v1 0d:19h:58m:58s

#s 83.0.0.0/8 12.1.10.10 1 UG---S-um--c v1 0d:19h:58m:58s
#d 91.1.10.0/24 91.1.10.62 1 U------um--- v9 0d:19h:58m:59s
#d 92.1.10.0/24 92.1.10.62 1 U------um--- v10 0d:19h:59m:2s
#d 93.1.10.0/24 93.1.10.62 1 U------um--- v11 0d:19h:59m:2s
(mb) MBGP, (mbe) MBGPExt, (mbi) MBGPInter, (mp) MPLS Lsp
(L) Matching LDP LSP, (l) Calculated LDP LSP, (m) Multicast
(P) LPM-routing, (R) Modified, (S) Static, (s) Static LSP
(T) Matching RSVP-TE LSP, (t) Calculated RSVP-TE LSP, (u) Unicast, (U) Up
(c) Compressed Route
Mask distribution:
2 default routes 12 routes at length 8

8 routes from Direct 18 routes from Static

LSP Route Handling

An LSP route is compressed only in the following circumstances:
● If the parent node of the LSP route is also an LSP route

● If the LSP nexthop of the parent node matches with this node


35 IPv6 Unicast Routing

● Overview of IPv6 Unicast Routing on page 893
● Router Interfaces on page 894
● Specifying IPv6 Addresses on page 895
● Neighbor Discovery Protocol on page 897
● Populating the Routing Table on page 899
● Configuring IP Unicast Routing on page 901
● Verifying the IP Unicast Routing Configuration on page 902
● IPv6 Forwarding Behavior on page 902
● Routing Configuration Example on page 903
● Tunnel Configuration Examples on page 905
This chapter assumes that you are already familiar with IPv6 unicast routing. If not, refer to the
following publications for additional information:
● RFC 2460—Internet Protocol, Version 6 (IPv6) Specification
● RFC 3513—Internet Protocol Version 6 (IPv6) Addressing Architecture
NOTE
For more information on interior gateway protocols, see Chapter 9, “RIPng,” or Chapter 11, “OSPFv3.”

The switch provides full Layer 3, IPv6 unicast routing. It exchanges routing information with other
routers on the network using either the IPv6 version of Routing Information Protocol (RIPng) or the
IPv6 version of Open Shortest Path First (OSPFv3) protocol. The switch dynamically builds and
maintains a routing table and determines the best path for each of its routes.
Beginning with ExtremeXOS version 12.0, support for route optimization via route compression has
been added, to reduce the number of routes to be installed in hardware. This helps with route
optimization and scaling. This feature allows you to install only less specific routes in the table when
overlapping routes with the same nexthop exist. For detailed information about route compression, see
“IP Route Compression” on page 885.
To enable this feature, use the command:
enable iproute ipv6 compression {vr <vrname>}
To disable this feature, use the command:
disable iproute ipv6 compression {vr <vrname>}

Beginning in release 11.2, ExtremeXOS can provide both IPv4 and IPv6 routing at the same time.
Separate routing tables are maintained for the two protocols. Most commands that require you to
specify an IP address can now accept either an IPv4 or IPv6 address and act accordingly. Additionally,
many of the IP configuration, enabling, and display commands have added tokens for IPv4 and IPv6 to
clarify the version required. For simplicity, existing commands affect IPv4 by default and require you to
specify IPv6, so configurations from an earlier release will still correctly configure an IPv4 network.
ACLs and routing policies also support IPv6. Use of an IPv6 address in a rule entry will automatically
use IPv6.
NOTE
IPv6 functionality is supported only on the virtual routers VR-Default and VR-Mgmt. VR-Mgmt supports only IPv6
addressing and static routing; VR-Default also supports dynamic routing.
Router Interfaces
The routing software and hardware routes IPv6 traffic between router interfaces. A router interface is
either a virtual LAN (VLAN) that has an IP address assigned to it, or, new for IPv6, a Layer 3 tunnel.
As you create VLANs and tunnels with IPv6 addresses, you can also choose to route (forward traffic)
between them. Both the VLAN switching and IP routing function occur within the switch.
An interface can have up to 255 IPv6 addresses, with at least one being a link local address. IPv4 and
IPv6 interfaces can coexist on the same VLAN, allowing both IPv4 and IPv6 networks to coexist on the
same Layer 2 broadcast domain.
NOTE
Each IP address and mask assigned to a VLAN must represent a unique IP subnet. You cannot configure the same
IP address and subnet on different VLANs within the same virtual router.
Tunnels
Layer 3 tunnels are a transition method, as networks change over from IPv4 to IPv6. ExtremeXOS
supports the use of IPv6-in-IPv4 tunnels (known as configured tunnels or 6in4 tunnels) and IPv6-to-
IPv4 tunnels (known as 6to4 tunnels). Both types of tunnels are used to connect regions of IPv6 routing
across a region of IPv4 routing. From the perspective of the router, the tunnel across the IPv4 region is
one hop, even if multiple IPv4 routers are traversed during transport.
A 6in4 tunnel connects one IPv6 region to one other IPv6 region. Multiple 6in4 tunnels can be
configured on a single router to connect with multiple IPv6 regions. Dynamic and static routing can be
configured across a 6in4 tunnel. Hosts in the IPv6 regions need not know anything about the configured
tunnel, since packets destined for remote regions are sent over the tunnel like any other type of routing
interface. The IPv6-in-IPv4 tunnel is created using the following command:
create tunnel <tunnel_name> ipv6-in-ipv4 destination <destination-address> source
<source-address>
where the source address refers to an existing address in the switch, and the destination address is a
remote destination accessible from the switch. A maximum of 255 IPv6-in-IPv4 tunnels can be
configured.

A 6to4 tunnel connects one IPv6 region with multiple IPv6 regions. Only one 6to4 tunnel can be
configured on a single router. A 6to4 tunnel is created using the following command
create tunnel <tunnel_name> 6to4 source <source-address>
where the source-address is an existing address in the switch.
Hosts in the IPv6 regions connected by 6to4 tunnels must be configured with 2002::/16 6to4 addresses
in order for traffic to flow. ExtremeXOS does not support 6to4 relay router functionality.
To delete a tunnel, use the following command:
delete tunnel <tunnel_name>
NOTE
The commands to create a tunnel are described above. Additional commands are described later in this chapter,
along with the instructions for defining the IPv6 configuration for VLANs.
To configure and unconfigure IPv6 addresses for the tunnels, use the following commands:
configure tunnel <tunnel_name> ipaddress [{eui64} [<ipv6_address_mask> | ipv6-link-

local]
unconfigure tunnel <tunnel_name> ipaddress <ipv6_address_mask>
To display tunnel information, use the following command:
show tunnel
Specifying IPv6 Addresses

IPv6 Addresses are 128 bits (16 bytes) when compared to the 32 bit IPv4 addresses. The ExtremeXOS
CLI accepts two standard representations for IPv6 addresses, as described in RFC 3513, section 2.2,
items 1, 2, and 3.
For example, the 128 bits of the address are represented by eight, four-digit hexadecimal numbers
separated by colons:
2000:af13:ee10:34c5:800:9192:ba89:2311
3f11:5655:2300:304:0000:0000:7899:acde
Leading zeros in a four-digit group can be omitted. There is a special use of a double colon (::) in an
address. The double colon stands for one or more groups of 16 bits of zeros and can only be used once
in an address. For example, the following addresses:
fe80:0:0:0:af34:2345:4afe:0
fe80:0:0:111:0:0:0:fe11
3c12:0:0:0:0:89:ff:3415
can be represented as:
fe80::af34:2345:4afe:0
fe80:0:0:111::fe11
3c12::89:ff:3415

Additionally, you can specify an address in a mixed IPv4/IPv6 mode that uses six, four-digit
hexadecimal numbers for the highest-order part of the address, and uses the IPv4 dotted decimal
representation for the lowest-order remaining portion. For example:
0:0:0:0:0:0:192.168.1.1
0:0:0:0:0:ffff:10.0.14.254
These can be represented as:
::192.168.1.1
::ffff:10.0.14.254
Both Global and Link-local IP addresses can be configured on a VLAN or tunnel interface, using the
following commands:
configure vlan <vlan_name> ipaddress [<ipaddress> {<ipNetmask>} | ipv6-link-local |

{eui64} <ipv6_address_mask>]
configure tunnel <tunnel_name> ipaddress [{eui64} [<ipv6_address_mask> | ipv6-link-
local]
where <ipaddress> refers to the address specified in the above format.
The IPv6 address configuration can be verified using the following commands:
router> | <vlan_name> stpd | security}show ipconfig ipv6 {vlan <vlan_name> | tunnel
<tunnelname>}
show ipconfig ipv6 {vlan <vlan_name> | tunnel <tunnelname>}
Duplicate Address Detection

When you configure an active interface with an IPv6 address, the interface must send out an
advertisement containing its address. All other interfaces on the subnet have the opportunity to respond
to the newly configured interface, and inform it that the address is a duplicate. Only after this process
occurs, can the interface use the newly configured address. If the interface receives a message that the
newly configured address is a duplicate, it cannot use the address.
Until the Duplicate Address Detection (DAD) process completes, the new address is considered
tentative, and will be shown as such in any display output. If the address is a duplicate, it will also be
labeled as such, and must be reconfigured. On an active interface, the DAD process should occur so
quickly that you would not see the address labeled as tentative. However, if you are configuring an
interface before enabling it, and you display the configuration, you will see that the address is currently
tentative. As soon as you enable the interface, the address should be ready to use, or labeled as
duplicate and must be reconfigured.
See RFC 2462, IPv6 Stateless Address Autoconfiguration, for more details.
Scoped Addresses
IPv6 uses a category of addresses called link-local addresses that are used solely on a local subnet.
Every IPv6 VLAN must have at least one link-local address. If a global IP address is configured on a
VLAN that has no link-local address, one is assigned automatically by the switch. The link-local
addresses start with the prefix fe80::/64. As a result, a switch can have the same link local address
assigned to different VLANs, or different neighbors on different links may be using the same link local

address. Because of this, there are cases where you need to specify an address and a VLAN/tunnel to
indicate which interface to use. For those cases, you can indicate the interface by using a scoped
address. To scope the address, append the VLAN/tunnel name to the end of the address, separated by
a percent sign (%). For example, to indicate the link local address fe80::2 on VLAN finance, use the
following form:
fe80::2%finance
Scoped addresses also appear in the outputs of display commands.
IPv6 Addresses Used in Examples

For the purposes of documentation, we follow RFC 3849, which indicates that the prefix 2001:db8::/32
can be used as a global unicast address prefix and will not be assigned to any end party.
Neighbor Discovery Protocol

The Neighbor Discovery Protocol, as defined in RFC 2461, defines mechanisms for the following
functions:
● Resolving link-layer addresses of the IPv6 nodes residing on the link.
● Locating routers residing on the attached link.
● Locating the address prefixes that are located on the attached link.
● Learn link parameters such as the link MTU, or Internet parameters such as the hop limit value that
has to be used in the outgoing packets.
● Automatic configuration of the IPv6 address for an interface.
● Detect whether the address that it wants to use is not already in use by another node, also known as
Duplicate Address Detection (DAD).
● Redirect the traffic to reach a particular destination through a better first-hop.
MAC Address Resolution

In IPv4, MAC address resolution is done by ARP. For IPv6, this functionality is handled by the
Neighbor Discovery Protocol. The router maintains a cache of IPv6 addresses and their corresponding
MAC addresses and allows the system to respond to requests from other nodes for the MAC address of
the IPv6 Addresses configured on the interfaces. You can also statically configure the MAC Address of
IPv6 destinations on the attached links using the following commands:
configure neighbor-discovery cache {vr <vr_name>} add [<ipv6address> |
<scoped_link_local>] <mac>
configure neighbor-discovery cache delete {vr <vr_name>} [<ipv6address> |
<scoped_link_local>]
Both statically configured and dynamic neighbor-discovery entries can be viewed using the following
command:
show neighbor-discovery cache ipv6 {{[<ipv6_addr> | <mac> | permanent] {vr <vr_name>}}

| vlan <vlan_name> | vr <vr_name>}

The neighbor-discovery entries that are learned dynamically can be cleared using the following
command:
clear neighbor-discovery cache ipv6 {<ipv6address> {vr <vr_name>} | vlan <vlan_name> |

vr <vr_name>}
Static neighbor discovery entries are never deleted, and are retained across system reboots.
Router Discovery
Also supported is router discovery—the ability to send out router advertisements that can be used by a
host to discover the router. The advertisements sent out contain the prefixes and configuration
parameters that allow the end nodes to auto-configure their addresses. The switch also responds to
requests from nodes for router advertisements.
The following settings can be configured on an interface to manage router advertisements:

● Settings to control the sending of router advertisements over the interface periodically and to control
responding to router solicitations
● The maximum time between sending unsolicited router advertisements
● The minimum time between sending unsolicited router advertisements
You can configure the following values, that are advertised by the switch:
● Managed Address Configuration Flag
● Other Stateful Configuration Flag
● Link MTU
● Retransmit Timer
● Current Hop Limit
● Default Lifetime
● Reachable Time
Additionally, you can configure the following values for each prefix on the prefix list associated with an
interface:
● Valid Lifetime of the Prefix
● On-Link Flag
● Preferred Lifetime of the Prefix
● Autonomous Flag
To enable router discovery on a VLAN, use the following command:
enable router-discovery {ipv6} vlan <vlan_name>
To configure the prefixes advertised by router discovery, use the following command to add prefixes:
configure vlan <vlan_name> router-discovery {ipv6} add prefix <prefix>
To display router discover settings, use the following command:
show router-discovery {ipv6} {vlan <vlan_name>}

NOTE
Unlike ExtremeWare, ExtremeXOS does not support host processing of neighbor router advertisements.
Populating the Routing Table

The switch maintains an IP routing table for both network routes and host routes. The table is
populated from the following sources:
● Dynamically, by way of routing protocol packets or by Internet Control Message Protocol (ICMP)
redirects exchanged with other routers
● Statically, by way of routes entered by the administrator:
■ Default routes, configured by the administrator
■ Locally, by way of interface addresses assigned to the system
■ By other static routes, as configured by the administrator
Once routes are populated using the above method, IPv6 forwarding needs to be enabled on the VLAN
enable ipforwarding ipv6 {vlan <vlan_name> | tunnel <tunnel-name> | vr <vr_name>}
NOTE
If you define a default route and subsequently delete the VLAN on the subnet associated with the default route, the
invalid default route entry remains. You must manually delete the configured default route.
Dynamic Routes
Dynamic routes are typically learned by way of RIPng or OSPFv3. Routers that use RIPng or OSPFv3
exchange information in their routing tables in the form of advertisements. Using dynamic routes, the
routing table contains only networks that are reachable.
Dynamic routes are aged out of the table when an update for the network is not received for a period of
time, as determined by the routing protocol. For details on configuration and behavior of dynamic
routes, please refer to the specific Chapters on RIPng and OSPFv3 in the Concepts Guide.
Static Routes
Static routes are manually entered into the routing table. Static routes are used to reach networks not
advertised by routers. Static IPv6 routes can be created using the following command:
configure iproute add <ipv6Netmask> [<ipv6Gateway> | <ipv6ScopedGateway>] {<metric>}
{vr <vrname>} {multicast-only | unicast-only}
Similar to IPv4 Unicast, IPv6 default and blackhole routes can also be configured. The IPv6 gateway can
be a global address or a scoped link-local address of an adjacent router.

Static routes can also be used for security reasons, to control which routes you want advertised by the
router. You configure, if you want all static routes to be advertised, using one of the following
commands:
enable ripng export [direct | ospfv3 | ospfv3-extern1 | ospfv3-extern2 | ospfv3-
inter | ospfv3-intra | static] [cost <number> {tag <number>} | policy <policy-name>]
or
disable ripng export [direct | ospfv3 | ospfv3-extern1 | ospfv3-extern2 | ospfv3-
inter | ospfv3-intra | static]
● enable ospfv3 {domain <domainName>} export [direct | ripng | static] [cost <cost>
type [ase-type-1 | ase-type-2] | <policy-map>]
or
disable ospfv3 {domain <domainName>} export [direct | ripng | static]
The default setting is disabled. Static routes are never aged out of the routing table.
A static route’s nexthop (gateway) must be associated with a valid IP subnet. An IP subnet is associated
with a single VLAN by its IP address and subnet mask. If the VLAN is subsequently deleted, the static
route entries using that subnet must be deleted manually.
The IPv6 routes can be viewed using the following command:
show iproute ipv6 {priority | vlan <vlan_name> | tunnel <tunnel-name> |

<ipv6Netmask> | summary {multicast | unicast}} {vr <vrname>}}
To view the IPv6 routes based on the type of the route, use the command:
show iproute ipv6 origin [direct | static | blackhole | ripng | ospfv3 | ospfv3-
intra | ospv3-inter | ospfv3-extern1 | ospfv3-extern2] {vr <vrname>}
Multiple Routes
When there are multiple, conflicting choices of a route to a particular destination, the router picks the
route with the longest matching network mask. If these are still equal, the router picks the route using
the following default criteria (in the order specified):
● Directly attached network interfaces
● Static routes
● ICMP redirects
● Dynamic routes
● Directly attached network interfaces that are not active.
NOTE
If you define multiple default routes, the route that has the lowest metric is used. If multiple default routes have the
same lowest metric, the system picks one of the routes.
You can also configure blackhole routes—traffic to these destinations is silently dropped.
The criteria for choosing from multiple routes with the longest matching network mask is set by
choosing the relative route priorities.

Configuring IP Unicast Routing
Relative Route Priorities

Table 107 lists the relative priorities assigned to routes depending on the learned source of the route.
NOTE
Although these priorities can be changed, do not attempt any manipulation unless you are expertly familiar with the
possible consequences.
Table 107: Relative route priorities

Route Origin Priority
Direct 10
BlackHole 50
Static 1100
ICMP 1200
OSPF3Intra 2200
OSPF3Inter 2300
RIPng 2400
OSPFv3 ASExt 3100
OSPFv3 Extern1 3200
OSPFv3 Extern2 3300
To change the relative route priority, use the following command:
configure iproute ipv6 priority [ripng | blackhole | icmp | static | ospfv3-intra |

ospfv3-inter | ospfv3-as-external | ospfv3-extern1 | ospfv3-extern2] <priority>
Configuring IP Unicast Routing

This section describes the commands associated with configuring IP unicast routing on the switch. To
configure routing:
1 Create and configure two or more VLANs.
2 Assign each VLAN that will be using routing an IP address using the following command:
Ensure that each VLAN has a unique IP address.
3 Configure a static route using the following command:
configure iproute add <ipv6Netmask> [<ipv6Gateway> | <ipv6ScopedGateway>]
{<metric>} {vr <vrname>} {multicast-only | unicast-only}
or
Configure a default route using the following command:
configure iproute add default [<ipv6Gateway> | <ipv6ScopedGateway>] {metric} {vr
<vrname>} {multicast-only | unicast-only}
Default routes are used when the router has no other dynamic or static route to the requested
destination.

4 Turn on IP routing for one or all VLANs using the following command:
enable ipforwarding ipv6 {vlan <vlan_name> | tunnel <tunnel-name> | vr <vr_name>}
5 Configure the routing protocol, if required. For a simple network using RIPng, the default
configuration may be acceptable.
6 Turn on RIPng or OSPFv3 using one of the following commands:
enable ripng
enable ospfv3 {domain <domainName>}
Verifying the IP Unicast Routing Configuration

The show ipstats ipv6 command displays the currently configured routes and includes how each
route was learned.
Additional verification commands include:

● show neighbor-discovery cache ipv6—Displays the neighbor discovery cache of the system.
● show ipconfig ipv6—Displays configuration information for one or more VLANs.
● show ipstats ipv6—Displays the IPv6 statistics for the switch or for the specified VLANs.
IPv6 Forwarding Behavior

Table 108 shows the Extreme switch platforms and the ExtremeXOS software versions that provide
hardware forwarding support for IPv6 unicast features.
Table 108: IPv6 Unicast Features

Switch Model ExtremeXOS Release Features
BlackDiamond 10808 ExtremeXOS 11.2 and later IPv6 Unicast forwarding
IPv6 tunneling
BlackDiamond 12000 series ExtremeXOS 11.4 and later IPv6 Unicast forwarding
IPv6 tunneling
Summit X450 a- and e- series ExtremeXOS 11.5 and later IPv6 Unicast forwarding
Summit X450 a- and e- series ExtremeXOS 12.0 and later IPv6 tunneling
BlackDiamond 8800 a- and e- series ExtremeXOS 11.6 and later IPv6 Unicast forwarding
BlackDiamond 8800 a- and e- series ExtremeXOS 12.0 and later IPv6 tunneling
BlackDiamond 12802 ExtremeXOS 12.0 and later IPv6 Unicast forwarding
IPv6 tunneling
Summit X250e ExtremeXOS 12.0 and later IPv6 Unicast forwarding
IPv6 tunneling
Hardware IPv6 Unicast Forwarding Support

With IPv6 routing enabled, traffic ingressing a port in any of the above modules is forwarded in
hardware. Any IP packets not matching a route or host in the hardware will be sent to the CPU for
software processing. For a route with a mask up to 64 bits, traffic is forwarded in hardware.

The port statistics indicate that traffic is forwarded at line-rate. The counters displayed in the show
ipstats ipv6 command output do not increment in this case.
Traffic ingressing on any older hardware modules will be routed via the routing tables in software. The
IPv6 packet forwarding statistics can be viewed from the show ipstats ipv6 command.
Hardware Tunnel Support

The platforms and software versions that provide hardware forwarding support for IPv6 traffic for both
IPv6-in-IPv4 and 6to4 tunnels for various platforms are shown in Table 108. In all the other platforms,
tunnel traffic is forwarded in software and the statistics can be viewed from the show ipstats ipv6
command.
NOTE
The MTU for IPv6 Tunnels is set to 1480 bytes, and is not user-configurable. Configuring jumbo frames has no
effect on the MTU size for IPv6 tunnels.

● Finance
■ Protocol-sensitive VLAN using IPv6 protocol
■ IP address 2001:db8:35::1/48.
● Personnel
■ Protocol-sensitive VLAN using the IPv6 protocol.
■ IP address 2001:db8:36::1/48.
● MyCompany

Figure 132: IPv6 Unicast routing configuration example
1 2 3 4 A B 5 6 7 8
2001:db8:35::1/48 2001:db8:36::1/48
MyCompany
2001:db8:35::/48 2001:db8:36::/48
Finance Personnel
1 2 3 4
IPv6 IPv6 = IPv6 traffic

NetBIOS NetBIOS = NetBIOS traffic
IPv6 IPv6
NetBIOS NetBIOS EX_106
The stations connected to the system generate a combination of IPv6 traffic and NetBIOS traffic. The
IPv6 traffic is filtered by the protocol-sensitive VLANs. All other traffic is directed to the VLAN
MyCompany.
In this configuration, all IPv6 traffic from stations connected to slots 1 and 3 have access to the router by
The example is configured as follows:
create vlan Finance

configure Finance protocol ipv6

configure Personnel protocol ipv6

configure Finance ipaddress 2001:db8:35::1/48

configure Personnel ipaddress 2001:db8:36::1/48
configure ripng add vlan Finance
configure ripng add vlan Personnel

Tunnel Configuration Examples
enable ipforwarding ipv6

enable ripng
6in4 Tunnel Configuration Example

Figure 133 illustrates a 6in4 tunnel configured between two IPv6 regions across an IPv4 region.
Figure 133: 6in4 Tunnel Example
Host A Router A
2001:db8:1::101/64
2001:db8:1::1/64
2
1 2001:db8:a::1/64
IPv6
192.168.1.1/24
IPv4
Router B
IPv6
2001:db8:a::2/64 1
10.2.0.1/24
2001:db8:2::1/64 2
Host B
2001:db8:2::101/64
EX_108
In Figure 133, Router A has an interface to an IPv4 region with the address 192.168.1.1 (for this example
we are using private IPv4 addresses, but to tunnel across the Internet, you would use a public address).
Router B has an IPv4 interface of 10.2.0.1. The IPv4 interface must be created before the tunnel is
configured and cannot be deleted until the tunnel is deleted.
This example has one subnet in each IPv6 region, 2001:db8:1::/64 for Router A and 2001:db8:2::/64 for
Router B. Hosts A and B are configured to use IPv6 addresses 2001:db8:1::101 and 2001:db8:2::101
respectively.

For traffic to move from one region to the other, there must be a route. In this example, a static route is
created, but you could enable RIPng or OSPFv3 on the tunnel interface.
In this example, we assume that the IPv4 network can route from Router A to Router B (in other words,
some IPv4 routing protocol is running on the public-ipv4 interfaces). For platforms on which hardware
based tunneling is supported (See Table 108 on page 902), IPv4 forwarding needs to be enabled on the
tunnel source VLAN. However, in platforms on which IPv6-in-IPv4 tunnels are supported in software
only, you do not need to enable IPv4 forwarding on the public interfaces in this example unless you are
also routing IPv4 traffic on them (in this example, it is assumed you are running no IPv4 traffic inside
your respective IPv6 networks, although you could).
When Host A needs to send a packet to 2001:db8:2::101 (Host B), it forwards it to Router A. Router A
receives an IPv6 packet from the IPv6 source address 2001:db8:1::101 to the destination 2001:db8:2::101.
Router A has the static route, for the route 2001:db8:2::/64 with next hop 2001:db8:a::2 (Router B)
through the tunnel interface. So Router A encapsulates the IPv6 packet inside an IPv4 header with the
source address 192.168.1.1 and destination address 10.2.0.1. The encapsulated IPv6 packet passes
through the IPv4 network and reaches the other end of the tunnel - Router B. Router B decapsulates the
packet and removes the IPv4 header. Router B then forwards the IPv6 packet to the destination host -
Host B.
NOTE
Each IPv6 packet is encapsulated inside an IPv4 header (20 bytes) before it is forwarded via a IPv6-in-IPv4 tunnel.
For example, a 66-byte packet from Host A will be encapsulated and forwarded as a 86-byte packet by the
Router A.
Router A
configure vlan default delete port all
create vlan public-ipv4
configure vlan public-ipv4 add port 1 untagged
configure vlan public-ipv4 ipaddress 192.168.1.1/24
create tunnel public6in4 ipv6-in-ipv4 destination 10.2.0.1 source 192.168.1.1
configure tunnel public6in4 ipaddress 2001:db8:a::1/64
enable ipforwarding ipv6 public6in4
create vlan private-ipv6
configure vlan private-ipv6 add port 2 untagged
configure vlan private-ipv6 ipaddress 2001:db8:1::1/64
enable ipforwarding ipv6 private-ipv6
configure iproute add 2001:db8:2::/64 2001:db8:a::2
enable ipforwarding public-ipv4
Router B
create tunnel public6in4 ipv6-in-ipv4 destination 192.168.1.1 source 10.2.0.1
configure tunnel public6in4 ipaddress 2001:db8:a::2/64
enable ipforwarding ipv6 public6in4

configure vlan private-ipv6 ipaddress 2001:db8:2::1/64

configure iproute add 2001:db8:1::/64 2001:db8:a::1
enable ipforwarding public-ipv4
6to4 Tunnel Configuration Example

Figure 134 illustrates a 6to4 tunnel configured between two IPv6 regions across an IPv4 region.
Figure 134: 6to4 Tunnel Configuration Example
Host 1 Router 1
2002:c0a8:101::204:96ff:fe1f:a52a/48 2002:c0a8:101::2/48
2
1 2002:c0a8:101::1/16
IPv6 192.168.1.1/24
IPv4
Router 2
IPv6
2002:a00:1::1/16 1
10.0.0.1/24 Host 2
2002:a00:1:1::1/64
2
2002:a00:1:1:204:96ff:fe1f:a432/64
2002:a00:1:2::1/64
3
Host 3
2002:a00:1:2:201:30ff:fe00:c200/64
EX_109
In Figure 134, Router 1 has an interface to an IPv4 region with the address 192.168.1.1 (for this example
we are using private IPv4 addresses, but to tunnel across the Internet, you would use a public address).
Router 2 has an IPv4 interface of 10.0.0.1. The IPv4 interface must be created before the tunnel is
configured and cannot be deleted until the tunnel is deleted.
The IPv6 endpoints of 6to4 tunnels must follow the standard 6to4 address requirement. The address
must be of the form 2002:<IPv4_source_endpoint>::/16, where <IPv4_source_endpoint> is replaced by
the IPv4 source address of the endpoint, in hexadecimal, colon separated form. For example, for a
tunnel endpoint located at IPv4 address 10.20.30.40, the tunnel address would be 2002:0a14:1e28::/16. In
hex, 10 is 0a, 20 is 14, 30 is 1e and 40 is 28.

This example shows a simple setup on the Router 1 side (one big /48 IPv6 routing domain with no
subnets), and a slightly more complex setup on the Router 2 side (two subnets :0001: and :0002: that are
/64 in length). Hosts 1, 2, and 3 will communicate using their global 2002: addresses.
The hosts in this example configure themselves using the EUI64 interface identifier derived from their
MAC addresses. Refer to your host OS vendor’s documentation for configuring IPv6 addresses and
routes.
In this example, we assume that the IPv4 network can route from Router 1 to Router 2 (in other words,
some IPv4 routing protocol is running on the public-ipv4 interfaces). However, you do not need to
enable IPv4 forwarding on the public interfaces in this example unless you are also routing IPv4 traffic
on them (in this example, it is assumed you are running no IPv4 traffic inside your respective IPv6
networks, although you could).
Router 1
create tunnel public6to4 6to4 source 192.168.1.1
configure tunnel public6to4 ipaddress 2002:c0a8:0101::1/16
enable ipforwarding ipv6 public6to4
configure vlan private-ipv6 ipaddress 2002:c0a8:0101::2/48
Router 2
create tunnel public6to4 6to4 source 10.0.0.1
configure tunnel public6to4 ipaddress 2002:0a00:0001::1/16
enable ipforwarding ipv6 public6to4
create vlan private-ipv6-sub1
configure vlan private-ipv6-sub1 add port 2 untagged
configure vlan private-ipv6-sub1 ipaddress 2002:0a00:0001:0001::1/64
enable ipforwarding ipv6 private-ipv6-sub1
create vlan private-ipv6-sub2
configure vlan private-ipv6-sub2 add port 3 untagged
configure vlan private-ipv6-sub2 ipaddress 2002:0a00:0001:0002::1/64
enable ipforwarding ipv6 private-ipv6-sub2
Host Configurations
The IPv6 addresses of these hosts are based on their MAC address-derived EUI64 interface identifiers
and the address prefixes for their subnets. Each host must also have a static route configured on it for
6to4 addresses.

Host 1:
● MAC address—00:04:96:1F:A5:2A
● IPv6 address—2002:c0a8:0101::0204:96ff:fe1f:a52a/48
● Static route—destination 2002::/16, gateway 2002:c0a8:0101::2
Host 2:
● MAC address—00:04:96:1F:A4:32
● IP address—2002:0a00:0001:0001:0204:96ff:fe1f:a432/64
● Static route—destination 2002::/16, gateway 2002:0a00:0001:0001::1
Host 3:
● MAC address—00:01:30:00:C2:00
● IP address—2002:0a00:0001:0002:0201:30ff:fe00:c200/64
● Static route—destination 2002::/16, gateway 2002:0a00:0001:0002::1


36 RIP

● Overview of RIP on page 912
● Route Redistribution on page 914
● RIP Configuration Example on page 915
● RFC 1058—Routing Information Protocol (RIP)
● RFC 1723—RIP Version 2
● Interconnections: Bridges and Routers
by Radia Perlman
ISBN 0-201-56332-0
Published by Addison-Wesley Publishing Company
Overview
The switch supports the use of two interior gateway protocols (IGPs); the Routing Information Protocol
(RIP), and the Open Shortest Path First (OSPF) protocol.
RIP is a distance-vector protocol, based on the Bellman-Ford (or distance-vector) algorithm. The
distance-vector algorithm has been in use for many years and is widely deployed and understood.
OSPF is a link-state protocol, based on the Dijkstra link-state algorithm. OSPF is a newer IGP and solves
a number of problems associated with using RIP on today’s complex networks.
NOTE
RIP and OSPF can be enabled on a single VLAN.
RIP is described in this chapter, and OSPF is described in Chapter 10, “OSPF.”
RIP Versus OSPF

The distinction between RIP and OSPF lies in the fundamental differences between distance-vector
protocols and link-state protocols. Using a distance-vector protocol, each router creates a unique routing
table from summarized information obtained from neighboring routers. Using a link-state protocol,
every router maintains an identical routing table created from information obtained from all routers in
the autonomous system (AS). Each router builds a shortest path tree, using itself as the root. The link-
state protocol ensures that updates sent to neighboring routers are acknowledged by the neighbors,
verifying that all routers have a consistent network map.

RIP
Advantages of RIP and OSPF

The biggest advantage of using RIP is that it is relatively simple to understand and to implement, and it
has been the de facto routing standard for many years.
RIP has a number of limitations that can cause problems in large networks, including the following:
● A limit of 15 hops between the source and destination networks
● A large amount of bandwidth taken up by periodic broadcasts of the entire routing table
● Slow convergence
● Routing decisions based on hop count; no concept of link costs or delay
● Flat networks; no concept of areas or boundaries
OSPF offers many advantages over RIP, including the following:

● No limitation on hop count
● Route updates multicast only when changes occur
● Faster convergence
● Support for load balancing to multiple routers based on the actual cost of the link
● Support for hierarchical topologies where the network is divided into areas
The details of RIP are explained later in this chapter.
Overview of RIP
RIP is an IGP first used in computer routing in the Advanced Research Projects Agency Network
(ARPAnet) as early as 1969. It is primarily intended for use in homogeneous networks of moderate size.
To determine the best path to a distant network, a router using RIP always selects the path that has the
least number of hops. Each router that data must traverse is considered to be one hop.
Routing Table
The routing table in a router using RIP contains an entry for every known destination network. Each
routing table entry contains the following information:
● IP address of the destination network
● Metric (hop count) to the destination network
● IP address of the next router
● Timer that tracks the amount of time since the entry was last updated
The router exchanges an update message with each neighbor every 30 seconds (default value), or when
there is a change to the overall routed topology (also called triggered updates). If a router does not
receive an update message from its neighbor within the route timeout period (180 seconds by default),
the router assumes the connection between it and its neighbor is no longer available.

Overview of RIP
Split Horizon
Split horizon is a scheme for avoiding problems caused by including routes in updates sent to the
router from which the route was learned. Split horizon omits routes learned from a neighbor in updates
sent to that neighbor.
Poison Reverse
Like split horizon, poison reverse is a scheme for eliminating the possibility of loops in the routed
topology. In this case, a router advertises a route over the same interface that supplied the route, but the
route uses a hop count of 16, which defines that router as unreachable.
Triggered Updates
Triggered updates occur whenever a router changes the metric for a route. The router is required to
send an update message immediately, even if it is not yet time for a regular update message to be sent.
This generally results in faster convergence, but may also result in more RIP-related traffic.
Route Advertisement of VLANs

Virtual LANs (VLANs) that are configured with an IP address but are configured to not route IP or are
not configured to run RIP, do not have their subnets advertised by RIP. RIP advertises only those
VLANs that are configured with an IP address, are configured to route IP, and run RIP.
RIP Version 1 Versus RIP Version 2

A new version of RIP, called RIP version 2, expands the functionality of RIP version 1 to include the
following:
● Variable-length subnet masks (VLSMs).
● Support for next-hop addresses, which allows for optimization of routes in certain environments.
● Multicasting.
RIP version 2 packets can be multicast instead of being broadcast, reducing the load on hosts that do
not support routing protocols.
NOTE
If you are using RIP with supernetting/Classless Inter-Domain Routing (CIDR), you must use RIPv2 only.

RIP
Route Redistribution
More than one routing protocol can be enabled simultaneously on the switch. Route redistribution
allows the switch to exchange routes, including static routes, between the routing protocols. Figure 135
is an example of route redistribution between an OSPF AS and a RIP AS.
Figure 135: Route redistribution
OSPF AS
Backbone Area
0.0.0.0
ABR
Area
121.2.3.4
ASBR ASBR
RIP AS
EX_046
Configuring Route Redistribution

Exporting routes from one protocol to another and from that protocol to the first one are discreet
configuration functions. For example, to run OSPF and RIP simultaneously, you must first configure
both protocols and then verify the independent operation of each. Then you can configure the routes to
export from OSPF to RIP and the routes to export from RIP to OSPF. Likewise, for any other
combinations of protocols, you must separately configure each to export routes to the other.

RIP Configuration Example
Redistributing Routes into RIP

Enable or disable the exporting of static, direct, BGP-learned, and OSPF-learned routes into the RIP
domain using the following commands:
enable rip export [bgp | direct | e-bgp | i-bgp | ospf | ospf-extern1 | ospf-extern2 |
ospf-inter | ospf-intra | static] [cost <number> {tag <number>} | policy <policy-
name>]
disable rip export [bgp | direct | e-bgp | i-bgp | ospf | ospf-extern1 | ospf-extern2
| ospf-inter | ospf-intra | static]
These commands enable or disable the exporting of static, direct, and OSPF-learned routes into the RIP
domain. You can choose which types of OSPF routes are injected, or you can simply choose ospf, which
will inject all learned OSPF routes regardless of type. The default setting is disabled.

● Finance
■ IP address 192.207.35.1.
● Personnel
■ IP address 192.207.36.1.
● MyCompany
This example does use protocol-sensitive VLANs that only admit IP packets. This is not a common
requirement for most networks. In most cases, VLANs will admit different types of packets to be
forwarded to the hosts and servers on the network.

RIP
Figure 136: RIP configuration example
1 2 3 4 A B 5 6 7 8
192.207.35.1 192.207.36.1
MyCompany
192.207.35.0 192.207.36.0
Finance Personnel
1 2 3 4
IP
IP = IP traffic
NetBIOS
NetBIOS
IP = NetBIOS traffic
IP
NetBIOS
NetBIOS EX_047
The stations connected to the system generate a combination of IP traffic and NetBIOS traffic. The IP
traffic is filtered by the protocol-sensitive VLANs. All other traffic is directed to the VLAN MyCompany.
In this configuration, all IP traffic from stations connected to slots 1 and 3 have access to the router by
More commonly, NetBIOS traffic would be allowed on the Finance and Personnel VLANs, but this
example shows how to exclude that traffic. To allow the NetBIOS traffic (or other type of traffic) along
with the IP traffic, remove the configure finance protocol ip and configure Personnel
protocol ip commands from the example.

The example in Figure 136 is configured as follows:
create vlan Finance

configure Finance protocol ip

configure Personnel protocol ip

configure Finance ipaddress 192.207.35.1

configure Personnel ipaddress 192.207.36.1
enable ipforwarding
configure rip add vlan all
enable rip

RIP

37 RIPng

● Overview of RIPng on page 920
● RIPng Configuration Example on page 922
● RFC 2080—RIPng for IPv6
Overview
Routing Information Protocol Next Generation (RIPng) is an interior gateway protocol (IGP) developed
for IPv6 networks. The analogous protocol used in IPv4 networks is called Routing Information
Protocol (RIP). Like RIP, RIPng is a relatively simple protocol for the communication of routing
information among routers. Many concepts and features of RIPng are directly parallel to those same
features in RIP.
RIPng is a distance-vector protocol, based on the Bellman-Ford (or distance-vector) algorithm. The
distance-vector algorithm has been in use for many years and is widely deployed and understood. The
other common IGP for IPv6 is OSPFv3, a link-state protocol.
NOTE
RIPng and OSPFv3 can be enabled on a single VLAN.
RIPng is described in this chapter, and OSPFv3 is described in Chapter 11, “OSPFv3.”
RIPng Versus OSPFv3

The distinction between RIPng and OSPFv3 lies in the fundamental differences between distance-vector
protocols and link-state protocols. Using a distance-vector protocol, each router creates a unique routing
table from summarized information obtained from neighboring routers. Using a link-state protocol,
every router maintains an identical routing table created from information obtained from all routers in
the autonomous system (AS). Each router builds a shortest path tree, using itself as the root. The link-
state protocol ensures that updates sent to neighboring routers are acknowledged by the neighbors,
verifying that all routers have a consistent network map.

RIPng
Advantages of RIPng and OSPFv3

The biggest advantage of using RIPng is that it is relatively simple to understand and to implement,
and it has been the de facto routing standard for many years.
RIPng has a number of limitations that can cause problems in large networks, including the following:
● A limit of 15 hops between the source and destination networks
● A large amount of bandwidth taken up by periodic broadcasts of the entire routing table
● Slow convergence
● Routing decisions based on hop count; no concept of link costs or delay
● Flat networks; no concept of areas or boundaries
OSPFv3 offers many advantages over RIPng, including the following:

● No limitation on hop count
● Route updates multicast only when changes occur
● Faster convergence
● Support for load balancing to multiple routers based on the actual cost of the link
● Support for hierarchical topologies where the network is divided into areas
The details of RIPng are explained later in this chapter.
Overview of RIPng
RIPng is primarily intended for use in homogeneous networks of moderate size.
To determine the best path to a distant network, a router using RIPng always selects the path that has
the least number of hops. Each router that data must traverse is considered to be one hop.
Routing Table
The routing table in a router using RIPng contains an entry for every known destination network. Each
routing table entry contains the following information:
● IP address and prefix length of the destination network
● Metric (hop count) to the destination network
● IP address of the next hop router, if the destination is not directly connected
● Interface for the next hop
● Timer that tracks the amount of time since the entry was last updated
● A flag that indicates if the entry is a new one since the last update
● The source of the route, for example, static, RIPng, OSPFv3, etc.
The router exchanges an update message with each neighbor every 30 seconds (default value), or when
there is a change to the overall routed topology (also called triggered updates). If a router does not
receive an update message from its neighbor within the route timeout period (180 seconds by default),
the router assumes the connection between it and its neighbor is no longer available.

Split Horizon
Split horizon is a scheme for avoiding problems caused by including routes in updates sent to the
router from which the route was learned. Split horizon omits routes learned from a neighbor in updates
sent to that neighbor.
Poison Reverse
Like split horizon, poison reverse is a scheme for eliminating the possibility of loops in the routed
topology. In this case, a router advertises a route over the same interface that supplied the route, but the
route uses a hop count of 16, which defines that router as unreachable.
Triggered Updates
Triggered updates occur whenever a router changes the metric for a route. The router is required to
send an update message immediately, even if it is not yet time for a regular update message to be sent.
This generally results in faster convergence, but may also result in more RIPng-related traffic.
Route Advertisement of VLANs

Virtual LANs (VLANs) that are configured with an IP address but are configured to not route IP or are
not configured to run RIP, do not have their subnets advertised by RIP. RIP advertises only those
VLANs that are configured with an IP address, are configured to route IP, and run RIP.
allows the switch to exchange routes, including static routes, between the routing protocols. Route
redistribution is also called route export.

configuration functions. For example, to run OSPFv3 and RIPng simultaneously, you must first
configure both protocols and then verify the independent operation of each. Then you can configure the
routes to export from OSPFv3 to RIPng and the routes to export from RIPng to OSPFv3. Likewise, for
any other combinations of protocols, you must separately configure each to export routes to the other.

RIPng
Redistributing Routes into RIPng

Enable or disable the exporting of static, direct, or other protocol-learned routes into the RIPng domain
using the following commands:
enable ripng export [direct | ospfv3 | ospfv3-extern1 | ospfv3-extern2 | ospfv3-

inter | ospfv3-intra | static] [cost <number> {tag <number>} | policy <policy-name>]
disable ripng export [direct | ospfv3 | ospfv3-extern1 | ospfv3-extern2 | ospfv3-inter

| ospfv3-intra | static]
These commands enable or disable the exporting of static, direct, and OSPF-learned routes into the
RIPng domain. You can choose which types of OSPF routes are injected, or you can simply choose
ospf, which will inject all learned OSPF routes regardless of type. The default setting is disabled.
RIPng Configuration Example

The following configuration is similar to the example in the RIP chapter, but uses IPv6 addresses. It
illustrates a BlackDiamond switch that has three VLANs defined as follows:
● Finance
■ IP address 2001:db8:35::1/48.
● Personnel
■ IP address 2001:db8:36::1/48.
● MyCompany
The stations connected to the system generate a combination of IPv6 traffic and NetBIOS traffic.
In this configuration, all traffic from stations connected to slots 1 and 3 have access to the router by way
of the VLAN Finance. Ports on slots 2 and 4 reach the router by way of the VLAN Personnel. All traffic
(NetBIOS and IPv6) is part of the VLAN MyCompany.

RIPng Configuration Example
The example is configured as follows:
create vlan Finance


configure Finance ipaddress 2001:db8:35::1/48

configure Personnel ipaddress 2001:db8:36::1/48

configure ripng add vlan Finance
configure ripng add vlan Personnel
enable ripng

RIPng

38 OSPF

● Overview of OSPF on page 925
● Configuring OSPF on page 934
● OSPF Configuration Example on page 936
● Displaying OSPF Settings on page 938
● RFC 2328—OSPF Version 2
● RFC 1765—OSPF Database Overflow
● RFC 2370—The OSPF Opaque LSA Option
● RFC 3101—The OSPF Not-So-Stubby Area (NSSA) Option
● RFC 3623—Graceful OSPF Restart
● Interconnections: Bridges and Routers
by Radia Perlman
ISBN 0-201-56332-0
Published by Addison-Wesley Publishing Company
Overview of OSPF
Open Shortest Path First (OSPF) is a link state protocol that distributes routing information between
routers belonging to a single IP domain; the IP domain is also known as an autonomous system (AS). In a
link-state routing protocol, each router maintains a database describing the topology of the AS. Each
participating router has an identical database maintained from the perspective of that router.
From the link state database (LSDB), each router constructs a tree of shortest paths, using itself as the
root. The shortest path tree provides the route to each destination in the AS. When several equal-cost
routes to a destination exist, traffic can be distributed among them. The cost of a route is described by a
single metric.
OSPF is an interior gateway protocol (IGP), as is the other common IGP, RIP. OSPF and RIP are
compared in Chapter 8, “RIP.”
Licensing
To use the complete OSPF functionality, you must have a Core license installed on your switch. The
BlackDiamond 10808 ships with a Core, or Advanced Core license. Other platforms can be upgraded to
a Core license. See “Software Licenses and Feature Packs” on page 37 for more information about
licensing.

OSPF
A subset of OSPF, called OSPF Edge Mode, is available with an Advanced Edge license.
OSPF Edge Mode

OSPF Edge Mode is a subset of OSPF available on platforms with an Advanced Edge license. There are
two restrictions on OSPF Edge Mode:
● At most, four Active OSPF VLAN interfaces are permitted. There is no restriction on the number of
Passive interfaces.
● The OSPF Priority on VLANs is 0, and is not configurable. This prevents the system from acting as a
DR or BDR
Link State Database

Upon initialization, each router transmits a link state advertisement (LSA) on each of its interfaces.
LSAs are collected by each router and entered into the LSDB of each router. After all LSAs are received,
the router uses the LSDB to calculate the best routes for use in the IP routing table. OSPF uses flooding
to distribute LSAs between routers. Any change in routing information is sent to all of the routers in the
network. All routers within an area have the exact same LSDB. Table 109 describes LSA type numbers.
Table 109: LSA type numbers
Type Number Description
1 Router LSA
2 Network LSA
3 Summary LSA
4 AS summary LSA
5 AS external LSA
7 NSSA external LSA
9 Link local—Opaque
10 Area scoping—Opaque
11 AS scoping—Opaque
Database Overflow
The OSPF database overflow feature allows you to limit the size of the LSDB and to maintain a
consistent LSDB across all the routers in the domain, which ensures that all routers have a consistent
view of the network.
Consistency is achieved by:

● Limiting the number of external LSAs in the database of each router
● Ensuring that all routers have identical LSAs

Overview of OSPF
To configure OSPF database overflow, use the following command:
configure ospf ase-limit <number> {timeout <seconds>}
Where:
● <number>—Specifies the number of external LSAs that the system supports before it goes into
overflow state. A limit value of 0 disables the functionality.
When the LSDB size limit is reached, OSPF database overflow flushes LSAs from the LSDB. OSPF
database overflow flushes the same LSAs from all the routers, which maintains consistency.
● timeout—Specifies the timeout, in seconds, after which the system ceases to be in overflow state. A
timeout value of 0 leaves the system in overflow state until OSPF is disabled and re-enabled.
Opaque LSAs
Opaque LSAs are a generic OSPF mechanism used to carry auxiliary information in the OSPF database.
Opaque LSAs are most commonly used to support OSPF traffic engineering.
Normally, support for opaque LSAs is autonegotiated between OSPF neighbors. In the event that you
experience interoperability problems, you can disable opaque LSAs across the entire system using the
following command:
disable ospf capability opaque-lsa
To re-enable opaque LSAs across the entire system, use the following command:
enable ospf capability opaque-lsa
If your network uses opaque LSAs, Extreme Networks recommends that all routers on your OSPF
network support opaque LSAs. Routers that do not support opaque LSAs do not store or flood them. At
minimum a well interconnected subsection of your OSPF network must support opaque LSAs to
maintain reliability of their transmission.
Graceful OSPF Restart

RFC 3623 describes a way for OSPF control functions to restart without disrupting traffic forwarding.
Without graceful restart, adjacent routers will assume that information previously received from the
restarting router is stale and won’t be used to forward traffic to that router. However, in many cases,
two conditions exist that allow the router restarting OSPF to continue to forward traffic correctly. The
first condition is that forwarding can continue while the control function is restarted. Most modern
router system designs separate the forwarding function from the control function so that traffic can still
be forwarded independent of the state of the OSPF function. Routes learned through OSPF remain in
the routing table and packets continue to be forwarded. The second condition required for graceful
restart is that the network remain stable during the restart period. If the network topology is not
changing, the current routing table remains correct. Often, networks can remain stable during the time
for restarting OSPF.
Restarting and Helper Mode

Routers involved with graceful restart fill one of two roles: the restarting router or the helper router.
With graceful restart, the router that is restarting sends out Grace-LSAs informing its neighbors that it is
in graceful restart mode, how long the helper router should assist with the restart (the grace period),

OSPF
and why the restart occurred. If the neighboring routers are configured to help with the graceful restart
(helper-mode), they will continue to advertise the restarting router as if it was fully adjacent. Traffic
continues to be routed as though the restarting router is fully functional. If the network topology
changes, the helper routers will stop advertising the restarting router. The helper router will continue in
helper mode until the restarting router indicates successful termination of graceful restart, the
Grace-LSAs expire, or the network topology changes. A router can be configured for graceful restart,
and for helper-mode separately. A router can be a helper when its neighbor restarts, and can in turn be
helped by a neighbor if it restarts.
Planned and Unplanned Restarts

Two types of graceful restarts are defined: planned and unplanned. A planned restart would occur if
the software module for OSPF was upgraded, or if the router operator decided to restart the OSPF
control function for some reason. The router has advance warning, and is able to inform its neighbors in
advance that OSPF is restarting. An unplanned restart would occur if there was some kind of system
failure that caused a remote reboot or a crash of OSPF, or an MSM failover occurs. As OSPF restarts, it
informs its neighbors that it is in the midst of an unplanned restart. You can decide to configure a
router to enter graceful restart for only planned restarts, for only unplanned restarts, or for both. Also,
you can separately decide to configure a router to be a helper for only planned, only unplanned, or for
both kinds of restarts.
Configuring Graceful OSPF Restart

To configure a router to perform graceful OSPF restart, use the following command:
configure ospf restart [none | planned | unplanned | both]
Since a router can act as a restart helper router to multiple neighbors, you will specify which neighbors
to help. To configure a router to act as a graceful OSPF restart helper, use the following command:
configure ospf [vlan [all | <vlan-name>] | area <area-identifier> | virtual-link

<router-identifier> <area-identifier>] restart-helper [none | planned | unplanned |
both]
The graceful restart period sent out to helper routers can be configured with the following command:
configure ospf restart grace-period <seconds>
By default, a helper router will terminate graceful restart if received LSAs would affect the restarting
router. This will occur when the restart-helper receives an LSA that will be flooded to the restarting
router or when there is a changed LSA on the restarting router's retransmission list when graceful
restart is initiated. To disable this behavior, use the following command:
disable ospf [vlan [all | <vlan-name>] | area <area-identifier> | virtual-link

<router-identifier> <area-identifier>] restart-helper-lsa-check
Areas
OSPF allows parts of a network to be grouped together into areas. The topology within an area is
hidden from the rest of the AS. Hiding this information enables a significant reduction in LSA traffic
and reduces the computations needed to maintain the LSDB. Routing within the area is determined
only by the topology of the area.

Overview of OSPF
The three types of routers defined by OSPF are as follows:

● Internal router (IR)—An internal router has all of its interfaces within the same area.
● Area border router (ABR)—An ABR has interfaces in multiple areas. It is responsible for exchanging
summary advertisements with other ABRs.
● Autonomous system border router (ASBR)—An ASBR acts as a gateway between OSPF and other
routing protocols, or other autonomous systems.
Backbone Area (Area 0.0.0.0)

Any OSPF network that contains more than one area is required to have an area configured as area
0.0.0.0, also called the backbone. All areas in an AS must be connected to the backbone. When designing
networks, you should start with area 0.0.0.0 and then expand into other areas.
NOTE
Area 0.0.0.0 exists by default and cannot be deleted or changed.
The backbone allows summary information to be exchanged between ABRs. Every ABR hears the area
summaries from all other ABRs. The ABR then forms a picture of the distance to all networks outside of
its area by examining the collected advertisements and adding in the backbone distance to each
advertising router.
When a VLAN is configured to run OSPF, you must configure the area for the VLAN. If you want to
configure the VLAN to be part of a different OSPF area, use the following command:
configure ospf vlan <vlan-name> area <area-identifier>
If this is the first instance of the OSPF area being used, you must create the area first using the
following command:
create ospf area <area-identifier>
Stub Areas
OSPF allows certain areas to be configured as stub areas. A stub area is connected to only one other area.
The area that connects to a stub area can be the backbone area. External route information is not
distributed into stub areas. Stub areas are used to reduce memory consumption and computational
requirements on OSPF routers. Use the following command to configure an OSPF area as a stub area:
configure ospf area <area-identifier> stub [summary | nosummary] stub-default-cost
<cost>
Not-So-Stubby-Areas
Not-so-stubby-areas (NSSAs) are similar to the existing OSPF stub area configuration option but have
the following two additional capabilities:
● External routes originating from an ASBR connected to the NSSA can be advertised within the
NSSA.
● External routes originating from the NSSA can be propagated to other areas, including the backbone
area.

OSPF
The command line interface (CLI) command to control the NSSA function is similar to the command
used for configuring a stub area, as follows:
configure ospf area <area-identifier> nssa [summary | nosummary] stub-default-cost

<cost> {translate}
The translate option determines whether type 7 LSAs are translated into type 5 LSAs. When
configuring an OSPF area as an NSSA, the translate should only be used on NSSA border routers,
where translation is to be enforced. If translate is not used on any NSSA border router in a NSSA, one
of the ABRs for that NSSA is elected to perform translation (as indicated in the NSSA specification). The
option should not be used on NSSA internal routers. Doing so inhibits correct operation of the election
algorithm.
Normal Area
A normal area is an area that is not:
● Area 0
● Stub area
● NSSA
Virtual links can be configured through normal areas. External routes can be distributed into normal
areas.
Virtual Links
In the situation when a new area is introduced that does not have a direct physical attachment to the
backbone, a virtual link is used. A virtual link provides a logical path between the ABR of the
disconnected area and the ABR of the normal area that connects to the backbone. A virtual link must be
established between two ABRs that have a common area, with one ABR connected to the backbone.
Figure 137 illustrates a virtual link.
NOTE
Virtual links cannot be configured through a stub or NSSA area.

Overview of OSPF
Figure 137: Virtual link using area 1 as a transit area
Virtual link
ABR ABR
Area 2 Area 1 Area 0

EX_044
Virtual links are also used to repair a discontiguous backbone area. For example, in Figure 138, if the
connection between ABR1 and the backbone fails, the connection using ABR2 provides redundancy so
that the discontiguous area can continue to communicate with the backbone using the virtual link.
Figure 138: Virtual link providing redundancy
Virtual link
Area 2
ABR 1 ABR 2
Area 1 Area 3
Area 0
EX_045

OSPF
Point-to-Point Support
You can manually configure the OSPF link type for a VLAN. Table 110 describes the link types.
Table 110: OSPF link types
Link Type Number of Routers Description
Auto Varies ExtremeXOS automatically determines the OSPF link type based on
the interface type. This is the default setting.
Broadcast Any Routers must elect a designated router (DR) and a backup designated
router (BDR) during synchronization. Ethernet is an example of a
broadcast link.
Point-to-point Up to 2 This type synchronizes faster than a broadcast link because routers do
not elect a DR or BDR. It does not operate with more than two routers
on the same VLAN. The Point-to-Point Protocol (PPP) is an example of
a point-to-point link. An OSPF point-to-point link supports only zero to
two OSPF routers and does not elect a designated router (DR) or
backup designated router (BDR). If you have three or more routers on
the VLAN, OSPF fails to synchronize if the neighbor is not configured.
Passive A passive link does not send or receive OSPF packets.
NOTE
The number of routers in an OSPF point-to-point link is determined per VLAN, not per link.
NOTE
All routers in the VLAN must have the same OSPF link type. If there is a mismatch, OSPF attempts to operate, but
it may not be reliable.

is an example of route redistribution between an OSPF AS and a RIP AS.
OSPF AS
Backbone Area
0.0.0.0
ABR
Area
121.2.3.4
ASBR ASBR
RIP AS
EX_046

OSPF

configuration functions. For example, to run OSPF and RIP simultaneously, you must first configure
both protocols and then verify the independent operation of each. Then you can configure the routes to
export from OSPF to RIP and the routes to export from RIP to OSPF. Likewise, for any other
combinations of protocols, you must separately configure each to export routes to the other.
Redistributing Routes into OSPF

To enable or disable the exporting of BGP, RIP, static, and direct (interface) routes to OSPF, use the
following commands:
enable ospf export [bgp | direct | e-bgp | i-bgp | rip | static] [cost <cost> type
[ase-type-1 | ase-type-2] {tag <number>} | <policy-map>]
disable ospf export [bgp | direct | e-bgp | i-bgp | rip | static]
These commands enable or disable the exporting of RIP, static, and direct routes by way of LSA to other
OSPF routers as AS-external type 1 or type 2 routes. The default setting is disabled.
The cost metric is inserted for all Border Gateway Protocol (BGP), RIP, static, and direct routes injected
into OSPF. If the cost metric is set to 0, the cost is inserted from the route. For example, in the case of
BGP export, the cost equals the multiple exit discriminator (MED) or the path length. The tag value is
used only by special routing applications. Use 0 if you do not have specific requirements for using a
tag. (The tag value in this instance has no relationship with IEEE 802.1Q VLAN tagging.)
The same cost, type, and tag values can be inserted for all the export routes, or policies can be used for
selective insertion. When a policy is associated with the export command, the policy is applied on every
exported route. The exported routes can also be filtered using policies.
Verify the configuration using the command:
show ospf
OSPF Timers and Authentication

Configuring OSPF timers and authentication on a per area basis is a shortcut to applying the timers and
authentication to each VLAN in the area at the time of configuration. If you add more VLANs to the
area, you must configure the timers and authentication for the new VLANs explicitly. Use the
command:
configure ospf vlan [<vlan-name> | all] timer <retransmit-interval> <transit-delay>
<hello-interval> <dead-interval> {<wait-timer-interval>}
Configuring OSPF
Each switch that is configured to run OSPF must have a unique router ID. Extreme Networks
recommends that you manually set the router ID of the switches participating in OSPF, instead of
having the switch automatically choose its router ID based on the highest interface IP address. Not

Configuring OSPF
performing this configuration in larger, dynamic environments could result in an older LSDB remaining
in use.
Configuring OSPF Wait Interval

ExtremeXOS allows you to configure the OSPF wait interval, rather than using the router dead interval.
CAUTION
Do not configure OSPF timers unless you are comfortable exceeding OSPF specifications. Non-standard settings may
not be reliable under all circumstances.
To specify the timer intervals, use the following commands:
configure ospf area <area-identifier> timer <retransmit-interval> <transit-delay>

configure ospf virtual-link <router-identifier> <area-identifier> timer <retransmit-

interval> <transit-delay> <hello-interval> <dead-interval>
configure ospf vlan [<vlan-name> | all] timer <retransmit-interval> <transit-delay>

OSPF Wait Interval Parameters

You can configure the following parameters:
● Retransmit interval—The length of time that the router waits before retransmitting an LSA that is
not acknowledged. If you set an interval that is too short, unnecessary retransmissions result. The
default value is 5 seconds.
● Transit delay—The length of time it takes to transmit an LSA packet over the interface. The transit
delay must be greater than 0.
● Hello interval—The interval at which routers send hello packets. Shorter times allow routers to
discover each other more quickly but also increase network traffic. The default value is 10 seconds.
● Dead router wait interval (Dead Interval)—The interval after which a neighboring router is declared
down because hello packets are no longer received from the neighbor. This interval should be a
multiple of the hello interval. The default value is 40 seconds.
● Router wait interval (Wait Timer Interval)—The interval between the interface coming up and the
election of the DR and BDR. This interval should be greater than the hello interval. If this time is
close to the hello interval, the network synchronizes very quickly but might not elect the correct DR
or BDR. The default value is equal to the dead router wait interval.
NOTE
The OSPF standard specifies that wait times are equal to the dead router wait interval.

OSPF
OSPF Configuration Example

Figure 140 is an example of an autonomous system using OSPF routers. The details of this network
follow.
Figure 140: OSPF configuration example
Area 0 IR 2 IR 1
10.0.1.1 10.0.1.2
10.0.3.2 10.0.2.2
2
Headquarters HQ
0_
_1
0_
0_
_1
0_
HQ
3
ABR 2 10.0.3.1 ABR 1 10.0.2.1
161.48.2.2
Los Angeles
LA
161.48.2.1
_1
160.26.25.1 160.26.26.1
26
61
_4
6_
8_
_2
Virtual link
2
60
i_1
Ch
160.26.26.2
160.26.25.2 Chicago
Area 5 Area 6 (stub)

EX_040
Area 0 is the backbone area. It is located at the headquarters and has the following characteristics:
● Two internal routers (IR1 and IR2)
● Two area border routers (ABR1 and ABR2)
● Network number 10.0.x.x
● Two identified VLANs (HQ_10_0_2 and HQ_10_0_3)
Area 5 is connected to the backbone area by way of ABR1 and ABR2. It is located in Chicago and has
the following characteristics:
● One identified VLAN (Chi_160_26_26)
● Two internal routers

OSPF Configuration Example
Area 6 is a stub area connected to the backbone by way of ABR1. It is located in Los Angeles and has
the following characteristics:
● One identified VLAN (LA_161_48_2)
● Three internal routers
● Uses default routes for inter-area routing
Two router configurations for the example in Figure 140 are provided in the following section.
Configuration for ABR1

The router labeled ABR1 has the following configuration:
create vlan HQ_10_0_2
create vlan HQ_10_0_3
create vlan LA_161_48_2
create vlan Chi_160_26_26
configure vlan HQ_10_0_2 ipaddress 10.0.2.1 255.255.255.0

configure vlan LA_161_48_2 ipaddress 161.48.2.2 255.255.255.0
configure vlan Chi_160_26_26 ipaddress 160.26.26.1 255.255.255.0
create ospf area 0.0.0.5

create ospf area 0.0.0.6
enable ipforwarding
configure ospf area 0.0.0.6 stub nosummary stub-default-cost 10

configure ospf add vlan LA_161_48_2 area 0.0.0.6
configure ospf add vlan Chi_160_26_26 area 0.0.0.5
configure ospf add vlan HQ_10_0_2 area 0.0.0.0
configure ospf add vlan HQ_10_0_3 area 0.0.0.0
configure ospf vlan LA_161_48_2 priority 10
configure ospf vlan Chi_160_26_26 priority 10
configure ospf vlan HQ_10_0_2 priority 5
configure ospf vlan HQ_10_0_3 priority 5
enable ospf
Configuration for IR1

The router labeled IR1 has the following configuration:
enable ipforwarding
configure ospf add vlan all area 0.0.0.0
configure ospf area 0.0.0.0 priority 10
enable ospf

OSPF
NOTE
In OSPF edge mode, the VLAN priority is “0” and cannot be set. (Refer to “OSPF Edge Mode” on page 926.) When
the license is upgraded to a Core license, the VLAN priority of “0” needs to be reset in order to participate in DR/
BDR election.
Displaying OSPF Settings

You can use a number of commands to display settings for OSPF. To show global OSPF information,
use the show ospf command with no options.
To display information about one or all OSPF areas, use the following command:
show ospf area {<area-identifier>}
The detail option displays information about all OSPF areas in a detail format.
To display information about OSPF interfaces for an area, a VLAN, or for all interfaces, use the
following command:
show ospf interfaces {vlan <vlan-name> | area <area-identifier>}
The detail option displays information about all OSPF interfaces in a detail format.
ExtremeXOS provides several filtering criteria for the show ospf lsdb command. You can specify
multiple search criteria, and only those results matching all of the criteria are displayed. This allows
you to control the displayed entries in large routing tables.
To display the current link-state database, use the following command:
show ospf lsdb {detail | stats} {area [<area-identifier> | all]} {{lstype} [<lstype> |
all]} {lsid <lsid-address>{<lsid-mask>}} {routerid <routerid-address> {<routerid-
mask>}} {interface[[<ip-address>{<ip-mask>} | <ipNetmask>] | vlan <vlan-name>]}
The detail option displays all fields of matching LSAs in a multiline format. The summary option
displays several important fields of matching LSAs, one line per LSA. The stats option displays the
number of matching LSAs but not any of their contents. If not specified, the default is to display in the
summary format.
A common use of this command is to omit all optional parameters, resulting in the following shortened
form:
show ospf lsdb
The shortened form displays LSAs from all areas and all types in a summary format.

39 OSPFv3

● Overview of OSPFv3 on page 939
Overview of OSPFv3
Open Shortest Path First (OSPF) is a link state protocol that distributes routing information between
routers belonging to a single IP domain; the IP domain is also known as an autonomous system (AS). In a
link-state routing protocol, each router maintains a database describing the topology of the AS. Each
participating router has an identical database for an area maintained from the perspective of that router.
From the link state database (LSDB), each router constructs a tree of shortest paths, using itself as the
root. The shortest path tree provides the route to each destination in the AS. When several equal-cost
routes to a destination exist, traffic can be distributed among them. The cost of a route is described by a
single metric.
OSPFv3 supports IPv6, and uses commands only slightly modified from that used to support IPv4.
OSPFv3 has retained the use of the 4-byte, dotted decimal numbers for router IDs, LSA IDs, and area
IDs.
OSPFv3 is an interior gateway protocol (IGP), as is the other common IGP for IPv6, RIPng. OSPFv3 and
RIPng are compared in Chapter 9, “RIPng.”
Licensing
To use OSPFv3, you must have a Core license installed on your switch. The BlackDiamond 10808 ships
with a Core, or Advanced Core license. Other platforms can be upgraded to a Core license. See the
section “Software Licenses and Feature Packs” on page 37 for more information about licensing.
Link State Database

Upon initialization, each router transmits a link state advertisement (LSA) on each of its interfaces.
LSAs are collected by each router and stored into the LSDB of each router. After all LSAs are received,
the router uses the LSDB to calculate the best routes for use in the IP routing table. OSPFv3 uses
flooding to distribute LSAs between routers. Any change in routing information is sent to all of the
routers in the network. All routers within an area have the exact same LSDB.

OSPFv3
Table 111 describes LSA type numbers.
Table 111: Selected OSPFv3 LSA types

Type Number Description
0x0008 Link LSA
0x2001 Router LSA
0x2002 Network LSA
0x2003 Inter-Area-Prefix LSA
0x2004 Inter-Area-Router LSA
0x2009 Intra-Area-Prefix LSA
0x4005 AS external LSA
Areas
OSPFv3 allows parts of a network to be grouped together into areas. The topology within an area is
hidden from the rest of the AS. Hiding this information enables a significant reduction in LSA traffic
and reduces the computations needed to maintain the LSDB. Routing within the area is determined
only by the topology of the area.
The three types of routers defined by OSPFv3 are as follows:

● Internal router (IR)—An internal router has all of its interfaces within the same area.
● Area border router (ABR)—An ABR has interfaces in multiple areas. It is responsible for exchanging
summary advertisements with other ABRs.
● Autonomous system border router (ASBR)—An ASBR acts as a gateway between OSPFv3 and other
routing protocols, or other autonomous systems.
Backbone Area (Area 0.0.0.0)

Any OSPFv3 network that contains more than one area is required to have an area configured as area
0.0.0.0, also called the backbone. All areas in an AS must be connected to the backbone. When designing
networks, you should start with area 0.0.0.0 and then expand into other areas.
NOTE
Area 0.0.0.0 exists by default and cannot be deleted or changed.
The backbone allows summary information to be exchanged between ABRs. Every ABR hears the area
summaries from all other ABRs. The ABR then forms a picture of the distance to all networks outside of
its area by examining the collected advertisements and adding in the backbone distance to each
advertising router.
When a VLAN is configured to run OSPFv3, you must configure the area for the VLAN. If you want to
configure the VLAN to be part of a different OSPFv3 area, use the following command:
configure ospfv3 {domain <domainName>} [vlan <vlan-name> | tunnel <tunnel-name>] area

<area-identifier>

Overview of OSPFv3
If this is the first instance of the OSPFv3 area being used, you must create the area first using the
following command:
create ospfv3 {domain <domainName>} area <area-identifier>
Stub Areas
OSPFv3 allows certain areas to be configured as stub areas. A stub area is connected to only one other
area. The area that connects to a stub area can be the backbone area. External route information is not
distributed into stub areas. Stub areas are used to reduce memory consumption and computational
requirements on OSPFv3 routers. To configure an OSPFv3 area as a stub area, use the following
command:
configure ospfv3 {domain <domainName>} area <area-identifier> stub [summary |
nosummary] stub-default-cost <cost>
Not-So-Stubby-Areas
Not-so-stubby-areas (NSSAs) are not supported currently in the ExtremeXOS implementation of
OSPFv3.
Normal Area
A normal area is an area that is not:
● Area 0
● Stub area
● NSSA
Virtual links can be configured through normal areas. External routes can be distributed into normal
areas.
Virtual Links
In the situation when a new area is introduced that does not have a direct physical attachment to the
backbone, a virtual link is used. A virtual link provides a logical path between the ABR of the
disconnected area and the ABR of the normal area that connects to the backbone. A virtual link must be
established between two ABRs that have a common area, with one ABR connected to the backbone.
Figure 141 illustrates a virtual link.
NOTE
Virtual links cannot be configured through a stub or NSSA area.

OSPFv3
Figure 141: Virtual link using area 1 as a transit area
Virtual link
ABR ABR
Area 2 Area 1 Area 0

EX_044
Virtual links are also used to repair a discontiguous backbone area. For example, in Figure 142, if the
connection between ABR1 and the backbone fails, the connection using ABR2 provides redundancy so
that the discontiguous area can continue to communicate with the backbone using the virtual link.
Figure 142: Virtual link providing redundancy
Virtual link
Area 2
ABR 1 ABR 2
Area 1 Area 3
Area 0
EX_045

Link-Type Support
You can manually configure the OSPFv3 link type for a VLAN. Table 112 describes the link types.
Table 112: OSPFv3 link types
Link Type Number of Routers Description
Auto Varies ExtremeXOS automatically determines the OSPFv3 link type based on
the interface type. This is the default setting.
Broadcast Any Routers must elect a designated router (DR) and a backup designated
router (BDR) during synchronization. Ethernet is an example of a
broadcast link.
Passive A passive link does not send or receive OSPFv3 packets.
NOTE
The number of routers in an OSPFv3 point-to-point link is determined per VLAN, not per link.
NOTE
All routers in the VLAN must have the same OSPFv3 link type. If there is a mismatch, OSPFv3 attempts to operate,
but it may not be reliable.
is an example of route redistribution between an OSPFv3 AS and a RIPng AS.

OSPFv3
OSPF AS
Backbone Area
0.0.0.0
ABR
Area
121.2.3.4
ASBR ASBR
RIP AS
EX_046

configuration functions. For example, to run OSPFv3 and RIPng simultaneously, you must first
configure both protocols and then verify the independent operation of each. Then you can configure the
routes to export from OSPFv3 to RIPng and the routes to export from RIPng to OSPFv3. Likewise, for
any other combinations of protocols, you must separately configure each to export routes to the other.

Redistributing Routes into OSPFv3

To enable or disable the exporting of RIPng, static, and direct (interface) routes to OSPFv3, use the
following commands:
enable ospfv3 {domain <domainName>} export [direct | ripng | static] [cost <cost> type
[ase-type-1 | ase-type-2] | <policy-map>]
disable ospfv3 {domain <domainName>} export [direct | ripng | static]
These commands enable or disable the exporting of RIPng, static, and direct routes by way of LSA to
other OSPFv3 routers as AS-external type 1 or type 2 routes. The default setting is disabled.
The cost metric is inserted for all RIPng, static, and direct routes injected into OSPFv3. If the cost metric
is set to 0, the cost is inserted from the route. The tag value is used only by special routing applications.
Use 0 if you do not have specific requirements for using a tag. (The tag value in this instance has no
relationship with IEEE 802.1Q VLAN tagging.)
The same cost, type, and tag values can be inserted for all the export routes, or policies can be used for
selective insertion. When a policy is associated with the export command, the policy is applied on every
exported route. The exported routes can also be filtered using policies.
Verify the configuration using the command:
show ospfv3 {domain <domainName>}
OSPFv3 Timers
Configuring OSPFv3 timers on a per area basis is a shortcut to applying the timers to each VLAN in the
area at the time of configuration. If you add more VLANs to the area, you must configure the timers for
the new VLANs explicitly. Use the command:
configure ospfv3 {domain <domainName>} [vlan <vlan-name> | tunnel <tunnel-name> |
[vlan | tunnel] all] timer {retransmit-interval} <retransmit-interval> {transit-delay}
<transit-delay> {hello-interval} <hello-interval> {dead-interval} <dead-interval>

OSPFv3
OSPFv3 Configuration Example

Figure 144 is an example of an autonomous system using OSPFv3 routers. The details of this network
follow.
Figure 144: OSPFv3 configuration example

Router 2
Area 0.0.0.0
2001:db8:4444:6666::2/64
to-r2
Router 1 2001:db8:4444:6666::1/64
2001:db8:3333:5555::1/64
Router 3
to-r3
2001:db8:3333:5555::2/64
Area 0.0.0.1
EX_107
In Figure 144 there are three Extreme Networks switches running ExtremeXOS images that have
support for OSPFv3. Router 1 is an area border router and is connected to two other switches Router 2
and Router 3. Router 1 runs OSPFv3 on both the links connecting it to Router 2 and Router 3.
The router configurations for the example in Figure 144 are provided in the following section. After
doing all the configurations, Router 1 will establish OSPFv3 adjacency with Router 2 and Router 3. They
will also exchange the various link state databases.

OSPFv3 Configuration Example
Configuration for Router 1

The router labeled Router 1 has the following configuration:
create vlan to-r2
create vlan to-r3
configure vlan to-r2 ipaddress 2001:db8:4444:6666::1/64
configure vlan to-r2 add port 1:1
configure ospfv3 routerid 0.0.0.1

configure ospfv3 add vlan to-r2 area 0.0.0.0
create ospfv3 area 0.0.0.1
enable ospfv3

create vlan to-r1

enable ospfv3

create vlan to-r1

enable ospfv3

OSPFv3

40 Border Gateway Protocol

● BGP Attributes on page 950
● BGP Communities on page 950
● BGP Features on page 951
This chapter describes how to configure the Border Gateway Protocol (BGP), an exterior routing
protocol available on the switch.
For more information on BGP, refer to the following documents:

● RFC 1771—Border Gateway Protocol version 4 (BGP-4)
● RFC 1965—Autonomous System Confederations for BGP
● RFC 1966—BGP Route Reflection
● RFC 1997—BGP Communities Attribute
● RFC 1745—BGP/IDRP for IP—OSPF Interaction
● RFC 2385—Protection of BGP Sessions via the TCP MD5 Signature Option
● RFC 2439—BGP Route Flap Damping
● RFC 2796—BGP Route Reflection - An Alternative to Full Mesh IBGP
● RFC 2842—Capabilities Advertisement with BGP-4
● RFC 2858—Multiprotocol Extensions for BGP-4
● RFC 2918—Route Refresh Capability for BGP-4
● draft_ieft_idr_restart_10.txt—Graceful Restart Mechanism for BGP
NOTE
ExtremeXOS supports BGP version 4 only.
NOTE
Although the CLI commands are available, this release of ExtremeXOS does not support the MBGP/Route-refresh
features.
Licensing
BGP requires a Core license, at a minimum. The BGP process will not spawn without the required
license level. The MSM-1XL is shipped with an Advanced Core license and the MSM-1 is shipped with
a Core license. Other platforms can be upgraded to a Core license. See “Software Licenses and Feature
Packs” on page 37 for more information about licensing.

Border Gateway Protocol
Overview
BGP is an exterior routing protocol that was developed for use in TCP/IP networks. The primary
function of BGP is to allow different autonomous systems (ASs) to exchange network reachability
information.
An AS is a set of routers that are under a single technical administration. This set of routers uses a
different routing protocol, for example Open Shortest Path First (OSPF), for intra-AS routing. One or
more routers in the AS are configured to be border routers, exchanging information with other border
routers (in different ASs) on behalf of all of the intra-routers.
BGP can be used as an exterior border gateway protocol (referred to as EBGP), or it can be used within
an AS as an interior border gateway protocol (referred to as IBGP).
BGP Attributes
The following BGP attributes are supported by the switch:
● Origin—Defines the origin of the route. Possible values are Interior Gateway Protocol (IGP), Exterior
Gateway Protocol (EGP), and incomplete.
● AS_Path—The list of ASs that are traversed for this route.
● Next_hop—The IP address of the next hop BGP router to reach the destination listed in the NLRI
field.
● Multi_Exit_Discriminator—Used to select a particular border router in another AS when multiple
border routers exist.
● Local_Preference—Used to advertise this router’s degree of preference to other routers within the
AS.
● Atomic_aggregate—Indicates that the sending border router has used a route aggregate prefix in the
route update.
● Aggregator—Identifies the BGP router AS number and IP address that performed route aggregation.
● Community—Identifies a group of destinations that share one or more common attributes.
● Cluster_ID—Specifies a 4-byte field used by a route reflector to recognize updates from other route
reflectors in the same cluster.
● Originator_ID—Specifies the router ID of the originator of the route in the local AS.\
BGP Communities
A BGP community is a group of BGP destinations that require common handling. ExtremeXOS
supports the following well-known BGP community attributes:
● no-export
● no-advertise
● no-export-subconfed

BGP Features
BGP Features
This section describes the following BGP features supported by ExtremeXOS:
● Route Reflectors on page 951
● Route Confederations on page 953
● Route Aggregation on page 956
● Using the Loopback Interface on page 956
● BGP Peer Groups on page 956
● BGP Route Flap Dampening on page 957
● BGP Route Selection on page 959
● BGP Static Network on page 960
● Graceful BGP Restart on page 961
Route Reflectors
Another way to overcome the difficulties of creating a fully meshed AS is to use route reflectors. Route
reflectors allow a single router to serve as a central routing point for the AS.
A cluster is formed by the route reflector and its client routers. Peer routers that are not part of the
cluster must be fully meshed according to the rules of BGP.
A BGP cluster, including the route reflector and its clients, is shown in Figure 145.
Figure 145: Route reflectors
AS 100
1.1.1.1
10.0.0.1 3.3.3.3
2.2.2.2 20.0.0.1
Non-client
Client
10.0.0.2
20.0.0.2
30.0.0.2
4.4.4.4
30.0.0.1
Route Reflector Client
Cluster
EX_042
The topology shown in Figure 145 minimizes the number of BGP peering sessions required in an AS by
using route reflectors.

In this example, although the BGP speakers 3.3.3.3 and 4.4.4.4 do not have a direct BGP peering session
between them, these speakers still receive routes from each other indirectly through 2.2.2.2. The router
2.2.2.2 is called a route reflector and is responsible for reflecting routes between its clients. Routes
received from the client 3.3.3.3 by the router 2.2.2.2 are reflected to 4.4.4.4 and vice-versa. Routes
received from 1.1.1.1 are reflected to all clients.
To configure router 1.1.1.1, use the following commands:

create vlan to_rr
configure vlan to_rr add port 1:1
configure vlan to_rr ipaddress 10.0.0.1/24
enable ipforwarding vlan to_rr
configure bgp router 1.1.1.1

configure bgp as-number 100
create bgp neighbor 10.0.0.2 remote-as 100
enable bgp
enable bgp neighbor all
To configure router 2.2.2.2, the route reflector, use the following commands:
create vlan to_nc
configure vlan to_nc add port 1:1
configure vlan to_nc ipaddress 10.0.0.2/24
enable ipforwarding vlan to_nc
create vlan to_c1

configure vlan to_c1 add port 1:2
configure vlan to_c1 ipaddress 20.0.0.2/24
enable ipforwarding vlan to_c1
create vlan to_c2

configure vlan to_c2 add port 1:2
configure vlan to_c2 ipaddress 30.0.0.2/24
enable ipforwarding vlan to_c2

configure bgp neighbor 20.0.0.1 route-reflector-client
configure bgp neighbor 30.0.0.1 route-reflector-client
enable bgp

create vlan to_rr

enable bgp

BGP Features

create vlan to_rr

enable bgp
Route Confederations
BGP requires networks to use a fully meshed router configuration. This requirement does not scale well,
especially when BGP is used as an IGP. One way to reduce the size of a fully meshed AS is to divide
the AS into multiple sub-ASs and to group these sub-ASs into a routing confederation. Within the
confederation, each sub-AS must be fully meshed. The confederation is advertised to other networks as
a single AS.
Route Confederation Example

Figure 146 shows an example of a confederation.
Figure 146: Routing confederation
AS 200
SubAS 65001
A B
EBGP 192.1.1.6/30 192.1.1.5/30 192.1.1.9/30
192.1.1.17/30 192.1.1.22/30
IBGP
192.1.1.18/30 192.1.1.21/30
C
EBGP
EBGP 192.1.1.13/30 192.1.1.14/30
IBGP 192.1.1.10/30
E D
SubAS 65002
EX_043
In this example, AS 200 has five BGP speakers. Without a confederation, BGP would require that the
routes in AS 200 be fully meshed. Using the confederation, AS 200 is split into two sub-ASs: AS65001
and AS65002. Each sub-AS is fully meshed, and IBGP is running among its members. EBGP is used

between sub-AS 65001 and sub-AS 65002. Router B and router D are EBGP peers. EBGP is also used
between the confederation and outside ASs.
To configure router A, use the following commands:
create vlan ab
configure vlan ab add port 1
configure vlan ab ipaddress 192.1.1.6/30
enable ipforwarding vlan ab
configure ospf add vlan ab area 0.0.0.0
create vlan ac
configure vlan ac add port 2
configure vlan ac ipaddress 192.1.1.17/30
enable ipforwarding vlan ac
configure ospf add vlan ac area 0.0.0.0
enable ospf

configure bgp routerid 192.1.1.17
configure bgp confederation-id 200
enable bgp
create bgp neighbor 192.1.1.5 remote-AS-number 65001

To configure router B, use the following commands:
create vlan ba
configure vlan ba add port 1
configure vlan ba ipaddress 192.1.1.5/30
enable ipforwarding vlan ba
configure ospf add vlan ba area 0.0.0.0
create vlan bc
configure vlan bc add port 2
configure vlan bc ipaddress 192.1.1.22/30
enable ipforwarding vlan bc
configure ospf add vlan bc area 0.0.0.0
create vlan bd
configure vlan bd add port 3
configure vlan bd ipaddress 192.1.1.9/30
enable ipforwarding vlan bd
configure ospf add vlan bd area 0.0.0.0
enable ospf

enable bgp


BGP Features
configure bgp add confederation-peer sub-AS-number 65002

To configure router C, use the following commands:
create vlan ca
configure vlan ca add port 1
configure vlan ca ipaddress 192.1.1.18/30
enable ipforwarding vlan ca
configure ospf add vlan ca area 0.0.0.0
create vlan cb
configure vlan cb add port 2
configure vlan cb ipaddress 192.1.1.21/30
enable ipforwarding vlan cb
configure ospf add vlan cb area 0.0.0.0
enable ospf

enable bgp

To configure router D, use the following commands:
create vlan db
configure vlan db add port 1
configure vlan db ipaddress 192.1.1.10/30
enable ipforwarding vlan db
configure ospf add vlan db area 0.0.0.0
create vlan de
configure vlan de add port 2
configure vlan de ipaddress 192.1.1.14/30
enable ipforwarding vlan de
configure ospf add vlan de area 0.0.0.0
enable ospf

enable bgp

configure bgp add confederation-peer sub-AS-number 65001
To configure router E, use the following commands:
create vlan ed
configure vlan ed add port 1
configure vlan ed ipaddress 192.1.1.13/30

enable ipforwarding vlan ed

configure ospf add vlan ed area 0.0.0.0
enable ospf

enable bgp

enable bgp neighbor 192.1.1.14
Route Aggregation
Route aggregation is the process of combining the characteristics of several routes so that they are
advertised as a single route. Aggregation reduces the amount of information that a BGP speaker must
store and exchange with other BGP speakers. Reducing the information that is stored and exchanged
also reduces the size of the routing table.
Using Route Aggregation

To use BGP route aggregation:
1 Enable aggregation using the following command:
enable bgp aggregation
2 Create an aggregate route using the following command:
configure bgp add aggregate-address {address-family [ipv4-unicast | ipv4-
multicast]} <ipaddress> {as-match | as-set} {summary-only} {advertise-policy
<policy>} {attribute-policy <policy>}
Using the Loopback Interface

If you are using BGP as your IGP, you may decide to advertise the interface as available, regardless of
the status of any particular interface. The loopback interface can also be used for EBGP multihop. Using
the loopback interface eliminates multiple, unnecessary route changes.
BGP Peer Groups

You can use BGP peer groups to group together up to 512 BGP neighbors. All neighbors within the peer
group inherit the parameters of the BGP peer group. The following mandatory parameters are shared
by all neighbors in a peer group:
● remote AS
● source-interface
● route-policy
● send-community
● next-hop-self

BGP Features
Each BGP peer group is assigned a unique name when it is created. To create or delete peer groups, use
create bgp peer-group <peer-group-name>

delete bgp peer-group <peer-group-name>
Changes made to the parameters of a peer group are applied to all neighbors in the peer group.
Modifying the following parameters will automatically disable and enable the neighbors before changes
take effect:
● remote-as
● timer
● source-interface
● soft-in-reset
● password
Adding Neighbors to a BGP Peer Group

To create a new neighbor and add it to a BGP peer group, use the following command:
create bgp neighbor <remoteaddr> peer-group <peer-group-name> {multi-hop}
The new neighbor is created as part of the peer group and inherits all of the existing parameters of the
peer group. The peer group must have remote AS configured.
To add an existing neighbor to a peer group, use the following command:
configure bgp neighbor [all | <remoteaddr>] peer-group [<peer-group-name> | none]

{acquire-all}
If you do not specify the acquire-all option, only the mandatory parameters are inherited from the
peer group. If you specify the acquire-all option, all of the parameters of the peer group are
inherited. This command disables the neighbor before adding it to the peer group.
To remove a neighbor from a peer group, use the peer-group none option.
When you remove a neighbor from a peer group, the neighbor retains the parameter settings of the
group. The parameter values are not reset to those the neighbor had before it inherited the peer group
values.
BGP Route Flap Dampening

Route flap dampening is a BGP feature designed to minimize the propagation of flapping routes across
an internetwork. A route is considered to be flapping when it is repeatedly available, then unavailable,
then available, then unavailable, and so on.
When a route becomes unavailable, a withdrawal message is sent to other connected routers, which in
turn propagate the withdrawal message to other routers. As the route becomes available again, an
advertisement message is sent and propagated throughout the network.
As a route repeatedly changes from available to unavailable, large numbers of messages propagate
throughout the network. This is a problem in an internetwork connected to the Internet because a route
flap in the Internet backbone usually involves many routes.

Minimizing the Route Flap

The route flap dampening feature minimizes the flapping problem as follows. Suppose that the route to
network 172.25.0.0 flaps. The router (in which route dampening is enabled) assigns network 172.25.0.0 a
penalty of 1000 and moves it to a “history” state in which the penalty value is monitored. The router
continues to advertise the status of the route to neighbors. The penalties are cumulative. When the route
flaps so often that the penalty exceeds a configurable suppress limit, the router stops advertising the
route to network 172.25.0.0, regardless of how many times it flaps. Thus, the route is dampened.
The penalty placed on network 172.25.0.0 is decayed until the reuse limit is reached, when the route is
again advertised. At half of the reuse limit, the dampening information for the route to network
172.25.0.0 is removed.
The penalty is decayed by reducing the penalty value by one-half at the end of a configurable time
period, called the half-life. Routes that flap many times may reach a maximum penalty level, or ceiling,
after which no additional penalty is added. The ceiling value is not directly configurable, but the
configuration parameter used in practice is the maximum route suppression time. No matter how often
a route has flapped, after it stops flapping, it will again be advertised after the maximum route
suppression time.
Configuring Route Flap Dampening

Using a routing policy, you enable BGP route flap dampening per BGP peer session, for a BGP peer
group, or for a set of routes.
To enable route flap dampening over BGP peer sessions, use the following command:
configure bgp neighbor [all | <remoteaddr>] {address-family [ipv4-unicast | ipv4-

multicast]} dampening {{half-life <half-life-minutes> {reuse-limit <reuse-limit-
number> suppress-limit <suppress-limit-number> max-suppress <max-suppress-minutes>} |
policy-filter [<policy-name> | none]}
To enable route flap dampening for a BGP peer group, use the following command:
configure bgp peer-group <peer-group-name> {address-family [ipv4-unicast | ipv4-

multicast]} dampening {{half-life <half-life-minutes> {reuse-limit <reuse-limit-
number> supress-limit <suppress-limit-number> max-suppress <max-suppress-minutes>}} |
policy-filter [<policy-name> | none]}
You can supply the dampening parameters directly through the command line interface (CLI)
command, or use the command to associate a policy that contains the desired parameters.
Disabling Route Flap Dampening

To disable route flap dampening for a BGP neighbor (disabling the dampening also deletes all the
configured dampening parameters), use the following command:
multicast]} no-dampening
To disable route flap dampening for a BGP peer group, use the following command:
configure bgp peer-group <peer-group-name> no-dampening

BGP Features
Viewing the Route Flap Dampening Configuration

To view the configured values of the route flap dampening parameters for a BGP neighbor, use the
following command:
show bgp [neighbor {detail} | {neighbor} <remoteaddr>]
To view the configured values of the route flap dampening parameters for a BGP peer group, use the
following command:
show bgp peer-group {detail | <peer-group-name> {detail}}
To display the dampened routes, use the following command:
show bgp neighbor <remoteaddr> {address-family [ipv4-unicast | ipv4-multicast]} flap-

statistics {detail} [all | as-path <path-expression> | community [no-advertise | no-
export | no-export-subconfed | number <community_num> | <AS_Num>:<Num> ] | network
[any / <netMaskLen> | <networkPrefixFilter>] {exact} ]
BGP Route Selection

BGP selects routes based on the following precedence (from highest to lowest):
● higher weight
● higher local preference
● shortest length (shortest AS path)
● lowest origin code
● lowest Multi Exit Discriminator (MED)
● route from external peer
● lowest cost to next hop
● lowest routerID
Stripping Out Private AS Numbers from Route Updates

Private AS numbers are AS numbers in the range 64512 through 65534. You can remove private AS
numbers from the AS path attribute in updates that are sent to external BGP (EBGP) neighbors. Possible
reasons for using private AS numbers include:
● The remote AS does not have officially allocated AS numbers.
● You want to conserve AS numbers if you are multihomed to the local AS.
Private AS numbers should not be advertised on the Internet. Private AS numbers can be used only
locally within an administrative domain. Therefore, when routes are advertised out to the Internet, the
routes can be stripped out from the AS paths of the advertised routes using this feature.
To configure private AS numbers to be removed from updates, use the following command:
enable bgp neighbor [<remoteaddr> | all] remove-private-AS-numbers
disable bgp neighbor [<remoteaddr> | all] remove-private-AS-numbers

BGP, OSPF, and RIP can be enabled simultaneously on the switch. Route redistribution allows the
switch to exchange routes, including static and direct routes, between any two routing protocols.
Exporting routes from OSPF to BGP and from BGP to OSPF are discrete configuration functions. To run
OSPF and BGP simultaneously, you must first configure both protocols and then verify the independent
operation of each. Then you can configure the routes to export from OSPF to BGP and the routes to
export from BGP to OSPF.

Exporting routes between any two routing protocols are discrete configuration functions. For example,
you must configure the switch to export routes from OSPF to BGP; and, if desired, you must configure
the switch to export routes from BGP to OSPF. You must first configure both protocols and then verify
the independent operation of each. Then you can configure the routes to export from OSPF to BGP and
the routes to export from BGP to OSPF.
You can use route maps to associate BGP attributes including Community, NextHop, MED, Origin, and
Local Preference with the routes. Route maps can also be used to filter out exported routes.
To enable or disable the exporting of OSPF, RIP, static, and direct (interface) routes to BGP, use the
following commands:
enable bgp export [direct | ospf | ospf-extern1 | ospf-extern2 | ospf-inter | ospf-

intra | rip | static] {address-family [ipv4-unicast | ipv4-multicast]} {export-policy
<policy-name>}
disable bgp export [direct | ospf | ospf-extern1 | ospf-extern2 | ospf-inter | ospf-

intra | rip | static] {address-family [ipv4-unicast | ipv4-multicast]}
Using the export command to redistribute routes complements the redistribution of routes using the
configure bgp add network command. The configure bgp add network command adds the route
to BGP only if the route is present in the routing table. The enable bgp export command redistributes
the specified routes from the routing table to BGP. If you use both commands to redistribute routes, the
routes redistributed using the network command take precedence over routes redistributed using the
export command.
BGP Static Network

ExtremeXOS BGP allows users to add static networks in BGP, which will be redistributed (advertised)
into the BGP domain if there is a corresponding active route in the IP routing table. Users can associate
a policy with the static BGP network to change or to set the route attributes before the route is
advertised to the BGP neighbors.
To create a static BGP network, use the following command:
configure bgp add network {address-family [ipv4-unicast | ipv4-multicast]} <ipaddr>/

<mask_len> {network-policy <policy>}
To delete a static BGP network, use the following command:
configure bgp delete network {address-family [ipv4-unicast | ipv4-multicast]} [all |

<ipaddress/mask length>]

BGP Features
Graceful BGP Restart

It is possible for BGP control functions to restart without disrupting traffic forwarding. Without graceful
restart, adjacent routers will assume that information previously received from the restarting router is
stale and won’t be used to forward traffic to that router. However, in many cases, two conditions exist
that allow the router restarting BGP to continue to forward traffic correctly. The first condition is that
forwarding can continue while the control function is restarted. Most modern router system designs
separate the forwarding function from the control function so that traffic can still be forwarded
independent of the state of the BGP function. Routes learned through BGP remain in the routing table
and packets continue to be forwarded. The second condition required for graceful restart is that the
network remain stable during the restart period. If the network topology is not changing, the current
routing table remains correct. Often, networks can remain stable during the time for restarting BGP.
With graceful BGP restart, routes received from a restarting router are marked as stale, but traffic
continues to be forwarded over these routes during the restart process.
Restarting and Receiver Mode

Routers involved with graceful restart fill one of two roles: the restarting router or the receiving router.
When a receiving router detects that the TCP connection with its peer is reset, it assumes that the peer
has entered graceful restart. If the neighboring routers are configured to help with the graceful restart
(receiver-mode), they will continue to advertise the restarting router as if it was fully functional. Traffic
continues to be routed as though the restarting router is fully functional. The receiver router will
continue in receiver mode until the restarting router re-establishes the TCP session, indicating
successful termination of graceful restart, the restart timers expire. A router can be configured for
graceful restart, and for receiver-mode separately. A router can be a receiver when its neighbor restarts,
and can in turn be helped by a neighbor receiver-mode router if it restarts.
Planned and Unplanned Restarts

Two types of graceful restarts are defined: planned and unplanned. A planned restart would occur if
the software module for BGP was upgraded, or if the router operator decided to restart the BGP control
function for some reason. An unplanned restart would occur if there was some kind of system failure
that caused a remote reboot or a crash of BGP, or an MSM failover occurs. You can decide to configure
a router to enter graceful restart for only planned restarts, for only unplanned restarts, or for both. Also,
you can decide to configure a router to be a receiver only, and not to do graceful restarts itself.
Configuring Graceful BGP Restart

To configure a router to perform graceful BGP restart, use the following command:
configure bgp restart [none | planned | unplanned | both | aware-only]
The address families participating in graceful restart are configured using the following command:
configure bgp restart [add | delete] address-family [ipv4-unicast | ipv4-multicast]
There are three timers that can be configured with the following commands:
configure bgp restart restart-time <seconds>

configure bgp restart stale-route-time <seconds>
configure bgp restart update-delay <seconds>

Graceful BGP Restart Configuration Example

In the following configuration example, EXOS-1 is the restarting BGP router, and EXOS-2 is the
receiving BGP router.
To configure router EXOS-1, use the following commands:
create vlan bgp-restart

configure vlan bgp-restart add port 2:2
configure vlan bgp-restart ipaddress 20.0.0.1/24
enable ipforwarding

configure bgp route-id 20.0.0.1
configure bgp restart both
enable bgp
To configure router EXOS-2, use the following commands:

create vlan bgp-restart
configure vlan bgp-restart add port 2:5
configure vlan bgp-restart ipaddress 20.0.0.2/24
enable ipforwarding

configure bgp route-id 20.0.0.2
configure bgp restart aware-only
enable bgp
You can use the following commands to verify that BGP graceful restart is configured:
show bgp
show bgp neighbor

41 Multicast Routing and Switching

● Configuring IP Multicast Routing Using PIM on page 969
● PIM Configuration Examples on page 970
● Multicast VLAN Registration on page 972
For more information on IP multicasting, refer to the following publications:

● RFC 1112—Host Extension for IP Multicasting
● RFC 2236—Internet Group Management Protocol, Version 2
● PIM-DM Version 2—draft_ietf_pim_v2_dm_03
● RFC 2362—Protocol-Independent Multicast - Sparse Mode (PIM-SM): Protocol Specification
The following URL points to the website for the IETF PIM Working Group:
http://www.ietf.org/html.charters/pim-charter.html
Overview
Multicast routing and switching is the functionality of a network that allows a single host (the multicast
server) to send a packet to a group of hosts. With multicast, the server is not forced to duplicate and
send enough packets for all the hosts in a group. Instead, multicast allows the network to duplicate the
packet where needed to supply the group. Multicast greatly reduces the bandwidth required to send
data to a group of hosts. IP multicast routing is a function that allows multicast traffic to be forwarded
from one subnet to another across a routing domain.
IP multicast routing consists of the following functions:

● A router that can forward IP multicast packets
● A router-to-router multicast routing protocol (for example, Protocol Independent Multicast (PIM))
● A method for the IP host to communicate its multicast group membership to a router (for example,
Internet Group Management Protocol (IGMP))
NOTE
You should configure IP unicast routing before you configure IP multicast routing.
Multicast Static Routes

You can configure the switch to support multicast static routes. Static routes are used to reach networks
not advertised by routers, and are manually entered into the routing table.

Multicast Routing and Switching
IP Multicast Static routes allow users to have multicast paths diverge from the unicast paths; for the
same source or destination tuple, IP multicast traffic can take a different path than IP unicast traffic.
Multicast Static Route Database and Relative Priority

All IP multicast static routes are stored and maintained in a separate route database. An EXOS switch
has three virtual sets of IP routes:
● Unicast routes
● MBGP multicast routes
● Multicast static route (configured manually by the network administrator)
When looking up a multicast source (performing reverse path forwarding), EXOS looks in all three IP
route databases and uses metric to select the route to use. These three lookups can result in three
different routes:
● Static Multicast Route

● Multicast (MBGP) routes
● Unicast Routes
EXOS calculates the priority of the routes by first using the metric. If there are multiple routes with the
same metric, then EXOS compares the table location. IP Multicast static routes rank highest, followed by
the MBGP route, and finally the unicast table. In other words, the relative priority of the route tables for
RPF computation is as follows, in the descending order of the priority:
● Multicast Static route

● MBGP route
● Unicast route
NOTE
Multicast static routes are supported in the IPv4 address family, but not the IPv6 address family.
See “Populating the Routing Table” on page 867 for more information about static routes, dynamic
routes, multiple routes, relative route priorities, and IP route sharing. See configure ipmroute add
and configure ipmroute delete for more information on adding and deleting muliticast static routes.
PIM Overview
The switch supports both dense mode and sparse mode operation. You can configure dense mode or
sparse mode on a per-interface basis. After they are enabled, some interfaces can run dense mode, while
others run sparse mode.
Licensing
To use the complete PIM functionality, you must have at least a Core license installed on your switch.
The BlackDiamond 10808 ships with a Core, or Advanced Core license. Other platforms can be
upgraded to a Core license. See “Software Licenses and Feature Packs” on page 37 for more information
about licensing.

Overview
A subset of PIM, called PIM Edge Mode, is available with an Advanced Edge license.
PIM Edge Mode

PIM Edge Mode is a subset of PIM available on platforms with an Advanced Edge license. There are
only three restrictions on PIM Edge Mode:
● The switch will not act as a candidate RP.
● The switch will not act as a candidate BSR.
● At most, two Active PIM-SM interfaces are permitted. There is no restriction on the number of
Passive interfaces (within the limit of the maximum IP interfaces).
● Only PIM Sparse Mode (PIM-SM) is supported in this mode.
Active PIM interfaces can have other PIM enabled routers on them. Passive interfaces should only have
hosts sourcing or receiving multicast traffic.
PIM Dense Mode

Protocol-Independent Multicast - Dense Mode (PIM-DM) is a multicast routing protocol. PIM-DM is a
broadcast and prune protocol, which allows you to prune and graft multicast routes.
PIM-DM routers perform reverse path multicasting (RPM). However, instead of exchanging its own
unicast route tables for the RPM algorithm, PIM-DM uses the existing unicast routing table for the
reverse path. As a result, PIM-DM requires less system memory.
PIM Sparse Mode

Unlike PIM-DM, Protocol-Independent Multicast - Sparse Mode (PIM-SM) is an explicit join and prune
protocol, and it supports shared trees as well as shortest path trees (SPTs). The routers must explicitly
join the group(s) in which they are interested in becoming a member, which is beneficial for large
networks that have group members that are sparsely distributed.
Using PIM-SM, the router sends a join message to the rendezvous point (RP). The RP is a central
multicast router that is responsible for receiving and distributing the initial multicast packets. You can
configure a dynamic or static RP.
When a router has a multicast packet to distribute, it encapsulates the packet in a unicast message and
sends it to the RP. The RP decapsulates the multicast packet and distributes it among all member
routers.
When a router determines that the multicast rate has exceeded a configured threshold, that router can
send an explicit join to the originating router. When this occurs, the receiving router gets the multicast
directly from the sending router and bypasses the RP.
NOTE
You can run either PIM-DM or PIM-SM per virtual LAN (VLAN).

PIM Mode Interoperation

An Extreme Networks switch can function as a PIM multicast border router (PMBR). A PMBR
integrates PIM-SM and PIM-DM traffic.
When forwarding PIM-DM traffic into a PIM-SM network, the PMBR acts as a virtual first hop and
encapsulates the initial traffic to RP. The PMBR forwards PIM-DM multicast packets to the RP, which,
in turn, forwards the packets to those routers that have joined the multicast group.
The PMBR also forwards PIM-SM traffic to a PIM-DM network, based on the (*.*.RP) entry. The PMBR
sends a (*.*.RP) join message to the RP, and the PMBR forwards traffic from the RP into the PIM-DM
network.
No commands are required to enable PIM mode interoperation. PIM mode interoperation is
automatically enabled when a dense mode interface and a sparse mode interface are enabled on the
same switch.
PIM Source Specific Multicast (PIM-SSM)

PIM-SM works well in many-to-many multicasting situations. For example, in video conferencing, each
participating site multicasts a stream that is sent to all the other participating sites. However, PIM-SM is
overly complex for one-to-many multicast situations, such as multimedia content distribution or
streaming stock quotes. In these and similar applications, the listener is silent and can know the source
of the multicast in advance, or can obtain it. In these situations, there is no need to join an RP, as the
join request can be made directly towards the source.
PIM Source Specific Multicast (PIM-SSM) is a special case of PIM-SM, in which a host will explicitly
send a request to receive a stream from a specific source, rather than from any source. The host must
use IGMPv3 for PIM-SSM, because the ability to request a stream from a specific source first became
available with IGMPv3. The PIM-SSM capable router interprets the IGMPv3 message to initiate a PIM-
SM join towards the source.
PIM-SSM has the following advantages:

● No overhead of switching to the source-specific tree and waiting for the first packet to arrive
● No need to learn and maintain an RP
● Fewer states to maintain on each router
● No need for the complex register mechanism from the source to the RP
● Better security, as stream is forwarded from sources known in advance
PIM-SSM has the following requirement:

● Any host that participates in PIM-SSM must use IGMPv3.
PIM-SSM is designed as a subset of PIM-SM and all messages are compliant with PIM-SM. PIM-SSM
and PIM-SM can coexist in a PIM network; only the last hop router need to be configured for PIM-SSM
if both source and receivers are present all the time. However, to avoid any JOIN delay, it is
recommended that you enable all routers along the (s,g) path for PIM-SSM.
PIM-SSM Address Range. A range of multicast addresses is used for PIM-SSM. Within that address
range, non-IGMPv3 messages will be ignored, and any IGMPv3 exclude messages will be ignored.
These messages will be ignored for all router interfaces, even those not configured for PIM-SSM. By
default there is no PIM-SSM range specified on the router. If you choose the default keyword in the CLI

Overview
when specifying the PIM-SSM range, you configure the range 232.0.0.0/8. You can also choose to
specify a different range for PIM-SSM by using a policy file.
To configure the PIM-SSM address range, use the following command:
configure pim ssm range [default | policy <policy-name>]
IGMP Overview
IGMP is a protocol used by an IP host to register its IP multicast group membership with a router. A
host that intends to receive multicast packets destined for a particular multicast address registers as a
member of that multicast address group. Periodically, the router queries the multicast group to see if
the group is still in use. If the group is still active, a single IP host responds to the query, and group
registration is maintained.
IGMPv2 is enabled by default on the switch, and beginning with release 11.2, ExtremeXOS supports
IGMPv3. However, the switch can be configured to disable the generation of periodic IGMP query
packets. IGMP should be enabled when the switch is configured to perform IP unicast or IP multicast
routing.
IGMPv3, specified in RFC 3376, adds support for source filtering. Source filtering is the ability for a
system to report interest in receiving packets only from specific source addresses, (filter mode include),
or from all sources except for specific addresses, (filter mode exclude). IGMPv3 is designed to be
interoperable with IGMPv1 and IGMPv2.
IGMP Snooping
IGMP snooping is a Layer 2 function of the switch; it does not require multicast routing to be enabled.
In IGMP snooping, the Layer 2 switch keeps track of IGMP reports and only forwards multicast traffic
to that part of the local network that requires it. IGMP snooping optimizes the use of network
bandwidth and prevents multicast traffic from being flooded to parts of the local network that do not
need it.
IGMP snooping is enabled by default on the switch. If IGMP snooping is disabled, all IGMP and IP
multicast traffic floods within a given VLAN. IGMP snooping expects at least one device on every
VLAN to periodically generate IGMP query messages.
When a port sends an IGMP leave message, the switch removes the IGMP snooping entry after 1000
milliseconds (the leave time is configurable, ranging from 0 to 10000 ms). The switch sends a query to
determine which ports want to remain in the multicast group. If other members of the VLAN want to
remain in the multicast group, the router ignores the leave message, but the port that requests removal
is removed from the IGMP snooping table.
If the last port within a VLAN sends an IGMP leave message and the router does not receive any
responses to the query, then the router immediately removes the VLAN from the multicast group.

Static IGMP
To receive multicast traffic, a host must explicitly join a multicast group by sending an IGMP report;
then, the traffic is forwarded to that host. In some situations, you would like multicast traffic to be
forwarded to a port where a multicast-enabled host is not available (for example, when you test
multicast configurations).
Static IGMP emulates a host or router attached to a switch port, so that multicast traffic is forwarded to
that port, and the switch will send a proxy join for all the statically configured IGMP groups when an
IGMP query is received. You can emulate a host to forward a particular multicast group to a port; and
you may emulate a router to forward all multicast groups to a port. Static IGMP is only available with
IGMPv2.
To emulate a host on a port, use the following command:
configure igmp snooping {vlan} <vlanname> ports <portlist> add static group <ip
address>
To emulate a multicast router on a port, use the following command:
configure igmp snooping {vlan} <vlanname> ports <portlist> add static router
To remove these entries, use the corresponding command:
configure igmp snooping {vlan} <vlanname> ports <portlist> delete static

group [<ip_address> | all]
configure igmp snooping vlan <vlanname> ports <portlist> delete static router
To display the IGMP snooping static groups, use the following command:
show igmp snooping vlan <name> static [group | router]
IGMP Snooping Filters

IGMP snooping filters allow you to configure a policy file on a port to allow or deny IGMP report and
leave packets coming into the port. (For details on creating policy files, see “Policy Manager” on
page 375.) IGMP snooping filters is available with IGMPv2 and IGMPv3.
For the policies used as IGMP snooping filters, all the entries should be IP address type entries, and the
IP address of each entry must be in the class-D multicast address space but should not be in the
multicast control subnet range (224.0.0.x/24).
Use the following template to create a snooping filter policy file that denies IGMP report and leave
packets for the 239.11.0.0/16 and 239.10.10.4/32 multicast groups:
#
# Add your group addresses between "Start" and "end"
# Do not touch the rest of the file!!!!
entry igmpFilter {
if match any {
#------------------ Start of group addresses ------------------
nlri 239.11.0.0/16;
nlri 239.10.10.4/32;
#------------------- end of group addresses -------------------

Configuring IP Multicast Routing Using PIM
} then {
deny;
}
}
entry catch_all {
if {
} then {
permit;
}
}
After you create a policy file, use the following command to associate the policy file and filter a set of
ports:
configure igmp snooping vlan <vlanname> ports <portlist> filter [<policy> | none]
To remove the filter, use the none option.
To display the IGMP snooping filters, use the following command:
show igmp snooping {vlan} <name> filter

To configure IP multicast routing:
1 Configure the system for IP unicast routing.
2 Enable multicast routing on the interface using the following command:
enable ipmcforwarding {vlan <name>}
3 Enable PIM on all IP multicast routing interfaces using the following command:
configure pim add vlan [<vlan-name> | all] {dense | sparse} {passive}
4 For PIM-SSM, specify the PIM-SSM range, specify IGMPv3, and enable PIM-SSM on the interfaces
using the following commands:
configure pim ssm range [default | policy <policy-name>]
enable igmp {vlan <vlan name>} {IGMPv1 | IGMPv2 | IGMPv3}
enable pim ssm vlan [<vlan_name> | all]
5 Enable PIM on the router using the following command:
enable pim

PIM Configuration Examples

Figure 147 and Figure 148 are used in Chapter 10, “OSPF,” to describe the Open Shortest Path First
(OSPF) configuration on a switch. See Chapter 10 for more information about configuring OSPF.
PIM-DM Configuration Example

In Figure 147, the system labeled IR 1 is configured for IP multicast routing, using PIM-DM.
Figure 147: IP multicast routing using PIM-DM configuration example
Area 0 IR 2 IR 1
10.0.1.1 10.0.1.2
10.0.3.2 10.0.2.2
2
HQ
Headquarters
0_
_
0_
10
_1
_0
HQ
_3
ABR 2 10.0.3.1 ABR 1 10.0.2.1
161.48.2.2
Los Angeles
LA
161.48.2.1
_1
160.26.25.1 160.26.26.1
26
61
_4
6_
8_
_2
Virtual link
2
60
i_1
Ch
160.26.26.2
160.26.25.2 Chicago

EX_040
The router labeled IR1 has the following configuration:

enable ipforwarding
enable ospf
enable ipmcforwarding
configure pim add vlan all dense
enable pim

PIM-SM Configuration Example

In Figure 148, the system labeled ABR1 is configured for IP multicast routing using PIM-SM.
Figure 148: IP multicast routing using PIM-SM configuration example
Area 0 IR 2 IR 1
10.0.1.1 10.0.1.2
10.0.3.2 10.0.2.2
HQ_10_10_4
2
HQ
Headquarters
0_
_1
0_
0_
_1
0_
HQ
3
ABR 2 10.0.3.1 ABR 1 10.0.2.1
Rendezvous
point
161.48.2.2
Los Angeles
LA
161.48.2.1
_1
160.26.25.1 160.26.26.1
26
61
_4
6_
8_
_2
Virtual link
2
60
i_1
Ch
160.26.26.2
160.26.25.2 Chicago

EX_062
The router labeled ABR1 has the following configuration:

configure vlan LA_161_48_2 ipaddress 161.48.2.2 255.255.255.0
configure vlan CHI_160_26_26 ipaddress 160.26.26.1 255.255.255.0
enable ipforwarding
configure pim add vlan all sparse
tftp TFTP_SERV -g -r rp_list.pol
configure pim crp HQ_10_0_3 rp_list 30
configure pim cbsr HQ_10_0_3 30

The policy file, rp_list.pol, contains the list of multicast group addresses serviced by this RP. This set
of group addresses are advertised as candidate RPs. Each router then elects the common RP for a group
address based on a common algorithm. This group to RP mapping should be consistent on all routers.
The following is a policy file which will configure the CRP for the address ranges 239.0.0.0/24 and
232.144.27.0:
entry extreme1 {
if match any {
}
then {
nlri 239.0.0.0/24 ;
nlri 232.144.27.0/24 ;
}
}
PIM-SSM Configuration Example

In the following example, the default PIM-SSM range of 232.0.0.0/8 is configured. For all interfaces,
non-IGMPv3 messages and IGMPv3 exclude messages will be ignored for addresses in this range. Hosts
that use IGMPv3 on VLAN v13 can request and receive source specific multicast streams for addresses
in the PIM-SSM range.
create vlan v12
create vlan v13
configure v12 add port 1
configure v13 add port 2
configure v12 ipaddress 12.1.1.1/24
configure v13 ipaddress 11.1.1.1/24
configure pim add vlan all sparse
enable ipforwarding
enable igmp IGMPv3
configure pim ssm range default
enable pim ssm vlan v13
enable pim
Multicast VLAN Registration

Multicast VLAN Registration (MVR) is designed to support distributing multicast streams for IPTV to
subscribers over a layer 2 network. In a standard layer 2 network, a multicast stream received on a
VLAN is not forwarded to another VLAN. The streams are confined to the layer 2 broadcast domain. In
an IGMP snooping environment, streams are forwarded only to interested hosts on a VLAN. For inter-
VLAN forwarding (routing) a multicast routing protocol, such as PIM/DVMRP must be deployed.
MVR breaks this basic rule, so that a stream received over layer 2 VLANs is forwarded to another
VLAN, eliminating the need for a layer 3 routing protocol. It simplifies the multicast stream distribution
and is a better solution for IPTV-like services. With MVR, a multicast stream is forwarded to all VLANs
containing interested hosts.

Figure 149: Standard VLAN Compared to an MVR VLAN
Stream to Group G Stream to Group G
Vlan1
Vlan1 (MVR Enabled)
Switch1 Switch1
H1 subscribed to G H1 subscribed to G
Vlan2 Vlan2
H2 subscribed to G H2 subscribed to G EX_142
In Figure 149, the left side shows a standard VLAN carrying a multicast stream. The host H1 receives
the multicast stream because it resides on VLAN Vlan1, but host H2 does not receive the multicast
traffic because no IP multicast routing protocol forwards the stream to VLAN Vlan2. On the right side
of the figure, H2 does receive the multicast stream. Because Vlan1 was configured as an MVR VLAN,
the multicast stream is forwarded to the other VLANs on the switch. containing hosts that have
requested the stream. To configure a VLAN as an MVR VLAN, use the following command:
configure mvr add vlan <vlan-name>
Typically, IGMP snooping is enabled, so only hosts that have requested a stream will see the multicast
traffic. For example, another host on Vlan2 would not receive the traffic unless it had sent an IGMP
request to be included in the group.
Notice that only Vlan1 is MVR enabled. Configure MVR only on the ingress VLAN. To enable MVR on
the switch, use the following command:
enable mvr
Basic MVR Deployment

Since MVR is primarily targeted for IPTV and similar applications, a basic deployment for that
application is shown in Figure 150. In the figure, an IPTV server is connected through a router to a
network of switches. Switch 1 has three customer VLANs, Vlan2, Vlan3, and Vlan4. The multicast
streams are delivered through the network core (Metro Ethernets), which often use a ring topology and
some kind of redundant protection to provide high availability. For example, McastVlan forms a ring
through switches Switch1 through Switch4. The link from Switch2 to Switch4 is shown as blocked, as it
would be if some form of protection, like EAPS were being used.

Figure 150: Basic MVR Deployment
IPTV server
Switch3
McastVlan
Switch4 Switch2
McastVlan
McastVlan
Switch1
Vlan2 Vlan4
H2 H3 H4 EX_143
Without MVR, there are two ways to distribute multicast streams in this topology:
● Extend subscriber VLANs (Vlan2, Vlan3, and Vlan4) to the network core, by tagging the ports
connecting the switches.
● Configure all VLANS with an IP address and run PIM or DVMRP on each switch.
There are problems with both of these approaches. In the first approach, multiple copies of the same
stream (IPTV channel) would be transmitted in the core, wasting bandwidth. In the second approach,
all switches in the network core would have to be layer 3 multicast aware, and would have to run a
multicast protocol. Typical network cores are layer 2 only.
MVR provides a simple solution to this problem. If McastVlan in Switch1 is configured with MVR, it
will leak the traffic into the local subscriber VLANs that contain hosts that request the traffic. For simple
cases, perform these configuration steps:
1 Configure MVR on McastVlan.
2 Configure an IP address and enable IGMP and IGMP snooping on the subscriber VLANs (by default
IGMP and IGMP snooping are enabled on Extreme Networks’ switches).
3 For all the multicast streams (IPTV channels), configure static IGMP snooping membership on the
router on McastVlan.
4 Enable MVR on the switches in the network.
This strategy conserves bandwidth in the core and does not require running PIM on the subscriber
switches.
In this topology, a host (for example, a cable box or desktop PC) will join a channel through an IGMP
join message. Switch1 snoops this message and adds the virtual port to corresponding cache's egress

list. This is possible because an MVR enabled VLAN can leak traffic to any other VLAN. When the user
switches to another channel, the host sends an IGMP leave for the old channel and a join for the new
channel. The corresponding virtual port is removed from the cache for the old channel and is added to
the cache for the new channel.
As discussed in “Static and Dynamic MVR,” McastVlan will also proxy IGMP joins learned on other
VLANs to the router. On an MVR network it is not mandatory to have a router to serve the multicast
stream. All that is required is to have a designated IGMP querier on McastVlan. The IPTV server could
also be directly connected to McastVlan.
Static and Dynamic MVR

Static MVR: In a typical IPTV network, there would be several high demand basic channels. At any
instant there would at least be one viewer for each of these channels (streams), and they should always
be available at the core. When user requests one of these channels, it is quickly pulled locally from the
multicast VLAN. These streams are always available on the core multicast VLAN through static
configuration on the streaming router. For those channels, the layer 2 switch will not send any IGMP
join messages towards the IGMP querier. These groups must be statically configured on the streaming
router so that these streams are always available on the ring. For example, on an Extreme Networks
router, you would use the following command:
configure igmp snooping {vlan} <vlanname> ports <portlist> add static group <ip
address>
If a multicast packet for a group in the static MVR range is received on an MVR enabled VLAN, it will
always be flooded on the MVR VLAN. This allows the neighbor switch in the ring to receive all the
static MVR streams.
Dynamic MVR: In contrast, since a video content provider would like to provide a variety of on-demand
and other premium channels, there are often many lower demand (fewer viewers) premium channels
that cannot all be made available simultaneously at the core network. These should be streamed from
the router only if requested by a host.
IGMP is the standard method used by a host to request a stream. However, IGMP packets are
constrained to a VLAN. Thus, subscribers' IGMP join requests on the VLAN cannot be forwarded onto
other VLANS. With MVR, a VLAN sends proxy IGMP join messages on behalf of all the subscriber
VLANs on the switch. Thus, in Figure 150, McastVlan would send join and leave messages on behalf of
Vlan2, Vlan3, and Vlan4. The router receives the messages on McastVlan and streams corresponding
channels onto the core network. This provides on-demand service, and an administrator doesn't need to
configure static IGMP on the router for each of these channels.
Configuring Static and Dynamic MVR: By default, all MVR streams are static. You can specify which
groups are static by using the following command:
configure mvr vlan <vlan-name> static group {<policy-name> | none}
Any other groups in the MVR address range are dynamic. You can specify the MVR address range by
configure mvr vlan <vlan-name> mvr-address {<policy-name> | none}
By using these two commands together, you can specify which groups are static and which are
dynamic. If you want all the groups to be dynamic, specify a policy file for the static group command
that denies all multicast addresses.

MVR Forwarding
The goal for MVR is to limit the multicast traffic in the core layer 2 network to only the designated
multicast VLAN. If the backbone layer 2 port is tagged with multiple VLANs, as shown in Figure 151, a
set of rules is needed to restrict the multicast streams to only one VLAN in the core.
Figure 151: Multiple VLANs in the Core Network
McastVlan, vc1, vc2
McastVlan, vc1, vc2

PC1 PC2
Switch1
Vlan2 p1 p2 vc2
H2 H3 H4 EX_144
In Figure 151, the core network has 2 more VLANs, vc1 and vc2, to provide other services. With MVR,
multicast traffic should be confined to McastVlan, and should not be forwarded to vc1 and vc2. It
should be noted that MVR is configured only on the ingress VLAN (McastVlan). MVR is not configured
on any other VLANs.
In the same way as the IGMP snooping forwarding rules, the multicast stream is forwarded onto
member ports and router ports on the VLAN. For a stream received on MVR enabled ports, this rule is
extended to extend membership and router ports to all other VLANs. This rule works well on the
topology in Figure 150. However, in a tagged core topology, this rule forwards traffic onto VLANs, such
as vc1 and vc2, on ports PC1 and PC2. This results in multiple copies of same stream on the core
network, thus reintroducing the problem that MVR was intended to solve.
To avoid multiple copies of the same stream, MVR forwards traffic with some special restrictions. MVR
traffic is not forwarded to another VLAN unless a host is detected on the port. On the ingress MVR
VLAN, packets are not duplicated to ports belonging to MVR VLANs. This is to prevent duplicate
multicast traffic streams on ingress ports. Streams belonging to static MVR groups will always be
forwarded on MVR VLANs so that any host can join such channels immediately. However, dynamic
groups will be streamed from the server only if some host is interested in them. A command is
provided to receive traffic on a port which is excluded by MVR. However, regular IGMP rules still
apply to these ports, so the ports must have a router connected or an IGMP host to receive the stream.
These rules are to prevent multicast packets from leaking into undesired virtual port, such as p2 on
VLAN pc2 in Figure 151. These rules also allow that, in most topologies, MVR could be deployed with
minimal configuration. However, unlike EAPS/STP, MVR is not intended to be a layer 2 protocol to
solve packet looping problems. Since multicast packets leak across VLANs, one can misconfigure and
end up with a multicast storm. MVR does not attempt to solve such problems.
NOTE
If a port were blocked by layer 2 protocols, that port would be removed from the egress list of the cache. This is
done dynamically per the port state.

For most situations, you will not need to manually configure ports to receive the MVR multicast
streams. But if one of the forwarding rules denies forwarding to a port that requires the streams, you
can manually receive the MVR multicast streams by using the following command:
configure mvr vlan <vlan-name> add receiver port <port-list>
Inter-Multicast VLAN Forwarding

In “Basic MVR Deployment” only simple topologies are considered, in which subscribers on different
VLANs access a multicast VLAN. There are topologies where streams need to be forwarded onto
another multicast VLAN, as shown in Figure 152. In this figure, a Multicast Service Provider (MSP)
multicast VLAN is attached to ports 1:1-2 on both switches, SW1 and SW2. On the customer side,
another multicast VLAN, delivers multicast streams to other switches around the ring.
Figure 152: Inter-Multicast VLAN Forwarding
MSP ring
1:1 1:2 1:1 1:2
SW1 SW2
1:4 1:5 1:5
Sw6
Customer ring
CustVlan
1:3 1:4
H2 H3 H4 EX_145
In this topology, a multicast stream could be leaked into the customer multicast network through either
switch SW1 or SW2. However, as described in “MVR Forwarding,” packets will not be forwarded to
router ports (ports 1:4 and 1:5 could be router ports if SW2 is an IGMP querier). To get around this,
MVR needs to be configured on CustVlan either on SW1 or SW2. Since the forwarding rules apply only
to non-MVR VLANs, traffic from one MVR VLAN will be leaked into the router ports of another
VLAN, if MVR is enabled on that.
In the topology above, the MSP multicast VLAN is carried on two switches that also carry the customer
multicast VLAN. When multiple switches carry both multicast VLANs, it is imperative that MVR is
configured on only one switch. Only that switch should be used as the transit point for multicast
streams from one multicast ring into another. Otherwise, duplicate packets will be forwarded. Also on
the non-MVR switches, the ring ports should be configured as static router ports, so that ring ports are
excluded from forwarding packets onto the customer ring. There is no mechanism to elect a designated
MVR forwarder, so it must be configured correctly.

MVR Configurations
MVR enables layer 2 network installations to deliver bandwidth intensive multicast streams. It is
primarily aimed at delivering IPTV over layer 2 networks, but would be valuable in many existing
EAPS or STP installations. This section explores a few possible deployment scenarios and configuration
details. Of course, real world networks could be lot different from these examples. This section is meant
to present some ideas on how to deploy MVR over existing networks, as well as to design new
networks that support MVR.
MVR with EAPS

Since MVR is designed with a layer 2 ring topology in mind, it is strongly recommended that it should
be deployed with EAPS. The MVR plus EAPS combination provides a superior solution for any triple
play network, where service provider intends to provide data, voice, and video services. EAPS is a
proven solution for providing sub-second SONET-like protection to layer 2 rings. For more detail on
EAPS refer to Chapter 19, “Ethernet Automatic Protection Switching.”
Consider a typical EAPS topology in Figure 153, where 3 VLANs on the core ring serve various clients
on each switch. To provide video service, one of the VLANs on the EAPS ring is designated as a
multicast VLAN. MVR is enabled only on this VLAN (mcastvlan). V1 is the control VLAN, and V2 is
another protected VLAN. A router serving the multicast feed would typically be running PIM on
mcastvlan, to support the static and dynamic IGMP membership on the VLAN.
Figure 153: MVR on an EAPS Ring
IPTV server
McastVlan
1:1
Router1 1:2 1:3
McastVlan, Switch3
V1, V2
Switch4 EAPS ring Switch2
McastVlan, V1, V2
1:1 1:2
Switch1
Vlan2 Vlan4
1:3 1:4
H2 H3 H4 EX_146

The following is a typical configuration for the router and switches.
Router1:
create vlan mcastvlan
configure mcastvlan add port 1:1
create vlan server
configure server add port 1:2
configure mcastvlan ipaddress 10.1.1.1/24
configure server ipaddress 11.1.1.1/24
configure igmp snooping mcastvlan port 1:1 add static group 239.1.1.1
enable ipforwarding
configure igmp snooping leave-timeout 2000
configure pim add vlan all
enable pim
Switch1:
create vlan mcastvlan
create vlan v1
create vlan v2
create vlan vlan2
configure vlan vlan2 add port 1:3
configure mcastvlan tag 20
configure mcastvlan add port 1:1,1:2 tag
configure mvr add vlan mcastvlan
configure v1 add port 1:1,1:2 tag
create eaps e1
configure eaps e1 mode transit
configure eaps e1 add control vlan v1
configure eaps e1 add protect vlan mcastvlan
configure eaps e1 add protect vlan v2
configure eaps port primary port 1:1
configure eaps port secondary port 1:2
enable eaps
enable mvr
Switch3:
create vlan McastVlan

create vlan v1
create vlan v2
configure mcastvlan tag 20
configure mcastvlan add port 1:2,1:3 tag
configure mcastvlan add port 1:1
configure mvr add vlan mcastvlan

create eaps e1
configure eaps e1 mode master
configure eaps e1 add control vlan v1
configure eaps e1 add protect vlan mcastvlan
configure eaps e1 add protect vlan v2
configure eaps port primary port 1:3
configure eaps port secondary port 1:2
enable eaps
enable mvr
NOTE
In this example, Switch3 is the EAPS master, but any of the other switches in the ring could have been configured
as the master.
MVR with STP

In a layer 2 ring topology, MVR works with STP as it works with EAPS. However, in other layer 2
topologies, additional configuration steps may be needed to make sure that multicast feeds reaches all
network segments. Extra configuration is required because all ports in the VLAN are part of an STP
domain, so that solely by examining the configuration it is not clear whether a port is part of bigger
ring or is just serving a few hosts. In EAPS this problem is solved by distinguishing between configured
primary or secondary ports from other VLAN ports. Consider a simplified layer 2 STP network as
shown in Figure 154.
Figure 154: MVR with STP
Switch4 MSP ring Switch2
MVlan, vc1 MVlan, vc1
Switch1
1:3 1:4
V1 V1
Vlan V1 cloud
EX_148
In this topology, subscribers are in a layer 2 cloud on VLAN V1. STP is configured for all ports of V1.
Since V1 spans on the ring as well, multicast cannot be forwarded on V1 blindly. Forwarding rules
(described in “MVR Forwarding”), dictates that multicast traffic will not be forwarded on STP enabled
ports. This is to make sure that multiple copies of multicast packets are not forwarded on the ring.

However, since other STP enabled ports on V1 (1:3,1:4) are not part of the ring multicast stream, they
need to be configured so that they get the packets. To configure the ports to receive packets, use the
following command (mentioned previously in “MVR Forwarding”):
configure mvr vlan <vlan-name> add receiver port <port-list>
NOTE
If the layer 2 cloud is connected back to ring ports, traffic may end up leaking into VLAN V1 in the ring. There is no
way to avoid that. So, such topologies must be avoided.
The following is a typical configuration:
Switch1:
create vlan v1
configure v1 tag 200
configure v1 add port 1:1, 1:2 tag
configure v1 add port 1:3, 1:4
create vlan mvlan
configure mvlan add port 1:1, 1:2
configure mvr add vlan mvlan
create stpd stp1
configure stp1 add vlan v1 port all
enable stpd stp1 port all
configure mvr vlan v1 add receiver port 1:3,1:4
enable mvr
MVR in a vMAN Environment

In the case of a vMAN, a packet is tagged with a vMAN tag in addition to a possible VLAN tag. This is
to provide VLAN aggregation for all customer traffic in the vMAN ring. Each customer would be given
its own VLAN, and traffic from all customers could be tunneled on single vMAN tag into the metro
ring to an outside Broadband Remote Access Server (BRAS). In a vMAN network, multicast traffic
could be distributed over a separate VLAN in the metro core. These packets are not subjected to vMAN
tunneling. Thus, IPTV service could be provided on this multicast VLAN on a vMAN network.
MVR deployment in a vMAN environment is not any different from that in an EAPS environment,
since a separate multicast VLAN on the metro ring is used for multicasting. However, it provides
interesting capabilities to MSPs for their video offerings. Different service bundles could be offered on
separate VLANs. Packets will not be forwarded to any metro link segments where a stream is not
required.
Figure 155 illustrates an example design for MVR in a vMAN environment. Any multicast packet
entering on MVlan will be forwarded on MVlan to the next switch. These multicast packets will not be
tunneled. When a packet reaches Switch1 on MVlan, the only way to forward them to a Customer Edge
(CE) port is to run PIM on MVlan and vMAN, or to configure MVR (the CE port belongs to the vMAN
and is an untagged port, other ports on ring are vMAN tagged ports). The drawback of running PIM is
that packets will be forwarded to all segments of the ring, and it will not scale well.
With MVR, switches on the vMAN do not have to run any routing protocol. If MVR is enabled on the
multicast VLAN, MVlan, traffic will be pulled from the IPTV server. Such multicast packets egressing
from the CE port will always be untagged. The downstream DSLAM will distribute untagged multicast
packets to the respective subscriber VLANs.

Figure 155: MVR in a vMAN Environment
IPTV server
MVlan
1:1
To BRAS
Router1 1:2 1:3
Switch3
Switch4 Provider's VMAN ring Switch2
MVlan, VMAN MVlan, VMAN

Tag VMAN port
1:1 1:2
Switch1
2:2 2:3
Untag CE port
User VMAN User VMAN

MVlan MVlan
IGMP
DSLAM DSLAM
EX_149
The following is a typical configuration.
Switch1:
create vman vman2
configure vman vman2 add port 2:2-2:3 untagged
configure vman vman2 add port 1:1,1:2 tagged
enable port 2:*
enable port 1:*
create vlan mvlan
configure vlan mvlan tag 200
configure vlan mvlan add port 1:1,1:2 tag
configure mvr add vlan mvlan
enable mvr

42 IPv6 Multicast

● MLD Overview on page 985
● MLD Snooping on page 985
● Static MLD on page 986
Overview
IPv6 multicast is a function that allows a single IPv6 host to send a packet to a group of IPv6 hosts.
NOTE
In the current release of ExtremeXOS IPv6 multicast packets are flooded to VLANs that receive the traffic; therefore,
MLD snooping is not supported. Also, ExtremeXOS does not currently support any IPv6 multicast routing protocols.
The hardware passes all IPv6 multicast packets to the software where forwarding is done via the slow path.
MLD Overview
MLD is a protocol used by an IPv6 host to register its IP multicast group membership with a router.
Periodically, the router queries the multicast group to see if the group is still in use. If the group is still
active, a single IP host responds to the query, and group registration is maintained.
MLD is the IPv6 equivalent to IGMP. MLDv1 is equivalent to IGMPv2, and MLDv2 is equivalent to
IGMPv3. MLDv1 is currently supported on the switch.
MLD Snooping
MLD snooping is a Layer 2 function of the switch; it does not require multicast routing to be enabled.
In MLD snooping, the Layer 2 switch keeps track of MLD reports and only forwards multicast traffic to
that part of the local network that requires it. MLD snooping optimizes the use of network bandwidth
and prevents multicast traffic from being flooded to parts of the local network that do not need it.
MLD snooping is enabled by default on the switch. If MLD snooping is disabled, all MLD and IP
multicast traffic floods within a given VLAN. MLD snooping expects at least one device on every
VLAN to periodically generate MLD query messages.
When a port sends an MLD done message, the switch removes the MLD snooping entry after 1000
milliseconds (the leave time is configurable, ranging from 0 to 10000 ms). The switch sends a query to
determine which ports want to remain in the multicast group. If other members of the VLAN want to
remain in the multicast group, the router ignores the done message, but the port that requests removal
is removed from the MLD snooping table.

IPv6 Multicast
If the last port within a VLAN sends an MLD done message and the router does not receive any
responses to the query, then the router immediately removes the VLAN from the multicast group.
Static MLD
To receive multicast traffic, a host must explicitly join a multicast group by sending an MLD report;
then, the traffic is forwarded to that host. In some situations, you would like multicast traffic to be
forwarded to a port where a multicast-enabled host is not available (for example, when you test
multicast configurations).
Static MLD emulates a host or router attached to a switch port, so that multicast traffic is forwarded to
that port, and the switch will send a proxy join for all the statically configured MLD groups when an
MLD query is received. You can emulate a host to forward a particular multicast group to a port; and
you may emulate a router to forward all multicast groups to a port.
To emulate a host on a port, use the following command:
configure mld snooping vlan ports add static group
To emulate a multicast router on a port, use the following command:
configure mld snooping vlan ports add static router
To remove these entries, use the corresponding command:
configure mld snooping vlan ports delete static group
configure mld snooping vlan ports delete static router
To display the MLD snooping static groups, use the following command:
show mld snooping vlan static

43 Multicast Source Discovery Protocol (MSDP)

● PIM Border Configuration on page 988
● MSDP Peers on page 988
● MSDP Mesh-Groups on page 990
● SA Cache on page 991
● Redundancy on page 992
● Scaling Limits on page 993
● SNMP MIBs on page 993
NOTE
For more information about MSDP, refer to RFC 3618.
Overview
MSDP is an interdomain multicast protocol used to connect multiple multicast routing domains that run
Protocol Independent Multicast-Sparse Mode (PIM-SM). MSDP speakers are configured on each PIM-
SM domain. These speakers establish a peering relationship with other MSDP speakers through secured
TCP connections. When the source sends traffic, the MSDP speaker learns about the source through its
Rendezvous Point (RP). In turn, the RP advertises the source to its peers through Source Advertisement
(SA) messages. The peers receive these advertisements and inform their RPs about the presence of the
active source in the other PIM-SM domain, which triggers the normal PIM operation in the
corresponding domains.
For example, as businesses expand and networks grow in size, it might become necessary to connect
PIM domains to allow multicast applications to reach other offices across the network. MSDP simplifies
this process by providing a mechanism to connect those multicast routing domains without
reconfiguring existing domains. Each PIM domain remains separate and has its own RP. The RP in each
domain establishes an MSDP peering relationship over a TCP connection either with RPs in other
domains or with border routers leading to other domains. When an RP learns about a new multicast
source in its own domain (using the normal PIM registration process), it then sends a Source-Active
(SA) message to all of its MSDP peers, letting them know about the new stream. In this way, the
network can receive multicast traffic from all over the network without having to reconfigure each
existing PIM domain.

Multicast Source Discovery Protocol (MSDP)
Supported Platforms
MSDP is supported on all platforms running a minimum software version of ExtremeXOS 12.0 with the
Core license (or higher).
Our implementation of MSDP is compliant with RFC 3618, and is compatible with other devices that are
compliant with this standard.
Limitations
The limitations of MSDP are as follows:
● There is no support for MSDP operating with SA cache disabled (transit node). MSDP will always
cache/store received SA messages.
● There is no support for anycast RP or logical RP.
● There is no support for MSDP on user-created virtual routers (VRs).
● A separate multicast routing table is not available in ExtremeXOS. So, MSDP has to rely on unicast
routes (or a unicast topology) to perform peer reverse-path-forwarding (RPF) checks.
● RIP routes are not used for peer-RPF checking. So, our implementation of MSDP does not exactly
conform to rule (iii) in section 10.1.3 of RFC-3618. However, our implementation of MSDP uses
BGP/OSPF for peer-RPF checking as per rule (iii) in section 10.1.3.
● Read-write/read-create access is not supported on MSDP MIB objects.
PIM Border Configuration

To create a PIM-SM domain for MSDP, you must restrict the reach of Bootstrap Router (BSR)
advertisements by defining a VLAN border. BSR advertisements are not sent out of a PIM interface
configured as a VLAN border, thereby defining a PIM domain for MSDP.
To configure a PIM VLAN border, use the following command:
configure pim <vlan_name> border
MSDP Peers
MSDP peers exchange messages to advertise active multicast sources. The peer with the higher IP
address passively listens to a well-known port number and waits for the side with the lower IP address
to establish a Transmission Control Protocol (TCP) connection on port 639. When a PIM-SM RP that is
running MSDP becomes aware of a new local source, it sends an SA message over the TCP connection
to its MSDP peer. When the SA message is received, a peer-RPF check is performed to make sure the
peer is toward the originating RP. If so, the RPF peer floods the message further. If not, the SA message
is dropped and the message is rejected.
To configure an MSDP peer, use the following command:
create msdp peer <remoteaddr> {remote-as <remote-AS>} {vr <vrname>}

MSDP Peers
To delete an MSDP peer, use the following command:
delete msdp peer [all | <remoteaddr>] {vr <vrname>}
To display configuration and run-time parameters about an MSDP peer, use the following command:
show msdp [peer {detail} | {peer} <remoteaddr>] {vr <vrname>}
MSDP Default Peers

You can configure a default peer to accept all SA messages. Configuring a default peer simplifies the
peer-RPF checking of SA messages. If no policy is specified, the current peer is the default RPF peer for
all SA messages.
When configuring a default peer, you can also specify an optional policy filter. If the peer-RPF check
fails, and a policy filter is configured, the default peer rule is applied to see if the SA message should be
accepted or rejected.
You can configure multiple default peers with different policies. However, all default peers must either
be configured with a default policy or not. A mix of default peers, with a policy and without a policy, is
not allowed.
To configure an MSDP default peer, and optional policy filter, use the following command:
configure msdp [{peer} <remoteaddr> | peer all] default-peer {default-peer-policy

<filter-name>} {vr <vrname>}
To remove the default peer, use the following command:
configure msdp [{peer} <remoteaddr> | peer all] no-default-peer {vr <vrname>}
To verify that a default peer is configured, use the following command:
Peer Authentication
MSDP supports TCP MD5 authentication (RFC-2385) to secure control messages between MSDP peers.
You must configure a secret password for an MSDP peer session to enable TCP MD5 authentication.
When a password is configured, MSDP receives only authenticated MSDP messages from its peers. All
MSDP messages that fail TCP MD5 authentication are dropped.
To configure TCP MD5 authentication on an MSDP peer, use the following command:
configure msdp [{peer} <remoteaddr> | peer all] password [none | {encrypted}

<tcpPassword>] {vr <vrname>}
To remove the password, use the following command:
configure msdp peer {all | <remoteaddr>} password none
The password displays in encrypted format and cannot be seen as simple text. Additionally, the
password is saved in encrypted format.

To display the password in encrypted format, use the following command:
Policy Filters
You can configure a policy filter to control the flow of SA messages going to or coming from an MSDP
peer. For example, policy filters can help mitigate state explosion during denial of service (DoS) or other
attacks by limiting what is propagated to other domains using MSDP.
To configure an incoming or outgoing policy filter for SA messages, use the following command:
configure msdp [{peer} <remoteaddr> | peer all] sa-filter [in | out] [<filter-name>
| none] {vr <vrname>}
To remove a policy filter for SA messages, use the none keyword:
configure msdp [{peer} <remoteaddr> | peer all] sa-filter [in | out] none
To verify that a policy filter is configured on an MSDP peer, use the following command:
SA Request Processing
You can configure the router to accept or reject SA request messages from a specified MSDP peer or all
peers. If an SA request filter is specified, only SA request messages from those groups permitted are
accepted. All others are ignored.
To configure the router to accept SA request messages from a specified MSDP peer or all peers, use the
following command:
enable msdp [{peer} <remoteaddr> | peer all] process-sa-request {sa-request-filter

<filter-name> } {vr <vrname>}
To configure the router to reject SA request messages from a specified MSDP peer or all peers, use the
following command:
disable msdp [{peer} <remoteaddr> | peer all] process-sa-request {vr <vrname>}
To display configuration and run-time parameters about MSDP peers, use the following command:
MSDP Mesh-Groups
MSDP can operate in a mesh-group topology. A mesh-group limits the flooding of SA messages to
neighboring peers. In a mesh-group, every MSDP peer must be connected directly to every other peer
in the group. In this fully-meshed topology, when a SA message is received from a member of the
mesh-group, the SA message is always accepted, but not flooded to other members of the same group.
Because MSDP peers are connected directly to every other peer in the mesh-group, an MSDP peer is not

SA Cache
required to forward SA messages to other members of the same mesh-group. However, SA messages
are flooded to members of other mesh-groups. An MSDP mesh-group is an easy way to implement
inter-domain multicast, as it relaxes the requirement to validate looping of MSDP control traffic (that is,
peer-RPF checking is not required). Consequently, SA messages do not loop in the network.
To configure an MSDP mesh-group, use the following command:
create msdp mesh-group <mesh-group-name> {vr <vrname>}
To remove an MSDP mesh-group, use the following command:
delete msdp mesh-group <mesh-group-name> {vr <vrname>}
To display information about an MSDP mesh-group, use the following command:
show msdp [mesh-group {detail} | {mesh-group} <mesh-group-name>] {vr <vrname>}
To configure a peer to be a member of an MSDP mesh-group, use the following command:
configure msdp [{peer} <remoteaddr> | peer all] mesh-group [<mesh-group-name> |

none] {vr <vrname>}
To remove a peer from an MSDP mesh-group, use the following command:
configure msdp [{peer} <remoteaddr> | peer all] mesh-group none {vr <vrname>}
SA Cache
As an MSDP router learns of new sources either through a PIM-SM Source-Register (SR) message or SA
message from its RPF peer, it creates an entry in SA cache (or refreshes the entry if it is already there)
and forwards this information to its peers. These entries are refreshed by periodic SA messages received
from the MSDP peers. If these entries are not refreshed within six minutes, they will time out. When a
PIM-SM RP detects that the source is no longer available it informs MSDP, which in turn removes the
SA information from the local database.
Caching makes it easy for local receivers to know immediately about inter-domain multicast sources
and to initiate building a source tree towards the source. However, maintaining a cache is heavy both in
CPU processing and memory requirements.
NOTE
Our implementation of MSDP does not support operating with local cache disabled.
To remove an SA cache server, use the following command:
unconfigure msdp sa-cache-server {vr <vrname>}
As MSDP uses the flood-and-join model to propagate information about sources, there is a restriction
that no more than two advertisements per cache entry will be forwarded per advertisement interval.
This is helpful in reducing an SA message storm and unnecessarily forwarding them to peers.

By default, the router does not send SA request messages to its MSDP peers when a new member joins
a group and wants to receive multicast traffic. The new member simply waits to receive SA messages,
which eventually arrive.
To configure the MSDP router to send SA request messages immediately to the MSDP peer when a new
member becomes active in a group, use the following command:
configure msdp sa-cache-server <remoteaddr> {vr <vrname>}
To purge all SA cache entries, use the following command:
clear msdp sa-cache {{peer} <remoteaddr> | peer all} {group-address <grp-addr>} {vr
<vrname>}
To display the SA cache database, use the following command:
show msdp [sa-cache | rejected-sa-cache] {group-address <grp-addr>} {source-address

<src-addr>} {as-number <as-num>} {originator-rp <originator-rp-addr>} {local} {peer
<remoteaddr>} {vr <vrname>}
Maximum SA Cache Entry Limit

You can configure a limit on the maximum number of SA cache entries that can be stored in the cache
database. Once the number of SA cache entries exceeds the pre-configured limit, any newly received
cache entries are discarded. You can configure the limit on a per-peer basis. By default, no SA message
limit is set. The router can receive an unlimited number of SA entries from an MSDP peer.
To configure a limit on the number of SA entries that can be stored in cache, use the following
command:
configure msdp [{peer} <remoteaddr> | peer all] sa-limit <max-sa> {vr <vrname>}
To allow an unlimited number of SA entries, use 0 (zero) as the value for <max-sa>.
To display the SA cache limit, use the following command:
Redundancy
Because the peering relationship between MSDP peers is based on TCP connections, after a failover
occurs the TCP connections need to be re-established again. All SA cache entries learned from the old
peering relationships must be flushed and relearned again on new TCP connections.
On a dual MSM system, MSDP runs simultaneously on both MSMs. During failover, the MSDP process
on the active MSM receives and processes all control messages. MSDP on the standby MSM is in a
down state, and doesn’t receive, transmit, or process any control messages. If the active MSM fails, the
MSDP process loses all state information and the standby MSM becomes active. However, the failover
from the active MSM to the standby MSM causes MSDP to loses all state information and dynamic data,
so it is not a hitless failover.
On fixed-configuration, stackable switches, an MSDP process failure brings down the switch.

Scaling Limits
Scaling Limits
Table 113: MSDP scaling limits
MSDP Peering
Connections
Platform (Active TCP Entries in Mesh-
Type Connections) SA Cache Groups
Chassis 32 16,000 8
Stackable 16 8,000 8
PC 12 8,000 4
SNMP MIBs
SNMP MIB access is not supported for MSDP.
Figure 156 shows two MSDP-speaking routers, MSDP-1 and MSDP-2. The example in this section shows
how to configure MSDP on each router to:
● Establish a peer session between MSDP-1 and MSDP-2. (To verify the session, enter the show msdp
peer command.)
● Exchange SA messages, if any, between MSDP-1 and MSDP-2. (To view the SA cache database, enter
the show msdp sa-cache command.)
Figure 156: MSDP configuration example
10.172.168.32 10.172.168.61
MSDP-1 MSDP-2
XOS006A

The minimum configuration required to establish an MSDP session between MSDP-1 and MSDP-2
follows.
Configuration for MSDP-1

# VLAN configuration
create vlan test
config vlan test ipaddress 10.172.168.32/24
config default del port 1:1
config vlan test add port 1:1
enable ipforwarding
# MSDP configuration
config msdp originator-id 10.172.168.32
create msdp peer 10.172.168.61
enable msdp peer 10.172.168.61
enable msdp
Configuration for MSDP-2

# VLAN configuration
create vlan test
config vlan test ipaddress 10.172.168.61/24
config default del port 1:1
config vlan test add port 1:1
enable ipforwarding
# MSDP configuration
config msdp originator-id 10.172.168.61
create msdp peer 10.172.168.32
enable msdp peer 10.172.168.32
enable msdp

3 Appendixes
A Software Upgrade and Boot Options
This appendix describes the following topics:

● Downloading a New Image on page 997
● Understanding Hitless Upgrade—Modular Switches Only on page 1006
● Saving Configuration Changes on page 1014
● Using TFTP to Upload the Configuration on page 1018
● Using TFTP to Download the Configuration on page 1019
● Synchronizing Nodes—Modular Switches and SummitStack Only on page 1020
● Accessing the Bootloader on page 1022
● Upgrading the BootROM—BlackDiamond 10808 Switch Only on page 1023
● Upgrading the BootROM—Summit Family of Switches and SummitStack Only on page 1023
● Upgrading the Firmware—BlackDiamond 8800 Series Switches Only on page 1024
Downloading a New Image

The core image file contains the executable code that runs on the switch and is preinstalled at the
factory. As new versions of this image are released, you should upgrade the software running on your
system. Modular software packages enhance the functionality of the ExtremeXOS core image currently
running on your switch. Modular software packages are not preinstalled at the factory.
On the BlackDiamond 10808 switch, the BlackDiamond 12804 switch, and the BlackDiamond 8800 series
switch with two MSMs installed, you can upgrade the images without taking the switch out of service.
Known as a hitless upgrade, this method of downloading and installing a new image minimizes
network interruption, reduces the amount of traffic lost, and maintains switch operation. For more
information, see “Understanding Hitless Upgrade—Modular Switches Only” on page 1006.
NOTE
The Hitless Upgrade feature is not supported on a SummitStack.

● Image Filename Prefixes on page 998
● Understanding the Image Version String on page 998
● Software Signatures on page 999
● Selecting a Primary or a Secondary Image on page 999
● Installing a Core Image on page 999
● Installing a Modular Software Package on page 1002

Software Upgrade and Boot Options
● Rebooting the Switch on page 1005

● Rebooting the Management Module—Modular Switches Only on page 1005
Image Filename Prefixes

The software image file can be a .xos file, which contains an ExtremeXOS core image, or a .xmod file,
which contains an ExtremeXOS modular software package.
You can identify the appropriate image or module for your platform based on the filename of the
image. Table 114 lists the filename prefixes for each platform.
Table 114: Filename prefixes

Platform Filename Prefixes
BlackDiamond 12800 series bd12K-
BlackDiamond 10808 bd10K-
BlackDiamond 8810 bd8800-
BlackDiamond 8806 bd8800-
Summit family summitX-
For example, if you have a BlackDiamond 8806 switch, download image filenames with the prefix
bd8800-. For additional installation requirements see the sections, “Installing a Core Image” on page 999
and “Installing a Modular Software Package” on page 1002.
Understanding the Image Version String

The image version string contains build information for each version of ExtremeXOS. You can use either
the show version or show switch command to display the ExtremeXOS version running on your
switch.
Depending on the command line interface (CLI) command, the output is structured as follows:
● show version
ExtremeXOS Version <major>.<minor>.<patch>.<build>
For example: ExtremeXOS version 10.1.2.16
● show switch
<major>.<minor>.<patch>.<build>
For example: 10.1.2.16

Table 115 describes the image version fields.
Table 115: Image version fields

Field Description
major Specifies the ExtremeXOS major version number.
minor Specifies the ExtremeXOS minor version number.
patch Identifies a specific patch release.
build Specifies the ExtremeXOS build number. This value is reset to 0 for each new major and
minor release.
The show version command also displays information about the firmware (BootROM) images running
on the switch. For more information, see “Displaying the BootROM and Firmware Versions” on
page 1027.
Software Signatures
Each ExtremeXOS image contains a unique signature. The BootROM checks for signature compatibility
and denies an incompatible software upgrade. In addition, the software checks both the installed
BootROM and software and also denies an incompatible upgrade.
Selecting a Primary or a Secondary Image

A switch can store up to two core images: a primary and a secondary. When downloading a new
image, you select which partition (primary or secondary) to install the new image. If you do not specify
a partition, the software image is downloaded and installed into the current (active) partition. If you
want to install the software image to the alternate partition, you must specify that partition before
downloading the image.
To view your current (active) partition, use the following command:
show switch
Output from this command includes the selected and booted images and if they are in the primary or
secondary partition.
If two MSMs are installed in a modular switch, the downloaded image is saved to the same location on
each one.
To select which image the switch will load on the next reboot, use the following command:
use image {partition} <partition> {msm <slotid>}
Installing a Core Image

Depending on your platform, you can upgrade the core image by using a download procedure from a
Trivial File Transfer Protocol (TFTP) server on the network or an external compact flash memory card
installed in the external compact flash slot of the Management Switch Fabric Module (MSM).

Beginning with ExtremeXOS 11.6, if you have an expired service contract and attempt to download a
new image, the system displays the following message:
Service contract expired, please renew it to be able to download the new software
image.
If you see this message, you must renew your service contract to download and install the new image.
For more information about service contracts, see “Software Licenses and Feature Packs” in Chapter
2, “ExtremeXOS Overview.”
The information in this section describes how to install a new software image.
For information about saving an existing or new switch configuration, see “Saving Configuration
Changes” on page 1014.
For information about installing a new BootROM image on the BlackDiamond 10808 switch or the
Summit family of switches, see the following sections:
● Upgrading the BootROM—BlackDiamond 10808 Switch Only on page 1023
● Upgrading the BootROM—Summit Family of Switches and SummitStack Only on page 1023
For information about installing a new firmware image on the BlackDiamond 8800 series switch or the
BlackDiamond 12800 series switches, see the following sections:
NOTE
Always refer to the most recent version of the ExtremeXOS Installation and Release Notes for the most current
instructions.
To download a new image:

1 Load the new image onto a TFTP server on your network (if you are using TFTP).
2 Load the new image onto an external compact flash memory card (if you are using the external
compact flash slot). This method is available only on modular switches.
Use a PC with appropriate hardware such as a compact flash reader/writer and follow the
manufacturer’s instructions to access the compact flash card and place the image onto the card.
For more information about installing the external compact flash memory card into the external
compact flash slot of the MSM, refer to the Extreme Networks Consolidated XOS Hardware Installation
Guide.
3 Verify which virtual router connects to your TFTP server.
If you loaded the image onto an external compact flash (modular switches only), proceed to step 4.
If you loaded the image onto a TFTP server, use one of the following ping commands to confirm
which virtual router reaches your TFTP server:
ping vr vr-Mgmt <host>
ping vr vr-Default <host>
At least one of these commands must successfully reach your TFTP server for you to download the
image. After verifying the virtual router that reaches your TFTP server, specify that virtual router
when you download the image.

4 Determine your booted and selected partition using the following command:
show switch
Output from this command indicates the selected and booted images and if they are in the primary
or the secondary partition. The selected image partition indicates which image will be used at the
next reboot. The booted image partition indicates the image used at the last reboot. It is the active
partition.
5 Select the partition to use when downloading an image. For more information, see “Selecting a
Primary or a Secondary Image” on page 999.
6 Download the new image to the switch using the following command:
download image [[<hostname> | <ipaddress>] <filename> {{vr} <vrname>} | memorycard
<filename>] {<partition>} {msm <slotid>}
● If you have an expired service contract and attempt to download a new image, the switch
displays the following message:
image.
If you see this message, you must renew your service contract to proceed.
● If you do not have an expired service contract, before the download begins the switch asks if you
want to install the image immediately after the download is finished.
■ If you download and install the image immediately to the active partition, the switch
automatically reboots after the image is installed.
■ If you download the image to the active partition and do not immediately install the image,
you do not need to reboot the switch until you use the image.
■ If you install the image to the inactive partition, you do not need to reboot the switch until
you use the image.
Enter y to install the image after download. Enter n to install the image at a later time.
If you download and install the software image on the active partition, the switch automatically
reboots after the download and installation is completed. The following message appears when
downloading and installing on the active partition:
Image will be installed to the active partition, a reboot required. Do you want
to continue? (y or n)
Enter y to continue the installation and reboot the switch. Enter n to cancel.
If you install the image at a later time, the image is still downloaded and saved to the switch, but
you must use the following command to install the software and reboot the switch:
install image <fname> {<partition>} {msm <slotid>} {reboot}
BlackDiamond 8800 Series Switch Only—When upgrading the switch from ExtremeXOS 11.4 or
earlier to ExtremeXOS 11.5, reboot the switch after downloading and installing the 11.5 image to
both installed MSMs.
NOTE
Unlike ExtremeWare, the download image command in ExtremeXOS causes the switch to use the newly
downloaded software image during the next switch reboot. To modify or reset the software image used during a
switch reboot, use the use image command.

NOTE
A secure method of upgrading the EXOS image uses SFTP or SCP2. Refer to Appendix A, “Configuration and Image
Commands” in the Extreme XOS 12.0 Concepts Guide.
Using EPICenter to Install a Core Image

Depending on your platform, you can use EPICenter to upgrade the core image. You can either load the
new image to the EPICenter server on your network or configure EPICenter to automatically poll and
download newly released images to the EPICenter server. Follow the instructions described in the
EPICenter documentation to appropriately configure the EPICenter server for your network
environment.
For more information about installing the EPICenter client and server, configuring EPICenter, and the
platforms EPICenter supports, refer to the EPICenter documentation that comes with the product or the
documentation available from the Extreme Networks website at
http://www.extremenetworks.com/services/documentation/swuserguides.asp.
Understanding Core Dump Messages

If you configure the switch to write core dump (debug) files to the internal memory card and attempt to
download a new software image, you might have insufficient space to complete the image download. If
this occurs, move or delete the core dump files from the internal memory. For example, if you have a
modular switch with an external memory card installed with space available, transfer the files to the
external memory card. On the Summit family of switches, transfer the files from the internal memory
card to a TFTP server. This frees up space on the internal memory card while keeping the core dump
files.
The switch displays a message similar to the following and prompts you to take action:
Core dumps are present in internal-memory and must be removed before this download can
continue. (Please refer to documentation for the “configure debug core-dumps” command
for additional information)
Do you want to continue with download and remove existing core dumps? (y/n)
Enter y to remove the core dump files and download the new software image. Enter n to cancel this
action and transfer the files before downloading the image.
Installing a Modular Software Package

In addition to the functionality available in the ExtremeXOS core image, you can add functionality to
your switch by installing modular software packages. Modular software packages are contained in files
named with the file extension.xmod, while the core images use the file extension.xos. Modular software
packages are built at the same time as core images and are designed to work in concert with the core
image, so the version number of a modular software package must match the version number of the
core image that it will be running with. For example, the modular software package for Secure Shell
(SSH) named as follows:
bd10K-11.2.0.18-ssh.xmod
Can run only with the core image named:
bd10K-11.2.0.18.xos

You can install a modular software package on the active partition or on the inactive partition. You
would install on the active partition if you want to add the package functionality to the currently
running core image without having to reboot the switch. You would install on the inactive partition if
you want the functionality available after a switch reboot.
To install the package, you use the same process that you use to install a new core image. Follow the
process described in the earlier section “Installing a Core Image.” On the BlackDiamond 10808 switch,
the BlackDiamond 12804 switch, and the BlackDiamond 8800 series switch, you can use hitless upgrade
to install the package. See “Understanding Hitless Upgrade—Modular Switches Only” on page 1006 for
more information.
You activate the installed modular software package either by rebooting the switch or by issuing the
following command:
run update
You can uninstall packages by issuing the following command:
uninstall image <fname> <partition> {msm <slotid>} {reboot}
NOTE
Do not terminate a process that was installed since the last reboot unless you have saved your configuration. If you
have installed a software module and you terminate the newly installed process without saving your configuration,
your module may not be loaded when you attempt to restart the process with the start process command.

● Guidelines for Activating SSL
● Upgrading a Modular Software Package
Guidelines for Activating SSL

Whether you are currently running SSH or downloading the SSH software module for the first time,
you must complete the following steps to activate the Secure Socket Layer (SSL) functionality that is
packaged with the SSH module. Before you can configure SSL on the switch, do the following:
● Download and install the SSH software module on the active partition
● Activate the SSH software module
● Gracefully terminate and re-start the thttpd process running on the switch
The following is an example of activating SSL on the switch:
download image 10.10.10.2 bd10K-11.2.0.18-ssh.xmod

run update
restart process thttpd
For more information about SSH and SSL, see Chapter 16, “Security.”
Upgrading a Modular Software Package

When Extreme Networks introduces a new core software image, a new modular software package is
also available. If you have a software module installed and upgrade to a new core image, you need to
upgrade to the corresponding modular software package.

Two methods are available to upgrade an existing modular software package on your switch.
Regardless of which method you choose, you must terminate and restart the processes associated with
the software module.
Method One.
1 Terminate the processes associated with the software module using the following command:
terminate process <name> [forceful | graceful] {msm <slot>}
2 Download the software module from your TFTP server or external compact flash memory card
<filename>] {<partition>} {msm <slotid>} {slot <slotid>}
3 Activate the installed modular package, if installed on the active partition, using the following
command:
run update
4 Restart the processes associated with the software module using the following command:
start process <name> {msm <slot>} {slot <slotid>}
Method Two.
1 Download the software module from your TFTP server or external compact flash memory card
<filename>] {<partition>} {msm <slotid>} {slot <slotid>}
2 Activate the installed modular package, if installed on the active partition, using the following
command:
run update
3 Terminate and restart the processes associated with the software module using the following
command:
Examples. The following examples upgrade the SSH module on the active partition and assume that
you have:
● Upgraded the switch to a new core image (see “Installing a Core Image” on page 999 for more
information)
● Downloaded the corresponding modular software package to your TFTP server. On a modular
switch, you can download the modular software package to an external compact flash memory card
(see “Downloading a New Image” on page 997 for more information)
The first example uses the terminate process and start process commands to terminate and restart
the processes associated with the software module that you are updating:
terminate process exsshd graceful

run update
start process exsshd

The second example uses the restart process command to terminate and restart the processes
associated with the software module that you are updating:

run update
restart process exsshd
Rebooting the Switch

To reboot the switch, use the following command:
reboot {time <month> <day> <year> <hour> <min> <sec>} {cancel} {msm <slot_id>} {slot
<slot-number> | node-address <node-address> | stack-topology {as-standby} }
Use this command to schedule a time to reboot the switch or to reboot the switch immediately. To
schedule a time to reboot the switch, use the following command:
reboot time <date> <time>
Where date is the date and time is the time (using a 24-hour clock format) when the switch will be
rebooted. The values use the following format:
mm dd yyyy hh mm ss
NOTE
When you configure a timed reboot of the switch, use the show switch command to see the scheduled time.
To reboot the switch immediately, use the following command:
reboot
If you do not specify a reboot time, the reboot occurs immediately following the command, and any
previously schedule reboots are cancelled. To cancel a previously scheduled reboot, use the cancel
option.
On SummitStack only. The reboot command, reboots active topology only. The reboot stack-topology
command reboots the entire stack topology. The reboot slot works only on active master node for other
slots, and on active non-master node for self slot. The reboot node-address will reboot the specified node
from any node. Use the reboot stack-topology as-standby to eliminate the dual master condition manually.
Rebooting the Management Module—Modular Switches Only

To reboot a management module in a specific slot, rather than rebooting the switch, use the following
command:
reboot {time <month> <day> <year> <hour> <min> <sec>} {cancel} {msm <slot_id>} {slot
<slot-number> | node-address <node-address> | stack-topology {as-standby} }
with the additional options available:

● slot_id— Specifies the slot where the module is installed
● msm-a—Specifies the MSM module installed in slot A
● msm-b—Specifies the MSM module installed in slot B

NOTE
When you configure a timed reboot of an MSM, use the show switch command to see the scheduled time.
For more information about all of the options available with the reboot command, see the ExtremeXOS
Understanding Hitless Upgrade—Modular Switches

Only
Hitless upgrade is a mechanism that allows you to upgrade the ExtremeXOS software running on the
MSMs without taking the switch out of service. Some additional benefits of using hitless upgrade
include:
● Minimizing network downtime
● Reducing the amount of traffic lost
Although any method of upgrading software can have an impact on network operation, including
interrupting Layer 2 network operation, performing a hitless upgrade can decrease that impact.
You must have two MSMs installed in your switch to perform a hitless upgrade. With two MSMs
installed in the switch, one assumes the role of primary and the other assumes the role of backup. The
primary MSM provides all of the switch management functions including bringing up and
programming the I/O modules, running the bridging and routing protocols, and configuring the
switch. The primary MSM also synchronizes its configurations with the backup MSM which allows the
backup to take over the management functions of the primary.
NOTE
The software on the I/O modules is not updated during a hitless upgrade, only the software on the MSMs. The I/O
module software is updated when the switch reboots or when a disabled slot is enabled.
NOTE
If you download an image to the backup MSM, the image passes through the primary MSM before the image is
downloaded to the backup MSM.
Understanding the I/O Version Number—BlackDiamond 8800

Series Switch Only
Each ExtremeXOS image comes bundled with an I/O module image and contains a unique upgrade
compatibility version number, known as the I/O version number. This number determines the
relationship between the I/O module image and the ExtremeXOS image and their support for hitless
upgrade. The I/O version number contains build information for each version of ExtremeXOS,
including the major and minor version numbers, and the I/O version number.

Understanding Hitless Upgrade—Modular Switches Only
Extreme Networks generates the I/O version number, and this number increases over time. Any
modifications to the I/O module image after a major software release changes the I/O version number.
For example, if Extreme Networks delivers a patch or service release that modifies the I/O module
image, the I/O version number increases.
When you initiate a hitless upgrade by using the run msm-failover {force} command on the backup
MSM, it checks the I/O version number to determine if a hitless upgrade is possible. Depending on the
currently running software, the switch performs, allows, or denies a hitless upgrade. The following
describes the switch behavior:
● If the new ExtremeXOS image supports hitless upgrade and is compatible with the current running
I/O module image, you can perform a hitless upgrade.
● If the new ExtremeXOS image supports hitless upgrade, is compatible with the current running I/O
image but with a degradation of functionality, you can perform a hitless upgrade with caveats. The
switch warns you that the upgrade is hitless; however, the downloaded software may result in a loss
of new functionality. You can either continue the upgrade with the modified functionality or cancel
the action.
To prevent a loss in functionality, schedule time to take the switch offline to perform the upgrade;
do not upgrade the software using hitless upgrade.
● If the new ExtremeXOS image supports hitless upgrade but is not compatible with the current
running I/O module image (the I/O version numbers do not match), you cannot perform a hitless
upgrade.
The switch warns you that the upgrade may not be hitless. You can either continue the upgrade or
cancel the action. If you continue the upgrade, the primary MSM downloads the new image to the
I/O module and reboots.
The following is a sample of the warning message displayed by the switch:
WARNING: The other MSM operates with a different version of I/O module image.
If you continue with the MSM failover, all I/O modules will be reset.
Are you sure you want to failover? (y/n)
Performing a Hitless Upgrade

The steps described in this section assume the following:
● You have received the new software image from Extreme Networks, and the image is on either a
TFTP server or an external compact flash memory card. For more information, see “Downloading a
New Image” on page 997.
● You are running a version of ExtremeXOS that supports hitless upgrade.
Review the following list to confirm that your system supports hitless upgrade:
■ BlackDiamond 10808 switch—Both MSMs are running ExtremeXOS 11.1 or later. Earlier versions
of ExtremeXOS do not support hitless upgrade.
■ BlackDiamond 8800 series switch with MSM-G8X and 8800 original modules installed—Both
MSMs are running ExtremeXOS 11.4 or later.
■ BlackDiamond 8800 series switch with a mix of MSM-G8X, 8800 original, and 8800 a- and e-series
modules installed—Both MSMs are running ExtremeXOS 11.5 or later.
■ BlackDiamond 8800 series switch with a mix of MSM-G8X, MSM-48, 8800 original, and 8800 a-
and e-series modules installed—Both MSMs are running ExtremeXOS 11.6 or later.

■ BlackDiamond 8800 series switch with a mix of MSM-48, 8800 original, and 8800 a- and e-series
■ BlackDiamond 12804 switch—Both MSMs are running ExtremeXOS 11.4 or later.
Hitless Upgrade Caveats for the BlackDiamond 8800 Series Switch Only
The following is a summary of hitless upgrade caveats for only the BlackDiamond 8800 series switch:
● If you are running ExtremeXOS 11.5 or earlier, do not attempt to install builds prior to 11.6.1 on an
MSM-48, and do not use the synchronize command if either the primary or secondary slot has a
pre-11.6.1 image and the secondary MSM is an MSM-48. On a BlackDiamond 8800 series switch with
MSM-48 modules installed, these versions of software are incompatible and cannot exist on MSMs
installed in the same chassis even during the hitless upgrade process.
CAUTION
If a pre-11.6 image is loaded onto an MSM-48, the MSM will no longer boot correctly and it may require a
system recovery process. If this occurs, please contact the Extreme Networks Technical Support Assistance
Center (TAC) for further instructions on recovering your system.
● If you attempt a hitless upgrade between major releases, the switch warns you that the upgrade is
not hitless. You can either continue the upgrade or cancel the action. If you continue the upgrade,
the primary MSM downloads the new image to the I/O module and reboots them.
● If you are running ExtremeXOS 11.4 or earlier, do not attempt to perform a hitless upgrade to
ExtremeXOS 11.5 or later. On the BlackDiamond 8800 series switch with MSM-G8X modules
installed, these versions of software are incompatible and cannot exist on MSMs installed in the
same chassis even during the hitless upgrade process. If attempted, the backup MSM enters the non-
operational state.
To recover from the non-operational state, do the following:
a From the primary MSM, use the synchronize command to return the MSMs to the same version
of software.
b To confirm the MSMs are synchronized, use the show switch command.
The current state indicates which MSM is the primary (displayed as MASTER), which MSM is the
backup (displayed as BACKUP), and if the backup MSM is synchronized with the primary MSM
(displayed as In Sync).
After you recover from the non-operational state and confirm the MSMs are synchronized, perform
the steps in “Installing a Core Image” on page 999 to install and upgrade the image on the switch.
NOTE
ExtremeXOS 11.6 introduced support for the BlackDiamond 8800 series MSM-48 module. If your switch is running
an earlier version of ExtremeXOS, you must upgrade to ExtremeXOS 11.6 prior to installing the MSM-48 in your
chassis.
ExtremeXOS 11.5 introduced support for the BlackDiamond 8800 a-series and e-series modules. If your switch is
running ExtremeXOS 11.4 or earlier, you must upgrade to ExtremeXOS 11.5 to operate the modules.
Hitless upgrade is not supported between major releases. Do not attempt to perform a hitless upgrade. To upgrade
the switch from ExtremeXOS 11.4 or earlier to ExtremeXOS 11.5 or later, reboot the switch after downloading and
installing the new image to both installed MSMs. For information about installing an image without using hitless
upgrade, see “Installing a Core Image” on page 999.

Summary of Tasks
To perform a hitless upgrade to install and upgrade the ExtremeXOS software on your system:
1 View the current switch information.
● Determine your selected and booted image partitions.
● Verify which MSM is the primary and which is the backup.
● Confirm that the MSMs are synchronized.
2 Select the partition to download the image to (and the partition to boot from after installing the
image).
3 Download and install the new ExtremeXOS software on the backup MSM. Reboot this MSM if
necessary.
4 Verify that the backup MSM comes up correctly and that the MSMs are synchronized.
5 Initiate failover from the primary MSM to the backup MSM. This MSM now becomes the new
backup MSM, and the original MSM becomes the primary MSM.
6 Verify that the MSMs come up correctly and that they are synchronized.
7 Download and install the new ExtremeXOS software on the original primary MSM (new backup
MSM). Reboot this MSM if necessary.
8 Verify that the new backup MSM comes up correctly and that the MSMs are synchronized.
9 Initiate failover from the new primary MSM to the new backup MSM.
This optional step restores the switch to the original primary and backup MSM.
10 Confirm that the failover is successful.
This optional step confirms which MSM is the primary or the backup.
Detailed Steps
To perform a hitless upgrade to install and upgrade the ExtremeXOS software on your system:
1 View current switch information using the following command:
show switch
Determine your selected and booted partition, verify which MSM is the primary and which is the
backup, and confirm that the MSMs are synchronized.
Output from this command indicates, for each MSM, the selected and booted images and if they are
in the primary or the secondary partition. The selected image partition indicates which image will be
used at the next reboot. The booted image partition indicates the image used at the last reboot. It is
the active partition.
2 Select the partition to download the image to and download and install the new ExtremeXOS
software on the backup MSM using the following command:
NOTE
If the backup MSM is installed in slot B, specify msm B. If the backup MSM is installed in slot A, specify msm
A.

● If you have an expired service contract and attempt to download a new image, you see the
following message:
image.
If you see this message, you must renew your service contract to proceed.
● If you do not have an expired service contract, before the download begins the switch asks if you
want to install the image immediately after the download is finished. If you download and install
the image immediately to the active partition, the switch automatically reboots the MSM.
■ If you download and install the software image on the active partition, the switch
automatically reboots the MSM after the image is installed. The following message appears
when downloading and installing on the active partition:
Image will be installed to the active partition, a reboot required. Do you
want to continue? (y or n)
Enter y to continue the installation and reboot the MSM. Enter n to cancel.
■ If you download and install the software image on the non-active partition, you need to
reboot the MSM manually before you proceed. To reboot the switch, use the following
command:
reboot {time <month> <day> <year> <hour> <min> <sec>} {cancel} {msm <slot_id>}
{slot <slot-number> | node-address <node-address> | stack-topology {as-standby}
}
Reboot only the backup MSM so the switch continues to forward traffic.
■ If you install the image at a later time, use the following command to install the software:
3 Verify that the backup MSM comes up correctly and that the MSMs are synchronized using the
following command:
show switch
4 Initiate failover from the primary MSM to the backup MSM using the following command:
run msm-failover
When you failover from the primary MSM to the backup MSM, the backup becomes the new
primary, runs the software on its active partition, and provides all of the switch management
functions.
If you have a BlackDiamond 8800 series switch and the new ExtremeXOS image supports hitless
upgrade but is not compatible with the current running I/O module image (the I/O version
numbers do not match), you cannot perform a hitless upgrade.
The switch displays a warning message similar to the following:
WARNING: The other MSM operates with a different version of I/O module image.
If you continue with the MSM failover, all I/O modules will be reset.
Are you sure you want to failover? (y/n)

You can either continue the upgrade or cancel the action. If you continue the upgrade, the primary
MSM downloads the new image to the I/O module and reboots.

5 Verify that the backup MSM comes up correctly and that the MSMs are synchronized using the
following command:
show switch
6 Select the partition to download the image to and download and install the new ExtremeXOS
software on the new backup MSM (this was the original primary MSM) using the following
command:
NOTE
If the new backup MSM is installed in slot A, specify msm A. If the new backup MSM is installed in slot B,
specify msm B.
Before the download begins, the switch asks if you want to install the image immediately after the
download is finished. If you download and install the image immediately to the active partition, the
switch automatically reboots the MSM.
● If you download and install the software image on the active partition, the switch automatically
reboots the MSM after the image is installed. The following message appears when downloading
and installing on the active partition:
Image will be installed to the active partition, a reboot required. Do you
want to continue? (y or n)
Enter y to continue the installation and reboot the MSM. Enter n to cancel.
● If you download and install the software image on the non-active partition, you need to reboot
the MSM manually before you proceed. To reboot the switch, use the following command:
reboot {time <month> <day> <year> <hour> <min> <sec>} {cancel} {msm <slot_id>}
{slot <slot-number> | node-address <node-address> | stack-topology {as-standby}
}
Reboot only the backup MSM so the switch continues to forward traffic.
● If you install the image at a later time, use the following command to install the software:
7 Verify that the new backup MSM comes up correctly and that the MSMs are synchronized using the
following command:
show switch
8 Optionally, initiate failover from the new primary MSM to the new backup MSM using the
following command:
run msm-failover
When you failover from the new primary MSM to the new backup MSM, this optional step restores
the switch to the original primary and backup MSM.

9 Optionally, confirm that the failover is successful by checking the current state of the MSMs using
show switch
You can also perform a hitless upgrade on ExtremeXOS modular software packages (.xmod files). To
perform a hitless upgrade of a software package, you must install the core software image first, and the
version number of the modular software package must match the version number of the core image
that it will be running with.
For more detailed information about modular software packages, see “Installing a Modular Software
Package” on page 1002. To perform a hitless upgrade, follow the steps described in the previous section,
“Performing a Hitless Upgrade.”
Hitless Upgrade Examples

This section provides examples for performing a hitless upgrade on the BlackDiamond 10808 and 12804
switch and the BlackDiamond 8800 series switch.
NOTE
Before you begin, make sure you are running a version of ExtremeXOS that supports hitless upgrade. For more
information, see the list on page 1007.
Examples on the BlackDiamond 10808 and 12804 Switches

Using the assumptions described below, the following examples perform a hitless upgrade for a core
software image on the BlackDiamond 10K switch (for the BlackDiamond 12804 switch, the software
image file name begins with bd12k-):
● You have received the new software image from Extreme Networks named bd10K-11.1.0.14.xos.
● You do not know your selected or booted partitions.
● You are currently using the primary partition.
● The image is on a TFTP server named tftphost.
● You are installing the new image immediately after download.
● The MSM installed in slot A is the primary.
● The MSM installed in slot B is the backup.
● You are running ExtremeXOS 11.1 or later on both MSMs.
Performing a Hitless Upgrade on the Inactive Partition

The following example shows the commands necessary to perform a hitless upgrade on the inactive
partition. In this example, the secondary partition is the inactive partition:
show switch
download image tftphost bd10K-11.1.0.14.xos secondary
show switch
reboot msm B
show switch
run msm-failover

reboot msm A
show switch
After executing these commands, MSM B will be the master, and the secondary partition will be the
active partition for both MSMs. The previously running software will reside on the inactive partition
(now, the primary partition).
Performing a Hitless Upgrade on the Current Partition

The following example shows the commands necessary to perform a hitless upgrade on the current
partition. In this example, the primary partition is the current partition:
show switch
download image tftphost bd10K-11.1.0.14.xos primary msm B
show switch
run msm-failover
show switch
download image tftphost bd10K-11.1.0.14.xos primary msm A
show switch
NOTE
If you download the image to the current partition, specifying the partition name is optional.
After executing these commands, MSM B will be the master, and the primary partition will be the active
partition for both MSMs. The previously running software was overwritten on the primary partition,
and is no longer on the switch.
Examples on the BlackDiamond 8800 Series Switch

Using the assumptions described below, the following examples perform a hitless upgrade for a core
software image on the BlackDiamond 8800 series switch:
● You have received the new software image from Extreme Networks named bd8800-11.4.0.12.xos.
● You do not know your selected or booted partitions.
● You are currently using the primary partition.
● The image is on a TFTP server named tftphost.
● You are installing the new image immediately after download.
● The MSM installed in slot A is the primary.
● The MSM installed in slot B is the backup.
● You are running ExtremeXOS 11.4 on both MSMs.
Performing a Hitless Upgrade on the Inactive Partition

The following example shows the commands necessary to perform a hitless upgrade on the inactive
partition. In this example, the secondary partition is the inactive partition:
show switch
download image tftphost bd8800-11.4.0.12.xos secondary
show switch
reboot msm B
show switch

run msm-failover
show switch
After executing these commands, MSM B will be the master, and the secondary partition will be the
active partition for both MSMs. The previously running software will reside on the inactive partition
(now, the primary partition).
Performing a Hitless Upgrade on the Current Partition

The following example shows the commands necessary to perform a hitless upgrade on the current
partition. In this example, the primary partition is the current partition and you restore the switch to the
original primary and backup MSM.
show switch
download image tftphost bd8800-11.4.0.12.xos primary msm B
show switch
run msm-failover
show switch
download image tftphost bd8800-11.4.0.12.xos primary msm A
run msm-failover
show switch
NOTE
If you download the image to the current partition, specifying the partition name is optional.
After executing these commands, MSM A is restored as the primary MSM, and the primary partition
will be the active partition for both MSMs. The previously running software was overwritten on the
primary partition, and is no longer on the switch.
Saving Configuration Changes

The configuration is the customized set of parameters that you have selected to run on the switch. As
you make configuration changes, the new settings are stored in run-time memory. Settings that are
stored in run-time memory are not retained by the switch when the switch is rebooted. To retain the
settings and have them loaded when you reboot the switch, you must save the configuration to
nonvolatile storage.
The switch can store multiple user-defined configuration files, each with its own filename. By default,
the switch has two prenamed configurations: a primary and a secondary configuration. When you save
configuration changes, you can select to which configuration you want the changes saved or you can
save the changes to a new configuration file. If you do not specify a filename, the changes are saved to
the configuration file currently in use. Or if you have never saved any configurations, you are asked to
save your changes to the primary configuration.
NOTE
Configuration files have a .cfg file extension. When you enter the name of the file in the CLI, the system
automatically adds the .cfg file extension.

If you have made a mistake or you must revert to the configuration as it was before you started making
changes, you can tell the switch to use the backup configuration on the next reboot.
Each filename must be unique and can be up to 32 characters long. Filenames are also case sensitive.
For information on filename restrictions, refer to the specific command in the ExtremeXOS Command
Reference Guide.
To save the configuration, use the following command:
save configuration {primary | secondary | <existing-config> | <new-config>}

● primary—Specifies the primary saved configuration
● secondary—Specifies the secondary saved configuration
● existing-config—Specifies an existing user-defined configuration (displays a list of available user-
defined configuration files)
● new-config—Specifies a new user-defined configuration
You are then prompted to save the changes. Enter y to save the changes or n to cancel the process.
To use the configuration, use the following command:
use configuration [primary | secondary | <file_name>]

● primary—Specifies the primary saved configuration
● secondary—Specifies the secondary saved configuration
● file_name—Specifies an existing user-defined configuration (displays a list of available user-defined
configuration files)
The configuration takes effect on the next reboot.
NOTE
If the switch is rebooted while in the middle of saving a configuration, the switch boots to factory default settings if
the previously saved configuration file is overwritten. The configuration that is not in the process of being saved is
unaffected.
Viewing a Configuration
You can view the current configuration on the switch by using the following command:
show configuration {<module-name>}
You can also view just the portion of the configuration that applies to a particular module (for example,
SNMP) by using the module-name parameter.
You can send output from the show configuration {<module-name>} command to the Extreme
Networks Technical Support department for problem-solving purposes. The output maintains the
command line interface (CLI) format of the current configuration on the switch.

Returning to Factory Defaults

To return the switch configuration to factory defaults, use the following command:
synchronize stacking
This command resets the entire configuration, with the exception of user accounts and passwords that
have been configured and the date and time. If used on a SummitStack configuration, the command
also preserves stacking-specific parameters so the stack will be formed when the reboot has occurred.
To unset the currently selected configuration image, reset all switch parameters, and reboot the switch,
unconfigure switch {all}
ASCII-Formatted Configuration Files

Beginning With ExtremeXOS 11.4, you can upload your current configuration in ASCII format to a
TFTP server. The uploaded ASCII file retains the CLI format and allows you do the following:
● View and modify the configuration using a text editor, and later download a copy of the file to the
same switch or to one or more different switches.
● Send a copy of the configuration file to Extreme Networks Technical Support for problem-solving
purposes.
Summary of Tasks
The following summary only describes the CLI involved to transfer the configuration and load it on the
switch; it is assumed that you know how to modify the configuration file with a text editor. As
previously described, to use these commands, use the .xsf file extension. These steps are not applicable
to configurations that use the .cfg file extension.
To work with an ASCII-formatted configuration file, complete the following tasks:

1 Upload the configuration to a network TFTP server using the following command:
upload configuration [<hostname> | <ipaddress>] <filename> {vr <vr-name>}
After the configuration file is on the TFTP server, use a text editor to enter the desired changes, and
rename the file if necessary so it has the .xsf extension.
2 Download the configuration from the TFTP server to the switch using one of the following
commands:
tftp [<host-name> | <ip-address>] -g -r <remote-file>
tftp get [<host-name> | <ip-address>] <remote-file>
3 Verify the configuration file is on the switch using the following command:
ls
4 Load and restore the new configuration file on the switch using the following command:
load script <script-file>
5 Save the configuration to the configuration database so the switch can reapply the configuration
after switch reboot using the following command:
save configuration {primary | secondary | <existing-config> | <new-config>}

When you save the configuration file, the switch automatically adds the .cfg file extension to the
filename. This saves the ASCII configuration as an XML-based configuration file.
Uploading the ASCII Configuration File To a TFTP Server

To upload the current switch configuration as an ASCII-based file to the TFTP server, use the upload
configuration command and save the configuration with the .xsf file extension.
For example, to transfer the current switch configuration as an ASCII-based file named
meg_upload_config1.xsf to the TFTP server with an IP address of 10.10.10.10, do the following:
upload configuration 10.10.10.10 meg_upload_config1.xsf
If you successfully upload the configuration to the TFTP server, the switch displays a message similar
to the following:
Uploading meg_upload_config1.xsf to 10.10.10.10 ... done!
Downloading the ASCII Configuration File to the Switch

To download the configuration from the TFTP server to the switch, use the tftp or tftp get
command. For example, to retrieve the configuration file named meg-upload_config1.xsf from a TFTP
server with an IP address of 10.10.10.10, you can use one of the following commands:
tftp 10.10.10.10 -g -r meg_upload_config1.xsf
tftp get 10.10.10.10 meg_upload_config1.xsf
If you successfully download the configuration to the switch, the switch displays a message similar to
the following:
Downloading meg_upload_config1.xsf to switch... done!
Verifying that the ASCII Configuration File is on the Switch

To confirm that the ASCII configuration file is on the switch, use the ls command. The file with a .xsf
extension is the ASCII configuration.
The following sample output contains an ASCII configuration file:
-rw-r--r-- 1 root 0 98362 Nov 2 13:53 Nov022005.cfg

-rw-r--r-- 1 root 0 117136 Dec 12 12:56 epicenter.cfg
-rw-r--r-- 1 root 0 68 Oct 26 11:17 mcastgroup.pol
-rw-r--r-- 1 root 0 21203 Dec 13 15:40 meg_upload_config1.xsf
-rw-r--r-- 1 root 0 119521 Dec 6 14:35 primary.cfg
-rw-r--r-- 1 root 0 96931 Nov 11 11:01 primary_11_11_05.cfg
-rw-r--r-- 1 root 0 92692 Jul 19 16:42 secondary.cfg
Loading the ASCII Configuration File

After downloading the configuration file, you must load the new configuration on the switch. To load
and restore the ASCII configuration file, use the load script <script-file> command. After issuing
this command, the ASCII configuration quickly scrolls across the screen.

The following is an example of the type of information displayed when loading the ASCII configuration
file:
script.meg_upload_config1.xsf.389 # enable snmp access

script.meg_upload_config1.xsf.390 # enable snmp traps
script.meg_upload_config1.xsf.391 # configure mstp region purple
script.meg_upload_config1.xsf.392 # configure mstp revision 3
script.meg_upload_config1.xsf.393 # configure mstp format 0
script.meg_upload_config1.xsf.394 # create stpd s0
Instead of entering each command individually, the script runs and loads the CLI on the switch.
Saving the Configuration

After you load the configuration, save it to the configuration database for use by the switch. This allows
the switch to reapply the configuration after a switch reboot. To save the configuration, use the save
configuration {primary | secondary | <existing-config> | <new-config>} command.
When you save the configuration file, the switch automatically adds the .cfg file extension to the
filename. This saves the ASCII configuration as an XML-based configuration file.
You can use any name for the configuration. For example, after loading the file meg_upload_config1.xsf,
you need to save it to the switch. To save the configuration as configuration1.cfg, use the following
command:
save configuration configuration1
Using TFTP to Upload the Configuration

You can upload the current configuration to a Trivial File Transfer Protocol (TFTP) server on your
network. Using TFTP, the uploaded configuration file retains your system configuration and is saved in
Extensible Markup Language (XML) format. This allows you to send a copy of the configuration file to
the Extreme Networks Technical Support department for problem-solving purposes.
To view your current switch configuration, use the show configuration {<module-name>} command
available on your switch. Do not use a text editor to view or modify your XML-based switch
configuration files.
ExtremeXOS 11.4 introduces support for ASCII-formatted configuration files. To view your current
switch configuration in ASCII-format, see “ASCII-Formatted Configuration Files” on page 1016 for more
information about uploading and downloading ASCII-formatted configuration files.
For more information about TFTP, see “Using the Trivial File Transfer Protocol” on page 76.
To upload the configuration from the switch to a TFTP server, you can use either the tftp or the tftp
put command:
● tftp [<host-name> | <ip-address>] -p -l <local-file> {-r <remote-file>}
■ host-name—Specifies the host name of the TFTP server
■ ip-address—Specifies the IP address of the TFTP server

Using TFTP to Download the Configuration
■ -p—Puts the specified file from the local host and copies it to the TFTP server
■ -l <local-file>—Specifies the name of the configuration file that you want to save to the TFTP
server
■ -r <remote-file>—Specifies the name of the configuration file on the TFTP server
● tftp put [<host-name> | <ip-address>] <local-file> {<remote-file>}
■ put—Puts the specified file from the local host and copies it to the TFTP server
■ host-name—Specifies the host name of the TFTP server
■ ip-address—Specifies the IP address of the TFTP server
■ <local-file>—Specifies the name of the configuration file that you want to save to the TFTP
server
■ <remote-file>—Specifies the name of the configuration file on the TFTP server
If you upload a configuration file and see the following message:
Error: No such file or directory
Check to make sure that you entered the filename correctly, including the .cfg extension, and that you
entered the correct host name or IP address for the TFTP server.
If your upload is successful, the switch displays a message similar to the following:
Uploading megtest1.cfg to TFTPhost ... done!
Beginning with ExtremeXOS 11.4, you can also upload the current configuration in ASCII format from
the switch to a TFTP server on your network. For more information, see “ASCII-Formatted
Configuration Files” on page 1016.
Using TFTP to Download the Configuration

You can download previously saved XML formatted XOS configuration files from a TFTP host to the
switch to modify the switch configuration. Do not use a text editor to view or modify your switch
configuration files; modify your switch configurations directly in the CLI.
ExtremeXOS 11.4 introduces support for ASCII-formatted configuration files. To view your current
switch configuration in ASCII-format, see “ASCII-Formatted Configuration Files” on page 1016 for more
information about uploading and downloading ASCII-formatted configuration files.
For more information about TFTP, see “Using the Trivial File Transfer Protocol” on page 76.
To download the configuration from a TFTP host to the switch, you can use either the tftp or the tftp
get command:
● tftp [<host-name> | <ip-address>] -g -r <remote-file> {-l <local-file>}
■ host-name—Is the host name of the TFTP server
■ ip-address—Is the IP address of the TFTP server
■ -g—Gets the specified file from the TFTP server and copies it to the local host

■ -r <remote-file>—Specifies the name of the configuration file that you want to retrieve from
the TFTP server
■ -l <local-file>—Specifies the name of the configuration file on the switch
● tftp get [host-name> | <ip-address>] <remote-file> {<local-file>} {force-overwrite}
■ get—Gets the specified file from the TFTP server and copies it to the local host
■ host-name—Is the host name of the TFTP server
■ ip-address—Is the IP address of the TFTP server
■ <remote_file>—Specifies the name of the configuration file that you want to retrieve from the
TFTP server
■ <local-file>—Specifies the name of the configuration file on the switch
■ force-overwrite—Specifies the switch to automatically overwrite an existing file
NOTE
Reference Guide.
If you download a configuration file and see the following message:
Error: Transfer timed out
Make sure that you entered the filename correctly, including the .cfg extension, and that you entered
the correct host name or IP address for the TFTP server.
If your download is successful, the switch displays a message similar to the following:
Downloading megtest2.cfg to switch... done!
Configurations are downloaded and saved into the switch nonvolatile memory. The configuration is
applied after you reboot the switch.
If the configuration currently running in the switch does not match the configuration that the switch
used when it originally booted, an asterisk (*) appears before the command line prompt when using the
CLI.
Beginning with ExtremeXOS 11.4, you can also download the current configuration in ASCII format
from a TFTP server on your network to the switch. For more information, see “ASCII-Formatted
Configuration Files” on page 1016.
Synchronizing Nodes—Modular Switches and

SummitStack Only
Before synchronizing nodes on a modular chassis or nodes on a SummitStack, review the following list
to confirm that your platform and both installed MSMs are running software that supports the
synchronize command:
● BlackDiamond 10808 switch—Both MSMs are running ExtremeXOS 11.0 or later.

Synchronizing Nodes—Modular Switches and SummitStack Only
● BlackDiamond 8810 switch with MSM-G8X and 8800 original modules installed—Both MSMs are
running ExtremeXOS 11.1 or later.
● BlackDiamond 8806 switch with MSM-G8X and 8800 original modules installed—Both MSMs are
running ExtremeXOS 11.3 or later.
● BlackDiamond 8800 series switch with a mix of MSM-G8X, 8800 original, and 8800 a- and e-series
● BlackDiamond 8800 series switch with a mix of MSM-G8X, MSM-48, 8800 original, and 8800 a- and
e-series modules installed—Both MSMs are running ExtremeXOS 11.6 or later.
● BlackDiamond 8800 series switch with a mix of MSM-48, 8800 original, and 8800 a- and e-series
● SummitStack—all nodes are running ExtremeXOS 12.0 or later.
CAUTION
BlackDiamond 8800 Series Only—Do not attempt to install builds prior to version 11.6.1 on an MSM-48, and do
not use the synchronize command if either the primary or secondary slot has a pre-11.6.1 image and the
secondary MSM is an MSM-48. If a pre-11.6 image is loaded onto an MSM-48, the MSM will no longer boot
correctly and it may require a system recovery process. If this occurs, please contact the Extreme Networks
Technical Support Assistance Center (TAC) for further instructions on recovering your system.
● BlackDiamond 12804 switch—Both MSMs are running ExtremeXOS 11.4 or later.
On a dual MSM system or a SummitStack with redundancy, you can take the primary node
configurations and images and replicate them on the backup node using the following command:
synchronize
CAUTION
During a synchronization, on a modular chassis, half of the switch fabric is lost. On a SummitStack, the active stack
will briefly alternate between a ring and daisy-chain topology. When the primary node finishes replicating its
configurations and images to the backup node, the full switch fabric or the stack ring is restored.
In addition to replicating the configuration settings and images, this command also replicates which
configuration or image the node should use on subsequent reboots. This command does not replicate
the run-time configuration. You must use the save configuration command to store the run-time
configuration first.
Additional Behavior on the BlackDiamond 8800 Series Switch

Only
On the BlackDiamond 8800 series switch (formerly known as Aspen), the I/O ports on the backup MSM
go down when you synchronize the MSMs. When the primary MSM finishes replicating its
configurations and images to the backup MSM, the I/O ports on the backup MSM come back up.
Automatic Synchronization of Configuration Files

On a dual MSM (node) modular chassis or on a SummitStack where redundancy is in use, ExtremeXOS
automatically synchronizes all of the configuration files from the primary node to the backup node if

the switch detects that the backup node’s configuration file contents are different from the primary
node. You do not configure this behavior.
The switch deletes the old configuration files on the backup node only upon a successful file
synchronization. If an error occurs, the switch does not delete the old configuration files on the backup
node. For example, if you install a backup node that contains different configuration files from the
primary node, the old configuration files are deleted after a successful bootup of the backup node.
To see a complete listing of the configuration files on your system, use the ls command.
For more detailed information, see “Replicating Data Between Nodes” on page 79.
Accessing the Bootloader

The Bootloader of the switch initializes certain important switch variables during the boot process. In
the event the switch does not boot properly, some boot option functions can be accessed through the
Bootloader.
Interaction with the Bootloader is required only under special circumstances and should be done only
under the direction of Extreme Networks Customer Support. The necessity of using these functions
implies a nonstandard problem which requires the assistance of Extreme Networks Customer Support.
To access the Bootloader menu:

1 Attach a serial cable to the console port of the switch.
2 Attach the other end of the serial cable to a properly configured terminal or terminal emulator,
power cycle the switch, and press the spacebar key on the keyboard of the terminal during the
bootup process.
NOTE
On the BlackDiamond 10808 switch, access the Bootloader by pressing the spacebar key immediately after a
power cycle of the MSM to get into the Bootloader application.
On the BlackDiamond 12800 series switches and the BlackDiamond 8800 series switches, when you see the
BootRom banner, press the spacebar key to get into the Bootloader application.
On the Summit family of switches, when you see the Bootloader banner, press the spacebar key to get into the
Bootloader application.
As soon as you see the BOOTLOADER> prompt (BlackDiamond 10808 switch or Summit family of
switches) or the BootRom -> prompt (BlackDiamond 8800 series switches and the BlackDiamond
12800 series switches), release the spacebar. You can issue a series of commands to:
■ View the installed images.
■ Select the image to boot from.
■ Select the configuration to use.
■ Load a recovery image over the management port.
■ Load a recovery image over the serial port (BlackDiamond 10808 switch only).
To see a list of available commands or additional information about a specific command, enter h or
type help.

Upgrading the BootROM—BlackDiamond 10808 Switch Only
The following describes some ways that you can use the Bootloader:
● Viewing images—To display a list of installed images, use the show image command.
● Selecting an image—To change the image that the switch boots from in flash memory, use the boot
{image number} command. If you specify image number, the specified image is booted. If you do
not specify an image number, the default image is booted.
● Selecting a configuration—To select a different configuration from the one currently running, use the
config {alt | default | <filename> | none} command. This command is useful if you
experience a problem with the current configuration and there is an alternate configuration available.
■ alt—Specifies the alternate configuration file
■ default—Specifies the default configuration file
■ filename—Specifies a configuration filename
■ none—Uses no configuration . This restores the switch to the default configuration. It may be
helpful if a password has been forgotten.
To view the current configuration, use this command without any arguments.
To exit the Bootloader, use the boot command. Specifying boot runs the currently selected ExtremeXOS
image.
Upgrading the BootROM—BlackDiamond 10808 Switch

Only
Upgrade the BootROM from a TFTP server on the network or an external compact flash memory card
installed in the compact flash slot of the MSM, after the switch has booted. For information about
loading an image to a TFTP server and verifying which virtual router connects to your TFTP server, see
“Installing a Core Image” on page 999. Upgrade the BootROM only when asked to do so by an Extreme
Networks technical representative.
To upgrade the BootROM, use the following command:
download bootrom [[<ipaddress> | <hostname>] <filename> {{vr} <vrname>} | memorycard

<filename>] {msm <slotid>}
Upgrading the BootROM—Summit Family of Switches

and SummitStack Only
The Summit family of switches have a two-stage BootROM. The first stage, called bootstrap, does basic
initialization of the switch processor and will load one of two second-stage bootloaders (called primary
and secondary).
If the switch does not boot properly, both the bootstrap and the bootloader allows the user to access the
boot options using the CLI.
If necessary, the bootloader can be updated after the switch has booted, using TFTP. You can upgrade
the BootROM from a TFTP server on the network after the switch has booted and only when asked to
do so by an Extreme Networks technical representative. For information about loading an image to a

TFTP server and verifying which virtual router connects to your TFTP server, see “Installing a Core
Image” on page 999.
To upgrade the BootROM, use the following command:
download bootrom [<ipaddress> | <hostname>] <filename> {{vr} <vrname>}
On a SummitStack, the bootrom can be centrally upgraded. Use the command above on the primary
(Master) stack node to download a bootrom to all stacking nodes. To download to a single stacking
node, you need to specify the slot parameter:
download bootrom [<ipaddress> | <hostname>] <filename> {slot <slot-number>} {{vr}

<vrname>}
NOTE
The Summit family of switches does not support user-created VRs.
Accessing the Bootstrap CLI on the Summit Family of Switches

The bootstrap CLI contains commands to support the selection of which bootloader to use. Interaction
with the bootstrap is required only under special circumstances and should be done only under the
direction of Extreme Networks Customer Support.
To access the bootstrap CLI:

1 Attach a serial cable to the serial console port of the switch.
2 Attach the other end of the serial cable to a properly configured terminal or terminal emulator.
3 Power cycle or reboot the switch.
4 As soon as you see the Bootstrap Banner, press the spacebar.
The BOOTSTRAP> prompt appears on the screen.
NOTE
If you accidentally enter the bootstrap CLI when you want to enter the Bootloader, at the BOOTSTRAP> prompt enter
the boot command.
For detailed information and instructions on accessing the bootloader, see “Accessing the Bootloader”
on page 1022.
Upgrading the Firmware—BlackDiamond 8800 Series

Switches Only
Firmware images are bundled with ExtremeXOS software images. The firmware contains BootROM
images for the MSM and I/O modules and associated firmware images for the backplane and PSU

Upgrading the Firmware—BlackDiamond 8800 Series Switches Only
controllers. ExtremeXOS automatically compares the existing firmware image flashed into the hardware
with the firmware image bundled with the ExtremeXOS image when you:
● Download a new version of ExtremeXOS to the active partition.
● Install a new module into an active chassis.
After a firmware image upgrade, messages are sent to the log.
You can configure the switch to automatically upgrade the firmware when a different image is detected,
or you can have the switch prompt you to confirm the upgrade process. To configure the switch’s
behavior during a firmware upgrade, use the following command:
configure firmware [auto-install | install-on-demand]

● auto-install—Specifies ExtremeXOS to automatically upgrade the firmware if the software detects
a newer firmware image is available. The switch does not prompt you to confirm the firmware
upgrade.
● on-demand—Specifies the switch to prompt you to upgrade the firmware when ExtremeXOS
determines that a newer firmware image is available. This is the default behavior.
You can use the install firmware command to install the firmware bundled with the ExtremeXOS
image. To install the new BootROM and firmware, wait until the show slot command indicates the
MSM and I/O modules are operational. When the modules are operational, use the install firmware
command.
If the bundled firmware image is newer than the existing firmware image, the switch prompts you to
confirm the upgrade. Enter y to upgrade the firmware. Enter n to cancel the firmware upgrade for the
specified hardware and continue scanning for other hardware that needs to be upgraded. Enter <cr> to
cancel the upgrade.
During the firmware upgrade, the switch also prompts you to save your configuration changes to the
current, active configuration. Enter y to save your configuration changes to the current, active
configuration. Enter n if you do not want to save your changes.
The new PSU controller firmware is used immediately after it is installed without rebooting the switch.
The new BootROM and firmware overwrite the older versions flashed into the hardware. In earlier
versions of ExtremeXOS, you are required to immediately reboot the system after a firmware upgrade.
In ExtremeXOS 11.3.3 or later, a reboot is also required, but not immediately after a firmware upgrade.
Use the reboot command to reboot the switch and activate the new BootROM and firmware.
During the firmware upgrade, do not cycle down or disrupt the power to the switch. If a power
interruption occurs, the firmware may be corrupted and need to be recovered. ExtremeXOS
automatically attempts to recover corrupted firmware; however, in some situations user intervention is
required.
Power over Ethernet (PoE) firmware is always automatically upgraded or downgraded to match the
operational ExtremeXOS code image. This configuration is not applicable to PoE firmware.

Upgrading the Firmware—BlackDiamond 12800 Series

Switches Only
Firmware images are bundled with ExtremeXOS software images. The firmware contains BootROM
images for the MSM and I/O modules and associated firmware images for the backplane and power
fan controllers (PFC). When you use the install firmware, ExtremeXOS compares the existing
firmware image flashed into the hardware with the firmware image bundled with the ExtremeXOS
image.
NOTE
Upgrading the firmware on a BlackDiamond 12800 series switch can take a minimum of 6 minutes per slot.
To install the new BootROM and firmware, wait until the show slot command indicates the MSMs and
I/O modules are operational. When the modules are operational, use the install firmware command
to install the BootROM and/or firmware only on modules that have older BootROM or firmware
images installed. After this installation is complete, use the reboot command to reboot the switch.
NOTE
When you use the install firmware command to install the new BootROM and firmware images, all of the I/O
modules enter the Down state. Upgrade the firmware when you can schedule a downtime for your network.
If the bundled firmware image is newer than the existing firmware image, the switch prompts you to
confirm the upgrade:
● Enter y to upgrade the firmware.
● Enter n to cancel the firmware upgrade for the specified hardware and continue scanning for other
hardware that needs to be upgraded.
● Enter <cr> to cancel the upgrade.
The PFC firmware is used immediately after it is installed without rebooting the switch. The new
BootROM and firmware overwrite the older versions flashed into the hardware. If you upgrade the
firmware image, a reboot is required. Use the reboot command to reboot the switch.
During the firmware upgrade, do not cycle down or disrupt the power to the switch. If a power
interruption occurs, the firmware may be corrupted and need to be recovered. ExtremeXOS
automatically attempts to recover corrupted firmware; however, in some situations user intervention is
required.
After a firmware image upgrade, messages are sent to the log.
Accessing the Bootstrap CLI on the BlackDiamond 12800 Series

Switches
The bootstrap CLI contains commands to support the selection of which bootloader to use.

Displaying the BootROM and Firmware Versions
To access the bootstrap CLI:

1 Attach a serial cable to the serial console port of the switch.
3 Power cycle or reboot the switch.
4 As soon as you see the Bootstrap Banner, press the spacebar.
The BootStrap> prompt appears on the screen.
NOTE
If you accidentally enter the bootstrap CLI when you want to enter the Bootloader, at the BootStap> prompt enter
the boot command.
For detailed information and instructions on accessing the bootloader, see “Accessing the Bootloader”
on page 1022.
Displaying the BootROM and Firmware Versions

To display the BootROM (firmware) version on the switch and on all of the modules and PSU
controllers installed in a modular switch, use the show version command.
The following is sample output from the Summit X450 series switch:
Switch : 800132-00-02 0512G00636 Rev 2.0 BootROM: 1.0.0.6 IMG: 11.4.0.15

XGM-2xn-1 :

on Fri Dec 30 11:05:42 PST 2005
BootROM : 1.0.0.6
The following is sample output from the BlackDiamond 10808 switch:
Chassis : 804300-00-09 0438F-00850 Rev 9.0

Slot-1 : 804403-00-08 0405F-00063 Rev 8.0 BootROM: 1.3.0.0 IMG: 11.4.0.23
Slot-2 : 804406-00-02 0511F-00605 Rev 2.0
Slot-3 : 804406-00-02 0511F-00602 Rev 2.0
Slot-4 :
Slot-5 :
Slot-6 :
Slot-7 :
Slot-8 :
MSM-A : 804301-00-09 0434F-00125 Rev 9.0 BootROM: 1.0.1.5 IMG: 11.4.0.25
MSM-B : 804301-00-09 0437F-00026 Rev 9.0 BootROM: 1.0.1.5 IMG: 11.4.0.23

on Fri Feb 24 11:02:17 PST 2006
BootROM : 1.0.1.5

The following is sample output from the BlackDiamond 8800 series switch:
Chassis : 800129-00-02 04344-00039 Rev 2.0

Slot-1 : 800114-00-04 04364-00021 Rev 4.0 BootROM: 1.0.1.7 IMG: 11.4.0.23
Slot-4 :
Slot-7 :
Slot-8 :
Slot-9 :
Slot-10 :
MSM-A : 800112-00-03 04334-00040 Rev 3.0 BootROM: 1.0.1.7 IMG: 11.4.0.23
MSM-B : 800112-00-03 04334-00004 Rev 3.0 BootROM: 1.0.1.7 IMG: 11.4.0.23
PSUCTRL-1 : 450117-00-01 04334-00021 Rev 1.0 BootROM: 2.13

on Thu Feb 16 12:47:41 PST 2006
BootROM : 1.0.1.7
The following is sample output from the BlackDiamond 12804 switch:
Chassis : 804023-00-09 06065-00549 Rev 9.0

Slot-1 : 804019-00-01 05234-00033 Rev 1.0 uC: 2.01 FPGA: 5.07
Slot-2 : 804019-00-02 05404-00055 Rev 2.0 uC: 2.01 FPGA: 5.07
Slot-5 : 804022-00-03 05474-00070 Rev 3.0 uC: 2.01 FPGA: 5.07
Slot-6 : 804019-00-01 05234-00026 Rev 1.0 uC: 2.01 FPGA: 5.07
MSM-A : 804017-00-02 05394-00007 Rev 2.0 BootROM: 1.0.0.2 IMG: 11.4.0.25 uC:
2.01 FPGA: 3.1f
MSM-B : 804017-00-02 05394-00011 Rev 2.0 BootROM: 1.0.0.2 IMG: 11.4.0.25 uC:
2.01

on Fri Feb 24 10:48:51 PST 2006
BootROM : 1.0.0.2

B Troubleshooting
This appendix describes some troubleshooting tips on the following topics:

● Troubleshooting Checklists on page 1030
● LEDs on page 1033
● Using the Command Line Interface on page 1034
● Using Standalone ELRP to Perform Loop Tests on page 1041
● Using the Rescue Software Image—Modular Switches Only on page 1044
● Debug Mode on page 1048
● Saving Debug Information on page 1048
● Evaluation Precedence for ACLs on page 1054
● TOP Command on page 1055
● TFTP Server Requirements on page 1055
● System Health Check on page 1055
● System Odometer on page 1057
● Temperature Operating Range on page 1059
● Unsupported Module Type—BlackDiamond 10808 Switch and BlackDiamond 8800 Series Switches
Only on page 1059
● Corrupted BootROM on the BlackDiamond 8800 Series Switch on page 1060
● Inserting Powered Devices in the PoE Module—BlackDiamond 8800 Series Switch Only on page 1060
● Modifying the Hardware Table Hash Algorithm—BlackDiamond 8800 Series Switch, SummitStack,
and the Summit Family of Switches Only on page 1060
● Untagged Frames on the 10 Gbps Module—BlackDiamond 10808 Switch Only on page 1062
● Understanding the Error Reading Diagnostics Message—Summit Family of Switches Only on
page 1062
● Running MSM Diagnostics from the Bootloader—BlackDiamond 10808 Switch Only on page 1062
● Contacting Extreme Networks Technical Support on page 1063
If you encounter problems when using the switch, this appendix may be helpful. If you have a problem
not listed here or in the release notes, contact Extreme Networks Technical Support.

Troubleshooting
Troubleshooting Checklists
This section provides simple troubleshooting checklists for Layer 1, Layer 2, and Layer 3. The
commands and recommendations described are applicable to both IPv4 and IPv6 environments unless
otherwise specified. If more detailed information about a topic is available, you are referred to the
applicable section in this appendix.
Layer 1
When troubleshooting Layer 1 issues, verify:
● The installation of cables and connectors.
● The behavior of LED status lights. For additional information about LEDs, see “LEDs” on page 1033.
● That the port is enabled, the link status is active, and speed and duplex parameters match the port
settings at the other end of the cable.
To display the configuration of one or more ports, use the show ports configuration command.
● That the packets are being received and transmitted.
To display the number of packets being received and transmitted, use the show ports {stack-
ports <stacking-port-list> | <port_list>} statistics {no-refresh} command.
● That there are no packet errors.
To display packet error statistics, use the following commands:
■ show ports {<port_list> | stack-ports <stacking-port-list>} rxerrors {no-
refresh}—Displays receive error statistics
■ show ports {stack-ports <stacking-port-list> | <port_list>} txerrors {no-
refresh}—Displays transmit error statistics
■ show ports {mgmt | <port_list>} collisions {no-refresh}—Displays collision statistics
Layer 2
● That the MAC addresses are learned, in the correct Virtual LAN (VLAN), and are not blackhole
entries.
To display FDB entries, use the show fdb command.
● Your VLAN configuration, including the VLAN tag, ports in the VLAN, and whether or not the
ports are tagged.
To display detailed information for each VLAN configured on the switch, use the show vlan
detail command.
For additional VLAN troubleshooting tips, see “VLANs” on page 1039.
● Your Spanning Tree Protocol (STP) configuration, including the STP domain (STPD) number, VLAN
assignment, and port state.
To display STP information, use the following commands:
■ show stpd detail—Displays the STP settings on the switch
■ show stpd ports—Displays the STP state of a port
■ show vlan stpd—Displays the STP configuration of the ports assigned to a specific VLAN

Troubleshooting Checklists
For additional STP troubleshooting tips, see “STP” on page 1039.
Layer 3
● The IP address assigned to each VLAN router interface.
To display summary information for all of the VLANs configured on the device, use the show vlan
command.
● That IP forwarding is enabled, the routing protocol is globally enabled, and the routing protocol is
enabled for a VLAN.
To display the configuration information for a specific VLAN, use one of the following commands:
■ show ipconfig {ipv4} {vlan <vlan_name>}—IPv4 environment
■ show ipconfig ipv6 {vlan <vlan_name> | tunnel <tunnelname>}—IPv6 environment
● Which destination networks are in the routing table and the source of the routing entry.
To display the contents of the routing table or the route origin priority, use one of the following
commands:
■ show iproute—IPv4 environment
■ show iproute ipv6—IPv6 environment
To display the contents of the routing table only for routes of a specified origin, use one of the
following commands:
■ show iproute origin—IPv4 environment
■ show iproute ipv6 origin—IPv6 environment
● That the IP Address Resolution Protocol (ARP) table has the correct entries.
NOTE
The ARP table is applicable only in IPv4 environments.
To display the contents of the IP ARP table, use the show iparp command.
● That the Neighbor Discovery (ND) cache has the correct entries.
NOTE
The ND cache is applicable only in IPv6 environments.
To display the contents of the ND cache, use the show neighbor-discovery cache ipv6
command.
● IP routing protocol statistics for the CPU of the switch.
Only statistics of the packets handled by the CPU are displayed. To display IP statistics for the CPU
of the switch, use one of the following commands:
■ show ipstats—IPv4 environment
■ show ipstats ipv6—IPv6 environment
● Your Open Shortest Path First (OSPF) configuration, including the OSPF area ID, router state, link
cost, OSPF timers, interface IP address, and neighbor list.

Troubleshooting
NOTE
OSPF is applicable only in IPv4 environments.
To display OSPF information, use the following commands:

■ show ospf—Displays global OSPF information for the switch
■ show ospf area—Displays information related to OSPF areas
■ show ospf area detail—Displays detailed information related to OSPF areas
■ show ospf interfaces detail—Displays detailed information about OSPF interfaces
● Your OSPFv3 configuration, including the OSPFv3 area ID, router state, link cost, OSPFv3 timers,
interface IP address, and neighbor list.
NOTE
OSPFv3 is applicable only in IPv6 environments.
To display OSPFv3 information, use the following commands:

■ show ospfv3—Displays global OSPFv3 information for the switch
■ show ospfv3 area—Displays information related to OSPFv3 areas
■ show ospfv3 interfaces—Displays detailed information about OSPFv3 interfaces
● Your Routing Information Protocol (RIP) configuration, including RIP poison reverse, split horizon,
triggered updates, transmit version, and receive version.
NOTE
RIP is applicable only in IPv4 environments.
To display detailed information about how you have RIP configured on the switch, use the show
rip command.
● RIP activity and statistics for all VLANs on the switch.
NOTE
RIP is applicable only in IPv4 environments.
To display RIP-specific statistics for all VLANs, use the show rip interface detail command.
● Your RIP next generation (RIPng) configuration, including RIPng poison reverse, split horizon,
triggered updates, transmit version, and receive version.
NOTE
RIPng is applicable only in IPv6 environments.
To display detailed information about how you have RIPng configured on the switch, use the show
ripng command.
● RIPng activity and statistics for all VLANs on the switch.

LEDs
NOTE
RIPng is applicable only in IPv6 environments.
To display RIPng-specific statistics for all VLANs, use the show ripng interface command.
● End-to-end connectivity.
To test for connectivity to a specific host, use the ping command.
● The routed path between the switch and a destination end station.
To verify and trace the routed path, use the traceroute command.
LEDs
Power LED does not light:
Check that the power cable is firmly connected to the device and to the supply outlet.
On powering-up, the MGMT LED lights yellow:
The device has failed its Power On Self Test (POST) and you should contact your supplier for advice.
A link is connected, but the Status LED does not light:
Check that:
● All connections are secure.
● Cables are free from damage.
● The devices at both ends of the link are powered-up.
● Both ends of the Gigabit link are set to the same autonegotiation state.
The Gigabit link must be enabled or disabled on both sides. If the two sides are different, typically
the side with autonegotiation disabled will have the link LED lit, and the side with autonegotiation
enabled will not be lit. The default configuration for a Gigabit port is autonegotiation enabled. Verify
by entering the following command:
show ports configuration
On power-on, some I/O modules do not boot:
Check the output of the show power budget command to see if all power supplies display the
expected input voltage. Also refer to the section “Power Management Guidelines” on page 89 for more
detailed information about power management.
ERR LED on the Management Switch Fabric Module (MSM) turns amber:
Check the syslog message for “critical” software errors. To reset the ERR LED and clear the log, use the
following command and reboot the switch:
clear log static
If you continue to see “critical” software errors or the ERR LED is still amber after issuing the clear
log static command and a switch reboot, contact Extreme Networks Technical support for further
assistance.

Troubleshooting
Status LED on the I/O module turns amber:
Check the syslog message for a related I/O module error. If the error is an inserted I/O module that
conflicts with the software configuration, use one of the following commands to reset the slot
configuration:
clear slot
configure slot <slot> module <module_type>
Otherwise, contact Extreme Networks Technical Support for further assistance.
ENV LED on the MSM turns amber:
Check each of the power supplies and all of the fans. Additionally, you display the status in the show
power and show fans displays.
Predictive Failure LED on the AC power supply blinks amber:
Check the current status of the power supply. If the speed of both fans is above 2000 RPM, the AC
power supply unit (PSU) is operating normally and no failure is imminent. To check and view the
health of the installed PSU, use the following command:
Switch does not power up:
All products manufactured by Extreme Networks use digital power supplies with surge protection. In
the event of a power surge, the protection circuits shut down the power supply.
To reset the power, unplug the switch for 1 minute, plug it back in, and attempt to power-up the
switch. If this does not work, try using a different power source (different power strip/outlet) and
power cord.
Using the Command Line Interface

This section describes helpful information for using and understanding the command line interface
(CLI). This section describes the following topics:
● General Tips and Recommendations on page 1035
● MSM Prompt—Modular Switches Only on page 1036
● Command Prompt on page 1037
● Port Configuration on page 1037
● Software License Error Messages on page 1038
● VLANs on page 1039
● STP on page 1039
● ESRP on page 1040
● VRRP on page 1041

General Tips and Recommendations

The initial welcome prompt does not display:
Check that:
● Your terminal or terminal emulator is correctly configured
● Your terminal or terminal emulator has the correct settings:
■ 9600 baud
■ 8 data bits
■ 1 stop bit
■ no parity
■ XON/OFF flow control enabled
For console port access, you may need to press [Return] several times before the welcome prompt
appears.
The SNMP Network Manager cannot access the device:
Check that:
● The Simple Network Management Protocol (SNMP) access is enabled for the system.
● The device IP address, subnet mask, and default router are correctly configured, and that the device
has been reset.
● The device IP address is correctly recorded by the SNMP Network Manager (refer to the user
documentation for the Network Manager).
● The community strings configured for the system and Network Manager are the same.
● The SNMPv3 USM, Auth, and VACM configured for the system and Network Manager are the
same.
The Telnet workstation cannot access the device:
Check that:
● The device IP address, subnet mask, and default router are correctly configured, and that the device
has been reset.
● You entered the IP address of the switch correctly when invoking the Telnet facility.
● Telnet access is enabled for the switch.
If you attempt to log in and the maximum number of Telnet sessions are being used, you should
receive an error message indicating so.
Traps are not received by the SNMP Network Manager:
Check that the SNMP Network Manager's IP address and community string are correctly configured,
and that the IP address of the Trap Receiver is configured properly on the system.
The SNMP Network Manager or Telnet workstation can no longer access the device:
Check that:
● Telnet access or SNMP access is enabled for the system.
● The port through which you are trying to access the device has not been disabled. If it is enabled,
check the connections and network cabling at the port.

Troubleshooting
● The port through which you are trying to access the device is in a correctly configured Virtual LAN
(VLAN).
● The community strings configured for the device and the Network Manager are the same.
Try accessing the device through a different port. If you can now access the device, a problem with the
original port is indicated. Re-examine the connections and cabling.
A network problem may be preventing you from accessing the device over the network. Try accessing
the device through the console port.
Permanent entries remain in the FDB:
If you have made a permanent entry in the FDB that requires you to specify the VLAN to which the
entry belongs and then deleted the VLAN, the FDB entry remains. Although this does not harm the
system, if you want to removed the entry, you must manually delete it from the FDB.
Default and static routes:
If you have defined static or default routes, those routes remain in the configuration independent of
whether the VLAN and VLAN IP address that used them remains. You should manually delete the
routes if no VLAN IP address is capable of using them.
You forget your password and cannot log in:
If you are not an administrator, another user having administrator access level can log in, delete your
user name, and create a new user name for you, with a new password.
Alternatively, another user having administrator access level can log in and initialize the device. This
will return all configuration information (including passwords) to the initial values.
In the case where no one knows a password for an administrator level user, contact your supplier.
MSM Prompt—Modular Switches Only

You do not know which MSM you are connected to:
If you use a console connection to access and configure the switch, you should connect to the console
port of the primary MSM, not the backup MSM. To determine which console port you are connected to
use the show switch command. The output displays both the primary and backup MSMs, if installed,
and an asterisk (*) appears to the right of the MSM you are connected to.
The following truncated sample output indicates that you are connected to MSM-A, the primary MSM:
MSM: MSM-A * MSM-B
You have user privileges, not administrator privileges, on the backup MSM:
If you establish a console connection to access the backup MSM, only user privileges are available. This
is true regardless of the privileges configured on the primary MSM. If you enter an administrator level
command on the backup MSM, the switch displays a message stating that the command is only
supported on the primary MSM.

Node Prompt—SummitStack Only

You do not know which Node you are connected to:
If you use a console connection to access and configure the switch, you should connect to the console
port of the master node. The word “Slot” followed by a hyphen and a single decimal digit slot number
will be inserted into the prompt in stacking mode. A sample of this prompt is:
* Slot-6 Stack.21 #
The sample indicates a changed configuration (*), the stackable is in stacking mode and is currently
using the slot number 6 in the active topology (“Slot-6 “), the system name is the default of “Stack”, the
command about to be executed is the 21st command, and the user is logged in as the administrator on
the Master node (#).
There will be no specific prompt that indicates the node role. Run show switch command to discover
the identities of the Master and Backup nodes. A successful login on a Standby node will show the “>”
character instead of the “#” character at the end of the prompt.
Command Prompt
You do not know if the switch configuration has been saved:
If an asterisk (*) precedes the command prompt, a new change to the switch configuration has not been
saved. To save the configuration, use the save configuration command. After you save the
configuration, the asterisk (*) no longer precedes the command prompt.
You do not know if you are logged in as an administrator or a user:
Observe the console prompt. If you are logged in as an administrator, the prompt ends with the hash
symbol(#). If you are logged in as a user, the prompt ends with a greater than sign (>).
The following is sample output from an administrator-level account:
BD-10808.1 #
The following is sample output from a user-level account:
BD-10808.1 >
Port Configuration
No link light on 10/100 Base port:
If patching from a switch to another switch, ensure that you are using a category 5 (CAT5) crossover
cable. This is a CAT5 cable that has pins 1 and 2 on one end connected to pins 3 and 6 on the other end.
Excessive RX CRC errors:
When a device that has autonegotiation disabled is connected to an Extreme Networks switch with
autonegotiation enabled, the Extreme Networks switch links at the correct speed, but in half-duplex
mode. The Extreme Networks switch 10/100 physical interface uses a method called parallel detection to
bring up the link. Because the other network device is not participating in autonegotiation (and does

Troubleshooting
not advertise its capabilities), parallel detection on the Extreme Networks switch is able only to sense
10 Mbps versus 100 Mbps speed and not the duplex mode. Therefore, the switch establishes the link in
half-duplex mode using the correct speed.
The only way to establish a full-duplex link is either to force it at both sides, or run autonegotiation on
both sides (using full-duplex as an advertised capability, which is the default setting on the Extreme
Networks switch).
NOTE
A mismatch of duplex mode between the Extreme switch and another network device causes poor network
performance. Viewing statistics using the show ports rxerrors command on the Extreme Networks switch may
display a constant increment of CRC errors. This is characteristic of a duplex mismatch between devices. This is
NOT a problem with the Extreme Networks switch.
Always verify that the Extreme Networks switch and the network device match in configuration for
speed and duplex.
No link light on Gigabit fiber port:
Check that:
● The transmit fiber goes to the receive fiber side of the other device and vice-versa. All Gigabit fiber
cables are of the crossover type.
● The Gigabit ports are set to Auto Off (using the command configure ports <port_list> auto
off speed [10 | 100 | 1000 | 10000] duplex [half | full]) if you are connecting the
Extreme Networks switch to devices that do not support autonegotiation.
By default, the Extreme Networks switch has autonegotiation set to On for Gigabit ports and set to
Off for 10 Gigabit ports.
● You are using multimode fiber (MMF) when using a 1000BASE-SX Gigabit Ethernet Interface
Connector (GBIC), and single-mode fiber (SMF) when using a 1000BASE-LX GBIC. 1000BASE-SX
technology does not work with SMF. The 1000BASE-LX technology works with MMF but requires
the use of a mode conditioning patchcord (MCP).
Software License Error Messages

You do not have the required software license:
If you attempt to execute a command and you do not have the required license, the switch returns the
following message:
Error: This command cannot be executed at the current license level.
You have reached the limits defined by the current software license level:
If you attempt to execute a command and you have reached the limits defined by the current license
level the switch returns the following message:
Error: You have reached the maximum limit for this feature at this license level.

VLANs
You cannot add a port to a VLAN:
If you attempt to add a port to a VLAN and get an error message similar to:
localhost:7 # configure vlan marketing add ports 1:1,1:2

Error: Protocol conflict when adding untagged port 1:1. Either add this port as tagged
or assign another protocol to this VLAN.
You already have a VLAN using untagged traffic on a port. Only one VLAN using untagged traffic can
be configured on a single physical port.
You verify the VLAN configuration using the following command:
The solution for this error using this example is to remove ports 1 and 2 from the VLAN currently
using untagged traffic on those ports. If this were the “default” VLAN, the command would be:
localhost:23 # configure vlan default delete ports 1:1,1:2
You can now re-enter the previous command without error:
localhost:26 # configure vlan marketing add ports 1:1,1:2
VLAN names:
There are restrictions on VLAN names. They cannot contain whitespaces and cannot start with a
numeric value.
VLANs, IP addresses, and default routes:
The system can have an IP address for each configured VLAN. You must configure an IP address
associated with a VLAN if you intend to manage (Telnet, SNMP, ping) through that VLAN or route IP
traffic.
You can also configure multiple default routes for the system. The system first tries the default route
with the lowest cost metric.
STP
You have connected an endstation directly to the switch and the endstation fails to boot correctly:
The switch has the Spanning Tree Protocol (STP) enabled, and the endstation is booting before the STP
initialization process is complete. Specify that STP has been disabled for that VLAN, or turn off STP for
the switch ports of the endstation and devices to which it is attempting to connect; then, reboot the
endstation.
Spanning Tree Domain names:
There are restrictions on Spanning Tree Domain (STPD) names. They cannot contain whitespaces and
cannot start with a numeric value.

Troubleshooting
You cannot add ports within a VLAN to the specified STPD:
Check to ensure that you are adding ports that already exist in the carrier VLAN.
If you see an error similar to the following:
Error: Cannot add VLAN default port 3:5 to STP domain
You might be attempting to add:

● Another 802.1D mode STP port to a physical port that already contains an 802.1D mode STP port
(only one 802.1D encapsulation STP port can be configured on a particular STP port).
● A carrier VLAN port to a different STP domain than the carrier VLAN belongs.
● A VLAN and/or port for which the carrier VLAN does not yet belong.
NOTE
This restriction is only enforced in an active STPD and when you enable STP to make sure you have a legal STP
configuration.
Only one carrier VLAN can exist in an STPD:
Only one carrier VLAN can exist in a given STPD although some of the ports on the carrier VLAN can
be outside the control of any STPD at the same time.
The StpdID must be identical to the VLANid of the carrier VLAN in that STPD.
The switch keeps aging out endstation entries in the switch FDB:
If the switch continues to age out endstation entries in the switch FDB:
● Reduce the number of topology changes by disabling STP on those systems that do not use
redundant paths.
● Specify that the endstation entries are static or permanent.
ESRP
ESRP names:
There are restrictions on Extreme Standby Router Protocol (ESRP) names. They cannot contain
whitespaces and cannot start with a numeric value.
You cannot enable an ESRP domain:
Before you enable a specific ESRP domain, it must have a domain ID. A domain ID is either a user-
configured number or the 802.1Q tag (VLANid) of the tagged master VLAN. The domain ID must be
identical on all switches participating in ESRP for that particular domain. If you do not have a domain
ID, you cannot enable ESRP on that domain.
Note the following on the interaction of tagging, ESRP, and ESRP domain IDs:
● If you have an untagged Master VLAN, you must specify an ESRP domain ID.
● If you have a tagged master VLAN, ESRP uses the 802.1Q tag (VLANid) of the master VLAN for the
ESRP domain ID. If you do not use the VLANid as the domain ID, you must specify a different
domain ID.

Using Standalone ELRP to Perform Loop Tests
You cannot delete the master VLAN from the ESRP domain:
If you attempt to remove the master VLAN before disabling the ESRP domain, you see an error
message similar to the following:
ERROR: Failed to delete master vlan for domain "esrp1" ; ESRP is enabled!
If this happens:
● Disable the ESRP domain using the disable esrp command.
● Remove the master VLAN from the ESRP domain using the configure esrp delete master
command.
VRRP
You cannot define VRRP virtual router parameters:
Before configuring any virtual router parameters for VRRP, you must first create the VRRP instance on
the switch. If you define VRRP parameters before creating the VRRP, you may see an error similar to
the following:
Error: VRRP VR for vlan vrrp1, vrid 1 does not exist.

Please create the VRRP VR before assigning parameters.
Configuration failed on backup MSM, command execution aborted!
If this happens,:
● Create a VRRP instance using the create vrrp vlan vrid command.
● Configure the VRRP instance’s parameters.

Having a tool to determine if the network has any loops is extremely useful. There are various other
protocols that can exploit this tool to prevent network loops. There are also situations where you might
want to check the topology for the existence or absence of a loop.
The Extreme Loop Recovery Protocol (ELRP) allows you to prevent, detect, and recover from Layer 2
loops in the network. You can use ELRP with other protocols such as ESRP, as described in “Using
ELRP with ESRP” on page 767. Other protocols such as Ethernet Automatic Protection Switching
(EAPS) requires that a network have a ring topology to operate. In this case you can use ELRP to ensure
that the network has a ring topology.
ELRP is used to detect network loops in a Layer 2 network. A switch running ELRP transmits multicast
packets with a special MAC destination address out of some or all of the ports belonging to a VLAN.
All of the other switches in the network treat this packet as a regular, multicast packet and flood it to all
of the ports belonging to the VLAN. If the packets transmitted by a switch are received back by that
switch, this indicates a loop in the Layer 2 network.
After a loop is detected through ELRP, different recovery actions can be taken such as blocking certain
ports to prevent loop or logging a message to system log. The action taken is largely dependent on the
protocol using ELRP to detect loops in the network.

Troubleshooting
Using ELRP with ESRP is one way you can use ELRP. For more information about configuring ESRP
and ELRP, see “Using ELRP with ESRP” on page 767. Another way to use ELRP is to invoke
“standalone” ELRP commands to determine whether a network has an Layer 2 loop or not. The
remaining sections describe how to configure standalone ELRP on your switch.

● About Standalone ELRP on page 1042
● Configuring Standalone ELRP on page 1042
● Displaying Standalone ELRP Information on page 1043
About Standalone ELRP

Standalone ELRP gives you the ability to send ELRP packets, either periodically or on an ad hoc “one-
shot” basis on a specified subset of VLAN ports. If any of these transmitted packets is received back
then standalone ELRP can perform a configured action such as sending a log message to the system log
file or sending a trap to the SNMP manager.
Standalone ELRP allows you to:

● Configure ELRP packet transmission on specified VLANs.
● Specify some or all the ports of VLAN for packet transmission.
NOTE
Reception of packets is not limited to any specific ports of the VLAN and cannot be configured.
● Configure transmission of ELRP packets on specified ports of a VLAN periodically with the added
ability to configure the interval between consecutive timings.
● Save and restore standalone ELRP configuration across reboots.
● Request periodic or non-periodic transmission of ELRP packets on specified ports of a VLAN.
For non-periodic ELRP requests:
● You can specify the number of times ELRP packets must be transmitted and the interval between
consecutive transmissions.
● A message is printed to the console and logged into the system log file indicating detection of
network loop when ELRP packets are received back or no packets are received within the
specified duration.
● There is no need to trap to the SNMP manager.
For periodic ELRP requests:
● If ELRP packets are received back, a message is printed to the system log file and a trap is sent to
the SNMP manager indicating detection of a network loop.
Configuring Standalone ELRP

This section describes configuring ELRP packet transmission to detect network loops.
The ELRP client (standalone ELRP) must be enabled globally in order for it to work on any VLANs. To
globally enable the ELRP client, use the following command:
enable elrp-client

The ELRP client can be disabled globally so that none of the ELRP VLAN configurations take effect. To
globally disable the ELRP client, use the following command:
disable elrp-client
To start one-time, non-periodic ELRP packet transmission on specified ports of a VLAN using a
particular count and interval, use one of the following commands:
● configure elrp-client one-shot <vlan_name> ports [<ports> | all] interval <sec>
retry <count> [log | print | print-and-log]—(This command is backward compatible with
Extreme Networks switches running the ExtremeWare software.)
● run elrp <vlan_name> {ports <ports>} {interval <sec>} {retry <count>}
These commands start one-time, non-periodic ELRP packet transmission on the specified ports of the
VLAN using the specified count and interval. If any of these transmitted packets is returned, indicating
loopback detection, the ELRP client can perform a configured action such as logging a message in the
system log file or printing a log message to the console. There is no need to trap to the SNMP manager
for non-periodic requests.
To start periodic ELRP packet transmission on specified ports of a VLAN using a particular interval,
configure elrp-client periodic <vlan_name> ports [<ports> | all] interval <sec> [log |
log-and-trap | trap]
This command starts periodic ELRP packet transmission on the specified ports of the VLAN using the
specified interval. If any of these transmitted packets is returned, indicating loopback detection, the
ELRP client can perform a configured action such as logging a message in the system log file and/or
sending a trap to the SNMP manager.
To disable a pending one-shot or periodic ELRP request for a specified VLAN, use the following
command:
unconfigure elrp-client <vlan_name>
Displaying Standalone ELRP Information

To display summary ELRP information, use the following command:
show elrp
The following information about ELRP appears:

● State of ELRP (enabled/disabled).
● Clients registered with ELRP
● ELRP packets transmitted
● ELRP packets received
For more detailed information about the output associated with the show elrp command, see the

Troubleshooting
Using the Rescue Software Image—Modular Switches

Only
WARNING!
The rescue image completely re-initializes the system. All data residing on the switch is cleared, including
configuration files, policy files, and other system-related files. Use this feature only with the guidance of Extreme
Networks Technical Support.
The concept of a rescue software image was introduced in ExtremeXOS 11.1. The rescue software image
recovers a switch that does not boot up by initializing the internal compact flash and installing the
ExtremeXOS software on both primary and secondary images of the compact flash. To use the rescue
software image, you must be running ExtremeXOS 11.1 or later. Earlier versions of ExtremeXOS do not
support the rescue software image.
Beginning with ExtremeXOS 11.3, the BlackDiamond 8800 series switch supports loading the rescue
image to the external compact flash memory card installed in the MSM. For more information see
“Obtaining the Rescue Image from an External Compact Flash Memory Card—BlackDiamond 8800
Series Switch Only” on page 1045.
Before you begin the recovery process, collect the following information:
● IP address, netmask, and gateway for the switch
● IP address of the Trivial File Transfer Protocol (TFTP) server that contains the ExtremeXOS image
● ExtremeXOS image filename (the image has a .xos filename extension)
NOTE
The rescue process initializes the primary and secondary images with the ExtremeXOS software image. No additional
software packages or configuration files are preserved or installed. This process takes a minimum of 7 minutes to
complete. To install additional modular software packages, BootROM images (BlackDiamond 10808 switch only),
and configuration files, see Appendix A, “Software Upgrade and Boot Options,” for more information.
Obtaining the Rescue Image from a TFTP Server

To recover the switch, you must enter the Bootloader and issue a series of commands. To access the
Bootloader:
1 Attach a serial cable to the console port of the MSM.
2 Attach the other end of the serial cable to a properly configured terminal or terminal emulator. The
terminal settings are:
● 9600 baud
● 8 data bits
● 1 stop bit
● no parity
● XON/OFF flow control enabled
3 Reboot the MSM and press the spacebar key on the keyboard of the terminal during the boot up
process.

Using the Rescue Software Image—Modular Switches Only
NOTE
You must press the spacebar key immediately after a power cycle of the MSM in order to get into the Bootloader
application.
On the BlackDiamond 12800 series switches and the BlackDiamond 8800 series switches, when you see the
BootRom banner, press the spacebar key to get into the Bootloader application.
As soon as you see the BOOTLOADER -> prompt (BlackDiamond 10808 switch) or the BootRom ->
prompt (BlackDiamond 8800 series switches and BlackDiamond 12800 series switches), release the
spacebar. From here, you can begin the recovery process.
To obtain the rescue image and recover the switch:

1 Provide the network information (IP address, netmask, and gateway) for the switch using the
following command:
configip ipaddress <ip-address>[/<netmask>] gateway <gateway-address>
● ip-address—Specifies the IP address of the switch
● netmask—Specifies the netmask of the switch
● gateway-address—Specifies the gateway of the switch
2 Download the ExtremeXOS image using the following command:
download image <tftp-address> <filename>
■ tftp-address—Specifies the IP address of the TFTP server that contains the ExtremeXOS
image
■ filename—Specifies the filename of the ExtremeXOS image
If you attempt to download a non-rescue image, the switch displays an error message and returns
you to the BOOTLOADER -> (BlackDiamond 10808 switch) or the BootRom -> (BlackDiamond 8800
series switches and BlackDiamond 12800 series switches) command prompt.
After you download the ExtremeXOS image file, the switch installs the software and reboots. After the
switch reboots, the switch enters an uninitialized state. At this point, configure the switch and save your
configuration. In addition, if you previously had modular software packages installed, you must re-
install the software packages to each switch partition. For more information about installing software
packages, see Appendix A, “Software Upgrade and Boot Options.”
If you are unable to recover the switch with the rescue image, or the switch does not reboot, contact
Extreme Networks Technical Support.
Obtaining the Rescue Image from an External Compact Flash

Memory Card—BlackDiamond 8800 Series Switch Only
In addition to recovering the switch using the internal compact flash and the management port,
ExtremeXOS 11.3 introduces support for loading the rescue image to the external compact flash
memory card installed in the MSM. The compact flash memory card must be file allocation table (FAT)
formatted. Use a PC with appropriate hardware such as a compact flash reader/writer and follow the
manufacturer’s instructions to access the compact flash card and place the image onto the card.

Troubleshooting
Before you remove or install any hardware, review the Extreme Networks Consolidated XOS Hardware
Installation Guide for correct handling instructions.
To recover the switch, you must remove power from the switch, install an appropriate compact flash
memory card into the MSM, and enter the Bootloader to issue a series of commands.
To access the Bootloader:

1 Remove all power cords from the power supplies switch. There should be no power to the switch.
2 Insert the FAT formatted compact flash memory card into the external compact flash slot of the
MSM installed in slot 5/A.
3 Remove the MSM installed in slot 6/B. Place the MSM in a safe location and do not re-install it until
you finish recovering the switch.
4 Attach a serial cable to the console port of the MSM installed in slot 5/A.
5 Attach the other end of the serial cable to a properly configured terminal or terminal emulator. The
terminal settings are:
● 9600 baud
● 8 data bits
● 1 stop bit
● no parity
● XON/OFF flow control enabled
6 Provide power to the switch by re-inserting the power cords into the power supplies.
7 Immediately press the spacebar until the BootRom -> prompt appears.
NOTE
application.
As soon as you see the BootRom -> prompt, release the spacebar. From here, you can begin the
recovery process.
To obtain the rescue image that you placed on the compact flash memory card and recover the switch:
1 Download the ExtremeXOS image that is already on the external compact flash memory card using
boot file <filename>
Where the filename specifies the image file for the BlackDiamond 8800 series switch.
2 At the BootRom -> prompt, press [Return]. The following message appears:
ok to continue
Type YES to begin the recovery process. This takes a minimum of 7 minutes.
3 After the process runs, the BootRom -> prompt displays the following message:
****press enter to reboot*****
Press [Return] to reboot the switch. The switch reboots and displays the login prompt. You have
successfully completed the setup from the external compact flash memory card.
4 Remove the external compact flash memory card installed in the MSM.

Using the Rescue Software Image—Modular Switches Only
configuration. In addition, if you previously had modular software packages installed, you must re-
install the software packages to each switch partition. For more information about installing software
Rescuing a Node in a SummitStack

You can use the rescue option on a node if it becomes unbootable due to a corrupt image. The rescue
operation is independent of the SummitStack feature.
To rescue a node:
1 Establish a terminal session using the console port of the node.

2 Reboot this node. While rebooting enter the bootrom program. To enter the bootrom, wait until you
see the message "Starting Default Bootloader ..." and then press and hold the space bar until the
bootrom prompt appears.
3 Provide the network information (IP address, netmask, and gateway) for the switch using the
following command:
config ipaddress <ip-address>[/<netmask>] gateway <gateway-address>
Where the:
■ ip-address—Specifies the IP address of the switch
■ netmask—Specifies the netmask of the switch
■ gateway-address—Specifies the gateway of the switch
4 Download the ExtremeXOS image using the following command:
download image <tftp-address> <filename>
Where the:
■ tftp-address—Specifies the IP address of the TFTP server that contains the ExtremeXOS image
■ filename—Specifies the filename of the ExtremeXOS image
If you attempt to download a non-rescue image, the switch displays an error message and returns you
to the BootRom command prompt.
configuration. In addition, if you previously had modular software packages installed, you must
reinstall the software packages to each switch partition. For more information about installing software
The rescue process:
● Does not affect stacking configuration parameters.

● Security configuration is set to factory default.

Troubleshooting
Debug Mode
The Event Management System (EMS) provides a standard way to filter and store messages generated
by the switch.With EMS, you must enable debug mode to display debug information. You must have
administrator privileges to use these commands. If you do not have administrator privileges, the switch
rejects the commands.
To enable or disable debug mode for EMS, use the following commands:
enable log debug-mode

disable log debug-mode
After debug mode has been enabled, you can configure EMS to capture specific debug information from
the switch. Details of EMS can be found in Chapter 5, “Status Monitoring and Statistics.”
Saving Debug Information

You can save switch data and statistics to an external memory card installed in the external compact
flash slot of an MSM (modular switches only), the internal memory card that comes preinstalled in the
switch, or a network TFTP server. With assistance from Extreme Networks Technical Support
personnel, you can configure the switch to capture troubleshooting information, such as a core dump
file, to the specified memory card or TFTP server.
The switch only generates core dump files in the following situations:
● If an ExtremeXOS process fails.
● When forced under the guidance of Extreme Networks Technical Support.
The core dump file contains a snapshot of the process when the error occurred.
NOTE
Use the commands described in this section only under the guidance of Extreme Networks Technical Support
personnel to troubleshoot the switch.

● Enabling the Switch to Send Debug Information to the Memory Card on page 1049
● Copying Debug Information to an External Memory Card—Modular Switches Only on page 1049
● Copying Debug Information to a TFTP Server on page 1050
● Managing Debug Files on page 1050
Modular Switches Only—Before you can enable and save process core dump information to the external
memory card, you must install an external memory card into the external compact flash slot of the
MSM. For more information about installing an external compact flash memory card, refer to the
Extreme Networks Consolidated XOS Hardware Installation Guide.

Enabling the Switch to Send Debug Information to the Memory

Card
To enable the switch to save process core dump information to the specified memory card, use the
following command:
configure debug core-dumps [internal-memory | memorycard | off]

● internal-memory—Specifies that saving debug information to the internal memory card is enabled.
Use this parameter only under the guidance of Extreme Networks Technical Support personnel.
● memorycard—Specifies that saving debug information to the external memory card is enabled. Use
this parameter only under the guidance of Extreme Networks Technical Support personnel. (This
● off—Specifies that saving debug information to the external memory card is disabled. This is the
default behavior.
Core dump files have a .gz file extension. The filename format is: core.<process-name.pid>.gz where
process-name indicates the name of the process that failed and pid is the numerical identifier of that
process. If you have a modular switch and save core dump files to the external memory card, the
filename also includes the affected MSM: MSM-A or MSM-B.
If you configure the switch to write core dump files to the internal memory card and attempt to
download a new software image, you might have insufficient space to complete the image download. If
this occurs, you must decide whether to continue the software download or move or delete the core
dump files from the internal memory. For example, if you have a modular switch with an external
memory card installed with space available, transfer the files to the external memory card. On the
Summit family of switches, transfer the files from the internal memory card to a TFTP server. This frees
up space on the internal memory card while keeping the core dump files.
Copying Debug Information to an External Memory Card—Modular

Switches Only
To save and copy debug information to the specified memory card, use the following command:
save debug tracefiles memorycard
After the switch writes a core dump file or other debug information to the external memory card, and
before you can view the contents on the card, you must ensure it is safe to remove the card from the
external compact flash slot on the MSM. Use the eject memorycard command to prepare the card for
removal. After you issue the eject memorycard command, you can manually remove the card from the
external compact flash slot on the MSM am read the data on the card.
To access and read the data on the card, use a PC with appropriate hardware such as a compact flash
reader/writer and follow the manufacturer’s instructions to access the compact flash card and read the
data.

Troubleshooting
Copying Debug Information to a TFTP Server

To save and copy debug information to the specified TFTP server, use the following command:
upload debug [<hostname> | <ipaddress>] {{vr} <vrname>}
Progress messages are displayed that indicate the file being copied and when the copying is finished.
Depending on your platform, the switch displays a message similar to the following:
The following files on have been uploaded:

Tarball Name: TechPubsLab_C_09271428.tgz
./primary.cfg
You can also use this command in conjunction with the show tech command. Prior to uploading debug
information files, the switch prompts you with the following message to run the show tech command
with the logto file option:
Do you want to run show tech logto file first? (y/n)
Enter y to run the show tech command before uploading debug information. If you enter y, the
show_tech.log.tgz file is included during the upload. Enter n to upload debug information without
running the show tech command.
After you upload the debug information, you should see a compressed TAR file on the TFTP server,
which contains the debug information.
Managing Debug Files

Using a series of commands, you can manage the files stored on the internal or external memory card.
For the purposes of this section, it is assumed that you have configured the switch to send core dump
information under the guidance of Extreme Networks Technical Support.
Managing the debug files might include any of the following tasks: renaming or copying a core dump
file, displaying a comprehensive list of files including core dump files, transferring core dump files, and
deleting a core dump file.
The following sections provide a brief overview of the available commands and describe the following
topics:
● Displaying Files on page 1051
● Moving or Renaming Files on page 1051
● Copying Files on page 1052
● Transferring Files on page 1052
● Deleting Files on page 1054
For information about managing the configuration or policy files stored on your system, see Chapter
5, “Managing the ExtremeXOS Software.”
NOTE
Filenames are case-sensitive. For information on filename restrictions, refer to the specific command in the

Displaying Files
To display a list of the files stored on your card, including core dump files, use the following command:
ls {internal-memory | memorycard}

● internal-memory—Lists the core dump files that are present and saved in the internal memory
card.
If the switch has not saved any debug files, no files are displayed.
● memorycard—Lists all files, including the core dump files, that are stored in the external compact
Output from this command includes the file size, date and time the file was last modified, and the file
name.
For more information about this command, including managing the configuration or policy files stored
on your system, see Chapter 5, “Managing the ExtremeXOS Software.”
Moving or Renaming Files

To move or rename an existing core dump file in the system, use the following command:
mv [internal-memory <old-name-internal> internal-memory <new-name-internal> |
<old-name-memorycard> memorycard <new-name-memorycard> | memorycard <new-name-
<new-name>]

● old-name-internal—Specifies the current name of the core dump file located on the internal
memory card.
● new-name-internal—Specifies the new name of the core dump file located on the internal memory
card.
● old-name-memorycard—Specifies the current name of the core dump file located on the external
compact flash memory card. (This parameter is available only on modular switches.)
● new-name-memorycard—Specifies the new name of the core dump file located on the external
compact flash memory card. (This parameter is available only on modular switches.)

Troubleshooting
Copying Files
The copy function allows you to make a copy of an existing file before you alter or edit the file. By
making a copy, you can easily go back to the original file if needed.
To copy an core dump file, use the following command:
cp [internal-memory <old-name-internal> internal-memory <new-name-internal> |

<old-name-memorycard> memorycard <new-name-memorycard> | memorycard <old-name-
<new-name>]

● old-name-internal—Specifies the name of the core dump file located on the internal memory card
that you want to copy.
● new-name-internal—Specifies the name of the newly copied core dump file located on the internal
memory card.
● old-name-memorycard—Specifies the name of the core dump file located on the external compact
● new-name-memorycard—Specifies the name of the newly copied core dump file located on the
external compact flash memory card. (This parameter is available only on modular switches.)
By making a copy of a core dump file, you can easily compare new debug information with the old file
if needed.
If you configure the switch to send core dump (debug) information to the internal memory card, specify
the internal-memory option to copy an existing core dump file. If you have a modular switch with an
external compact clash memory card installed, you can copy the core dump file to that card.
Transferring Files
TFTP allows you to transfer files to and from the switch, internal memory card, and on a modular
switch, the external memory card.
To transfer a core dump file, use the tftp, tftp get, and tftp put commands:



● host-name—Specifies the name of the remote host.
● ip-address—Specifies the IP address of the TFTP server.
NOTE
created VRs.
● -g—Gets the specified file from the TFTP server and copies it to the local host. (This parameter is
● get—Gets the specified file from the TFTP server and copies it to the local host. (This is part of the
tftp get command.)
● -p—Puts the specified file from the local host and copies it to the TFTP server. (This parameter is
● put—Puts the specified file from the local host and copies it to the TFTP server. (This is part of the
tftp put command.)
card.
● local-file-memcard—Specifies the name of the file on the external compact flash card. (This
● force-overwrite—Specifies the switch to automatically overwrite an existing file. (This parameter
is available only on the tftp get command.)
NOTE
Reference Guide.
If you configure the switch to send core dump information to the internal memory card, specify the
internal-memory option to transfer an existing core dump file from the internal memory card to the
TFTP server. If you have a modular switch with an external compact flash memory card installed,
specify the memorycard option to transfer an existing core dump file from the external memory card to
the TFTP server.
For more information about TFTP, see Chapter 4, “Managing the Switch.” For more information about
managing the configuration or policy files stored on your system, see Chapter 5, “Managing the

Troubleshooting
ExtremeXOS Software.” For detailed information about downloading software image files, BootROM
files, and switch configurations, see Appendix A, “Software Upgrade and Boot Options.”
Deleting Files
To delete a core dump file from your card, use the following command:
rm {internal-memory | memorycard} <file-name>

If you have core dump files stored in the internal memory and you attempt to download a new
software image, you might have insufficient space to complete the image download. If this happens,
delete core dump files from the internal memory.
For example, if you have a modular switch with an external memory card installed with space
available, transfer the files to the external memory card. On the Summit family switch, transfer the
files from the internal memory card to a TFTP server. This frees up space on the internal memory
card while keeping the core dump files.
● file-name—Specifies the name of the core dump file to delete.
If you delete a core dump file from the system, that file is unavailable.
You can use the * wildcard to delete core dump files from the internal memory card.
Evaluation Precedence for ACLs

The ACLs on a port are evaluated in the following order:
● Persistent dynamic ACLs
● Host-integrity permit ACLs
● MAC source address deny ACLs
● Source IP lockdown source IP permit ACLs
● Source IP lockdown deny all ACLs
● ARP validation CPU ACLs
● ACLs created using the CLI
● DoS Protect-installed ACLs
● Sentriant-installed ACLs
● CFM-installed ACLs
● MAC-in-MAC installed ACLs
● ACLs applied with a policy file (see Chapter 12, “Access Lists (ACLs),” for precedence among these
ACLs)

TOP Command
For information on policy files and ACLs, see Chapter 11, “Policy Manager,” and Chapter 12, “Access
Lists (ACLs).”
TOP Command
The top command is a UNIX-based command that displays real-time CPU utilization information by
process. The output contains a list of the most CPU-intensive tasks and can be sorted by CPU usage,
memory usage, and run time. For more detailed information about the top command, refer to your
UNIX documentation.
TFTP Server Requirements

Extreme Networks recommends using a TFTP server that supports blocksize negotiation (as described
in RFC 2348, TFTP Blocksize Option), to enable faster file downloads and larger file downloads.
System Health Check

This section provides a brief overview the system health check functionality of the following platforms:
● Summit family of switches
For all platforms, system health check errors are reported to the syslog. If you see an error, contact
For more detailed information about the system health checker, including configuration examples for
the modular platforms, see Chapter 5, “Status Monitoring and Statistics.”
Overview of the System Health Checker

Two modes of health checking are available on the modular platforms: polling and backplane
diagnostic packets. Only polling is available on the Summit family of switches. These system health
check methods are briefly described for each platform.
BlackDiamond 10808 Switch and the BlackDiamond 12800 Series Switches

● Polling is always enabled on the system and occurs every 60 seconds by default. The system health
checker polls and tracks the ASIC counters that collect correctable and uncorrectable packet memory
errors, check sum errors, and parity errors on a per ASIC basis. By reading and processing the
registers, the system health check detects and associates faults to specific system ASICs.
● Backplane diagnostic packets are disabled by default. After this feature is enabled, the system health
checker tests the packet path for a specific I/O module every 6 seconds by default. The Management

Troubleshooting
Switch Fabric Module (MSM) sends and receives diagnostic packets from the I/O module to
determine the state and connectivity. (The other I/O modules with backplane diagnostic packets
disabled continue polling every 60 seconds by default.)
BlackDiamond 8800 Series Switch

● Polling is always enabled on the system and occurs every 5 seconds by default. The polling value is
not a user-configured parameter. The system health check polls the control plane health between
MSMs and I/O modules, monitors memory levels on the I/O module, monitors the health of the
I/O module, and checks the health of applications and processes running on the I/O module. If the
system health checker detects an error, the health checker notifies the MSM.
checker tests the data link for a specific BlackDiamond 8800 original I/O module every 5 seconds by
default. The MSM sends and receives diagnostic packets from the I/O module to determine the state
and connectivity.
Beginning with ExtremeXOS 11.5 for the BlackDiamond 8800 e-series and a-series I/O modules, if
you enable this feature, the system health checker tests and verifies the 10 Gbps links between the
switch fabric and the ports on the specified I/O module. The MSM sends and receives diagnostic
packets from the I/O module to determine the state and connectivity. Similar to the BlackDiamond
8800 original I/O modules, this test occurs every 5 seconds by default.
Summit Family of Switches

On the Summit family of switches, the system health checker polls and reads the switch fabric and CPU
registers. Unlike the modular platforms, only polling is available on the Summit family of switches.
Polling is always enabled on the system and occurs in the background every 10 seconds; the polling
value is not a user-configured parameter.
Enabling and Disabling Backplane Diagnostic Packets on the

Switch—Modular Switches Only
To enable backplane diagnostic packets, use the following command:
enable sys-health-check slot <slot>
checker tests the packet path every 6 seconds for the specified slot.
● BlackDiamond 8800 series switch—By default, the system health checker tests the data link
(BlackDiamond 8800 original modules) or the 10 Gbps links (BlackDiamond e-series and a-series
modules) every 5 seconds for the specified slot.
NOTE
Enabling backplane diagnostic packets increases CPU utilization and competes with network traffic for resources.
To disable backplane diagnostic packets, use the following command:
disable sys-health-check slot <slot>

System Odometer
● BlackDiamond 10808 and the BlackDiamond 12800 series switches—By default, the system health
checker discontinues sending backplane diagnostic packets and returns the polling frequency to
60 seconds on the specified slot. Only polling is enabled.
● BlackDiamond 8800 series switches—By default, the system health checker discontinues sending
backplane diagnostic packets to the specified slot. Only polling is enabled.
Configuring Backplane Diagnostic Packets on the Switch—

To configure the frequency of sending backplane diagnostic packets, use the following command:
configure sys-health-check interval <interval>
NOTE
Extreme Networks does not recommend configuring an interval of less than the default interval. Doing so can cause
System Odometer
Each field replaceable component contains a system odometer counter in EEPROM. The show
odometers command displays an approximate days of service duration for an individual component
since the component was manufactured.
Monitored Components
On a modular switch, the odometer monitors the following components:
● Chassis
● MSMs
● I/O modules
● Power controllers
On the Summit family of switches, the odometer monitors the following components:
● Switch
● XGN-2xn card
Recorded Statistics
The following odometer statistics are collected by the switch:
● Service Days—The amount of days that the component has been running
● First Recorded Start Date—The date that the component was powered-up and began running
Depending on the software version running on your switch, the modules installed in your switch, and
the type of switch you have, additional or different odometer information may be displayed.

Troubleshooting
The following is sample output from a BlackDiamond 10808 switch:
Service First Recorded

Field Replaceable Units Days Start Date
----------------------- ------- --------------
Chassis : BD-10808 107 Feb-23-2004
Slot-1 : G60X 99 Dec-10-2003
Slot-2 : G60X 74 Mar-22-2004
Slot-3 : G60X 151 Jan-12-2004
Slot-4 :
Slot-5 : 10G6X 49 Apr-09-2004
Slot-6 :
Slot-7 : G60T 184 Dec-03-2003
Slot-8 : 10G6X 146 Jan-12-2004
MSM-A : MSM-1XL 62 Apr-21-2004
MSM-B : MSM-1XL 172 Dec-14-2003
PSUCTRL-1 : 152 Mar-17-2004
PSUCTRL-2 :
The following is sample output from the BlackDiamond 8800 series switch:

------------------------- ------- --------------
Chassis : BD-8810 209 Dec-07-2004
Slot-1 : G48T 208 Dec-07-2004
Slot-2 : 10G4X 219 Nov-02-2004
Slot-3 : G48T 228 Oct-26-2004
Slot-4 : G24X 226 Oct-19-2004
Slot-5 : G8X 139 Dec-07-2004
Slot-6 :
Slot-7 : 10G4X 160 Dec-16-2004
Slot-8 : 10G4X 133 Dec-14-2004
Slot-9 : G48P 111 Nov-04-2004
Slot-10 :
MSM-A : MSM-G8X 137 Dec-07-2004
MSM-B :
PSUCTRL-1 : 209 Dec-07-2004
PSUCTRL-2 : 208 Dec-07-2004
The following is sample output from a BlackDiamond 12804 switch:

------------------------- ------- --------------
Chassis : BD-8904 40 Sep-22-2005
Slot-1 : G20XT 20 Nov-10-2005
Slot-2 :
Slot-3 :
Slot-4 :
Slot-5 :
Slot-6 : G20T 20 Nov-10-2005
Slot-7 :
Slot-8 :
MSM-A : MSM-5 20 Nov-10-2005
MSM-B : MSM-5 20 Nov-10-2005

Temperature Operating Range
PSUCTRL-1 : 0 Nov-05-2005
PSUCTRL-2 : 0 Nov-05-2005
The following is sample output from a Summit X450 series switch:

------------------------- ------- --------------
Switch : SummitX450-24t 7 Dec-08-2004
XGM-2xn-1 :
Temperature Operating Range

ExtremeXOS has its own temperature operating range of -10° to 50° C.
On a modular switch, any module in the switch that is reported outside this range is automatically shut
down. ExtremeXOS specifically performs a reboot on any MSM that falls outside the expected range.
This behavior is expected and not indicative of a problem. If you experience this behavior more than
once, contact Extreme Networks Technical Support.
On the Summit X450 family switch, if the switch runs outside the expected range, the switch logs an
error message, generates a trap, and continues running. No components are shutdown. To verify the
state of the switch, use either the show switch or show temperature commands. If the temperature
exceeds the maximum limit, the show switch output indicates the switch in an OPERATIONAL
(Overheat) mode, and the show temperature output indicates an error state due to overheat.
Unsupported Module Type—BlackDiamond 10808

Switch and BlackDiamond 8800 Series Switches Only
BlackDiamond 10808 Switch—If you are running ExtremeXOS 11.2 or earlier and install a module that
is not supported by the currently running software image, the console displays a message similar to the
following:
Msg on boot up "HEEELLLPPPP!!! Unknown cardtype" due to image does not support the
blade.
If you see this message, the switch does not checkpoint the configuration or synchronize the primary
and backup MSMs until you remove the unsupported module.
BlackDiamond 8800 Series Switches—If you install a module that is not supported by the currently
running software image, the console displays a message similar to the following:
“Unable to read cardtype for slot 4”
If you see this message, the switch checkpoints the configuration and synchronizes the primary and
backup MSMs; however, the unsupported module is not brought up and is not available for use. If you
have an unsupported module installed and use the show slot command, the slot state appears as
Empty.

Troubleshooting
Corrupted BootROM on the BlackDiamond 8800 Series

Switch
If your default BootROM image becomes corrupted, you can force the MSM to boot from an alternate
BootROM image, by inserting a pen into the Alternate (A) and Reset (R) holes on the BlackDiamond
8800 MSM and applying pressure. The alternate BootROM image also prints boot progress indicators,
and you can later use this alternate image to re-install a new default BootROM image. Finally, a
corrupted compact flash can be recovered from either the Alternate or Default BootROM.
For more information, refer to the Extreme Networks Consolidated XOS Hardware Installation Guide.
Inserting Powered Devices in the PoE Module—

BlackDiamond 8800 Series Switch Only
powered devices (PDs) are not powered when the actual delivered power for the module is within
approximately 19 W of the configured inline power budget for that slot. However, actual aggregate
power can be delivered up to the configured inline power budget for the slot (for example, when
delivered power from ports increases or when the configured inline power budget for the slot is
reduced).
Modifying the Hardware Table Hash Algorithm—

BlackDiamond 8800 Series Switch, SummitStack, and
the Summit Family of Switches Only
With hardware forwarding, the switch stores addresses in the hardware table to quickly forward
packets to their destination. The switch uses a hash algorithm to decide where to store the addresses in
the hardware table. The standard, default hash algorithm works well for most systems; however, for
some addresses with certain patterns, the hardware may attempt to store address information in the
same section of the hardware. This can cause an overflow of the hardware table even though there is
enough room to store addresses.
Error Messages Displayed With ExtremeXOS 11.4 and Earlier. If you experience a full hardware table that
affects Layer 2, IP local host, and IP multicast forwarding, you see messages similar to the following in
the log:
<Info:HAL.IPv4Adj.Info> : adj 136.159.188.109: IP add error is Table full for new or
newly resolved ARP, egress valid
<Info:HAL.IPv4Adj.Info> : adj 136.159.188.109: returned -17 for L3 table bucket 181
<Warn:HAL.IPv4Mc.Warning> : Could not allocate a hardware S,G,V entry
(889f4648,effffffa,70) - hardware table resource exceeded (rv=-17).

Modifying the Hardware Table Hash Algorithm—BlackDiamond 8800 Series Switch, SummitStack, and the Summit Family
Error Messages Displayed With ExtremeXOS 11.5 and Later. If you experience a full hardware table that
affects Layer 2, IP local host, and IP multicast forwarding, you see messages similar to the following in
the log:
<HAL.IPv4Adj.L3TblFull> MSM-A: IPv4 unicast entry not added. Hardware L3 Table full.
<Card.IPv4Adj.Warning> Slot 4: IPv4 unicast entry not added. Hardware L3 Table full.
<HAL.IPv4Mc.GrpTblFullEnt> MSM-A: IPv4 multicast entry (10.0.0.1,224.1.1.1,vlan 1) not
added. Hardware Group Table full.
<Card.IPv4Mc.Warning> Slot-4: IPv4 multicast entry not added. Hardware L3 Table full.
Configuring the Hash Algorithm

If you experience a full hardware table, you can configure a different hash algorithm to select a different
section of the hardware to store addresses. You must save your configuration and reboot the switch to
modify the hash algorithm used by the hardware table.
NOTE
Modify the hardware table hash algorithm only with the guidance of Extreme Networks technical personnel.
To modify the hardware table utilization, use the following command:
configure forwarding hash-algorithm [crc16 | crc32]

● In ExtremeXOS 11.5, the default hash algorithm is crc32.
● In ExtremeXOS 11.4 and earlier, the default hash algorithm is crc16.
After you enter the command, the switch displays a message similar to the following:
Warning: This command will only take effect after a save and reboot
To use the new hash algorithm, save your switch configuration and reboot the switch.
Upgrading to ExtremeXOS 11.5. When you upgrade to ExtremeXOS 11.5, the hash algorithm
automatically becomes crc32. For example, if you saved a configuration using an image from
ExtremeXOS 11.4 or earlier with the hash algorithm set to crc16, when ExtremeXOS 11.5 loads on the
switch, the hash algorithm becomes crc32. To change the hash algorithm to crc16, use the configure
forwarding hash-algorithm crc16 command and save your switch configuration.
Viewing the Hash Algorithm Setting

To view the hardware table settings on the switch, including the configured hash algorithm and the
current hash algorithm, use the following command:
show forwarding configuration
Forwarding table hash algorithm

Configured hash: crc16
Current hash: crc16

Troubleshooting
Untagged Frames on the 10 Gbps Module—

BlackDiamond 10808 Switch Only
On the BlackDiamond 10808 switch, the 10 Gbps module must have the serial number 804405-00-09 or
higher to support untagged frames. To display the serial number of the module, use the show slot
<slot_number> command. (All the modules for the BlackDiamond 8800 series switch support tagged
and untagged frames.)
Understanding the Error Reading Diagnostics

Message—Summit Family of Switches Only
If you have never run diagnostics on the switch or stack ports and use the show diagnostics
command, the switch displays a message similar to the following:
Result: FAIL
Test date run is invalid. Please run Diagnostics.
Error reading diagnostics information.
This message is normal and expected if you have never run diagnostics on the switch. After running
diagnostics, you should see information about the executed test using the show diagnostics
command.
Running MSM Diagnostics from the Bootloader—

BlackDiamond 10808 Switch Only
If you experience problems with your MSM module, or you are unable to use the run diagnostics
command, you can enter the Bootloader and issue a series of commands to run diagnostics on the MSM.
To access the Bootloader:

1 Attach a serial cable to the console port of the MSM.
3 Reboot the MSM and press the spacebar key on the keyboard of the terminal during the boot up
process.
NOTE
application.
As soon as you see the BOOTLOADER> prompt, release the key. From here, you can run the diagnostics
on the MSM.

Contacting Extreme Networks Technical Support
To run diagnostics on the MSM:

1 Identify the currently running software images by using the show images command.
2 Run diagnostics on the MSM by using the boot [1-4] command.
The numbers 1 through 4 correlate to specific images and diagnostics on the MSM:
● 1—XOS primary image
● 2—XOS secondary image
● 3—Diagnostics for image 1 (initiates diagnostics for the primary image)
● 4—Diagnostics for image 2 (initiates diagnostics for the secondary image)
For example, to run diagnostics on the primary image, use the following command:
boot 3
When the test is finished, the MSM reboots and runs the ExtremeXOS software.
Contacting Extreme Networks Technical Support

If you have a network issue that you are unable to resolve, contact Extreme Networks technical support.
Extreme Networks maintains several Technical Assistance Centers (TACs) around the world to answer
networking questions and resolve network problems.
You can contact technical support by phone at:

● (800) 998-2408
● (408) 579-2826
Or by email at:
● [email protected]
You can also visit the support website at:
http://www.extremenetworks.com/services/resources/
From the support website, you can download software updates (requires a service contract) and
documentation (including a .pdf version of this manual).

Troubleshooting

C CNA Agent
The entire CNA software package consists of multiple parts. The Extreme Networks devices run only
the CNA Agent. You must have the entire package; you cannot use the CNA Agent without the CNA
software from Avaya. The user interface is a combination of a Java applet hosted from the CNA Server
and a command line interface (CLI). You configure and manage the CNA Agent using the CLI.
NOTE
Contact your Avaya representative to obtain the rest of the CNA software and the associated documentation.
If you are using Avaya CNA solutions, you download a separate software module to add the CNA
Agent to ExtremeXOS 11.2 software; this software module runs on all platforms.
NOTE
You must download and install the SSH software module before downloading and installing the CNA Agent software
module.
This appendix covers the following topics regarding the CNA Agent:
● Downloading the CNA Agent Software Module on page 1066
● Running the Tests on page 1066
● Configuring the CNA Agent on page 1067
Overview
The CNA Agent accepts requests (from the CNA Server) to run tests for measuring and verifying
network performance, and the CNA Agent reports back the results to the CNA Server. The CNA Agent
functions as a UDP service accepting authenticated requests from and returning test results to the CNA
Server.
You download the CNA Agent software as a separate software module onto your device. If you do not
have it already, you must also download the separate SSH software module, which contains SSL.
After you enable the software, the CNA Agent registers with the CNA Server and exchanges openSSL
keys, using a 128-bit encryption key. All following communication between the CNA Agent and the
CNA Server is encrypted.
The CNA Server communicates with the CNA Agent to request specific tests and to schedule those
tests. The CNA Agent runs the requested tests and sends the results back to the CNA Server.
The CNA Agent can register with one CNA Server, talk to one CNA Server at a time, and respond to
only one test request at a time. The other test requests are queued up.

CNA Agent
The system sends messages to the syslog for each new connection, the status of that connection,
received test requests, and reported test results. Also EMS messages are generated after successful
registration.
Redundancy—Modular Switches and SummitStack Only

The CNA Agent software offers resiliency and redundancy on the BlackDiamond 10808 switch, the
BlackDiamond 12800 series switches, SummitStack, and the BlackDiamond 8800 series switch. On
systems with a primary and a backup node, the CNA Agent software fails over to the backup node
when the primary node goes down. Also, the node restarts the CNA Agent software if the CNA Agent
goes down or gets into a looped condition.
Downloading the CNA Agent Software Module

To use the CNA Agent functionality, you download the separate Extreme Networks software module
(cna.xmod) following the instructions outlined in Appendix A, “Software Upgrade and Boot Options.”
Because the CNA Agent software uses openSSL, you must first download the separate Extreme
Networks SSH software module that contains Secure Sockets Layer (SSL).
NOTE
You must download the Secure Shell (SSH) module before downloading the CNA module.
If you attempt to download the CNA software module and you have not already downloaded the SSH
software module, the system returns an error.
Running the Tests

The CNA Server requests the CNA Agent to run a specific test, and the CNA Agent runs the test and
reports the results back to the CNA Server.
The CNA Agent can run the following tests:

● Traceroute—Measures per-hop round-trip delays to a target IP address by sending a sequence of
hop-limited UPD messages, each with a time-to-live (TTL) value that is one greater than that of the
preceding message
● RTP—Emulates VoIP traffic to measure delay, packet loss and jitter to another CNA Agent by
sending a simulated RTP stream that is echoed back
● Ping—Sends an ICMP echo message to a target IP address and reports whether or not a response
was returned
● TCPconnect—Attempts to establish a TCP connection to a specified port at a target IP address and
reports whether the attempt succeeded or failed
● Merge—Used by CNA software to identify a single device with multiple IP addresses and to merge
its multiple appearances into one in the network topology map

Configuring the CNA Agent
The CNA Agent starts the specified test within 100 ms after it receives an authenticated and correctly
formatted test request from the CNA Server. The CNA Agent sends the test results to the CNA Server
within 100 ms of test completion.
Configuring the CNA Agent

To run the tests, configure the following:
● Enable the CNA Agent.
● Configure the IP address the CNA Agent uses to communicate with the CNA Server.
● Configure the VLAN interface for the CNA Server
■ This step is optional. By default, the CNA Server uses the primary IP address on the switch’s
Default VLAN.
● Clear the counters.
■ This step is optional.
● Display the test results.
To save all these configurations, use the save configuration command.
Enabling the CNA Agent

To enable the CNA Agent (test plug), use the following command:
enable cna-testplug
After you enable the CNA Agent, you register the CNA Agent with the CNA Server, and the CNA
Agent performs the requested network tests and reports the results.
To disable the CNA Agent, use the following command:
disable cna-testplug
Connecting to the CNA Server

After you enable the CNA Agent, you must configure this parameter to allow the CNA Agent to
register with the CNA Server. Use the following command:
configure cna-testplug scheduler ipaddress <ip_address>
You enter the IP address of the CNA Server. (With ExtremeXOS software version 11.2, this must be an
IPv4 address.)
After you configure the IP address for the CNA server, the CNA Agent and the CNA Server exchange
SSL keys and establish encryption. After successful registration, you can connect with more than one
CNA Server, in which case you share the encryption key you negotiated with your initial CNA Server
connection.
At this time, the system establishes the socket connection to the CNA Server using this IP address. The
CNA Agent listens for instructions for the testing from this IP address.

CNA Agent
Configuring the Interface

By default, the Extreme Networks device uses the default VLAN as the interface that the CNA Agent
(test plug) uses to receive test requests, conduct the tests, and send the results to the CNA Server. (The
default VLAN belongs to the default virtual router: VR-Default).
If you want to change this interface, use the following command:
configure cna-testplug vlan <vlan_name>
If the VLAN has more than one IP address, the system uses the primary IP address.
NOTE
Extreme Networks recommends that you put IP telephones on the same virtual router.
Clearing the Counters

To clear the CNA Agent (test plug) counters, use the following command:
clear cna-testplug counters
This command clears the CNA Agent counters on the Extreme Networks devices and resets those
counters to 0.
You can also use the clear counters command, which clears all the counters on the device including
those associated with the CNA Agent.
Displaying CNA Agent Information

To display summaries of completed tests as well as the CNA Server (scheduler) connections and status
for the CNA Agent (test plug), use the following command:
enable cna-testplug
Troubleshooting
If the CNA Agent is not able to register with the CNA Server, check the following items:
● Ensure the time on the Extreme Networks device is set correctly.
■ To display the time, use the show switch command.
■ To reset the time, use the configure time <month> <day> <year> <hour> <min> <sec>
command.
● Ensure that the openSSL certificate on the CNA Server has not expired.

D Supported Protocols, MIBs, and Standards
This appendix provides a list of software standards and protocols supported by ExtremeXOS. This
appendix includes the following topics:
● General Routing and Switching on page 1069

● Virtual LANS (VLANs), Virtual MANs (vMANs) and MAC in MAC on page 1069
● Quality of Service (QoS) and Policies on page 1070
● Routing Information Protocol (RIP) on page 1070
● Open Shortest Path First (OSPF) on page 1070
● Border Gateway Protocol 4 (BGP4) on page 1070
● Power over Ethernet (PoE) on page 1070
● IP Multicast on page 1070
● Management - SNMP & MIBs on page 1071
● Management - Other on page 1072
● Security on page 1072
● IPv6 on page 1072
● MIB Support Details on page 1073
General Routing and Switching

RFC 1812 Requirements for IP Version 4 Routers RFC 826 Ethernet Address Resolution ProtocolRFC
2338 Virtual Router Redundancy Protocol
RFC 1519 An Architecture for IP Address Allocation
with CIDR RFC 3619 Ethernet Automatic Protection Switching
(EAPS) Version 1 and EAPS Version 2
RFC 1256 ICMP Router Discovery Messages
IEEE 802.1ab Link Layer Discovery Protocol (LLDP)
RFC 1122 Requirements for Internet Hosts -
Communication Layers IEEE 802.1D-1998 Spanning Tree Protocol
RFC 768 User Datagram Protocol IEEE 802.1W - 2001 Rapid Spanning Tree Protocol
RFC 791 Internet Protocol Extreme Standby Router Protocol (ESRP)
RFC 792 Internet Control Message Protocol IEEE 802.1Q - 1998 Virtual Bridged Local Area
Networks
RFC 793 Transmission Control Protocol
Virtual LANS (VLANs), Virtual MANs (vMANs) and MAC in MAC

IEEE 802.1Q VLAN Tagging Protocol-sensitive VLANs
IEEE 802.3ad Static load sharing configuration and Multiple STP domains per VLAN
LACP-based dynamic configuration
Virtual MANs

Supported Protocols, MIBs, and Standards
Quality of Service (QoS) and Policies

IEEE 802.1D -1998 (802.1p) Packet Priority RFC 2597 Assured Forwarding
RFC 2474 Definition of the Differentiated Services Field RFC 2475 An Architecture for Differentiated Service
(DS Field) in the IPv4 and IPv6 Headers (Core and Edge Router Functions)
RFC 2598 DiffServ Expedited Forwarding
Routing Information Protocol (RIP)

RFC 1058 Routing Information Protocol v1 RFC 2453 RIP Version 2
Open Shortest Path First (OSPF)

RFC 2328 OSPF Version 2 RFC 1765 OSPF Database Overflow
RFC 1587 The OSPF NSSA Option RFC 2370 The OSPF Opaque LSA Option
RFC 3623 Graceful OSPF Restart
Border Gateway Protocol 4 (BGP4)

RFC 1771 A Border Gateway Protocol 4 (BGP-4) RFC 1745 BGP4/IDRP for IP---OSPF Interaction
RFC 1965 Autonomous System Confederations for BGP RFC 2385 Protection of BGP Sessions via the TCP MD5
Signature Option
RFC 2796 BGP Route Reflection - An Alternative to
Full Mesh IBGP RFC 2439 BGP Route Flap Dampening
RFC 1997 BGP Communities Attribute
Power over Ethernet (PoE)

RFC 3621 Power over Ethernet MIB IEEE 802.3af standard
IP Multicast
RFC 1112 Host extensions for IP multicasting (Internet RFC 2362 Protocol Independent Multicast-Sparse Mode
Group Management Protocol version 1) (PIM-SM): Protocol Specification
RFC 2236 IGMP Version 2 PIM-DM Draft IETF PIM Dense Mode t
RFC 3376 IGMP Version 3 RFC 3618 MSDP
IGMP Snooping with Configurable Router Registration
Forwarding

Management - SNMP & MIBs
RFC 1155 Structure and identification of management RFC 2575 View-based Access Control Model (VACM) for
information for TCP/IP-based internets the Simple Network Management Protocol
RFC 1157 Simple Network Management Protocol (SNMP) RFC 2576 Coexistence between Version 1, Version 2,
and Version 3 of the Internet-standard Network
RFC 1212 Concise MIB definitions
Management Framework
RFC 1213 Management Information Base for Network
RFC 1757 Remote Network Monitoring Management
Management of TCP/IP-based internets: MIB-II
Information Base
RFC 1215 Convention for defining traps for use with
RFC 2021 Remote Network Monitoring Management
the Simple Network Management Protocol (SNMP)
Information Base Version 2 using SMIv2
RFC 2233 Evolution of the Interfaces Group of MIB-II
RFC 1724 RIP Version 2 MIB Extension
RFC 1901 Introduction to Community-based SNMPv2
RFC 1850 OSPF Version 2 Management Information Base
RFC 1902 Structure of Management Information for
BGP4-V2-MIB draft-ietf-idr-bgp4-mibv2-02.txt (in lieu
Version 2 of the Simple Network Management Protocol
of RFC 1657)
(SNMPv2)
RFC 1493 Definitions of Managed Objects for Bridges
RFC 1903 Textual Conventions for Version 2 of the
Simple Network Management Protocol (SNMPv2) Definition of Managed Objects for Bridges with Rapid
Spanning Tree Draft-ietf-bridge-rstpmib-03.txt
RFC 1904 Conformance Statements for Version 2 of the
Simple Network Management Protocol (SNMPv2) RFC 2465 Management Information Base for IP
Version 6: Textual Conventions Group
RFC 1905 Protocol Operations for Version 2 of the
Simple Network Management Protocol (SNMPv2) RFC 2466 Management Information Base for IP
Version 6: ICMPv6 Group
RFC 1906 Transport Mappings for Version 2 of the
Simple Network Management Protocol (SNMPv2) RFC 2665 Definitions of Managed Objects for the
Ethernet-like Interface Types
RFC 1907 Management Information Base for Version 2
of the Simple Network Management Protocol (SNMPv2) RFC 2668 Definitions of Managed Objects for IEEE
802.3 Medium Attachment Units (MAUs)
RFC 1908 Coexistence between Version 1 and Version 2
of the Internet-standard Network Management Framework RFC 2787 Definitions of Managed Objects for the
RFC 2570 Introduction and Applicability Statements for
Internet-Standard Management Framework RFC 2737 Entity MIB (Version 2)
RFC 2571 An Architecture for Describing Simple Network RFC 3621 Power over Ethernet MIB (BlackDiamond
Management Protocol (SNMP) Management Frameworks 8800 family of switches only)
RFC 2572 Message Processing and Dispatching for the PIM MIB draft-ietf-pim-mib-v2-01.txt
Simple Network Management Protocol (SNMP)
IEEE-8021-PAE-MIB
RFC 2573 Simple Network Management Protocol
IEEE8021X-EXTENSIONS-MIB
(SNMP) Applications
ExtremeWare vendor MIBs (includes statistics, STP,
RFC 2574 User-based Security Model (USM) for version 3
CPU monitoring, EAPS, ACL, CLEAR-Flow, and others)
of the Simple Network Management Protocol (SNMPv3)
EAPS MIB supports get functions
IEEE 802.1 AB LLDP Basic MIB
IEEE 802.1 AB LLDP-EXT-DOT1-MIB
IEEE 802.1 AB LLDP-EXT-DOT3-MIB

Management - Other
RFC 854 Telnet Protocol Specification BSD System Logging Protocol (SYSLOG), with Multiple
Syslog Servers
Telnet client and server
Local Messages (criticals stored across reboots)
Secure Shell 2 (SSH2) client and server
RFC 2030 Simple Network Time Protocol (SNTP)
Secure Copy 2 (SCP2) client and server
Version 4 for IPv4 and OSI
Configuration logging
Multiple Images, Multiple Configs
Security
Routing protocol authentication IEEE 802.1x Port Based Network Access Control
RFC 1492 An Access Control Protocol, Sometimes RFC 2138 Remote Authentication Dial In User Service
Called TACACS (RADIUS)
Secure Shell (SSHv2) & Secure Copy (SCPv2) with RFC 2139 RADIUS Accounting
encryption/authentication
Access Control Lists (ACLs)
Secure Socket Layer (SSL)
RFC 2406 IP Encapsulating Security Payload (ESP)
IPv6
RFC 2460, Internet Protocol, Version 6 (IPv6) RFC 3587, Global Unicast Address Format
Specification
RFC 2710, IPv6 Multicast Listener Discovery v1
RFC 2461, Neighbor Discovery for IP Version 6, (IPv6) (MLDv1) Protocol
RFC 2462, IPv6 Stateless Address Auto configuration - RFC 3810, IPv6 Multicast Listener Discovery v2
Router Requirements (MLDv2) Protocol
RFC 2463, Internet Control Message Protocol (ICMPv6) RFC 2740, OSPF for IPv6
for the IPv6 Specification
RFC 2080, RIPng
RFC 2464, Transmission of IPv6 Packets over Ethernet
RFC 2893, Configured Tunnels
Networks
RFC 3056, 6to4
RFC 2465, IPv6 MIB, General Group and Textual
Conventions Static Unicast routes for IPv6
RFC 2466, MIB for ICMPv6 Telnet server over IPv6 transport
RFC 1981, Path MTU Discovery for IPv6, August 1996 SSH-2 server over IPv6 transport
- Router requirements
Ping over IPv6 transport
RFC 3513, Internet Protocol Version 6 (IPv6)
Traceroute over IPv6 transport
Addressing Architecture

MIB Support Details
MIB Support Details

The following sections describes the MIB support provided by the ExtremeXOS SNMP agent residing
on Extreme Networks devices running ExtremeXOS. This support includes standard MIBs (RFC based
and Internet draft based) listed in the “Management - SNMP & MIBs” table on on page 1071 as well as
Extreme Networks ‘private’ MIBs.
Where applicable, the document notes how the implementation differs from the standards or from the
private MIBs.
Standard MIBs
RFC 1213 (MIB-II)

The following tables, groups, and variables are supported in this MIB.
Table/Group Supported Variables Comments

System group scalars All objects The object 'sysServices' will always return the value '79'.
Interfaces group Supported as per RFC 2233.
IP Group scalars All objects
ipAddrTable All objects
ipRouteTable All objects Supported as read only. Routes are indexed by prefix only.
ipNetToMediaTable All objects
ICMP group All objects
TCP group scalars All objects
tcpConnTable All objects
UDP group scalars All objects
udpTable All objects
EGP Group Not supported
SNMP group All objects
At group All objects Supported as read only.
RFC 2233 (IF-MIB)


IfNumber
ifTable ifIndex The ifIndex for ports is calculated as ((slot * 1000) + port).
For VLANs, the ifIndex starts from 1000001.
ifDescr
ifType Only the following values are supported:
{other, ethernetCsmacd, softwareLoopback, propVirtual}
ifMtu


ifSpeed
ifPhysAddress
ifAdminStatus The testing state is not supported.
ifOperStatus
ifLastChange
ifInOctets Updated every time SNMP queries this counter.
ifInUcastPkts Updated every time SNMP queries this counter.
IfInNUcastPkts(deprecated) Though deprecated, this object will return a value, if the
system keeps a count. Updated every time SNMP queries
this counter.
ifInDiscards No count is kept of this object, so it will always return 0.
ifInErrors Updated every time SNMP queries this counter.
ifInUnknownProtos No count is kept of this object, so it will always return 0.
ifOutOctets Updated every time SNMP queries this counter.
ifOutUcastPkts Updated every time SNMP queries this counter.
IfOutNUcastPkts(deprecated) Though deprecated, this object will return a value, if the
this counter.
ifOutDiscards Updated every time SNMP queries this counter.
ifOutErrors Updated every time SNMP queries this counter.
IfOutQLen(deprecated) Though deprecated, this object will return a value, if the
this counter.
IfSpecific Not implemented. Will always return “iso.org.dod.internet”.
ifXTable All objects Only port interfaces will return non-zero values for the
counter objects in this table.
The object 'ifPromiscuousMode' is supported read-only.
‘ifCounterDiscontinuityTime’ is not implemented.
All the statistics counters in the ifXTable are updated every
time SNMP queries them.
ifStackTable Not supported
IfTestTable Not supported
ifRcvAddressTable All objects The ‘ifRcvAddressTable’ is supported read-only. Also, only
entries for physical ports will appear in it.
snmpTraps linkDown
linkUp

MIB Support Details
RFC 1215
This MIB defines an SMI for SNMPv1 traps, and some traps themselves. Of these, the following are
supported.
Traps Comments
coldStart The system cannot distinguish between a cold and warm reboot, so the warmStart
trap is always sent. This is the same behavior as in Extremeware.
warmStart
authenticationFailure
linkDown
linkUp
RFC 1493 (BRIDGE-MIB) and draft-ietf-bridge-rstpmib-03.txt

The BRIDGE-MIB has been augmented with draft-ietf-bridge-rstpmib-03.txt for 802.1w support. Objects
below that are defined in the latter are marked as such.

dot1dBase group scalars dot1dBaseBridgeAddress
dot1dBaseNumPorts This object returns the number of ports in STP
domain ‘s0’, not the total number of ports on
the switch.
dot1dBaseType
dot1dBasePortTable Partial dot1dBasePortIfIndex is supported.
dot1dBasePortCircuit will always have the value {
0 0 } since there is a one to one correspondence
between a physical port and its ifIndex.
dot1dBasePortDelayExceededDiscards and
dot1dBasePortMtuExceededDiscards are not
supported and always return 0.


dot1dStp group scalars dot1dStpProtocolSpecification
dot1dStpPriority
dot1dStpTimeSinceTopologyChang
e
dot1dStpTopChanges
dot1dStpDesignatedRoot
dot1dStpRootCost Values for these objects will be returned for the
STP domain 's0' only. For other domains, see
dot1dStpRootPort
the EXTREME-STPEXTENSTIONS-MIB.
dot1dStpMaxAge
dot1dStpHelloTime
dot1dStpHoldTime
dot1dStpForwardDelay
dot1dStpBridgeMaxAge
dot1dStpBridgeHelloTime
dot1dStpBridgeForwardDelay
dot1dStpVersion This object is not present in the original
RFC1493, but is defined in the Internet draft
‘draft-ietf-bridge-rstp-mib-03.txt’.
dot1dStpTxHoldCount This object is not present in the original
This object not supported; it always returns a
value of (1). Attempting to set it yields an
error.
dot1dStpPathCostDefault This object is not present in the original
For this object only 8021d1998(1) is
supported at this time, not stp802112001(2).
Attempting to set (2) yields an error.
dot1dStpExtPortTable All objects This object is not present in the original
The object ‘dot1dStpPortProtocolMigration’ is
not supported; it always returns a value of (2).
Attempting to set it yields an error.
dot1dStpPortTable All objects
STP Traps newRoot
topologyChange

MIB Support Details

dot1dTpFdbTable Supported The object dot1dTpFdbTable displays ports and
FDB mac addresses. They include both the
static and dynamic FDB entries on the switch.
The MIB does not provide a way to identify the
VLAN on which the entry was learned. The
ports numbers are assumed to be 1 to 128 on
Slot 1, and 128 to 255 on Slot 2, etc. (that is,
with a total of 128 ports on each of the slots
on a Chassis system).
dot1dTpPortTable Supported
dot1dStatic group Supported
Dot1dStaticAllowedToGoTo Not supported
RFC 1724 (RIPv2-MIB)


rip2Globals rip2GlobalRouteChanges
rip2GlobalQueries
rip2IfStatTable All objects
rip2IfConfTable All objects
rip2PeerTable Not supported

RFC 1850 (OSPF-MIB)


ospfGeneralGroup All objects
ospfAreaTable All objects
ospfStubAreaTable All objects
ospfLsdbTable All objects
ospfAreaRangeTable All objects
ospfHostTable All objects
ospfIfTable All objects
Only entries for the default VR are supported.
ospfIfMetricTable All objects
ospfVirtIfTable All objects
ospfNbrTable All objects
ospfVirtNbrTable All objects
ospfExtLsdbTable All objects
ospfAreaAggregateTable All objects
ospfTrap All traps
BGP4-V2-MIB (draft-ietf-idr-bgp4-mibv2-02.txt)
This MIB is supported in lieu of RFC 1657(BGP4-MIB). The following tables, groups, and variables are
supported in this MIB.
Table/Group Supported Variables Comment

bgpM2BaseNotifications All notifications
bgpM2VersionTable bgpM2VersionIndex
bgpM2VersionSupported Supports BGP version 4 only.
bgpM2SupportedAuthTable bgpM2SupportedAuthCode Does not support any authentication.
bgpM2SupportedAuthValue
bgpM2SupportedCapabilities bgpM2CapabilitySupportAvailable Supports capabilities advertisement
(BGP-MP, Route-Refresh and Cisco-
Route-Refresh).
bgpM2SupportedCapabilitiesTable bgpM2SupportedCapabilityCode Supports following capabilities:
BGP-MP — 1
Route-Refresh — 2
Cisco Router-Refresh — 128
bgpM2SupportedCapability
bgpM2BaseScalars bgpM2AsSize Implements AS Size of twoOctet only.
bgpM2LocalAs
bgpM2LocalIdentifier
bgpM2BaseScalarRouteReflectExts bgpM2RouteReflector
bgpM2ClusterId

MIB Support Details

bgpM2BaseScalarASConfedExts bgpM2ConfederationRouter
bgpM2ConfederationId
bgpM2BaseScalarConfiguration bgpM2CfgBaseScalarStorageType This object is not implemented.
bgpM2CfgLocalAs
bgpM2CfgLocalIdentifier
bgpM2CfgBaseScalarReflectorExts bgpM2CfgRouteReflector Supported read-only.
bgpM2CfgClusterId
bgpM2CfgBaseScalarASConfedExts bgpM2CfgConfederationRouter Supported read-only.
bgpM2CfgConfederationId
bgpM2PeerTable bgpM2PeerIdentifier
bgpM2PeerState
bgpM2PeerStatus
bgpM2PeerConfiguredVersion
bgpM2PeerNegotiatedVersion
bgpM2PeerLocalAddrType Only IPV4 address types are supported.
bgpM2PeerLocalAddr
bgpM2PeerLocalPort
bgpM2PeerLocalAs
bgpM2PeerRemoteAddrType Only IPV4 address types are supported.
bgpM2PeerRemoteAddr
bgpM2PeerRemotePort
bgpM2PeerRemoteAs
bgpM2PeerIndex
bgpM2PeerErrorsTable bgpM2PeerLastErrorReceived
bgpM2PeerLastErrorSent
bgpM2PeerLastErrorReceivedTime
bgpM2PeerLastErrorSentTime
bgpM2PeerLastErrorReceivedText
bgpM2PeerLastErrorSentText
bgpM2PeerLastErrorReceivedData Not implemented in BGP. Always
sends NULL data.
bgpM2PeerLastErrorSentData Not implemented in BGP. Always
sends NULL data.
bgpM2PeerAuthTable bgpM2PeerAuthSent
bgpM2PeerAuthSentCode
bgpM2PeerAuthSentValue These entries are always NULL, as
ExtremeXOS BGP does not implement
bgpM2PeerAuthRcvd authentication.
bgpM2PeerAuthRcvdCode
bgpM2PeerAuthRcvdValue
bgpM2PeerEventTimesTable bgpM2PeerFsmEstablishedTime


bgpM2PeerInUpdatesElapsedTime
bgpM2PeerConfiguredTimersTable bgpM2PeerConnectRetryInterval
bgpM2PeerHoldTimeConfigured
bgpM2PeerKeepAliveConfigured
bgpM2PeerMinASOrigInterval
bgpM2PeerMinRouteAdverInterval
bgpM2PeerNegotiatedTimersTable bgpM2PeerHoldTime
bgpM2PeerKeepAlive
bgpM2PeerCapsAnnouncedTable bgpM2PeerCapAnnouncedCode
bgpM2PeerCapAnnouncedIndex
bgpM2PeerCapAnnouncedValue
bgpM2PeerCapsReceivedTable bgpM2PeerCapReceivedCode
bgpM2PeerCapReceivedIndex
bgpM2PeerCapReceivedValue
bgpM2PeerCountersTable bgpM2PeerInUpdates
bgpM2PeerOutUpdates
bgpM2PeerInTotalMessages
bgpM2PeerOutTotalMessages
bgpM2PeerFsmEstablishedTrans
bgpM2PrefixCountersTable bgpM2PrefixCountersAfi
bgpM2PrefixCountersSafi
bgpM2PrefixInPrefixes
bgpM2PrefixInPrefixesAccepted
bgpM2PrefixInPrefixesRejected
bgpM2PrefixOutPrefixes
bgpM2PeerReflectorClientTable bgpM2PeerReflectorClient “meshedClient” type is not supported.
bgpM2PeerConfedMemberTable bgpM2PeerConfedMember
bgpM2CfgPeerAdminStatusTable bgpM2CfgPeerAdminStatus
bgpM2CfgPeerNextIndex
bgpM2CfgPeerTable bgpM2CfgPeerConfiguredVersion Only supported value is 4.
bgpM2CfgAllowVersionNegotiation Supported as read-only object.
bgpM2CfgPeerLocalAddrType Only IPV4 address types are supported.
bgpM2CfgPeerLocalAddr
bgpM2CfgPeerLocalAs Local AS per neighbor cannot be
configured. This object is
implemented as read-only.
bgpM2CfgPeerRemoteAddrType Only IPV4 address types are supported.
bgpM2CfgPeerRemoteAddr
bgpM2CfgPeerRemotePort Implemented as read-only.
bgpM2CfgPeerRemoteAs

MIB Support Details

bgpM2CfgPeerEntryStorageType Not supported.
bgpM2CfgPeerError Not supported.
bgpM2CfgPeerBgpPeerEntry Not supported.
bgpM2CfgPeerRowEntryStatus
bgpM2CfgPeerIndex
bgpM2CfgPeerStatus
bgpM2CfgPeerAuthTable bgpM2CfgPeerAuthEnabled
bgpM2CfgPeerAuthCode This table is implemented as read-only.
bgpM2CfgPeerAuthValue
bgpM2CfgPeerTimersTable bgpM2CfgPeerConnectRetryInterval Implemented as read-only.
bgpM2CfgPeerHoldTimeConfigured
bgpM2CfgPeerKeepAliveConfigured
bgpM2CfgPeerMinASOrigInterval Implemented as read-only.
bgpM2CfgPeerMinRouteAdverInter Implemented as read-only.
bgpM2CfgPeerReflectorClientTable bgpM2CfgPeerReflectorClient “meshedClient” type is not supported.
bgpM2CfgPeerConfedMemberTable bgpM2CfgPeerConfedMember
bgpM2NlriTable bgpM2NlriIndex Not implemented. Always returns 0.
bgpM2NlriAfi
bgpM2NlriSafi
bgpM2NlriPrefix
bgpM2NlriPrefixLen
bgpM2NlriBest
bgpM2NlriCalcLocalPref Local preference is returned.
bgpM2PathAttrIndex
bgpM2NlriOpaqueType Not implemented. Returns a default
value.
bgpM2NlriOpaquePointer Not implemented. Returns a default
value.
bgpM2AdjRibsOutTable bgpM2AdjRibsOutIndex Not implemented. Always returns 0.
bgpM2AdjRibsOutRoute
bgpM2Rib bgpM2PathAttrCount
bgpM2PathAttrTable bgpM2PathAttrOrigin
bgpM2PathAttrNextHopAddrType
bgpM2PathAttrNextHop
bgpM2PathAttrMedPresent
bgpM2PathAttrMed
bgpM2PathAttrLocalPrefPresent
bgpM2PathAttrLocalPref
bgpM2PathAttrAtomicAggregate
bgpM2PathAttrAggregatorAS


bgpM2PathAttrAggregatorAddr
bgpM2AsPathCalcLength
bgpM2AsPathString
bgpM2AsPathIndex PathAttrIndex is used for the
corresponding AS-Path index too.
bgpM2AsPath4byteTable Not supported This table is not supported.
bgpM2AsPathTable bgpM2AsPathSegmentIndex
bgpM2AsPathElementIndex
bgpM2AsPathType
bgpM2AsPathElementValue
bgpM2PathAttrUnknownTable Not supported This table is not implemented by
ExtremeXOS BGP.
bgpM2PathAttrOriginatorIdTable bgpM2PathAttrOriginatorId
bgpM2PathAttrClusterTable bgpM2PathAttrClusterIndex
bgpM2PathAttrClusterValue
bgpM2PathAttrCommTable bgpM2PathAttrCommIndex
bgpM2PathAttrCommValue
bgpM2PathAttrExtCommTable Not supported ExtremeXOS BGP does not implement
Extended Communities.
bgpM2LinkLocalNextHopTable Not supported This table is not supported (this is for
BGP extension for IPV6).
RFC 2668 (MAU-MIB)


ifMauTable All objects
ifJackTable All objects
ifMauAutoNegTable All objects Setting auto-negotiation via SNMP is not
supported.
The following new Extreme proprietary Mau types have been added to the ifMauType textual
convention:
extremeMauType1000BaseWDMHD OBJECT IDENTIFIER

::= { extremeMauType 7 }
"Gigabit WDM, half duplex"
extremeMauType1000BaseWDMFD OBJECT IDENTIFIER
"Gigabit WDM, full duplex"
extremeMauType1000BaseLX70HD OBJECT IDENTIFIER
"Gigabit LX70, half duplex"
extremeMauType1000BaseLX70FD OBJECT IDENTIFIER

MIB Support Details
"Gigabit LX70, full duplex"

extremeMauType1000BaseZXHD OBJECT IDENTIFIER
"Gigabit ZX, half duplex"
extremeMauType1000BaseZXFD OBJECT IDENTIFIER
"Gigabit ZX, full duplex"
Corresponding Mau Type List Bits values have been added:
extreme_ifMauTypeListBits_b1000baseWDMHD-- 64
extreme_ifMauTypeListBits_b1000baseWDMFD-- 65
extreme_ifMauTypeListBits_b1000baseLX70HD-- 66
extreme_ifMauTypeListBits_b1000baseLX70FD-- 67
extreme_ifMauTypeListBits_b1000baseZXHD-- 68
extreme_ifMauTypeListBits_b1000baseZXFD-- 69
The following standards-based additions have been made as a 'Work in Progress', as per draft-ietf-
hubmib-mau-mib-v3-02.txt.
1. A new enumeration 'fiberLC(14)' for the JackType textual convention.
2. New Mau types:
dot3MauType10GigBaseX OBJECT-IDENTITY
STATUS current
DESCRIPTION "X PCS/PMA (per 802.3 section 48), unknown PMD."
::= { dot3MauType 31 }
dot3MauType10GigBaseLX4 OBJECT-IDENTITY
STATUS current
DESCRIPTION "X fiber over WWDM optics (per 802.3 section 53)"
dot3MauType10GigBaseR OBJECT-IDENTITY
STATUS current
DESCRIPTION "R PCS/PMA (per 802.3 section 49), unknown PMD."
dot3MauType10GigBaseER OBJECT-IDENTITY
STATUS current
DESCRIPTION "R fiber over 1550 nm optics (per 802.3 section 52)"
dot3MauType10GigBaseLR OBJECT-IDENTITY
STATUS current
dot3MauType10GigBaseSR OBJECT-IDENTITY
STATUS current
dot3MauType10GigBaseW OBJECT-IDENTITY
STATUS current

DESCRIPTION "W PCS/PMA (per 802.3 section 49 and 50), unknown PMD."
dot3MauType10GigBaseEW OBJECT-IDENTITY
STATUS current
DESCRIPTION "W fiber over 1550 nm optics (per 802.3 section 52)"
dot3MauType10GigBaseLW OBJECT-IDENTITY
STATUS current
dot3MauType10GigBaseSW OBJECT-IDENTITY
STATUS current
3. Corresponding new Mau Type List bit values:
b10GbaseX(31) — 10GBASE-X
b10GbaseLX4(32) — 10GBASE-LX4
b10GbaseR(33) — 10GBASE-R
b10GbaseER(34 — 10GBASE-ER
b10GbaseLR(35) — 10GBASE-LR
b10GbaseSR(36) — 10GBASE-SR
b10GbaseW(37) — 10GBASE-W
b10GbaseEW(38) — 10GBASE-EW
b10GbaseLW(39) — 10GBASE-LW
b10GbaseSW(40) — 10GBASE-SW
RFC 2787 (VRRP-MIB)


vrrpOperations vrrpNodeVersion
vrrpNotificationCntl
vrrpStatistics vrrpRouterChecksumErrors
vrrpRouterVersionErrors
vrrpRouterVrIdErrors
vrrpOperTable All objects Creation of a new row or modifying an existing
row requires vrrpOperAdminState to be set to
'down'; otherwise any kind of set will fail on this
table.
vrrpOperAuthType does not support
'ipAuthenticationHeader'.
vrrpAssoIpAddrTable All objects

MIB Support Details

vrrpRouterStatsTable All objects
vrrpNotifications vrrpTrapNewMaster
vrrpTrapAuthFailure
PIM-MIB (draft-ietf-pim-mib-v2-01.txt)
This MIB is superset of RFC 2934.

pimInterfaceTable pimInterfaceIfIndex
pimInterfaceAddress
pimInterfaceNetMask
pimInterfaceMode
pimInterfaceDR
pimInterfaceHelloInterval
pimInterfaceStatus
pimInterfaceJoinPruneInterval
pimInterfaceCBSRPreference
pimInterfaceTrigHelloInterval Not supported.
pimInterfaceHelloHoldtime
pimInterfaceLanPruneDelay
pimInterfacePropagationDelay
pimInterfaceOverrideInterval
pimInterfaceGenerationID These objects are supported as read only.
pimInterfaceJoinPruneHoldtime
pimInterfaceGraftRetryInterval
pimInterfaceMaxGraftRetries
pimInterfaceSRTTLThreshold
pimInterfaceLanDelayEnabled
pimInterfaceSRCapable
pimInterfaceDRPriority This object is supported as read only.
pimNeighborTable pimNeighborAddress
pimNeighborIfIndex
pimNeighborUpTime
pimNeighborExpiryTime
pimNeighborMode Feature unsupported, only default value is
returned.
pimNeighborLanPruneDelay Feature unsupported, only default value is
returned.
pimNeighborOverrideInterval


pimNeighborTBit Feature unsupported so only default value is
returned.
pimNeighborSRCapable Feature unsupported so only default value is
returned.
pimNeighborDRPresent Feature unsupported so only default value is
returned.
pimIpMRouteTable pimIpMRouteUpstreamAssertTimer
pimIpMRouteAssertMetric
pimIpMRouteAssertMetricPref
pimIpMRouteAssertRPTBit
pimIpMRouteFlags
pimIpMRouteRPFNeighbor
pimIpMRouteSourceTimer
pimIpMRouteOriginatorSRTTL Feature unsupported so only default value is
returned.
pimIpMRouteNextHopTable pimIpMRouteNextHopPruneReason
pimIpMRouteNextHopAssertWinner
pimIpMRouteNextHopAssertTimer
pimIpMRouteNextHopAssertMetric Not supported.
pimIpMRouteNextHopAssertMetricPref Not supported.
pimIpMRouteNextHopJoinPruneTimer Not supported.
pimRPSetTable pimRPSetGroupAddress
pimRPSetGroupMask
pimRPSetAddress
pimRPSetHoldTime
pimRPSetExpiryTime
pimRPSetComponent
pimCandidateRPTable pimCandidateRPGroupAddress
pimCandidateRPGroupMask
pimCandidateRPAddress
pimCandidateRPRowStatus
pimComponentTable pimComponentIndex
pimComponentBSRAddress
pimComponentBSRExpiryTime
pimComponentCRPHoldTime This object is supported as read only.
pimComponentStatus
Scalars pimJoinPruneInterval

MIB Support Details

pimSourceLifetime
pimStateRefreshInterval State Refresh feature is not supported, so
pimStateRefreshLimitInterval these variables are set to defaults.
pimStateRefreshTimeToLive
PIM Traps pimNeighborLoss Not supported.
SNMPv3 MIBs
The XOS SNMP stack fully supports the SNMPv3 protocol and therefore implements the MIBs in the
SNMPv3 RFCs. Specifically, the MIBs in following RFCs are fully supported.
● RFC 2570 – Introduction to Version 3 of the Internet-standard Network Management Framework
● RFC 2571 – An Architecture for describing SNMP Management Frameworks
● RFC 2572 – Message Processing and Dispatching for the Simple Network Management Protocol
(SNMP)
● RFC 2573 – SNMPv3 Applications.
● RFC 2574 – User-based Security Model (USM) for version 3 of the Simple Network Management
Protocol (SNMPv3)
● RFC 2575 – View-based Access Control Model (VACM) for the Simple Network Management
Protocol (SNMP)
● RFC 2576 – Coexistence between Version 1, Version 2, and Version 3 of the Internet-standard
Network Management Framework
RFC 2737 (ENTITY-MIB)


entityPhysicalTable entPhysicalndex
entPhysicalDescr
entPhysicalVendorType
entPhysicalContainedIn
entPhysicalClass
entPhysicalParentRelPos
entPhysicalName
entPhysicalHardwareRev
entPhysicalFirmwareRev
entPhysicalSoftwareRev
entPhysicalSerialNum
entPhysicalMfgName
entPhysicalModelName


entPhysicalAlias
entPhysicalAssetID
entPhysicalIsFRU
RFC 3621 (PoE-MIB)


pethPsePortTable All objects Objects in this table are read-only.
pethMainPseTable All objects Objects in this table are read-only.
pethNotifications pethPsePortOnOffNotification
pethMainPowerUsageOnNotification
pethMainPowerUsageOffNotification
IEEE-8021-PAE-MIB
This MIB contains objects for the 802.1X protocol draft D10 of the 802.1X standard. The following tables,
groups, and variables are supported in this MIB.

dot1xPaeSystemAuthControl
dot1xPaePortTable All objects
dot1xAuthConfigTable Not supported In lieu of these tables, Extreme Networks supports the
per-station based versions which are present in the
dot1xAuthStatsTable Not supported IEEE8021X- EXTENSIONS-MIB.
dot1xAuthDiagTable Not supported This table has been deprecated in the drafts subsequent
to the 2001 version of the 802.1X standard.
dot1xAuthSessionStatsTable Not supported
dot1xSuppConfigTable None These tables are not applicable to the switch since they
are for a supplicant.
dot1xSuppStatsTable None
IEEE8021X-EXTENSIONS-MIB

dot1xAuthStationTable All objects
dot1xAuthConfigTable All objects
dot1xAuthStatsTable All objects

MIB Support Details
RFC 1757 (RMON-MIB)


etherStatsTable All objects, except
etherStatsDropEvents
historyControlTable All objects
etherHistoryTable All objects, except
etherHistoryDropEvents
alarmTable All objects
eventTable All objects
logTable All objects
RFC 2021 (RMON2-MIB)


probeCapabilities
probeSoftwareRev
probeHardwareRev
probeDateTime
probeResetControl
trapDestTable All objects
RFC 2465 (IPV6 MIB)


ipv6Forwarding All objects
ipv6DefaultHopLimit All objects
ipv6Interfaces All objects
ipv6IfTableLastChange All objects
ipv6IfTable All objects except
ipv6IfEffectiveMtu
ipv6IfStatsTable All objects
ipv6AddrPrefixTable All objects
ipv6AddrTable All objects
ipv6RouteNumber All objects


ipv6DiscardedRoutes All objects
ipv6RouteTable All objects
ipv6NetToMediaTable All objects
RFC 2466 (IPV6 ICMP MIB)


ipv6IfIcmpTable All objects
IEEE 802.1AB (LLDP-MIB)

All tables and variables of this MIB are supported.
IEEE 802.1AB (LLDP-EXT-DOT1-MIB)

IEEE 802.1AB (LLDP-EXT-DOT3-MIB)

RFC 2665 (EtherLike-MIB)


Dot3StatsTable dot3StatsIndex
dot3StatsAlignmentErrors
dot3StatsFCSErrors
dot3StatsSingleCollisionFrames
dot3StatsMultipleCollisionFrames
dot3StatsSQETestErrors Not supported
dot3StatsDeferredTransmissions
dot3StatsLateCollisions
dot3StatsExcessiveCollisions
dot3StatsInternalMacTransmitErrors
dot3StatsCarrierSenseErrors Not supported
dot3StatsFrameTooLongs
dot3StatsInternalMacReceiveErrors

MIB Support Details

dot3StatsSymbolErrors Not supported
dot3StatsEtherChipSet
dot3StatsDuplexStatus
dot3CollTable dot3CollCount
dot3CollFrequencies
dot3ControlTable Not supported
dot3PauseTable Not supported
Other unsupported Tables and nodes in EtherLike MIB: dot3ControlTable, dot3PauseTable, dot3Tests – all
nodes under this, dot3Errors, etherConformance, etherGroups, etherCompliance, dot3Compliance
Extreme Networks Proprietary MIBs
EXTREME-SYSTEM-MIB

extremeSaveConfiguration
extremeSaveStatus
extremeCurrentConfigInUse
extremeConfigToUseOnReboot
extremeOverTemperatureAlarm
extremePrimaryPowerOperational Not supported: always returns
True.
extremePowerStatus Not supported: always returns
presentOK
extremePowerAlarm Not supported: always returns
False.
extremeRedundantPowerStatus
extremeRedundantPowerAlarm Not supported: always returns
presentOK
extremeInputPowerVoltage Contains the input voltage of
the latest power-supply to
power-on in a system with
multiple power-supplies.
extremePrimarySoftwareRev


extremeSecondarySoftwareRev
ExtremeImageToUseOnReboot
extremeSystemID
extremeSystemBoardID
extremeSystemLeftBoardID
Not supported.
extremeSystemRightBoardID
extremeRmonEnable
extremeBootROMVersion Returns information for the
current MSM only.
extremeDot1dTpFdbTableEnable
extremeHealthCheckErrorType
extremeHealthCheckAction
extremeHealthCheckMaxRetries
extremeCpuUtilRisingThreshold
extremeCpuTaskUtilPair Not supported.
extremeCpuAggregateUtilization
extremeCpuUtilRisingThreshold
extremeAuthFailSrcAddr
extremeCpuTransmitPriority
extremeImageBooted
extremeMasterMSMSlot This will return the internal
slot number assigned to the
MSM: 9,10 on the
BlackDiamond 10808;
11,12 on the BlackDiamond
8810;
7,8 on the BlackDiamond
8806
extremeChassisPortsPerSlot Returns the max ports per slot
for the system.
ExtremeMsmFailoverCause
extremeFanStatusTable All objects Fans are now within fan-trays,
so they are numbered to be
able to identify the tray
number: 101-10N for tray1,
201-20N for tray2 etc.
FanNo = (TrayNum*100+FanNum)
extremeCpuTaskTable All objects Not supported.
extremeCpuTask2Table All objects Not supported.
extremeSlotTable All objects Cards are currently not
configurable via SNMP.
extremePowerSupplyTable extremePowerSupplyStatus,
extremePowerSupplyInputVoltage

MIB Support Details

extremeImageTable extremeImageNumber This table contains image
information for all images
installed on the device.
extremeImageNumber values
are not compatible with EW
releases: Current image has
value 3 instead of 0.
extremeMajorVersion
extremeSubMajorVersion
extremeMinorVersion
extremeBuildNumber
extremeTechnologyReleaseNumber
extremeSustainingReleaseNumber
extremeBranchRevisionNumber
extremeImageType
extremeImageDescription Description of image contains
image version, including major
version, submajor version,
minor version, build version,
build branch, build-master
login, build date.
extremePatchVersion
extremeCpuMonitorInterval
extremeCpuMonitorTotalUtilization
extremeCpuMonitorTable
extremeCpuMonitorSlotId
extremeCpuMonitorProcessName
extremeCpuMonitorProcessId
extremeCpuMonitorProcessState
extremeCpuMonitorUtilization5secs
extremeCpuMonitorUtilization1min
extremeCpuMonitorUtilization5mins
extremeCpuMonitorUtilization30mins
extremeCpuMonitorUtilization1hour
extremeCpuMonitorUserTime
extremeCpuMonitorSystemTime
extremeCpuMonitorSystemTable
extremeCpuMonitorSystemSlotId
extremeCpuMonitorSystemUtilization5secs


extremeCpuMonitorSystemUtilization1min
extremeCpuMonitorSystemUtilization5mins
extremeCpuMonitorSystemUtilization30mins
extremeCpuMonitorSystemUtilization1hour
extremeCpuMonitorSystemMaxUtilization
extremeMemoryMonitorSystemTable
extremeMemoryMonitorSystemSlotId
extremeMemoryMonitorSystemTotal
extremeMemoryMonitorSystemFree
extremeMemoryMonitorSystemUsage
extremeMemoryMonitorUserUsage
extremeMemoryMonitorTable
extremeMemoryMonitorSlotId
extremeMemoryMonitorProcessName
extremeMemoryMonitorUsage
extremeMemoryMonitorLimit
extremeMemoryMonitorZone
extremeMemoryMonitorGreenZoneCount
extremeMemoryMonitorYellowZoneCount
extremeMemoryMonitorOrangeZoneCount
extremeMemoryMonitorRedZoneCount
extremeMemoryMonitorGreenZoneThreshold
extremeMemoryMonitorYellowZoneThreshold
extremeMemoryMonitorOrangeZoneThreshold
extremeMemoryMonitorRedZoneThreshold
EXTREME-VLAN-MIB

extremeVirtualGroup extremeNextAvailableVirtIfIndex
extremeVlanIfTable extremeVlanIfIndex While creating a new row in the
extremeVlanIfTable, the value of the
object extremeVlanIfDescr MUST be
specified.
For all tables in this MIB that contain
objects with RowStatus semantics, the
only values supported are:
{ active, createAndGo, destroy }
extremeVlanIfDescr
extremeVlanIfType

MIB Support Details

extremeVlanIfGlobalIdentifier Not supported.
extremeVlanIfStatus
extremeVlanIfIgnoreStpFlag Not supported.
extremeVlanIfIgnoreBpduFlag Not supported.
extremeVlanIfLoopbackModeFlag
extremeVlanIfVlanId
extremeGlobalMappingTable Not supported
extremeVlanEncapsTable Not supported
extremeVlanIpTable All objects For all tables in this MIB that contain
extremeVlanProtocolTable Not supported
extremeVlanProtocolBindingTable Partial For all tables in this MIB that contain
New to ExtremeXOS: Association of a
protocol filter and VLAN.
extremeVlanProtocolVlanTable Not supported
extremeVlanProtocolDefTable Partial For all tables in this MIB that contain
New to ExtremeXOS: This is a new table
introduced to add a protocol filter with
etype values during creation of filter
itself.
extremeVlanOpaqueTable All objects This is a read only table. For adding ports
to a VLAN or deleting ports to a vlan, use
extremeVlanOpaqueControlTable.
extremeVlanOpaqueControlTable All objects For all tables in this MIB that contain
New to ExtremeXOS:
extremeVlanOpaqueControlTable is a write
only table and cannot be used to read.
This is used to add/delete ports on a
VLAN.
extremeVlanStackTable All objects Not supported.
extremeVlanL2StatsTable All objects Not supported.
This table contains per VLAN information
about the number of packets sent to the
CPU, the number of packets learnt, the
number of Igmp control packets snooped
and the number of Igmp data packets
switched. This is the same information
that is available via the CLI command
'show l2stats'.

EXTREME-TRAPPOLL-MIB

extremeSmartTrapFlushInstanceTableIndex This object acts as a flush control
for the
extremeSmartTrapInstanceTable.
Setting this object can flush the
matching entries from the
extremeSmartTrapInstanceTable
based on certain rules as defined
in the MIB.
extremeSmartTrapRulesTable The entries created in the
extremeSmartTrapRulesTable
define the rules that are used to
generate Extreme smart traps.
The object
extremeSmartTrapRulesDesiredOID
supports OID values whose prefix is
among the following:
ipAddrTable
ifMauTable
extremeSlotTable
extremeVlanGroup
extremeVirtualGroup
extremeVlanProtocolTable
extremeVlanProtocolVlanTable
extremeVlanOpaqueTable
extremeStpDomainTable
extremeStpPortTable
extremeStpVlanPortTable
pethPsePortTable
extremePethSystem
extremePethPseSlotTable
extremePethPsePortTable
extremeSmartTrapInstanceTable extremeSmartTrapInstanceTable is
a read-only table that stores the
information about which variables
have changed according to rules
defined in the
extremeSmartTrapRulesTable.
EXTREME-ESRP-MIB

extremeEsrpDomainTable All objects
extremeEsrpDomainMemberTable All objects
extremeEsrpDomainNeighborTable All objects
extremeEsrpDomainAwareTable All objects
extremeEsrpDomainStatsTable All objects
extremeEsrpNotifications extremeEsrpDomainStateChange

MIB Support Details
EXTREME-EDP-MIB

extremeEdpTable All objects
extremeEdpNeighborTable All objects
extremeEdpPortTable All objects
EXTREME-OSPF-MIB

extremeOspfInterfaceTable All objects
EXTREME-FDB-MIB

extremeFdbPermFdbTable All objects
Support SNMP get and get next operations only.
extremeFdbMacFdbCounterTable All objects
EXTREME-TRAP-MIB
This MIB defines the following Extreme-specific SNMPv1 traps generated by Extreme Networks
devices.
Trap Comments
extremeOverheat
extremeFanFailed
extremeFanOK
extremeInvalidLoginAttempt
extremePowerSupplyGood
extremePowerSupplyFail
extremeEdpNeighborAdded
extremeEdpNeighborRemoved
extremeModuleStateChanged

EXTREME-V2TRAP-MIB
This MIB defines the following Extreme-specific SNMPv2c traps generated by Extreme Networks
devices.
Trap Comments
extremeHealthCheckFailed
extremeMsmFailoverTrap
extremeBgpM2PrefixReachedThreshold
extremeBgpM2PrefixMaxExceeded
extremeEapsStateChange Send on master/transit nodes.
extremeEapsFailTimerExpFlagSet
extremeEapsFailTimerExpFlagSet
extremeEapsLinkDownRingComplete
extremeEapsLastStatusChangeTime Send on master/transit nodes. Provides a general indication of a status
change using a 10 second timer.
extremeEapsPortStatusChange Send on master/transit nodes.
extremeEapsConfigChange Send on master/transit nodes. This trap has a granularity of 30
seconds.
extremeEapsSharedPortStateChange Send on controller/partner nodes.
extremeEapsRootBlockerStatusChange Send on controller/partner nodes.
extremeNMSInventoryChanged These traps are not generated by the ExtremeXOS SNMP agent but by
the EPICenter NMS.
extremeNMSTopologyChanged
EXTREME-STPEXTENSIONS-MIB

extremeStpDomainTable All objects
extremeStpPortTable All objects These tables contain information on all STP
domains, whereas the BRIDGE-MIB contains
extremeStpVlanPortTable All objects information only of the STP domain 's0'.
EXTREME-ENTITY-MIB

extremeEntityFRUTable entPhysicalndex
extremeEntityStartTime
extremeEntityOdometer
extremeEntityOdometerUnit

MIB Support Details
EXTREME-CLEARFLOW-MIB
This MIB defines the following Extreme-specific traps generated by Extreme Networks devices.
Trap Comments
extremeClearflowMessage
The varbinds supported in this trap are:
extremeClearflowMsgId
extremeClearflowMsg
extremeClearflowPolicyName
extremeClearflowRuleName
extremeClearflowRuleValue
extremeClearflowRuleThreshold
extremeClearflowRuleInterval
extremeClearflowVlanName
extremeClearflowPortName
EXTREME-PoE-MIB

extremePethSystemAdminEnable
extremePethSystemDisconnectPrecedence
extremePethSystemUsageThreshold
Objects are supported read-only.
extremePethSystemPowerSupplyMode
extremePethPseSlotTable All objects
extremePethPsePortTable All objects

EXTREME-RMON-MIB

extremeRtStatsTable extremeRtStatsIndex
extremeRtStatsIntervalStart
extremeRtStatsCRCAlignErrors
extremeRtStatsUndersizePkts
extremeRtStatsOversizePkts
All objects are supported read-only.
extremeRtStatsFragments
extremeRtStatsJabbers
extremeRtStatsCollisions
extremeRtStatsTotalErrors
extremeRtStatsUtilization
EXTREME-QOS-MIB

extremeQosProfileTable extremeQosProfileIndex The extremeQosProfileTable is
table is read only.
extremeQosProfileName
extremeQosProfileMinBw
extremeQosProfileMaxBw
extremeQosProfilePriority
extremeQosProfileRowStatus
extremeIQosProfileTable extremeIQosProfileIndex ExtremeXOS does not support
global QoS profile settings in CLI;
it supports per port settings only.
Walks of this table display the
default values with which the
ports are initialized.
extremeIQosProfileName
extremeIQosProfileMinBwType
extremeIQosProfileMinBw
extremeIQosProfileMaxBwType
extremeIQosProfileMaxBw
extremeIQosProfileRED
extremeIQosProfileMaxBuf
extremePerPortQosTable extremePerPortQosIndex
extremePerPortQosMinBw
extremePerPortQosMaxBw

MIB Support Details
ExtremePerPortQosPriority
extremePerPortQosRowStatus
extremeQosByVlanMappingTable extremeVlanIfIndex Shows mapping of VLAN to
queues for untagged packets. For
tagged packets, the vpri field
determines which queue the
packet should be using.
extremeQosByVlanMappingQosProfileIndex


Index of Commands
C configure dos-protect type l3-protect alert-

threshold, 546
check policy, 376 configure dos-protect type l3-protect notify-
clear access-list counter, 403 threshold, 546
clear cna-testplug counters, 1068 configure dot1p type, 448
clear counters, 285, 554, 675, 684 configure eaps add control vlan, 667
clear eaps counters, 675, 684 configure eaps add protect vlan, 667
clear elsm ports auto-restart, 262 configure eaps config-warnings off, 670
clear elsm ports counters, 269 configure eaps config-warnings on, 670
clear inline-power stats ports, 225 configure eaps failtime, 665
clear log counters, 285 configure eaps failtime expiry-action, 665
clear log static, 1033 configure eaps fast-convergence, 668
clear msdp sa-cache peer, 992 configure eaps hellotime, 665
clear session, 51, 75 configure eaps mode, 664
clear slot, 126, 1034 configure eaps primary port, 666
clear sys-recovery-level, 253, 259 configure eaps secondary port, 666
clear vlan dhcp-address-allocation, 534 configure eaps shared-port domain, 679
configure access-list, 378, 403 configure eaps shared-port mode, 678
configure account, 51 configure eaps shared-port segment-timeout, 679
configure banner, 51 configure edp advertisement-interval, 158
configure bgp add aggregate-address, 956 configure elrp-client one-shot, 1043
configure bgp add network, 960 configure elrp-client periodic, 1043
configure bgp delete network, 960 configure elsm ports hellotime, 266
configure bgp import-policy, 378, 432 configure elsm ports hold-threshold, 266
configure bgp neighbor dampening, 958 configure esrp add master, 757
configure bgp neighbor no-dampening, 958 configure esrp add member, 758
configure bgp neighbor peer-group, 957 configure esrp add track-environment failover,
configure bgp neighbor route-policy, 378, 432 759
configure bgp peer-group dampening, 958 configure esrp add track-iproute, 760
configure bgp peer-group no dampening, 958 configure esrp add track-ping, 760
configure bgp peer-group route-policy, 378, 432 configure esrp add track-vlan, 760
configure bootprelay add, 879 configure esrp delete elrp-poll ports, 769
configure bootprelay delete, 879 configure esrp delete master, 758, 1041
configure bootprelay dhcp-agent information configure esrp delete member, 758
check, 880 configure esrp delete track-iproute, 760
configure bootprelay dhcp-agent information configure esrp delete track-ping, 760
option, 880 configure esrp delete track-vlan, 760
configure cli max-sessions, 66 configure esrp domain-id, 749, 757
configure cna-testplug vlan, 1068 configure esrp election-policy, 754, 759
configure core-dumps, 1049 configure esrp elrp-master-poll disable, 769
configure diffserv examination code-point, 451 configure esrp elrp-premaster-poll disable, 769
configure diffserv replacement, 452 configure esrp elrp-premaster-poll enable, 768
configure dns-client add, 62 configure esrp mode, 748
configure dns-client default-domain, 62 configure esrp port-mode ports, 764
configure dos-protect acl-expire, 546 configure esrp ports mode, 763
configure dos-protect interval, 546 configure esrp ports no-restart, 762
configure dos-protect trusted-ports, 546 configure esrp ports restart, 762
configure esrp timer premaster, 753

Index of Commands
configure failsafe-account, 58 configure mpls rsvp-te add profile, 855

configure fdb agingtime, 360 configure mpls rsvp-te delete lsp, 856, 857
configure firmware, 1025 configure mpls rsvp-te delete path, 853
configure forwarding hash-algorithm, 1061 configure mpls rsvp-te delete profile, 856
configure igmp snooping add static group, 968 configure mpls rsvp-te lsp add path, 856
configure igmp snooping add static router, 968, configure mpls rsvp-te path add ero, 853
986 configure mpls rsvp-te path delete ero, 854
configure igmp snooping delete static group, 968, configure mpls rsvp-te vlan, 852
986 configure msdp default-peer, 989
configure igmp snooping delete static router, 968, configure msdp mesh-group, 991
986 configure msdp mesh-group none, 991
configure igmp snooping filter, 969 configure msdp no-default-peer, 989
configure inline-power budget slot, 221 configure msdp password, 989
configure inline-power disconnect-precedence, configure msdp password none, 989
217, 222 configure msdp sa-cache-server, 992
configure inline-power label ports, 224 configure msdp sa-filter, 990
configure inline-power operator-limit ports, 220, configure msdp sa-filter none, 990
224 configure msdp sa-limit, 992
configure inline-power priority ports, 217, 223 configure mstp format, 723
configure inline-power usage-threshold, 219, 223 configure mstp region, 723
configure iparp add proxy, 870 configure mstp revision, 723
configure ip-mtu vlan, 136, 138 configure netlogin add mac-list, 514
configure iproute add default, 67, 72, 871, 901 configure netlogin base-url, 506
configure iproute priority, 869, 901 configure netlogin delete mac-list, 514
configure jumbo-frame size, 136 configure netlogin dot1x guest-vlan, 500
configure lldp med fast-start repeat-count, 182, configure netlogin dotix timers, 501
200 configure netlogin dynamic-vlan, 520
configure lldp ports management-address, 198 configure netlogin dynamic-vlan uplink-ports, 520
configure lldp ports port-description, 198 configure netlogin local-user, 494, 495
configure lldp ports system-capabilities, 198 configure netlogin move-fail-action, 486
configure lldp ports system-name, 198 configure netlogin port mode, 518
configure lldp ports vendor-specific, 198 configure netlogin port no-restart, 522
configure lldp reinitialize-delay, 196 configure netlogin port restart, 522
configure lldp snmp-notification-interval, 197 configure netlogin redirect-page, 507
configure lldp transmit-delay], 196 configure node slot priority, 78
configure lldp transmit-hold, 196 configure ospf area external-filter, 378, 432
configure lldp transmit-interval, 196 configure ospf area interarea-filter, 378, 432
configure log filter, 279, 281 configure ospf area nssa, 930
configure log filter events match, 282 configure ospf area stub, 929, 941
configure log target, 277 configure ospf area timer, 935
configure log target filter, 277 configure ospf ase-limit, 927
configure log target format, 283 configure ospf restart, 928
configure log target match, 280 configure ospf timer, 935
configure log target syslog, 276 configure ospf virtual-link timer, 935
configure mac-lockdown-timeout ports aging- configure ospf vlan area, 929, 940
time, 532 configure ospf vlan timer, 934, 935, 945
configure meter, 473 configure pim add vlan, 969
configure mld snooping add static group, 986 configure pim border, 988
configure mpls add vlan, 814, 852 configure ports auto off, 51, 131, 1038
configure mpls delete vlan, 814, 852 configure ports auto on, 131
configure mpls hello-hold-time, 815 configure ports auto-polarity off, 133
configure mpls php, 801 configure ports auto-polarity on, 133
configure mpls rsvp-te add lsp, 855, 856 configure ports limit-learning, 526
configure mpls rsvp-te add path, 853 configure ports lock-learning, 528

Index of Commands
configure ports preferred-medium, 162 configure ssh2 key, 51, 561

configure ports qosprofile, 454 configure ssh2 key pregenerated, 561
configure ports redundant, 160 configure ssl certificate, 570
configure ports unlock-learning, 529 configure ssl certificate pregenerated, 572
configure powersupply, 90 configure ssl privkeyword pregenerated, 572
configure protocol add, 306 configure stpd, 692
configure qosprofile ingress, 464 configure stpd add vlan, 698, 729, 732
configure radius server client-ip, 548 configure stpd default-encapsulation, 697
configure radius shared-secret, 548, 550 configure stpd delete vlan, 699, 700
configure radius timeout, 548 configure stpd max-hop-count, 728
configure radius-accounting, 549, 558 configure stpd mode, 696
configure radius-accounting timeout, 549 configure stpd ports cost, 691
configure rip import-policy, 378, 432 configure stpd ports edge-safeguard disable, 712
configure rip trusted-gateway, 378, 432 configure stpd ports edge-safeguard enable, 712
configure rip vlan route-policy, 378, 432 configure stpd ports link-type, 711
configure sflow agent, 289 configure stpd ports mode, 697
configure sflow collector, 289 configure stpd priority, 692
configure sflow max-cpu-sample-limit, 292 configure stpd tag, 733
configure sflow poll-interval, 291 configure sys-health-check interval, 247, 1057
configure sflow ports sample-rate, 291 configure sys-recovery-level, 51, 250
configure sflow sample-rate, 291 configure sys-recovery-level slot, 254
configure sharing add ports, 148 configure sys-recovery-level switch, 251
configure sharing address-based, 142 configure tacacs server client-ip, 556
configure sharing delete ports, 148 configure tacacs shared-secret, 556, 558
configure slot module, 51, 126, 1034 configure tacacs timeout, 556
configure snmp add community, 93 configure tacacs-accounting timeout, 558
configure snmp add trapreceiver community, 93 configure telnet access-profile, 75
configure snmp delete trapreceiver, 93 configure telnet port, 73
configure snmpv3 add access, 97 configure telnet vr, 73
configure snmpv3 add filter subtree type, 101 configure time, 51
configure snmpv3 add filter-profile param, 101 configure timezone, 51, 102, 103
configure snmpv3 add group user, 97 configure traffic queue, 473
configure snmpv3 add mib-view, 99 configure trusted-ports trust-for dhcp-server, 537
configure snmpv3 add mib-view subtree, 99 configure trusted-servers vlan add server trust-for
configure snmpv3 add notify tag, 101 dhcp-server, 536
configure snmpv3 add target-addr param configure trusted-servers vlan delete server trust-
ipaddress, 100 for dhcp-server, 537
configure snmpv3 add target-params, 96 configure vlan add esrp elrp-poll ports, 769
configure snmpv3 add user, 97 configure vlan add ports, 698
configure snmpv3 delete access, 98 configure vlan dhcp-address-range, 534
configure snmpv3 delete filter, 101 configure vlan dhcp-lease-timer, 534
configure snmpv3 delete filter-profile, 101 configure vlan dhcp-options, 534
configure snmpv3 delete group user, 98 configure vlan esrp elrp-master-poll enable, 769
configure snmpv3 delete mib-view, 99 configure vlan ipaddress, 51, 71, 871, 901
configure snmpv3 delete notify, 102 configure vlan name, 308
configure snmpv3 delete target-addr, 100 configure vlan qosprofile, 454
configure snmpv3 delete target-params, 100 configure vpls add service, 830
configure snmpv3 delete user, 97 configure vr add ports, 371
configure snmpv3 engine-boots, 96 configure vr add protocol, 371
configure snmpv3 engine-id, 96 configure vr delete ports, 370
configure snmpv3 target-params user mp-model, configure vr delete protocol, 371
100 configure vrrp vlan vrid add track-iproute, 780
configure sntp-client, 104 configure vrrp vlan vrid add track-ping, 780
configure sntp-client update-interval, 104 configure vrrp vlan vrid add track-vlan, 779

Index of Commands
configure vrrp vlan vrid delete track-iproute, 780 disable elsm ports, 267
configure vrrp vlan vrid delete track-ping, 780 disable elsm ports auto-restart, 267
configure vrrp vlan vrid delete track-vlan, 779 disable esrp, 754, 758, 1041
cp, 110, 116, 1052 disable flooding, 364
create account, 51, 57 disable idletimeout, 52
create bgp neighbor peer-group, 957 disable inline-power, 215, 221
create bgp peer-group, 957 disable inline-power legacy, 219, 224
create cfm domain, 209 disable inline-power ports, 221
create eaps, 663 disable inline-power slot, 221
create eaps shared-port, 678 disable ipforwarding, 554
create esrp, 747, 756 disable iproute ipv6 compression, 893
create log filter, 279 disable ip-security arp gratuitous-protection vlan,
create meter, 473, 474 542
create msdp mesh-group, 991 disable ip-security arp learning learn-from-arp
create msdp peer, 988 vlan ports, 540
create netlogin local-user, 492, 493 disable ip-security arp learning learn-from-dhcp
create ospf area, 929, 941 vlan ports, 541
create protocol, 306 disable ip-security arp validation, 544
create stpd, 693, 732 disable ip-security dhcp-snooping vlan ports, 536
create traffic queue, 473 disable ip-security source-ip-lockdown ports, 539
create virtual-router, 370 disable learning port, 362
create vlan, 51, 372 disable lldp ports, 195
disable log debug-mode, 1048
D disable log target, 275
disable mac-lockdown-timeout ports, 533
delete account, 51, 58 disable msdp process-sa-request, 990
delete bgp peer-group, 957 disable netlogin, 486
delete eaps, 664 disable netlogin dot1x guest-vlan ports, 501
delete eaps shared-port, 678, 679 disable netlogin logout-privilege, 507
delete esrp, 757 disable netlogin ports vlan, 486
delete fdbentry, 526 disable netlogin session-refresh, 507
delete msdp mesh-group, 991 disable ospf capability opaque-lsa, 927
delete msdp peer, 989 disable ospf export, 934, 945
delete netlogin local-user, 495 disable ospf export static, 868, 900
delete stpd, 693 disable port, 52, 129
delete traffic queue, 473 disable radius, 549
delete virtual router, 370 disable radius-accounting, 550
delete vlan, 51 disable rip export, 915
delete vpls, 829 disable rip export static, 868, 900
disable access-list refresh blackhole, 377 disable ripng export, 922
disable bgp export, 960 disable rmon, 296
disable bgp neighbor remove-private-as-numbers, disable sflow, 290
959 disable sflow ports, 290
disable bootp vlan, 51, 70 disable sharing, 148
disable clear-flow, 574 disable smartredundancy, 160
disable cli-config-logging, 52, 286 disable snmp access, 92
disable clipaging, 52 disable snmp traps lldp, 197
disable cpu-monitoring, 121 disable ssh2, 52
disable dhcp ports vlan, 533 disable sys-health-check slot, 248, 1056
disable dhcp vlan, 70 disable tacacs, 557
disable dos-protect, 545 disable tacacs-accounting, 559
disable eaps, 668 disable telnet, 52, 75
disable edp ports, 157 disable udp-echo-server, 883
disable elrp-client, 1043 disable vpls, 830

Index of Commands
disable web http, 560 enable ip-security source-ip-lockdown ports, 539

disable web https, 570 enable jumbo-frame ports, 136
download bootrom, 61, 1023, 1024 enable license, 52
download image, 61, 1001, 1009, 1011 enable lldp ports, 195
download ssl certificate, 571 enable log debug-mode, 285, 1048
download ssl privkey, 571 enable log target, 274
enable log target console, 283
E enable log target session, 283
enable mac-lockdown-timeout ports, 532
edit policy, 73, 376, 563 enable msdp process-sa-request, 990
eject memorycard, 1049 enable netlogin, 486
enable access-list refresh blackhole, 377 enable netlogin dot1x guest-vlan ports, 501
enable bgp aggregation, 956 enable netlogin logout-privilege, 507
enable bgp export, 960 enable netlogin ports, 486
enable bgp neighbor remove-private-as-numbers, enable netlogin session-refresh, 507
959 enable ospf, 871, 902
enable bootp vlan, 52, 70 enable ospf capability opaque-lsa, 927
enable bootprelay, 879 enable ospf export, 934, 945
enable clear-flow, 574 enable ospf export static, 868, 900
enable cli-config-logging, 52, 286 enable pim, 969
enable clipaging, 52 enable port, 129
enable cna-testplug, 1067, 1068 enable radius, 549
enable cpu-monitoring, 121 enable radius-accounting, 550
enable dhcp ports vlan, 533 enable rip, 871, 902
enable dhcp vlan, 70 enable rip export, 915
enable diffserv replacement ports, 452 enable rip export static, 868, 900
enable dos-protect, 545 enable ripng export, 922
enable dos-protect simulated, 545 enable rmon, 296
enable dot1p replacement ports, 448 enable sflow, 289
enable eaps, 668 enable sflow ports, 290
enable edp ports, 158 enable sharing grouping, 148
enable elrp-client, 1042 enable smartredundancy, 160
enable elsm ports, 265 enable snmp access, 92
enable elsm ports auto-restart, 262, 267 enable snmp traps lldp, 196
enable esrp, 758 enable snmp traps lldp-med, 183
enable flooding, 364 enable sntp-client, 104
enable idletimeout, 52 enable ssh2, 52, 563, 565
enable inline-power, 215, 220 enable stpd, 733
enable inline-power legacy, 219, 224 enable stpd auto-bind, 699
enable inline-power ports, 220 enable stpd rapid-root-failover, 700
enable inline-power slot, 220 enable sys-health-check slot, 247, 1056
enable ipforwarding, 871, 901 enable tacacs, 557
enable ipmcforwarding, 969 enable tacacs-accounting, 559
enable iproute ipv6 compression, 893 enable telnet, 52, 75
enable ip-security arp gratuitous-protection vlan, enable udp-echo-server, 883
542 enable vpls, 829
enable ip-security arp learning learn-from-arp vlan enable web http, 560
ports, 540 enable web https, 570
enable ip-security arp learning learn-from-dhcp
vlan ports, 540 H
enable ip-security arp validation vlan violation-
action, 544 history, 50, 52
enable ip-security dhcp-snooping vlan ports
violation-action, 536

Index of Commands
I show clear-flow port, 574

show clear-flow rule-all, 574
install image, 1001, 1010, 1011 show clear-flow rule-triggered, 574
show clear-flow vlan, 574
L show configuration, 116, 134, 1015
show cpu-monitoring, 122
load script, 1016, 1017 show dhcp-client state, 70
logout, 72 show dhcp-server, 534
ls, 77, 111, 112, 116, 117, 1016, 1017, 1051 show diagnostics, 245
ls memorycard, 112 show diagnostics slot, 245
show diffserv, 452
M show dos-protect, 546
show eaps, 670
mv, 109, 116, 1051
show eaps shared-port, 677, 679
show edp, 158
N show elrp, 770, 1043
nslookup, 62 show elsm, 268
show elsm ports, 262, 268
P show esrp, 747, 761, 767, 770
show esrp counters, 767
ping, 55, 61, 62, 63 show fans, 273, 1034
ping mac, 211 show fdb, 361
show forwarding configuration, 1061
Q show igmp snooping filter, 969
show igmp snooping static group, 968
quit, 72 show inline-power, 221, 222, 223, 224, 225
show inline-power configuration ports, 223, 224,
R 225, 228
show inline-power info ports, 218, 228
reboot, 80, 81, 1005
show inline-power slot, 222, 226
refresh policy, 377
show inline-power stats ports, 229
reset inline-power ports, 218, 225
show inline-power stats slot, 227
restart process, 119
show iparp, 872
rm, 115, 116, 1054
show ipconfig, 872, 902
run diagnostics, 237, 238
show iproute, 871
run elrp, 1043
show iproute ipv6, 902
run msm-failover, 79, 81, 1010, 1011
show ip-security arp learning vlan, 541
run update, 1003
show ip-security arp validation vlan, 544
show ip-security dhcp-snooping entries vlan, 537
S show ip-security dhcp-snooping vlan, 537
save configuration, 1015, 1016, 1021 show ip-security gratuitous-protection vlan, 543
save debug tracefiles memorycard, 1049 show ip-security source-ip-lockdown, 539
scp2, 567 show log, 284
show access-list, 378, 383, 403 show log components, 277
show access-list counter, 385, 403 show log configuration filter, 279
show account, 58 show log configuration target, 276
show accounts, 58 show log counters, 284
show banner, 52 show log events, 278
show bgp peer-group, 959 show mac-lockdown-timeout fdb ports, 533
show bootprelay, 881 show mac-lockdown-timeout ports, 533
show checkpoint-data, 80, 81 show management, 75, 94, 297, 554, 563
show clear-flow, 574 show memory process, 120
show clear-flow acl-modified, 574 show mirroring, 157
show clear-flow any, 574 show mld snooping static group, 986

Index of Commands
show mpls ldp, 818 show switch, 79, 80, 103, 104, 248, 251, 252,
show mpls rsvp-te, 857 554, 998, 999, 1001, 1009
show mpls rsvp-te lsp, 858 show temperature, 272
show mpls rsvp-te path, 857 show traffic mode, 467
show mpls rsvp-te profile, 857 show version, 998
show msdp mesh-group, 991 show virtual-router, 372
show msdp peer, 989 show vlan, 311, 512, 521, 1039
show neighbor-discovery cache, 902 show vlan dhcp-address-allocation, 534
show netlogin, 521 show vlan dhcp-config, 534
show netlogin guest-vlan, 501 show vlan security, 527
show netlogin local-users, 494, 495 show vlan stpd, 741
show netlogin mac-list, 515 show vman, 330
show netlogin vlan, 487 show vrrp, 780
show node, 81 ssh2, 567
show odometers, 1057 start process, 119
show ospf, 934, 938, 945 synchronize, 79, 81, 485, 656, 703, 751, 778,
show ospf area, 938 1020, 1021
show ospf interfaces, 938
show ospf lsdb, 938 T
show ospf lsdb area lstype, 938
show policy, 383 telnet, 61, 70
show ports configuration, 1033 terminate process, 118
show ports info detail, 527 tftp, 73, 76, 116, 117, 376, 564, 1016, 1017,
show ports information, 365, 454, 464 1018, 1019
show ports qosmonitor, 455, 464 tftp get, 1016, 1017, 1020
show ports rxerrors, 233, 1038 tftp put, 1019
show ports sharing, 151 top, 1055
show ports statistics, 232 traceroute, 61, 62, 63
show ports txerrors, 232 traceroute mac, 211
show power, 91, 273, 1034
show power budget, 91 U
show power controller, 91
unconfigure access-list, 378, 403
show process, 117
unconfigure bootprelay dhcp-agent information
show protocol, 312
check, 880
show qosprofile, 456
unconfigure bootprelay dhcp-agent information
show qosprofile ports, 464
option, 880
show rmon memory, 297
unconfigure eaps primary port, 669
show session, 75
unconfigure eaps secondary port, 669
show sflow, 289, 293
unconfigure eaps shared-port link-id, 679
show sflow statistics, 293
unconfigure eaps shared-port mode, 679
show slot, 92, 126, 258
unconfigure elrp-client, 1043
show snmpv3 access, 97
unconfigure inline-power budget slot, 216, 221
show snmpv3 filter, 101
unconfigure inline-power disconnect-precedence,
show snmpv3 filter-profile, 101
217, 222
show snmpv3 group, 97
unconfigure inline-power operator-limit ports,
show snmpv3 mib-view, 99
220, 224
show snmpv3 notify, 102
unconfigure inline-power priority ports, 217, 223
show snmpv3 target-addr, 100
unconfigure inline-power usage-threshold, 219,
show snmpv3 target-params, 100
223
show snmpv3 user, 97
unconfigure lldp, 201
show sntp-client, 104
unconfigure mpls, 816
show ssl, 571, 572
unconfigure msdp sa-cache-server, 991
show stpd, 700, 739
unconfigure mstp region, 723
show stpd ports, 712, 740

Index of Commands
unconfigure netlogin dot1x guest-vlan, 501

unconfigure port redundant, 160
unconfigure sflow, 292
unconfigure sflow agent, 289
unconfigure sflow collector, 289
unconfigure stpd ports link-type, 712
unconfigure switch, 52, 1016
unconfigure trusted-ports trus-for dhcp-server,
537
unconfigure vlan dhcp, 534
unconfigure vlan dhcp-address-range, 534
unconfigure vlan dhcp-options, 534
uninstall image, 1003
upload configuration, 117, 1016, 1017
upload debug, 1050
upload log, 284
use configuration, 1015
use image, 999
V
virtual-router, 372

Glossary
A
AAA Authentication, authorization, and accounting. A system to control
which computer resources specific users can access and to keep track
of the activity of specific users over the network.
ABR Area border router. In OSPF, an ABR has interfaces in multiple areas,
and it is responsible for exchanging summary advertisements with
other ABRs.
ACL Access Control List. ACLs are a mechanism for filtering packets at the
hardware level. Packets can be classified by characteristics such as the
source or destination MAC, IP addresses, IP type, or QoS queue. Once
classified, the packets can be forwarded, counted, queued, or dropped.
In Extreme Networks XOS software, you configure ACLs by creating a
file, called a policy file (with a .pol file extension). The system parses
the policy file and loads the ACL into the hardware.
alternate port In RSTP, the alternate port supplies an alternate path to the root
bridge and the root port.
AP Access point. In wireless technology, access points are the devices that
connect to the regular wired network and forward and receive the
radio signals that transmit wireless data.
area In OSPF, an area is a logical set of segments connected by routers. The

topology within an area is hidden from the rest of the AS.
ARP Address Resolution Protocol. ARP is part of the TCP/IP suite used to
dynamically associate a device’s physical address (MAC address) with
its logical address (IP address). The system broadcasts an ARP
request, containing the IP address, and the device with that IP address
sends back its MAC address so that traffic can be transmitted.
AS Autonomoous system. In OSPF, an AS is a connected segment of a

network topology that consists of a collection of subnetworks (with
hosts attached) interconnected by a set of routes. The subnetworks and
the routers are expected to be under the control of a single
administration. Within an AS, routers may use one or more interior
routing protocols and sometimes several sets of metrics. An AS is
expected to present to other ASs an appearance of a coherent interior
routing plan and a consistent picture of the destinations reachable
through the AS. An AS is identified by a unique 16-bit number.
ASBR Autonomous system border router. In OSPF, an ASBR acts as a

gateway between OSPF and other routing protocols or other ASs.

Glossary
A (Continued)
autobind In STP, autobind, when enabled, automatically adds or removes ports
from the STPD. If ports are added to the carrier VLAN, the member
ports of the VLAN are automatically added to the STPD. If ports are
removed from the carrier VLAN, those ports are also removed from
the STPD.
autonegotation As set forth in IEEE 802.3u, autonegotation allows each port on the
switch—in partnership with its link partner—to select the highest
speed between 10 Mbps and 100 Mbps and the best duplex mode.
B
backbone area In OSPF, a network that has more than one area must have a
backbone area, configured as 0.0.0.0. All areas in an AS must connect
to the backbone area.
backup port In RSTP, the backup port supports the designated port on the same
attached LAN segment. Backup ports exist only when the bridge is
connected as a self-loop or to a shared media segment.
backup router In VRRP, the backup router is any VRRP router in the VRRP virtual
router that is not elected as the master. The backup router is available
to assume forwarding responsibility if the master becomes
unavailable.
BDR Backup designated router. In OSPF, the system elects a DR and a BDR.
The BDR smooths the transition to the DR, and each multiaccess
network has a BDR. The BDR is adjacent to all routers on the network
and becomes the DR when the previous DR fails. The period of
disruption in transit traffic lasts only as long as it takes to flood the
new LSAs (which announce the new DR). The BDR is elected by the
protocol; each hello packet has a field that specifies the BDR for the
network.
BGP Border Gateway Protocol. BGP is a router protocol in the IP suite

designed to exchange network reachability information with BGP
systems in other ASs. You use a fully meshed configuration with BGP.
BGP provides routing updates that include a network number, a list of
ASs that the routing information passed through, and a list of other
path attributes. BGP works with cost metrics to choose the best
available path; it sends updated router information only when one
host has detected a change, and only the affected part of the routing
table is sent. BGP communicates within one AS using Interior BGP
(IBGP) because BGP does not work well with IGP. The routers inside
the AS thus maintain two routing tables: one for the IGP and one for
IBGP. BGP uses exterior BGP (EBGP) between different ASs.
bi-directional rate shaping This is a hardware-based technology that allows you to manage
bandwidth on Layer 2 and Layer 3 traffic flowing to each port on the
switch and to the backplane, per physical port on the I/O module.
The parameters differ across platforms and modules.

B (Continued)
blackhole In the Extreme Networks implementation, you can configure the
switch so that traffic is silently dropped. Although this traffic appears
as received, it does not appear as transmitted (because it is dropped).
BOOTP Bootstrap Protocol. BOOTP is an Internet protocol used by a diskless

workstation to discover its own IP address, the IP address of a BOOTP
server on the network, and a file that can be loaded into memory to
boot the machine. Using BOOTP, a workstation can boot without a
hard or floppy disk drive.
BPDU Bridge protocol data unit. In STP, a BPDU is a packet that initiates
communication between devices. BPDU packets contain information
on ports, addresses, priorities, and costs and ensure that the data ends
up where it was intended to go. BPDU messages are exchanged across
bridges to detect loops in a network topology. The loops are then
removed by shutting down selected bridge interfaces and placing
redundant switch ports in a backup, or blocked, state.
bridge In conventional networking terms, bridging is a Layer 2 function that

passes frames between two network segments; these segments have a
common network layer address. The bridged frames pass only to
those segments connected at a Layer 2 level, which is called a
broadcast domain (or VLAN). You must use Layer 3 routing to pass
frames between broadcast domains (VLANs).
In wireless technology, bridging refers to forwarding and receiving
data between radio interfaces on APs or between clients on the same
radio. So, bridged traffic can be forwarded from one AP to another AP
without having to pass through the switch on the wired network.
broadcast A broadcast message is forwarded to all devices within a VLAN,

which is also known as a broadcast domain. The broadcast domain, or
VLAN, exists at a Layer 2 level; you must use Layer 3 routing to
communicate between broadcast domains, or VLANs. Thus, broadcast
messages do not leave the VLAN. Broadcast messages are identified
by a broadcast address.
C
carrier VLAN In STP, carrier VLANs define the scope of the STPD, including the
physical and logical ports that belong to the STPD as well as the
802.1Q tags used to transport EMISTP- or PVST+-encapsulated
BPDUs. Only one carrier VLAN can exist in any given STPD.
CCM In CFM, connectivity check messages are multicast CFM frames

transmitted periodically by a MEP to ensure connectivity across the
maintenance entities to which the transmitting MEP belongs. The
CCM messages contain a unique ID for the specified domain. Because
a failure to receive a CCM indicates a connectivity fault in the
network, CCMs proactively check for network connectivity.
CFF In CFM, this means CFM filter function. You apply CFFs to specified
ports within the CFM MA to block CCM messages at or inferior to
that MA level.

Glossary
C (Continued)
CFM Connectivity Fault Management allows an ISP to proactively detect
faults in the network for each customer service instance individually
and separately. CFM comprises capabilities for detecting, verifying,
and isolating connectivity failures in virtual bridged LANs.
checkpointing Checkpointing is the process of copying the active state configurations

from the primary MSM to the backup MSM on modular switches.
CIDR Classless Inter-Domain Routing. CIDR is a way to allocate and specify

the Internet addresses used in interdomain routing more flexibly than
with the original system of IP address classes. This address
aggregation scheme uses supernet addresses to represent multiple IP
destinations. Rather than advertise a separate route for each
destination, a router uses a supernet address to advertise a single
route representing all destinations. RIP does not support CIDR; BGP
and OSPF support CIDR.
CIST Common and Internal Spanning Tree. In an MSTP environment, the

CIST is a single spanning tree domain that connects MSTP regions.
The CIST is responsible for creating a loop-free topology by
exchanging and propagating BPDUs across MSTP regions. You can
configure only one CIST on each switch.
CIST root port In an MSTP environment, the port on the CIST regional root bridge
that connects to the CIST root bridge is the CIST root port. The CIST
root port is the master port for all MSTIs in that MSTP region, and it is
the only port that connects the entire region to the CIST root bridge.
CIST regional root bridge Within an MSTP region, the bridge with the lowest path cost to the
CIST root bridge is the CIST regional root bridge If the CIST root
bridge is inside an MSTP region, that same bridge is the CIST regional
root for that region because it has the lowest path cost to the CIST
root. If the CIST root bridge is outside an MSTP region, all regions
connect to the CIST root through their respective CIST regional roots.
CIST root bridge In an MSTP environment, the bridge with the lowest bridge ID
becomes the CIST root bridge. The bridge ID includes the bridge
priority and the MAC address. The CIST root bridge can be either
inside or outside an MSTP region. The CIST root bridge is unique for
all regions and non-MSTP bridges, regardless of its location.
CLEAR-Flow CLEAR-Flow allows you to specify certain types of traffic to perform

configured actions on. You can configure the switch to take an
immediate, preconfigured action to the specified traffic or to send a
copy of the traffic to a management station for analysis. CLEAR-Flow
is an extension to ACLs, so you must be familiar with ACL policy files
to apply CLEAR-Flow.
CLI Command line interface. You use the CLI to monitor and manage the
switch.
cluster In BGP, a cluster is formed within an AS by a route reflector and its

client routers.

C (Continued)
CNA Converged Network Analyzer. This application suite, available from
Avaya, allows the server to determine the best possible network path.
The CNA Agent is a software piece of the entire CNA application that
you install on Extreme Networks devices. You use the CNA Agent
software only if you are using the Avaya CNA solution, and the CNA
Agent cannot function unless you also obtain the rest of the CNA
application from Avaya.
combo port Combination port. On some Extreme Networks devices (such as the
Summit X450 switch), certain ports can be used as either copper or
fiber ports.
common link In EAPS, the common link is the physical link between the controller
and partner nodes in a network where multiple EAPS share a
common link between domains.
controller node In EAPS, the controller node is that end of the common line that is
responsible for blocking ports if the common link fails, thereby
preventing a superloop.
control VLAN In EAPS, the control VLAN is a VLAN that sends and receives EAPS
messages. You must configure one control VLAN for each EAPS
domain.
CoS Class of Service. Specifying the service level for the classified traffic
type.
CRC Cyclic redundancy check. This simple checksum is designed to detect

transmission errors. A decoder calculates the CRC for the received
data and compares it to the CRC that the encoder calculated, which is
appended to the data. A mismatch indicates that the data was
corrupted in transit.
CRC error Cyclic redundancy check error. This is an error condition in which the
data failed a checksum test used to trap transmission errors. These
errors can indicate problems anywhere in the transmission path.
CSPF Constrained shortest path first. An algorithm based on the shortest

path first algorithm used in OSPF, but with the addition of multiple
constraints arising from the network, the LSP, and the links. CSPF is
used to minimize network congestion by intelligently balancing traffic.
D
DA Destination address. The DA is the IP or MAC address of the device
that is to receive the packet.
DAD Duplicate Address Detection. IPv6 automatically uses this process to

ensure that no duplicate IP addresses exist.

Glossary
D (Continued)
default encapsulation mode In STP, default encapsulation allows you to specify the type of BPDU
encapsulation to use for all ports added to a given STPD, not just to
one individual port. The encapsulation modes are:
● 802.1d—This mode is used for backward compatibility with
previous STP versions and for compatibility with third-party
switches using IEEE standard 802.1d.
● EMISTP—Extreme Multiple Instance Spanning Tree Protocol
(EMISTP) mode is an extension of STP that allows a physical port
to belong to multiple STPDs by assigning the port to multiple
VLANs.
● PVST+—This mode implements PVST+ in compatibility with third-
party switches running this version of STP.
designated port In STP, the designated port provides the shortest path connection to
the root bridge for the attached LAN segment. Each LAN segment has
only one designated port.
Device Manager The Device Manager is an Extreme Networks-proprietary process that

runs on every node and is responsible for monitoring and controlling
all of the devices in the system. The Device Manager is useful for
system redundancy.
DF Don’t fragment bit. This is the don’t fragment bit carried in the flags
field of the IP header that indicates that the packet should not be
fragmented. The remote host will return ICMP notifications if the
packet had to be split anyway, and these are used in MTU discovery.
DHCP Dynamic Host Configuration Protocol. DHCP allows network

administrators to centrally manage and automate the assignment of IP
addresses on the corporate network. DHCP sends a new IP address
when a computer is plugged into a different place in the network. The
protocol supports static or dynamic IP addresses and can dynamically
reconfigure networks in which there are more computers than there
are available IP addresses.
DiffServ Differentiated Services. Defined in RFC 2474 and 2475, DiffServ is an

architecture for implementing scalable service differentiation in the
Internet. Each IP header has a DiffServ (DS) field, formerly known as
the Type of Service (TOS) field. The value in this field defines the QoS
priority the packet will have throughout the network by dictating the
forwarding treatment given to the packet at each node. DiffServ is a
flexible architecture that allows for either end-to-end QoS or
intradomain QoS by implementing complex classification and
mapping functions at the network boundary or access points. In the
Extreme Networks implementation, you can configure the desired QoS
by replacing or mapping the values in the DS field to egress queues
that are assigned varying priorities and bandwidths.
DNS Domain Name Server. This is system is used to translate domains

names to IP addresses. Although the Internet is based on IP addresses,
names are easier to remember and work with. All these names must
be translated back to the actual IP address and the DNS servers do so.

D (Continued)
domain In CFM, a maintenance domain is the network, or part of the network,
that belongs to a single administration for which connectivity faults
are managed.
DoS attack Denial of service attacks occur when a critical network or computing
resource is overwhelmed so that legitimate requests for service cannot
succeed. In its simplest form, a DoS attack is indistinguishable from
normal heavy traffic. ExtremeWare XOS software has configurable
parameters that allow you to defeat DoS attacks.
DR Designated router. In OSPF, the DR generates an LSA for the

multiaccess network and has other special responsibilities in the
running of the protocol. The DR is elected by the OSPF protocol.
dropped packets These are packets that the switch received but does not transmit.
E
EAPS Extreme Automatic Protection Switching. EAPS is an Extreme
Networks-proprietary protocol that prevents looping Layer 2 of the
network. This feature is discussed in RFC 3619.
EAPS domain An EAPS domain consists of a series of switches, or nodes, that

comprise a single ring in a network. An EAPS domain consists of a
master node and transit nodes. The master node consists of one
primary and one secondary port. EAPS operates by declaring an EAPS
domain on a single ring.
EAPS link ID Each common link in the EAPS network must have a unique link ID.
The controller and partner shared ports belonging to the same
common link must have matching link IDs, and not other instance in
the network should have that link ID.
EBGP Exterior Border Gateway Protocol. EBGP is a protocol in the IP suite

designed to exchange network reachability information with BGP
systems in other ASs. EBGP works between different ASs.
ECMP Equal Cost Multi Paths. In OSPF, this routing algorithm distributes
network traffic across multiple high-bandwidth links to increase
performance. The Extreme Networks OSPF implementation supports
multiple equal cost paths between points and divides traffic evenly
among the available paths. As many as four links may be involved in
an ECMP link, and traffic is shared on the basis of IP source/
destination address session.
edge ports In STP, edge ports connect to non-STP devices such as routers,
endstations, and other hosts.
EDP Extreme Discovery Protocol. EDP is a protocol used to gather

information about neighbor Extreme Networks switches. Extreme
Networks switches use EDP to exchange topology information.
EEPROM Electrically erasable programmable read-only memory. EEPROM is a

memory that can be electronically programmed and erased but does
not require a power source to retain data.

Glossary
E (Continued)
EGP Exterior Gateway Protocol. EGP is an Internet routing protocol for
exchanging reachability information between routers in different ASs.
BGP is a more recent protocol that accomplishes this task.
election algorithm In ESRP, this is a user-defined criteria to determine how the master
and slave interact. The election algorithm also determines which
device becomes the master or slave and how ESRP makes those
decisions.
ELSM Extreme Link Status Monitoring. ELSM is an Extreme Networks-

proprietary protocol that monitors network health. You can also use
ELSM with Layer 2 control protocols to improve Layer 2 loop
recovery in the network.
ELRP Extreme Loop Recovery Protocol. ELRP is an Extreme Networks-

proprietary protocol that allows you to detect Layer 2 loops.
EMISTP Extreme Multiple Instance Spanning Tree Protocol. This Extreme

Networks-proprietary protocol uses a unique encapsulation method
for STP messages that allows a physical port to belong to multiple
STPDs.
EMS Event Management System. This Extreme Networks-proprietary

system saves, displays, and filters events, which are defined as any
occurrences on a switch that generate a log message or require action.
encapsulation mode Using STP, you can configure ports within an STPD to accept specific
BPDU encapsulations. The three encapsulation modes are:
● 802.1D—This mode is used for backward compatibility with
previous STP versions and for compatibility with third-party
switches using IEEE standard 802.1D.
● EMISTP—Extreme Multiple Instance Spanning Tree Protocol mode
is an extension of STP that allows a physical port to belong to
multiple STPDs by assigning the port to multiple VLANs.
● PVST+—This mode implements PVST+ in compatibility with third-
party switches running this version of STP.
EPICenter EPICenter is an Extreme Networks-proprietary graphical user interface

(GUI) network management system.
ESRP Extreme Standby Router Protocol. ESRP is an Extreme Networks-

proprietary protocol that provides redundant Layer 2 and routing
services to users.
ESRP-aware device This is an Extreme Networks device that is not running ESRP itself but
that is connected on a network with other Extreme Networks switches
that are running ESRP. These ESRP-aware devices also fail over.
ESRP domain An ESRP domain allows multiple VLANs to be protected under a

single logical entity. An ESRP domain consists of one domain-master
VLAN and zero or more domain-member VLANs.

E (Continued)
ESRP-enabled device An ESRP-enabled device is an Extreme Networks switch with an ESRP
domain and ESRP enabled. ESRP-enabled switches include the ESRP
master and slave switches.
ESRP groups An ESRP group runs multiple instances of ESRP within the same
VLAN (or broadcast domain). To provide redundancy at each tier, use
a pair of ESRP switches on the group.
ESRP instance You enable ESRP on a per domain basis; each time you enable ESRP is
an ESRP instance.
ESRP VLAN A VLAN that is part of an ESRP domain, with ESRP enabled, is an
ESRP VLAN.
Ethernet This is the IEEE 802.3 networking standard that uses carrier sense
multiple access with collision detection (CSMA/CD). An Ethernet
device that wants to transmit first checks the channel for a carrier, and
if no carrier is sensed within a period of time, the device transmits. If
two devices transmit simultaneously, a collision occurs. This collision
is detected by all transmitting devices, which subsequently delay their
retransmissions for a random period. Ethernet runs at speeds from 10
Mbps to 10 Gbps on full duplex.
extended mode ESRP extended mode supports and is compatible only with switches
running ExtremeWare XOS software exclusively.
F
Fast Convergence In EAPS, Fast Convergence allows convergence in less than 50
milliseconds. You configure this parameter for the entire switch, not
by EAPS domain.
fast path This term refers to the data path for a packet that traverses the switch
and does not require processing by the CPU. Fast path packets are
handled entirely by ASICs and are forwarded at wire speed rate.
FDB Forwarding database. The switch maintains a database of all MAC

address received on all of its ports and uses this information to decide
whether a frame should be forwarded or filtered. Each FDB entry
consists of the MAC address of the sending device, an identifier for
the port on which the frame was received, and an identifier for the
VLAN to which the device belongs. Frames destined for devices that
are not currently in the FDB are flooded to all members of the VLAN.
For some types of entries, you configure the time it takes for the
specific entry to age out of the FDB.
FIB Forwarding Information Base. On the BlackDiamond 8800 family of

switches and the Summit X450 switch, the Layer 3 routing table is
referred to as the FIB.
frame This is the unit of transmission at the data link layer. The frame
contains the header and trailer information required by the physical
medium of transmission.

Glossary
F (Continued)
full-duplex This is the communication mode in which a device simultaneously
sends and receives over the same link, doubling the bandwidth. Thus,
a full-duplex 100 Mbps connection has a bandwidth of 200 Mbps, and
so forth. A device either automatically adjusts its duplex mode to
match that of a connecting device or you can configure the duplex
mode; all devices at 1 Gbps or higher run only in full-duplex mode.
G
GBIC Gigabit Interface Connector. These devices, available in a variety of
fiber modes and physical shapes, provide the physical interface to a
gigabit Ethernet connection.
Gigabit Ethernet This is the networking standard for transmitting data at 1000 Mbps or
1 Gbps. Devices can transmit at multiples of gigabit Ethernet as well.
H
HA Host Attach. In ExtremeWare XOS software, HA is part of ESRP that
allows you to connect active hosts directly to an ESRP switch; it allows
configured ports to continue Layer 2 forwarding regardless of their
ESRP status.
half-duplex This is the communication mode in which a device can either send or
receive data, but not simultaneously. (Devices at 1 Gbps or higher do
not run in half-duplex mode; they run only in full-duplex mode.)
header This is control information (such as originating and destination

stations, priority, error checking, and so forth) added in front of the
data when encapsulating the data for network transmission.
hitless failover In the Extreme Networks implementation on modular switches, hitless

failover means that designated configurations survive a change of
primacy between the two MSMs with all details intact. Thus, those
features run seamlessly during and after control of the system changes
from one MSM to another.
I
IBGP Interior Border Gateway Protocol. IBGP is the BGP version used
within an AS.
ICMP Internet Control Message Protocol. ICMP is the part of the TCP/IP
protocol that allows generation of error messages, test packets, and
operating messages. For example, the ping command allows you to
send ICMP echo messages to a remote IP device to test for
connectivity. ICMP also supports traceroute, which identifies
intermediate hops between a given source and destination.

I (Continued)
IEEE Institute of Electrical and Electronic Engineers. This technical
professional society fosters the development of standards that often
become national and international standards. The organization
publishes a number of journals and has many local chapters and
several large societies in special areas.
IETF Internet Engineering Task Force. The IETF is a large, open,

international community of network designers, operators, vendors,
and researchers concerned with the evolution of the Internet
architecture and the smooth operation of the Internet. The technical
work of the IETF is done in working groups, which are organized by
topic.
IGMP Internet Group Management Protocol. Hosts use IGMP to inform local
routers of their membership in multicast groups. Multicasting allows
one computer on the Internet to send content to multiple other
computers that have identified themselves as interested in receiving
the originating computer's content. When all hosts leave a group, the
router no longer forwards packets that arrive for the multicast group.
IGMP snooping This provides a method for intelligently forwarding multicast packets
within a Layer 2 broadcast domain. By “snooping” the IGMP
registration information, the device forms a distribution list that
determines which endstations receive packets with a specific multicast
address. Layer 2 switches listen for IGMP messages and build
mapping tables and associated forwarding filters. IGMP snooping also
reduces IGMP protocol traffic.
IGP Interior Gateway Protocol. IGP refers to any protocol used to exchange
routing information within an AS. Examples of Internet IGPs include
RIP and OSPF.
inline power According to IEEE 802.3 af, inline power refers to providing an AC or
DC power source through the same cable as the data travels. It allows
phones and network devices to be placed in locations that are not near
AC outlets. Most standard telephones use inline power.
IP Internet Protocol. The communications protocol underlying the

Internet, IP allows large, geographically diverse networks of
computers to communicate with each other quickly and economically
over a variety of physical links; it is part of the TCP/IP suite of
protocols. IP is the Layer 3, or network layer, protocol that contains
addressing and control information that allows packets to be routed.
IP is the most widely used networking protocol; it supports the idea of
unique addresses for each computer on the network. IP is a
connectionless, best-effort protocol; TCP reassembles the data after
transmission. IP specifies the format and addressing scheme for each
packet.
IPv6 Internet Protocol version 6. IPv6 is the next-generation IP protocol.

The specification was completed in 1997 by IETF. IPv6 is backward-
compatible with and is designed to fix the shortcomings of IPv4, such
as data security and maximum number of user addresses. IPv6
increases the address space from 32 to 128 bits, providing for an
unlimited (for all intents and purposes) number of networks and
systems; IPv6 is expected to slowly replace IPv4, with the two existing
side by side for many years.

Glossary
I (Continued)
IP address IP address is a 32-bit number that identifies each unique sender or
receiver of information that is sent in packets; it is written as four
octets separated by periods (dotted-decimal format). An IP address
has two parts: the identifier of a particular network and an identifier
of the particular device (which can be a server or a workstation)
within that network. You may add an optional subnetwork identifier.
Only the network part of the address is looked at between the routers
that move packets from one point to another along the network.
Although you can have a static IP address, many IP addresses are
assigned dynamically from a pool. Many corporate networks and
online services economize on the number of IP addresses they use by
sharing a pool of IP addresses among a large number of users. (The
format of the IP address is slightly changed in IPv6.)
IPTV Internal Protocol television. IPTV uses a digital signal sent via
broadband through a switched telephone or cable system. An
accompanying set top box (that sits on top of the TV) decodes the
video and converts it to standard television signals.
IR Internal router. In OSPF, IR is an internal router that has all interfaces

within the same area.
IRDP Internet Router Discovery Protocol. Used with IP, IRDP enables a host
to determine the address of a router that it can use as a default
gateway. In Extreme Networks implementation, IP multinetting
requires a few changes for the IRDP.
ISO This abbreviation is commonly used for the International Organization

for Standardization, although it is not an acronym. ISO was founded
in 1946 and consists of standards bodies from more than 75 nations.
ISO had defined a number of important computer standards,
including the OSI reference model used as a standard architecture for
networking.
ISP An Internet Service Provider is an organization that provides access to

the Internet. Small ISPs provide service via modem and ISDN while
the larger ones also offer private line hookups (T1, fractional T1, etc.).
Customers are generally billed a fixed rate per month, but other
charges may apply. For a fee, a Web site can be created and
maintained on the ISP's server, allowing the smaller organization to
have a presence on the Web with its own domain name.
ITU-T International Telecommunication Union-Telecommunication. The ITU-

T is the telecommunications division of the ITU international
standards body.
J
jumbo frames These are Ethernet frames that are larger that 1522 bytes (including the
4 bytes in the CRC). The jumbo frame size is configurable on Extreme
Networks devices; the range is from 1523 to 9216 bytes.

L
LACP Link Aggregation Control Protocol. LACP is part of the IEEE 802.3ad
and automatically configures multiple aggregated links between
switches.
LAG Link aggregation group. A LAG is the logical high-bandwidth link

that results from grouping multiple network links in link aggregation
(or load sharing). You can configure static LAGs or dynamic LAGs
(using the LACP).
Layer 2 Layer 2 is the second, or data link, layer of the OSI model, or the MAC
layer. This layer is responsible for transmitting frames across the
physical link by reading the hardware, or MAC, source and
destination addresses.
Layer 3 Layer 3 is the third layer of the OSI model. Also known as the
network layer, Layer 3 is responsible for routing packets to different
LANs by reading the network address.
LED Light-emitting diode. LEDs are on the device and provide information
on various states of the device’s operation. See the Consolidated XOS
Hardware Installation Guide for a complete explanation of the LEDs on
devices running ExtremeWare XOS.
license ExtremeWare XOS version 11.1 introduces a licensing feature to the

ExtremeWare XOS software. You must have a license, which you
obtain from Extreme Networks, to apply the full functionality of some
features.
link aggregation Link aggregation, also known as trunking or load sharing, conforms to
IEEE 802.3ad. This feature is the grouping of multiple network links
into one logical high-bandwidth link.
link type In OSPF, there are four link types that you can configure: auto,
broadcast, point-to-point, and passive.
LLDP Link Layer Discovery Protocol. LLDP conforms to IEEE 802.1ab and is
a neighbor discovery protocol. Each LLDP-enabled device transmits
information to its neighbors, including chassis and port identification,
system name and description, VLAN names, and other selected
networking information. The protocol also specifies timing intervals in
order to ensure current information is being transmitted and received.
load sharing Load sharing, also known as trunking or link aggregation, conforms to
IEEE 802.3ad. This feature is the grouping of multiple network links
into one logical high-bandwidth link. For example, by grouping four
100 Mbps of full-duplex bandwidth into one logical link, you can
create up to 800 Mbps of bandwidth. Thus, you increase bandwidth
and availability by using a group of ports to carry traffic in parallel
between switches.
LFS Link Fault Signal. LFS, which conforms to IEEE standard 802.3ae-2002,
monitors 10 Gbps ports and indicates either remote faults or local
faults.

Glossary
L (Continued)
loop detection In ELRP, loop detection is the process used to detect a loop in the
network. The switch sending the DLRP PDU waits to receive its
original PDU back. If the switch received this original PDU, there is a
LSA Link state advertisement. An LSA is a broadcast packet used by link

state protocols, such as OSPF. The LSA contains information about
neighbors and path costs and is used by the receiving router to
maintain a routing table.
LSDB Link state database. In OSPF, LSDB is a database of information about

the link state of the network. Two neighboring routers consider
themselves to be adjacent only if their LSDBs are synchronized. All
routing information is exchanged only between adjacent routers.
M
MAC address Media access control address. The MAC address, sometimes known as
the hardware address, is the unique physical address of each network
interface card on each device.
MAN Metropolitan area network. A MAN is a data network designed for a

town or city. MANs may be operated by one organization such as a
corporation with several offices in one city, or be shared resources
used by several organizations with several locations in the same city.
MANs are usually characterized by very high-speed connections.
master node In EAPS, the master node is a switch, or node, that is designated the
master in an EAPS domain ring. The master node blocks the
secondary port for all non-control traffic belonging to this EAPS
domain, thereby avoiding a loop in the ring.
master router In VRRP, the master router is the physical device (router) in the VRRP
virtual router that is responsible for forwarding packets sent to the
VRRP virtual router and for responding to ARP requests. The master
router sends out periodic advertisements that let backup routers on
the network know that it is alive. If the VRRP IP address owner is
identified, it always becomes the master router.
master VLAN In ESRP, the master VLAN is the VLAN on the ESRP domain that
exchanges ESRP-PDUs and data between a pair of ESRP-enabled
devices. You must configure one master VLAN for each ESRP domain,
and a master VLAN can belong to only one ESRP domain.
MED Multiple exit discriminator. BGP uses the MED metric to select a
particular border router in another AS when multiple border routers
exist.
MEP In CFM, maintenance end point is and end point for a single domain,
or maintenance association. The MEP may be either outward-facing or
inward facing.

M (Continued)
metering In QoS, metering monitors the traffic pattern of each flow against the
traffic profile. For out-of-profile traffic the metering function interacts
with other components to either re-mark or drop the traffic for that
flow. In the Extreme Networks implementation, you use ACLs to
enforce metering.
member VLAN In ESRP, you configure zero or more member VLANs for each ESRP
domain. A member VLAN can belong to only one ESRP domain. The
state of the ESRP device determines whether the member VLAN is in
forwarding or blocking state.
MIB Management Information Base. MIBs make up a database of

information (for example, traffic statistics and port settings) that the
switch makes available to network management systems. MIB names
identify objects that can be managed in a network and contain
information about the objects. MIBs provide a means to configure a
network device and obtain network statistics gathered by the device.
Standard, minimal MIBs have been defined, and vendors often have
private enterprise MIBs.
MIP In CFM, the maintenance intermediate point is intermediate between

endpoints. Each MIP is associated with a single domain, and there
may be more than one MIP in a single domain.
mirroring Port mirroring configures the switch to copy all traffic associated with
one or more ports to a designated monitor port. The monitor port can
be connected to an network analyzer or RMON probe for packet
analyzer.
MMF Multimode fiber. MMF is a fiber optic cable with a diameter larger
than the optical wavelength, in which more than one bound mode can
propagate. Capable of sending multiple transmissions simultaneously,
MMF is commonly used for communications of 2 kilometers or less.
MSM Master Switch Fabric Module. This Extreme Networks-proprietary

name refers to the module that holds both the control plane and the
switch fabric for switches that run the ExtremeWare XOS software on
modular switches. One MSM is required for switch operation; adding
an additional MSM increases reliability and throughput. Each MSM
has two CPUs. The MSM has LEDs as well as a console port,
management port, modem port, and compact flash; it may have data
ports as well. The MSM is responsible for upper-layer protocol
processing and system management functions. When you save the
switch configuration, it is saved to all MSMs.
MSDP Multicast Source Discovery Protocol. MSDP is used to connect

multiple multicast routing domains. MSDP advertises multicast
sources across Protocol Independent Multicast-Sparse Mode (PIM-SM)
multicast domains or Rendezvous Points (RPs). In turn, these RPs run
MSDP over TCP to discover multicast sources in other domains.

Glossary
M (Continued)
MSTI Multiple Spanning Tree Instances. MSTIs control the topology inside
an MSTP region. An MSTI is a spanning tree domain that operates
within a region and is bounded by that region; and MSTI does not
exchange BPDUs or send notifications to other regions. You can map
multiple VLANs to an MSTI; however, each VLAN can belong to only
one MSTI.You can configure up to 64 MSTIs in an MSTP region.
MSTI regional root bridge In an MSTP environment, each MSTI independently elects its own root
bridge. The bridge with the lowest bridge ID becomes the MSTI
regional root bridge. The bridge ID includes the bridge priority and
the MAC address.
MSTI root port In an MSTP environment, the port on the bridge with the lowest path
cost to the MSTI regional root bridge is the MSTI root port.
MSTP Multiple Spanning Tree Protocol. MSTP, based on IEEE 802.1Q-2003

(formerly known as IEEE 892.1s), allows you to bundle multiple
VLANs into one spanning tree (STP) topology, which also provides
enhanced loop protection and better scaling. MSTP uses RSTP as the
converging algorithm and is compatible with legacy STP protocols.
MSTP region An MSTP region defines the logical boundary of the network.
Interconnected bridges that have the same MSTP configuration are
referred to as an MSTP region. Each MSTP region has a unique
identifier, is bound together by one CIST that spans the entire
network, and contains from 0 to 64 MSTIs. A bridge participates in
only one MSTP region at one time. An MSTP topology is individual
MSTP regions connected either to the rest of the network with 802.1D
and 802.1w bridges or to each other.
MTU Maximum transmission unit. This term is a configurable parameter

that determines the largest packet than can be transmitted by an IP
interface (without the packet needing to be broken down into smaller
units).
Note: Packets that are larger than the configured MTU size are
dropped at the ingress port. Or, if configured to do so, the system can
fragment the IPv4 packets and reassemble them at the receiving end.
multicast Multicast messages are transmitted to selected devices that specifically

join the multicast group; the addresses are specified in the destination
address field. In other words, multicast (point-to-multipoint) is a
communication pattern in which a source host sends a message to a
group of destination hosts.
multinetting IP multinetting assigns multiple logical IP interfaces on the same

circuit or physical interface. This allows one bridge domain (VLAN) to
have multiple IP networks.

M (Continued)
MVR Multicast VLAN registration. MVR allows a subscriber on a port to
subscribe and unsubscribe to a multicast stream on the network-wide
multicast VLAN; it allows the single multicast VLAN to be shared in
the network while subscribers remain in separate VLANs. MVR
provides the ability to continuously send multicast streams in the
multicast VLAN, but to isolate the The application from the subscriber
VLANs for bandwidth and security reasons. MVR allows a multicast
stream received over a Layer 2 VLAN to be forwarded to another
VLAN, eliminating the need for a Layer 3 routing protocol; this
feature is often used for IPTV applications.
N
NAT Network Address Translation. This is a network capability that
enables a group of computers to dynamically share a single incoming
IP address. NAT takes the single incoming IP address and creates a
new IP address for each client computer on the network.
netlogin Network login provides extra security to the network by assigning

addresses only to those users who are properly authenticated. You can
use web-based, MAC-based, or IEEE 802.1x-based authentication with
network login. The two modes of operation are campus mode and ISP
mode.
netmask A netmask is a string of 0s and 1s that mask, or screen out, the

network part of an IP address, so that only the host computer part of
the address remains.
neutral state/switch In ESRP, the neutral state is the initial state entered by the switch. In a
neutral state, the switch waits for ESRP to initialize and run. A neutral
switch does not participate in ESRP elections.
NLRI Network layer reachability information. In BGP, the system sends

routing update messages containing NLRI to describe a route and how
to get there. A BGP update message carries one or more NLRI prefixes
and the attributes of a route for each NLRI prefix; the route attributes
include a BGP next hop gateway address, community values, and
other information.
node In the Extreme Networks implementation, a node is a CPU that runs

the management application on the switch. Each MSM on modular
switches installed in the chassis is a node.
In general networking terms, a node is a device on the network.
Node Manager The Node Manager performs the process of node election, which
selects the master, or primary, MSM when you have two MSMS
installed in the modular chassis. The Node Manager is useful for
system redundancy.

Glossary
N (Continued)
NSSA Not-so-stubby area. In OSPF, NSSA is a stub area, which is connected
to only one other area, with additional capabilities:
● External routes originating from an ASBR connected to the NSSA
can be advertised within the NSSA.
● External routes originating from the NSSA can be propagated to
other areas.
O
odometer In the Extreme Networks implementation, each field replaceable
component contains a system odometer counter in EEPROM.
On modular switches, using the CLI, you can display how long each
following individual component has been in service:
● chassis
● MSMs
● I/O modules
● power controllers
On stand-alone switches, you display the days of service for the
switch.
option 82 This is a security feature that you configure as part of BOOTP/DHCP.

Option 82 allows a server to bind the client’s port, IP address, and
MAC number for subscriber identification.
OSI Open Systems Interconnection. OSI is the international standard

computer network architecture known for its 7-layer reference model.
OSI reference model The 7-layer standard model for network architecture is the basis for
defining network protocol standards and the way that data passes
through the network. Each layer specifies particular network
functions; the highest layer is closest to the user, and the lowest layer
is closest to the media carrying the information. So, in a given message
between users, there will be a flow of data through each layer at one
end down through the layers in that computer and, at the other end,
when the message arrives, another flow of data up through the layers
in the receiving computer and ultimately to the end user or program.
This model is used worldwide for teaching and implementing
networking protocols.
OSPF Open Shortest Path First. This is an IGP. OSPF, a routing protocol for
TCP/IP networks, uses a link state routing algorithm that calculates
routes for packets based on a number of factors, including least hops,
speed of transmission lines, and congestion delays. You can also
configure certain cost metrics for the algorithm. This protocol is more
efficient and scalable than vector-distance routing protocols. OSPF
features include least-cost routing, ECMP routing, and load balancing.
Although OSPF requires CPU power and memory space, it results in
smaller, less frequent router table updates throughout the network.
This protocol is more efficient and scalable than vector-distance
routing protocols.

O (Continued)
OSPFv3 OSPFv3 is one of the routing protocols used with IPV6 and is similar
to OSPF.
OUI The Organizational Unique Identifier is the first 24 bits of a MAC

address for a network device that indicate a specific vendor as
assigned by IEEE.
P
packet This is the unit of data sent across a network. Packet is a generic term
used to describe units of data at all levels of the protocol stack, but it
is most correctly used to describe application data units. The packet is
a group of bits, including data and control signals, arranged in a
specific format. It usually includes a header, with source and
destination data, and user data. The specific structure of the packet
depends on the protocol used.
partner node In EAPS, the partner node is that end of the common link that is not a
controller node; the partner node does not participate in any form of
blocking.
PD Powered device. In PoE, the PD is the powered device that plugs into
the PoE switch.
PDU Protocol data unit. A PDU is a message of a given protocol comprising

payload and protocol-specific control information, typically contained
in a header.
PIM-DM Protocol-Independent Multicast - Dense mode. PIM-DM is a multicast

protocol that uses Reverse Path Forwarding but does not require any
particular unicast protocol. It is used when recipients are in a
concentrated area.
PIM-SM Protocol-Independent Multicast - Sparse mode. PIM-SM is a multicast

protocol that defines a rendezvous point common to both sender and
receiver. Sender and receiver initiate communication at the
rendezvous point, and the flow begins over an optimized path. It is
used when recipients are in a sparse area.
ping Packet Internet Groper. Ping is the ICMP echo message and its reply
that tests network reachability of a device. Ping sends an echo packet
to the specified host, waits for a response, and reports success or
failure and statistics about its operation.
PMBR PIM multicast border router. A PIMBR integrates PIM-DM and PIM-
SM traffic.
PoE Power over Ethernet. The PoE standard (IEEE 802.3af) defines how
power can be provided to network devices over existing Ethernet
connections.

Glossary
P (Continued)
policy files You use policy files in ExtremeWare XOS to specify ACLs and
policies. A policy file is a text file (with a .pol extension) that specifies
a number of conditions to test and actions to take. For ACLs, this
information is applied to incoming traffic at the hardware level.
Policies are more general and can be applied to incoming routing
information; they can be used to rewrite and modify routing
advertisements.
port mirroring Port mirroring configures the switch to copy all traffic associated with
one or more ports to a designated monitor port. A packet bound for or
heading away from the mirrored port is forwarded onto the monitor
port as well. The monitor port can be connected to a network analyzer
or RMON probe for packet analysis. Port mirroring is a method of
monitoring network traffic that a network administrator uses as a
diagnostic tool or debugging feature; it can be managed locally or
remotely.
POST Power On Self Test. On Extreme Networks switches, the POST runs
upon powering-up the device. If the MGMT LED is yellow after the
POST completes, contact your supplier for advice.
primary port In EAPS, a primary port is a port on the master node that is
designated the primary port to the ring.
protected VLAN In STP, protected VLANs are the other (other than the carrier VLAN)
VLANs that are members of the STPD but do not define the scope of
the STPD. Protected VLANs do not transmit or receive STP BPDUs,
but they are affected by STP state changes and inherit the state of the
carrier VLAN. Also known as non-carrier VLANs, they carry the data
traffic.
In EAPS, a protected VLAN is a VLAN that carries data traffic
through an EAPS domain. You must configure one or more protected
VLANs for each EAPS domain. This is also known as a data VLAN.
proxy ARP This is the technique in which one machine, usually a router, answers
ARP requests intended for another machine. By masquerading its
identity (as an endstation), the router accepts responsibility for routing
packets to the real destination. Proxy ARP allows a site to use a single
IP address with two physical networks. Subnetting is normally a
better solution.
PVST+ Per VLAN Spanning Tree +. This implementation of STP has a 1:1
relationship with VLANs. The Extreme Networks implementation of
PVST+ allows you to interoperate with third-party devices running
this version of STP. PVST is a earlier version of this protocol and is
compatible with PVST+.
Q
QoS Quality of Service. Policy-enabled QoS is a network service that
provides the ability to prioritize different types of traffic and to
manage bandwidth over a network. QoS uses various methods to
prioritize traffic, including IEEE 802.1p values and IP DiffServ values.

R
RADIUS Remote Authentication Dial In User Service. RADIUS is a client/
server protocol and software that enables remote access servers to
communicate with a central server to authenticate dial-in users and
authorize their access to the requested system or service. RADIUS
allows a company to maintain user profiles in a central database that
all remote servers can share. It provides better security, allowing a
company to set up a policy that can be applied at a single
administered network point. With RADIUS, you can track usage for
billing and for keeping network statistics.
RARP Reverse ARP. Using this protocol, a physical device requests to learn
its IP address from a gateway server's ARP table. When a new device
is set up, its RARP client program requests its IP address from the
RARP server on the router. Assuming that an entry has been set up in
the router table, the RARP server will return the IP address to the
machine which can store it for future use.
RFC Request for Comment. The IETF RFCs describe the definitions and
parameters for networking.
RIP Routing Information Protocol. This IGP vector-distance routing

protocol is part of the TCP/IP suite and maintains tables of all known
destinations and the number of hops required to reach each. Using
RIP, routers periodically exchange entire routing tables. RIP is suitable
for use only as an IGP.
RIPng RIP next generation. RIPng is one of the routing protocols used with
IPV6 and is similar to RIP.
RMON Remote monitoring. RMON is a standardized method to make switch

and router information available to remote monitoring applications. It
is an SNMP network management protocol that allows network
information to be gathered remotely. RMON collects statistics and
enables a management station to monitor network devices from a
central location. It provides multivendor interoperability between
monitoring devices and management stations. RMON is described in
several RFCs (among them IETF RFC 1757 and RFC 2201). Network
administrators use RMON to monitor, analyze, and troubleshoot the
network. A software agent can gather the information for presentation
to the network administrator with a graphical user interface (GUI).
The administrator can find out how much bandwidth each user is
using and what Web sites are being accessed; you can also set alarms
to be informed of potential network problems.
root bridge In STP, the root bridge is the bridge with the best bridge identifier
selected to be the root bridge. The network has only one root bridge.
The root bridge is the only bridge in the network that does not have a
root port.
root port In STP, the root port provides the shortest path to the root bridge. All
bridges except the root bridge contain one root port.
route aggregation In BGP, you can combine the characteristics of several routes so they
are advertised as a single route, which reduces the size of the routing
tables.

Glossary
R (Continued)
route flapping A route is flapping when it is repeatedly available, then unavailable,
then available, then unavailable. In the ExtremeWare XOS BGP
implementation, you can minimize the route flapping using the route
flap dampening feature.
route reflector In BGP, you can configure the routers within an AS such that a single
router serves as a central routing point for the entire AS.
routing confederation In BGP, you can configure a fully meshed AS into several sub-ASs and
group these sub-ASs into a routing confederation. Routing
confederations help with the scalability of BGP.
RSTP Rapid Spanning Tree Protocol. RSTP, described in IEEE 802.1w, is an

enhanced version of STP that provides faster convergence. The
Extreme Networks implementation of RSTP allows seamless
interoperability with legacy STP.
S
SA Source address. The SA is the IP or MAC address of the device issuing
the packet.
SCP Secure Copy Protocol. SCP2, part of SSH2, is used to transfer

configuration and policy files.
secondary port In EAPS, the secondary port is a port on the master node that is
designated the secondary port to the ring. The transit node ignores the
secondary port distinction as long as the node is configured as a
transit node.
sFlow sFlow allows you to monitor network traffic by statistically sampling

the network packets and periodically gathering the statistics. The
sFlow monitoring system consists of an sFlow agent (embedded in a
switch, router, or stand-alone probe) and an external central data
collector, or sFlow analyzer.
SFP Small form-factor pluggable. SFP is a specific type of GBIC, designed

for use with small form factor connectors. These transceivers offer
high speed and physical compactness.
6in4 tunnels The 6in4 tunnels, which encapsulate IPv6 packets into IPv4 packets,
use dynamic routing protocols to establish connectivity between
several IPv6 networks through over the intervening IPv4 network. The
6in4 tunnel must be created at each tunnel endpoint. The IPv6 routing
protocol considers the 6in4 tunnel as a single IPv6 hop, even if the
tunnel comprises many IPv4 hops; the IPv6 protocol considers the
6in4 tunnel as a normal IPv6 point-to-point link. You can have many
6in4 tunnels per VR.

S (Continued)
6to4 tunnels The 6to4 tunnels are one way to send IPv6 packets over IPv4
networks. This transition mechanism provides a way to connect IPv6
end-site networks by automatically tunnelling over the intervening
IPv4 Internet. A special IPv6 routing prefix is used to indicate that the
remaining part of the external routing prefix contains the IPv4
endpoint address of a boundary IPv6 router for that site that will
process IPv6-in-IPv4-encapsulated packets. You can have only one
6to4 tunnel per VR.
slow path This term refers to the data path for packets that must be processed by
the switch CPU, whether these packets are generated by the CPU,
removed from the network by the CPU, or simply forwarded by the
CPU.
SMF Single-mode fiber. SMF is a laser-driven optical fiber with a core

diameter small enough to limit transmission to a single bound mode.
SMF is commonly used in long distance transmission of more than 3
miles; it sends one transmission at a time.
SNMP Simple Network Management Protocol. SNMP is a standard that uses

a common software agent to remotely monitor and set network
configuration and runtime parameters. SNMP operates in a
multivendor environment, and the agent uses MIBs, which define
what information is available from any manageable network device.
You can also set traps using SNMP, which send notifications of
network events to the system log.
SNTP Simple Network Time Protocol. SNTP is used to synchronize the

system clocks throughout the network. An extension of the Network
Time Protocol, SNTP can usually operate with a single server and
allows for IPv6 addressing.
SSH Secure Shell, sometimes known as Secure Socket Shell, is a UNIX-

based command interface and protocol of securely gaining access to a
remote computer. With SSH commands, both ends of the client/server
connection are authenticated using a digital certificate, and passwords
are protected by being encrypted. At Extreme Networks, the SSH is a
separate software module, which must be downloaded separately.
(SSH is bundled with SSL in the software module.)
SSL Secure Sockets Layer. SSL is a protocol for transmitting private

documents using the Internet. SSL works by using a public key to
encrypt data that is transferred over the SSL connection. SSL uses the
public-and-private key encryption system, which includes the use of a
digital certificate. At Extreme Networks, SSL is bundled with the SSH
software module, which must be downloaded separately. SSL used for
other applications than SSH, CNA at Extreme Networks for example.
standard mode Use ESRP standard mode if your network contains switches running
ExtremeWare and switches running ExtremeWare XOS, both
participating in ESRP.

Glossary
S (Continued)
STP Spanning Tree Protocol. STP is a protocol, defined in IEEE 802.1d,
used to eliminate redundant data paths and to increase network
efficiency. STP allows a network to have a topology that contains
physical loops; it operates in bridges and switches. STP opens certain
paths to create a tree topology, thereby preventing packets from
looping endlessly on the network. To establish path redundancy, STP
creates a tree that spans all of the switches in an extended network,
forcing redundant paths into a standby, or blocked, state. STP allows
only one active path at a time between any two network devices (this
prevents the loops) but establishes the redundant links as a backup if
the initial link should fail. If STP costs change, or if one network
segment in the STP becomes unreachable, the spanning tree algorithm
reconfigures the STP topology and re-establishes the link by activating
the standby path.
STPD Spanning Tree Domain. An STPD is an STP instance that contains one
or more VLANs. The switch can run multiple STPDs, and each STPD
has its own root bridge and active path. In the Extreme Networks
implementation of STPD, each domain has a carrier VLAN (for
carrying STP information) and one or more protected VLANs (for
carrying the data).
STPD mode The mode of operation for the STPD. The two modes of operation are:
● 802.1d—Compatible with legacy STP and other devices using the
IEEE 802.1d standard.
● 802.1w—Compatible with Rapid Spanning Tree (RSTP).
stub areas In OSPF, a stub area is connected to only one other area (which can be
the backbone area). External route information is not distributed to
stub areas.
superloop In EAPS, a superloop occurs if the common link betwee two EAPS
domains goes down and the master nodes of both domains enter the
falied state putting their respective secondary ports into the
forwarding state. If there is a data VLAN spanning both EAPS
domains, this action forms a loop between the EAPS domains.
system health check The primary responsibility of the system health checker is to monitor
and poll error registers. In addition, the system health checker can be
enabled to periodically send diagnostic packets. System health check
errors are reported to the syslog.
T
TACACS+ Terminal Access Controller Access Control System. Often run on
UNIX systems, the TACAS+ protocol provides access control for
routers, network access servers, and other networked computing
devices via one or more centralized servers. TACACS+ provides
separate authentication, authorization, and accounting services. User
passwords are administered in a central database rather than in
individual routers, providing easily scalable network security
solutions.

T (Continued)
tagged VLAN You identify packets as belonging to the same tagged VLAN by
putting a value into the 12-bit (4 octet) VLAN ID field that is part of
the IEEE 802.1Q field of the header. Using this 12-bit field, you can
configure up to 4096 individual VLAN addresses (usually some are
reserved for system VLANs such as management and default VLANs);
these tagged VLANs can exist across multiple devices. The tagged
VLAN can be associated with both tagged and untagged ports.
TCN Topology change notification. The TCN is a timer used in RSTP that
signals a change in the topology of the network.
TCP Transmission Control Protocol. Together with Internet Protocol (IP),

TCP is one of the core protocols underlying the Internet. The two
protocols are usually referred to as a group, by the term TCP/IP. TCP
provides a reliable connection, which means that each end of the
session is guaranteed to receive all of the data transmitted by the other
end of the connection, in the same order that it was originally
transmitted without receiving duplicates.
TFTP Trivial File Transfer Protocol. TFTP is an Internet utility used to

transfer files, which does not provide security or directory listing. It
relies on UDP.
transit node In EAPS, the transit node is a switch, or node, that is not designated a
master in the EAPS domain ring.
U
UDP User Datagram Protocol. This is an efficient but unreliable,
connectionless protocol that is layered over IP (as is TCP). Application
programs must supplement the protocol to provide error processing
and retransmitting data. UDP is an OSI Layer 4 protocol.
unicast A unicast packet is communication between a single sender and a

single receiver over a network.
untagged VLAN A VLAN remains untagged unless you specifically configure the IEEE
802.1Q value on the packet. A port cannot belong to more than one
untagged VLAN using the same protocol.
USM User-based security model. In SNMPv3, USM uses the traditional

SNMP concept of user names to associate with security levels to
support secure network management.
V
virtual link In OSPF, when a new area is introduced that does not have a direct
physical attachment to the backbone, a virtual link is used. Virtual
links are also used to repair a discontiguous backbone area.

Glossary
V (Continued)
virtual router In the Extreme Networks implementations, virtual routers allow a
single physical switch to be split into multiple virtual routers. Each
virtual router has its own IP address and maintains a separate logical
forwarding table. Each virtual router also serves as a configuration
domain. The identity of the virtual router you are working in
currently displays in the prompt line of the CLI. The virtual routers
discussed in relation to Extreme Networks switches themselves are not
the same as the virtual router in VRRP.
In VRRP, the virtual router is identified by a virtual router (VRID) and
an IP address. A router running VRRP can participate in one or more
virtual routers. The VRRP virtual router spans more than one physical
router, which allows multiple routers to provide redundant services to
users.
virtual router MAC address In VRRP, RFC 2338 assigns a static MAC address for the first five
octets of the VRRP virtual router. These octets are set to 00-00-5E-00-
01. When you configure the VRRP VRID, the last octet of the MAC
address is dynamically assigned the VRID number.
VLAN Virtual LAN. The term VLAN is used to refer to a collection of devices
that communicate as if they are on the same physical LAN. Any set of
ports (including all ports on the switch) is considered a VLAN. LAN
segments are not restricted by the hardware that physically connects
them. The segments are defined by flexible user groups you create
with the CLI.
VLSM Variable-length subnet masks. In OSPF, VLSMs provide subnets of

different sizes within a single IP block.
vMAN Virtual MAN. In ExtremeWare XOS software, vMANs are a bi-

directional virtual data connection that creates a private path through
the public network. One vMAN is completely isolated from other
vMANs; the encapsulation allows the vMAN traffic to be switched
over Layer 2 infrastructure. You implement vMAN using an
additional 892.1Q tag and a configurable EtherType; this feature is
also known as Q-in-Q switching.
VPN Virtual private network. A VPN is a private network that uses the
public network (Internet) to connect remote sites and users. The VPN
uses virtual connections routed through the Internet from a private
network to remote sites or users. There are different kinds of VPNs,
which all serve this purpose. VPNs also enhance security.
VoIP Voice over Internet Protocol is an Internet telephony technique. With

VoIP, a voice transmission is cut into multiple packets, takes the most
efficient path along the Internet, and is reassembled when it reaches
the destination.
VR-Control This virtual router is part of the embedded system in Extreme

Networks BlackDiamond 10K switches. The VR-Control is used for
internal communications between all the modules and subsystems in
the switch. It has no ports, and you cannot assign any ports to it. It
also cannot be associated with VLANs or routing protocols. (Referred
to as VR-1 in earlier ExtremeWare XOS software versions.)

V (Continued)
VR-Default This virtual router is part of the embedded system in Extreme
Networks BlackDiamond 10K switches. The VR-Default is the default
virtual router on the system. All data ports in the switch are assigned
to this virtual router by default; you can add and delete ports from
this virtual router. Likewise, this virtual router contains the default
VLAN. Although you cannot delete the default VLAN from this
virtual router, you can add and delete any user-created VLANs. One
instance of each routing protocol is spawned for this virtual router,
and they cannot be deleted. (Referred to as VR-2 in earlier
ExtremeWare XOS software versions.)
VRID In VRRP, the VRID identifies the VRRP virtual router. Each VRRP
virtual router is given a unique VRID. All the VRRP routers that
participate in the VRRP virtual router are assigned the same VRID.
VR-Mgmt This virtual router is part of the embedded system in Extreme

Networks BlackDiamond 10K switches. The VR-Mgmt enables remove
management stations to access the switch through Telnet, SSH, or
SNMP sessions; and it owns the management port. The management
port cannot be deleted from this virtual router, and no other ports can
be added.The Mgmt VLAN is created is in this virtual router, and it
cannot be deleted; you cannot add or delete any other VLANs or any
routing protocols to this virtual router. (Referred to as VR-0 in earlier
ExtremeWare XOS software versions.)
VRRP Virtual Router Redundancy Protocol. VRRP specifies an election

protocol that dynamically assigns responsibility for a virtual router to
one of the VRRP routers on a LAN. The VRRP router controlling the
IP address(es) associated with a virtual router is called the master
router, and forwards packets sent to these IP addresses. The election
process provides dynamic failover in the forwarding responsibility
should the master router become unavailable. In case the master
router fails, the virtual IP address is mapped to a backup router's IP
address; this backup becomes the master router. This allows any of the
virtual router IP addresses on the LAN to be used as the default first-
hop router by end-hosts. The advantage gained from using VRRP is a
higher availability default path without requiring configuration of
dynamic routing or router discovery protocols on every host. VRRP is
defined in RFC 2338.
VRRP router Any router that is running VRRP. A VRRP router can participate in
one or more virtual routers with VRRP; a VRRP router can be a
backup router for one or more master routers.
XENPAK Pluggable optics that contain a 10 Gigabit Ethernet module. The

XENPAKs conform to the IEEE 802.3ae standard.

Glossary

Index
Symbols accounting server

RADIUS, 549
, 205 TACACS+, 558
! prompt, 56, 252, 255 accounts
# prompt, 55, 56 creating, 57
* prompt, 56 default, 57
.cfg file, 1014 deleting, 58
.gz file, 1049 failsafe, 58
.pol file, 375 viewing, 58
.xmod file, 1002 ACL match conditions, 386, 392
.xos file, 1002 ACL-based traffic, QoS, 446
> prompt, 55, 56 ACLs
.pol file, 375
Numerics action modifiers, 384
actions, 384
10 gigabit ports, 132 counters, 403
802.1ad, 313, 343 description, 380
802.1ag, 203 editing, 376
802.1ah, 322 evaluation precedence, 1054
802.1D, 689 examples, 403–405
802.1D-2004, 689 metering, 457
802.1Q refreshing, 377
encapsulation, TLS, 824 rule entry, 381
802.1Q tagging, 303 rule syntax, 381
802.1Q-2003, 721 rules, 400
802.1s, 721 smart refresh, 377
802.1w, 709 transferring to the switch, 376
802.1x troubleshooting, 375, 457, 1054
and NAP, 502 acquired node, 593
802.1x authentication action modifiers
advantages, 483 ACL, 384
co-existence with web-based, 482 vMAN ACLs, 393
configuration, example, 497 action statements, policy, 431
disadvantages, 483 actions
interoperability requirements, 496 ACL, 384
methods, 496 vMAN ACLs, 392
requirements, 482 active interfaces, 965
VLAN movement, post-authentication, 501 active node, 592
802.3af, 214 active topology, 592
Address Resolution Protocol. See ARP
A address-based load sharing, 141
address-based load-sharing, 141, 142
access levels, 54
admin account, 57
accessing the switch, 53
Adspec, 839
account types
advertisement interval, EDP, 158
admin, 55
advertising labels, 812
user, 54
agent, local, 289
agent, RMON, 293

Index
aging entries, FDB, 358 Gigabit ports, 132

alarm actions, 296 off, 132
Alarms, RMON, 295 on, 131
area 0 possible settings, 132
OSPF, 929 support, 131
OSPFv3, 940 autonomous system
areas description, 950
OSPF, 928 expressions, 429
OSPFv3, 940 autopolarity, 133
ARP
and IP multinetting, 875 B
communicating with devices outside subnet,
871 backbone area, OSPF, 929
configuring proxy ARP, 870 backbone area, OSPFv3, 940
displaying system table, 872 backbone bridge. See MAC-in-MAC
gratuitous ARP protection, 541 backplane diagnostics
incapable device, 870 configuring, 247
proxy ARP between subnets, 871 disabling, 248
proxy ARP, description of, 870 enabling, 247
responding to ARP requests, 870 backup node role, 593
ARP learning bandwidth restriction, 460
adding permanent entries, 540 banner
configuring, 539 string, 51
DHCP secured ARP, 540 warning, 55
displaying information, 541 base URL, network login, 506
overview, 539 BGP, 797
ARP validation and IP multinetting, 876
configuring, 543 attributes, 950
displaying information, 544 autonomous system, 950
AS numbers, private, 959 autonomous system path, 950
ASCII-formatted configuration file cluster, 951
downloading, 1017 community, 950
loading, 1017 description, 950
support, 117 examples
troubleshooting, 1016 route confederations, 953–956
uploading, 1017 route reflector, 952–953
verifying, 1017 features, 951
authentication loopback interface, 956
local database, 492 peer groups
RADIUS, 488, 547 creating, 957
TACACS+, 555 deleting, 957
authentication methods description, 956
802.1x, 496 mandatory parameters, 956
MAC-based, 513 neighbors, 957
web-based, 506 private AS numbers, 959
AuthnoPriv, 98 redistributing to OSPF, 960
AuthPriv, 98 route aggregation
autobind ports, 699 description, 956
automatic failover, 161 using, 956
automatic restart, ELSM, 266 route confederations, 953
autonegotiation route flap dampening
description, 130 configuring, 958
displaying setting, 163 description, 957
flow control, 131 viewing, 959

Index
route reflectors, 951 CFM

route selection, 959 and MAC-in-MAC, 205
static networks, 960 and vMANs, 205
bi-directional rate shaping CCM messages, 205
configuring, 464 CFM messages, 205
description, 461 configuring CCMs, 211
maximum bandwidth settings, 463 configuring domains, 207, 208
maximum committed rate, 463 configuring MAs, 207, 209
maximum ingress queues, 462 configuring MEPs, 210
minimum bandwidth settings, 463 configuring MIPs, 210
binding labels, description of, 795 configuring ping, 211
BlackDiamond 10808 switch, 35 configuring traceroute, 211
BlackDiamond 12804 switch, 35 displaying, 211
BlackDiamond 8800 original modules, 35 domain format, 208
BlackDiamond 8800 series switch, 36, 127 domains, 204
BlackDiamond 8806 example, 212
MSM slots, 127 implementation, 203
BlackDiamond 8810 limits, 207
MSM slot, 126 MA formats, 209
blackhole entries, FDB, 359, 526 MA levels, 204
Bootloader MAC addresses, 203
accessing, 1022 mapping, 204, 205
exiting, 1023 MEPs, 205
prompt, 1022 MIPs, 206
BOOTP relay MPs, 205
configuring, 879 number of ports, 205
viewing, 881 ping, 206
BOOTP server, 70 traceroute, 206
BOOTP, using, 70 troubleshooting, 203
BootROM TTL, 206
displaying, 1027 verifying, 211
prompt, 1022 VLAN association with MAs, 205
upgrading CFM Ethernet types, 203
BlackDiamond 10808, 1023 checkpointing
Summit X450 family, 1023 bulk, 80
Bootstrap Protocol. See BOOTP dynamic, 81
bootstrap, accessing statistics, displaying, 81
BlackDiamond 12804, 1027 CIST
Summit X450 family, 1024 See also MSTP
Border Gateway Protocol. See BGP BPDUs
bulk checkpointing, 80 CIST records, 724
M-records, 724
C configuring, 724
definition, 724
cabling enabling, 725
10/10/1000BASE-T ports, 133 regional root bridge, 724
crossover cables, 133 root bridge, 724
cabling for redundancy, 162 root port, 725
calculated, 797 CLEAR-Flow
calculated LSP, 796 configuring, 573
campus mode, 484 enabling and disabling, 574
candidate node, 592 overview, 573
carrier vlan, STP, 694 rule types, 577

Index
CLI deleting, 115

! prompt, 56, 252, 255 description, 1014
# prompt, 55, 56 displaying, 111
* prompt, 56 downloading, 1019
> prompt, 55, 56 managing, 108
access levels, 54 overview, 116
command shortcuts, 47 relaying from primary to backup, 80
configuration access, 55 renaming, 109
history, 50 saving changes, 1015
limits, 48 selecting, 1015
line-editing keys, 50 uploading, 1018
named components, 47 using, 1015
prompt line, 56 configuration mode, XML, 117
starting up, 59 configuration, change log, 286
symbols, 48 configuring
syntax, 45 label advertisement filters, 815
syntax helper, 46 LDP session timers, 815
syntax symbols (table), 48 PHP, 801
users resetting parameters, 802
adding, 57 VPLS domain, 829
deleting, 58 configuring PoE, 220
viewing, 58 configuring traceroute, 211
using, 45–46 connectivity, 62
cluster, 951 Connectivity Fault Management. See CFM
collector, remote, 289 conservative label retention mode, 813
combination ports, 161 console
combo ports, 161 connection, 66
command maximum sessions, 66
history, 50 control path, 592
prompts, 55, 56 control VLAN, EAPS, 666
shortcuts, 47 controlling Telnet access, 72
command line interface. See CLI conventions, guide
command syntax, understanding, 45 notice icons, 30
Common and Internal Spanning Tree. See CIST text, 31
common commands (table), 51–52 core dump file
communicating with devices outside subnet, 871 .gz file, 1049
community strings copying, 1052
private, 94 copying to the switch, 1049
public, 94 copying to the tftp server, 1050
read, 94 deleting, 1054
read-write, 94 description, 1048
compatibility version number, 1006 displaying, 1051
components, EMS, 277 renaming, 1051
conditions, EMS, 278 sending to the switch, 1049
configuration core image. See image
primary and secondary, 1015 Core license, 38
returning to factory default, 1016 CPU monitoring
viewing current, 1015 description, 121
configuration command prompt, 55, 56 disabling, 121
configuration domain, virtual routers, 369 enabling, 121
configuration file overview, 108
.cfg file, 1014 troubleshooting, 121
ASCII-formatted, 117 cpu utilization, history, 122
copying, 110 CPU utilization, TOP command, 1055

Index
D MSM, 236
running, 237
daisy chain topology, 601 BlackDiamond 12804 switch
data port, 593 I/O module, 237
database applications, and QoS, 439 MSM, 237
database overflow, OSPF, 926 running, 238
debug information, 1048 BlackDiamond 8800 series switch
debug mode, 285, 1048 I/O module, 236
See also EMS MSM, 236
default running, 237
accounts, 57 displaying, 232, 245
gateway, 778, 865 LEDs, 239
passwords, 59 slot, 235
port status, 129 Summit family of switches, 238
returning to factory settings, 1016 system, 235
users, 57 DiffServ
denial of service protection See also QoS
configuring, 546 and virtual routers, 453
description, 544 code point, 451
disabling, 545 configuring, 450
displaying settings, 546 examining, 451
enabling, 545 disabling route advertising
destination VLAN, network login, 493 RIP, 913
DHCP RIPng, 921
disabling, 533 displaying MPLS information, 803
displaying settings, 534 distance-vector protocol, description, 911, 919
enabling, 533 DNS
network login and, 482 configuring, 62
requirement for web-based network login, 482 description, 61
DHCP bindings database, 535 documentation, using, 31
DHCP relay Domain Name Service. See DNS
and IP multinetting, 877 domains, CFM, 204
configuring, 879 domains, EAPS, 656
viewing, 881 domains, ESRP, 749
DHCP secured ARP, 540 domains, STP, 693
DHCP server downloading
and IP multinetting, 877 ASCII-formatted configuration, 1017
configuring, 533 configuration, 1019
description, 533 downstream unsolicited (DU), definition of, 790
DHCP snooping downstream unsolicited mode, 812
configuring, 536 downstream-on-demand mode, 812
disabling, 536 dual master situation, 645
displaying information, 537 duplex setting, ports, 131
overview, 535 duplex, displaying setting, 163
DHCP trusted ports dynamic checkpointing, 81
configuring, 537 dynamic entries, FDB, 358, 526
overview, 537 Dynamic Host Configuration Protocol. See DHCP
DHCP trusted server Dynamic MVR, 975
configuring, 536 dynamic routes, 867
displaying information, 537 dynamic routes, IPv6, 899
overview, 535 dynamic VLANs. See netlogin
diagnostics
BlackDiamond 10808 switch
I/O module, 236

Index
E 680
spatial reuse, 660
EAPOL and DHCP, 482 status information, displaying, 670, 679
EAPS switch mode, defining, 664
and IP multinetting, 877 transit node, 654, 664
and MVR, 978 troubleshooting, 657, 666
common link, 675 Easy-Setup, 594
configuring, 662 edge safeguard
control VLAN, 666 description, 712
description, 654, 655 disabling, 712
disabling enabling, 712
domain, 668 EDP
loop protection, 670 advertisement interval, 158
on a switch, 668 clearing counters, 158
EAPS domain default, 157
creating and deleting, 663 description, 157
enabling, 662 disabling, 157
domain, 668 enabling, 158
loop protection, 670 timeout interval, 158
on a switch, 668 viewing information, 158, 163
failed state, 657, 665 egress flooding
failtime expiry action, 658, 665 displaying, 365
failtimer, 658, 665 guidelines, 364
Fast Convergence, 655, 668 egress traffic rate limiting, 459
FDB, 658 election algorithms, ESRP, 754
hardware layer, 658 ELRP
health-check packet, 658, 665 and ESRP, 768
hellotime, 665 behavior
hitless failover support, 655 ESRP master switch, 768
licensing, 653 ESRP pre-master switch, 768
link down message, 657 description, 767
loop protection messages, 670 loop detection, 1041
master node, 654, 664 standalone, 1041
multiple domains per switch, 659 with ESRP, overview, 744
names, 47 without ESRP, 1041
polling, 658 ELSM
polling timers, configuring, 665 and Layer 2 protocols, 270
primary port, 654, 666 automatic restart, 266
process, 657 clearing counters, 269
protected VLAN, 667 configuration example, 270
ring port, unconfiguring, 669 configuring
ring restoration, 658 hello timer, 266
rings and a common link, 661 hold threshold, 266
secondary port, 654, 656, 666 description, 261
shared port disabling, 267
common link failure, 676 displaying information, 268
configuration rules, 684 ELSM link state, 263
configuring the domain ID, 678 enabling, 265
creating and deleting, 678 fault detection, 261
defining the mode, 678 hello messages, 261
description, 675 hello transmit states, 261
show eaps display fields (table), 671, 680 hold threshold, 266
show eaps shared-port display fields (table), link state, 262

Index
port states entries, FDB, 357

down, 262 EPICenter support, 67
down-stuck, 262 EPS-600 PSU, 214
down-wait, 262 error, unsupported module, 1059
up, 262 ESRP
sticky threshold, 267 802.1Q tag, 749
timers and ELRP, 744
down, 264 and IP multinetting, 775, 877
hello, 264 and load sharing, 763
hellorx, 265 and OSPF, 753
up, 265 and STP, 775
EMISTP and VRRP, 775, 782, 788
description, 696 auto toggle, 744, 749
example, 706 basic topology, 746
rules, 707 description, 744
EMS direct link, 750
and dual MSM systems, 275 displaying data, 767
configuring targets domain ID, 749
components, 277 domains, description, 749
conditions, 278 don’t count, 763
description, 275 election algorithms, 754
severity, 277 environment tracking, 759
subcomponents, 277 ESRP-aware, 746
debug mode, 285 ESRP-aware portlist, 765
description, 273 examples, 770–774
displaying messages extended mode
console, 283 description, 744, 748
session, 283 differences between standard mode, 748
event message formats, 283 failover time, 753
expressions groups, 764
matching, 280 hitless failover support, 750
regular, 280 host attach, 762
filtering event messages, 275 licensing, 743
filters linking switches, 750
configuring, 279 load sharing and, 763
creating, 279 master
viewing, 279 behavior, 752
log target definition, 745
default, 274 determining, 751
disabling, 275 electing, 753
enabling, 274 election algorithms, 754
types, 274 multiple VLANs sharing a host port, 750
logs neutral state, behavior, 752
displaying, 284 ping tracking, 760
displaying counters, 284 port restart, 762
uploading, 284 port weight, 748
parameters pre-master
behavior, 282 behavior, 752
matching, 281 timeout, 753
severity levels, 276 reasons to use, 745
viewing components and subcomponents, 277 restarting ports, 762
viewing conditions, 278 route table tracking, 760
encapsulation modes, 696 selective forwarding, 765
See also STP

Index
slave mode dynamic, 358

behavior, 752 limiting, 362
definition, 745 multicast with multiport entries, 365
standard mode non-aging, 359
description, 744, 748 permanent, 359
differences between extended mode, 748 prioritizing, 362
tracking static, 359
description, 759 error message, 359
example, 761 prioritizing entries, 525
troubleshooting, 745, 747, 1040 troubleshooting, 359
VLANid, 749 feature pack, 41
ESRP-aware, description, 746 enabling, 42
Ether type, 330 MPLS, 41
Ethernet Automatic Protection Switching. See feature packs
EAPS displaying, 43
evaluation precedence, ACLs, 1054 features, platform-specific, 30
Event Management System. See EMS FEC
Events, RMON, 295 binding labels, 795
EXP field, 793 definition of, 789, 790
explicit packet marking, QoS, 446 propagating labels, 812
explicit route, 845 file server applications, and QoS, 439
extended mode, ESRP domain, 744, 748 file syntax, policy, 427
Extreme Discovery Protocol. See EDP file system administration, 107
Extreme Link Status Monitoring. See ELSM filename requirements, 108, 1050
Extreme Loop Recovery Protocol. See ELRP filenames, troubleshooting, 108, 1050
Extreme Multiple Instance Spanning Tree files
Protocol. See EMISTP copying, 110, 1052
Extreme Standby Router Protocol. See ESRP deleting, 115, 1054
displaying, 111, 1051
F renaming, 109, 1051
filter profiles and filters, SNMPv3, 101
failover, 79, 593 filters
failsafe account, 58 label advertisement, 815
configuring on a stack, 619 filters, protocol, 305
family, 238 firmware
fan tray information, 271 displaying, 1027
Fast Convergence, EAPS, 655 upgrading
fault protection, 654 BlackDiamond 12804, 1026
FDB BlackDiamond 8800 series, 1025
configuring aging time, 360 fixed filter reservation style, 841
contents, 357 flooding, 363
creating a permanent entry example, 360 flooding, displaying, 163
description, 357 flow control
disabling MAC learning, 362 displaying setting, 163
displaying, 361 Gigabit Ethernet ports, 131
dynamic entries Forwarding Database. See FDB
limiting, 526 Forwarding Equivalence Class. See FEC
lock down, 528 forwarding rules, MVR, 976
egress flooding, 363
entries G
adding, 358
aging, 358 graceful OSPF restart, 927
blackhole, 359 gratuitous ARP
description, 357 description, 541

Index
enabling, 543 tasks

Greenwich Mean Time Offsets (table), 104 detailed, 1009
groups summary, 1009
ESRP, 764 understanding, 1006
SNMPv3, 97 hold threshold, ELSM, 266
guest VLAN host attach, ESRP, 762
creating, 500 HQoS
description, 498 bandwidth mode, 466
disabling, 501 bandwidth mode example, 480
enabling, 501 configuring egress queues, 474
guidelines, 500 configuring ingress queues, 473
scenarios, 499 CoS value pairs, 466
settings, 501 CoS values, 469
troubleshooting, 500 description, 465
unconfiguring, 501 displaying, 476
displaying queue CoS values, 477
H egress, 465
guidelines, 472
hardware recovery ingress, 465
clearing the shutdown state, 253 MEF-14, 465
configuring, 251 meters, 457, 468
description, 251 modes, 466
displaying, 252 modifying rates, 470
hardware support, 35 modifying traffic queues, 475
hardware table ports, 468
modifying utilization, 1061 statistics, 471, 476, 478
sample error messages, 1060 strict priority
troubleshooting, 1060 configuring CoS values, 469
viewing settings, 1061 strict priority example, 478
helper-mode, 928 strict priority mode, 465, 466
Hierarchical QoS. sub-queues, 466
See HQoS traffic mode, 476
History, RMON, 294 traffic queues, 467, 476
hitless failover, 593 troubleshooting, 42, 472
caveats utilization, 476
BlackDiamond 10808, 87 vMANs, 465
BlackDiamond 12804, 87 displaying flood control, 477
BlackDiamond 8800 series, 87 flood control, 471
modular switches, all, 86 HTTP
description, 82 disabling, 560
EAPS, 655 enabling, 560
ESRP, 750 overview, 560
I/O version number, 1006 Hypertext Transfer Protocol. See HTTP
network login, 485
platform support, 85 I
PoE, 213
protocol support, 82 I/O module
STP, 702 MSM, same module on BlackDiamond 8800
VRRP, 778 MSM, 126
hitless upgrade, 593 power management, 87
caveats, BlackDiamond 8800 only, 1008 I/O ports
performing, 1007 BlackDiamond 8806 switch, 127
software support, 1007 BlackDiamond 8810 switch, 126

Index
same module on BlackDiamond 8800 MSM, description, 963

126 example, 970
I/O version number, 1006 IGMP
IEEE 802.1ad, 143, 313, 343 description, 967
IEEE 802.1ag, 203 snooping, 967, 985
IEEE 802.1ah, 322 snooping filters, 968
IEEE 802.1D, 689 PIM mode interoperation, 966
IEEE 802.1D-2004, 689 PIM multicast border router (PMBR), 966
IEEE 802.1Q, 303 PIM-DM, 965
IEEE 802.1Q-2003, 721 PIM-SM, 965
IEEE 802.1s, 721 IP multinetting
IEEE 802.1w, 709 and ESRP, 775
IEEE 802.1x, 496 configuring, 878
IEEE 802.3af, 214 description, 873
IGMP example, 879
and IP multinetting, 876 interface, 874
description, 967 interoperability with
snooping, 967, 985 ARP, 875
snooping filters, 968 BGP, 876
static, 968 DHCP relay, 877
image DHCP server, 877
.xos file, 1002 EAPS, 877
definition, 997 ESRP, 877
downloading, 1001 IGMP, IGMP snooping, 876
EPICenter, using, 1002 IRDP, 875
expired service contract, 1000 OSPF, 876
primary and secondary, 999 PIM, 877
rescue, 1044 RIP, 876
selecting a partition, 999 STP, 877
upgrading, 999 VRRP, 877
version string, 998 recommendations, 873
implicit NULL labels, 795 topology, 874
independent LSP control, 813 IP parameters, configuring, 70
ingress rate shaping. See bi-directional rate IP security
shaping or QoS ARP learning, 539
ingress rates, 461 ARP validation, 543
inheriting ports, MSTP, 700 dependencies, 535
Input/Output module. See I/O module DHCP bindings database, 535
interfaces DHCP snooping, 535
active, 965 gratuitous ARP, 541
IP multinetting, 874 source IP lockdown, 538
IPv6 router, 894 trusted DHCP server, 535
passive, 965 IP unicast routing
router, 866 BOOTP relay, 879
Internet Group Management Protocol. See IGMP configuration examples, 872
Internet Router Discovery Protocol. See IRDP configuring, 871
interoperability requirements, 802.1x default gateway, 865
authentication, 496 DHCP relay, 879
IP address enabling, 871
entering, 71 multinetting
VLANs, 309 description, 873
IP fragmentation, 137 example, 879
IP multicast routing proxy ARP, 870
configuring, 969 relative priorities, 869

Index
router interfaces, 866 label stack

routing table definition of, 790
dynamic routes, 867 encapsulation, 794
multiple routes, 868 Label Switch Path. See LSP
populating, 867 Label Switch Router. See LSR
static routes, 868, 899 labels
verifying the configuration, 871 advertising, 812
IPv6 advertising modes, 812
displaying VLANs, 312 binding, 795
link aggregation, 141 configuring label advertisement filters, 815
ping, 63 definition of, 789, 790
VLANs, 299, 310 locally assigned, 795
IPv6 addresses, scoped, 897 propagating, 812
IPv6 multicast routing, MLD, 985 remotely assigned, 795
IPv6 unicast routing retention modes, 813
configuration examples, 903 space partitioning, 795
configuring, 901 swapping, defined, 790
enabling, 901 LACP.
relative priorities, 900 See link aggregation
router interfaces, 894 LAG.
routing table See link aggregation
dynamic routes, 899 latestReceivedEngineTime, 96
routing table IPv6 Layer 1, troubleshooting, 1030
multiple routes, 900 Layer 2 protocols and ELSM, 270
populating, 899 Layer 2, troubleshooting, 1030
verifying the configuration, 902 Layer 3, troubleshooting, 1031
IRDP, and IP multinetting, 875 LDP
ISP mode, 484 definition of, 790, 811
hello-adjacency, 811
J message exchange, 812
neighbor discovery protocol, 811
jumbo frames session timers, configuring, 815
BlackDiamond 8800 and Summit X450, 136 LEDs
configuring MPLS modules, 795 SummitStack, 619
description, 135 LEDs, during diagnostics, 239
enabling, 136 legacy powered devices.
IP fragmentation, 137 See PoE
path MTU discovery, 137 LER
troubleshooting, 136 definition of, 790
viewing port settings, 163 described, 792
vMANs, 135 LFS
description, 132
K troubleshooting, 132
liberal label retention mode, 813
keys
license voucher, 40
line-editing, 50
licensing
port monitoring, 234
displaying, 43
EAPS, 653
L enabling, 40
Label Edge Router. See LER ESRP, 743
label object, 843 license voucher, 40
label retention modes, 813 more than one switch, 40
label space partitioning (table), 795 ordering, 40
security license, 41

Index
sFlow, 287 default TLVs, 187

SSH2, 41 EMS messages, 186
SummitStack, 604 enabling, 195
trial, 38 EtherType, 183
upgrades, 38 IP address advertisement, 186
verifying, 40 length limit, 183
VRRP, 777 LLDPDU, 183
licensing, level restriction, 606 LLDP-MED fast start, 182
licensing, upgrades, 606 LLDP-MED TLVs, 182
limit, sFlow maximum CPU sample limit, 292 LLDP-MED traps, 183
limiting entries, FDB, 362 mandatory TLVs, 189
line-editing keys, 50 MED information, 202
link aggregation messages received, 188
See also load sharing messages sent, 187
adding or deleting ports, 148 multicast address, 183
and control protocols, 140 neighbor information, 202
and IPv6, 141 overview, 181
and software-controlled redundant ports, 140 port configuration information, 201
and vMANs, 140 receive only TLVs, 185
broadcast, multicast, and unknown packets, received TLVs, 188
145 repeated TLVs, 187
description, 139 restoring defaults, 201
displaying, 151 SNMP traps, 186
dynamic, 140 standards-mandated TLVs, 187
example, 150 statistics, 201
LACP system description TLV, 191
active and standby ports, 143 timers, 185
algorithms, 149 transmitted TLVs, 187
and ELSM, 148 troubleshooting, 184, 194, 200
configuring, 148 unconfiguring, 195, 201
defaulted port action, 145 load sharing
displaying, 151 See also link aggregation
LAG, 148 address-based, 141
master port, 148 algorithms, 141, 142
maximum links, 149 and ESRP don’t count, 763
verifying configuration, 149 and ESRP host attach, 763
maximum ports and groups, 145 and VLANs, 141, 150
restrictions, 147 BlackDiamond 10808 switch, 147
static, 140 BlackDiamond 12804 switch, 147
troubleshooting, 139, 140, 148 BlackDiamond 8800 series switch, 146
Link Fault Signal. configuring, 148
See LFS displaying, 163
Link Layer Discovery Protocol. See LLDP master port, 148
link types maximum ports and groups, 145
configuring in MSTP, 711 port-based, 142
configuring in RSTP, 711 Summit X450 switch, 145
link-state advertisement. See LSA troubleshooting, 150
link-state database. See LSDB local agent, 289
link-state protocol, description, 911, 919 local database authentication
LLDP description, 492
and 802.1x, 186 password, 492
Avaya-Extreme information, 202 user name, 492
avaya-extreme TLVs, 182 local netlogin account
configuring, 195 creating, 492

Index
deleting, 495 MAC lockdown

destination VLAN configuring, 528
creating, 493 displaying entries, 529
modifying, 495 unconfiguring, 529
displaying, 495 MAC lockdown timer
modifying, 495 configuring, 532
local routing database, 837 disabling, 533
locally assigned labels, 795 displaying entries, 533
lockdown timer, MAC, 530 displaying the configuration, 533
locked entries, 528 enabling, 532
log target, EMS examples
disabling, 275 active device, 530
enabling, 274 disconnecting devices, 531
logging configuration changes, 286 inactive device, 530
logging in, 59, 603 port movement, 532
logging messages. See EMS reconnecting devices, 531
logout privilege, network login, 507 overview, 529
loop detection understanding, 530
using ELRP and ESRP, 768 MAC-based
using standalone ELRP, 1041 security, 362, 525
loop tests VLANs, network login, 517
using ELRP and ESRP, 768 MAC-based authentication
using standalone ELRP, 1041 advantages, 483
loopback interface, 956 configuration, example, 516
LPS, tunnel, definition of, 791 configuration, secure MAC, 515
LSA type numbers (table) description, 513
OSPF, 926 disabling, 514
OSPFv3, 940 disadvantages, 483
LSA, description, 926, 939 enabling, 514
LSDB, description, 926, 939 MAC-in-MAC
LSP and ACLs, 324
and PW, 824 and vMAN frames, 322
calculated, 796 BVLANs, 323
control modes, 813 description, 322
definition of, 790 Ethertype, 324
matching next hop, 796 SVLANs, 323, 324
next hops (figure), 796 tag description, 323
routing, 796 troubleshooting, 324
scaling, 851 management access, 54
LSR, 799 management accounts, displaying, 61
definition of, 790 Management Information Base. See MIBs
egress, definition of, 792 management port, 67
functions (table), 793 Management Switch Fabric Module. See MSM
ingress, definition of, 792 manually bind ports, 698
LER, description of, 792 master node role, 593
locally assigned labels, 795 master port, load sharing, 148
remotely assigned labels, 795 match conditions, ACL, 386, 392
types (figure), 792 match conditions, policy, 429
LW XENPAK, 134 matching, 797
matching expressions, EMS, 280
M matching LSP next hop, 796
matching parameters, EMS, 281
MAC learning, FDB, 362 maximum CPU sample limit, sFlow, 292
maximum transmission unit (MTU), 847

Index
memory protection, 108, 120 platforms supported, 988

metering, 457, 461 policy filter, 990
mgmt VLAN, 67 redundancy, 992
MIBs, supported, 93, 1071 SA cache, 991
mixed media modules, 127 SA cache entry limit, 992
MLD SA request processing, 990
description, 985 scaling limits, 993
static, 986 MSM
modular switch console sessions, 66
jumbo frames, 135 reboot, 1005
load sharing, configuring, 148 MSM module
monitor port, 152 BlackDiamond 8806 switch, 127
port number, 49, 129 BlackDiamond 8810 switch, 126
port-mirroring, 152, 154 MSM prompt, troubleshooting, 1036
slot configuration, 125 MSTI
virtual port, 154 See also MSTP
module configuring, 725
enabling and disabling, 126 enabling, 726
type and number of, 126 identifier, 725
module recovery regional root bridge, 726
actions, 255, 256 root port, 726
clearing the shutdown state, 259 MSTI ID, 725
configuring, 254 MSTP
description, 253 See also RSTP
displaying, 258 advantages of, 721
troubleshooting, 260 boundary ports, 726
monitor port, port-mirroring, 152 common and internal spanning tree, 724
monitoring command prompt, 55, 56 configuring, 728
monitoring the switch, 231 edge safeguard, 712
MPLS enabling, 728
configuration example (figure), 807 hop count, 728
Data Plane, 791 identifiers, 697
definition of, 791 inheriting ports, 700
displaying information, 803 link types
feature pack, 41 auto, 711
introduction, 789 broadcast, 711
label stack (figure), 794 configuring, 711
resetting configuration parameters, 802 description, 711
sample network (figure), 790 edge, 711
shim header, 793 point-to-point, 711
shim header (figure), 793 multiple spanning tree instances, 725
shim layer, 793 operation, 730
terms and acronyms (table), 790 overview, 721
troubleshooting, 41 port roles
unicast frame on Ethernet (figure), 794 alternate, 710
MSDP, 987, 1070 backup, 710
configuration example, 993 designated, 710
default peers, 989 disabled, 710
limitations, 988 master, 727
mesh-groups, 990 root, 710
MIBs, 993 region
peer authentication, 989 configuring, 723
peers, 988 description, 722
PIM border configuration, 988 identifiers, 723

Index
unconfiguring, 723 netlogin. See network login

multicast netlogin-only
FDB static entry, 365 disabled, 488
vMANs, 318 enabled, 488
Multicast Source Discovery Protocol (MSDP) Network Access Protection. See NAP
See MSDP network login
Multicast VLAN Registration. authenticating users, 487
See MVR authentication methods, 482
Multicast, Source Specific, 966 campus mode, 484
multinetting. See IP multinetting configuration examples
multiple routes, 868 802.1x, 497
multiple routes IPv6, 900 MAC-based, 516
Multiple Spanning Tree Instances. See MSTI web-based, 510
Multiple Spanning Tree Protocol. See MSTP description, 481
multiple supplicants, network login support, 483 disabling, 486
MultiProtocol Label Switching. See MPLS disabling, port, 486
MVR enabling, 486
and EAPS, 978 exclusions and limitations, 487
and STP, 980 guest VLAN, 498
dynamic, 975 hitless failover support, 485
forwarding rules, 976 ISP mode, 484
in a vMAN environment, 982 local account
static, 975 creating, 492
deleting, 495
N displaying, 495
modifying, 495
names local account, destination VLAN
character types, 47 creating, 493
conventions, 47 modifying, 495
maximum length of, 47 local database authentication, 492
switch, 56 logout privilege, 507
VLAN, 307 MAC-based VLANs, 517
VLAN, STP, EAPS, 47 move fail action, 486
NAP multiple supplicants, 483
and 802.1x, 502 port, enabling, 486
and ACLs, 505 RADIUS attributes, 489
overview, 502 RADIUS authentication, 488
sample scenarios, 503 redirect page, 507
VSA definitions, 504 secure MAC, 514
native VLAN, PVST+, 709 session refresh, 507
neighbor discovery protocol, LDP, 811 settings, displaying, 487
netlogin user
dynamic VLANs netlogin-only disabled, 488
description, 519 netlogin-only enabled, 488
displaying, 521 user accounts, 488
enabling, 520 web-based authentication, user login, 511
example, 521 Next Hop Label Forward Entry (NHLFE), definition
uplink ports, 520 of, 791
port restart noAuthnoPriv, 98
description, 521 node, 592
disabling, 522 node election
displaying, 522 configuring priority, 78
enabling, 522 determining primary, 78
guidelines, 522 overview, 78

Index
node ID, 593 stub area, 929

node role, 593 timers, 934
node role election, 593 virtual link, 930
node role election priority, 593 wait interval, configuring, 935
node states, 81 OSPFv3
node status, viewing, 81 advantages, 920
non-aging entries, FDB, 359 area 0, 940
normal area areas, 940
OSPF, 930 authentication, 945
OSPFv3, 941 backbone area, 940
notification tags, SNMPv3, 101 configuration example, 946–947
notification, SNMPv3, 99 description, 919, 939
Not-So-Stubby-Area. See NSSA enabling, 902
NSSA, 929, 941 link type, 943
See also OSPF LSA, 939
See also OSPFv3 LSDB, 939
normal area, 941
O NSSA, 941
redistributing routes
opaque LSAs, OSPF, 927 configuring, 921, 944
Open Shortest Path First IPv6. See OSPFv3 description, 921, 943
Open Shortest Path First. See OSPF enabling or disabling, 945
operational node, 594 router types, 940
ordered LSP control, 813 stub area, 941
OSPF timers, 945
advantages, 912 virtual link, 941
and ESRP, 753
and IP multinetting, 876 P
area 0, 929
areas, 928 partition, 999
authentication, 934 passive interfaces, 965
backbone area, 929 password security
configuration example, 936–937 configuring, 60
consistency, 926 displaying, 61
database overflow, 926 password, encrypted, 59
description, 911, 925 passwords, 59
display filtering, 938 creating, 59
enabling, 871 default, 59
graceful restart, 927 displaying, 61
link type, 932 failsafe account, 58
LSA, 926 forgetting, 59
LSDB, 926 local database authentication, 492
normal area, 930 security, 59
NSSA, 929 shared secret, RADIUS, 548
opaque LSAs, 927 shared secret, RADIUS accounting, 550
point-to-point links, 932 shared secret, TACACS+, 556
redistributing routes shared secret, TACACS+ accounting, 558
configuring, 914, 934 troubleshooting, 59
description, 914, 933 path error message, 840
enabling or disabling, 934 path message, 839
redistributing to BGP, 960 path MTU discovery, 137
restart, 927 peer groups, 956
router types, 929 Penultimate Hop Popping. See PHP
settings, displaying, 938 Per VLAN Spanning Tree. See PVST+

Index
permanent entries, FDB, 359 usage threshold, 223

permit-established, 404 PoE features, 214
PHP poison reverse, RIP, 913
configuring, 801 poison reverse, RIPng, 921
definition of, 791, 794 policies
implicit NULL labels, 795 action statements, 431
PIM autonomous system expressions, 429
and IP multinetting, 877 examples
mode interoperation, 966 translating a route map, 434
multicast border router (PMBR), 966 translating an access profile, 432
Source Specific Multicast, 966 file syntax, 427
PIM-DM rule entry, 428
description, 965 Policy Based Routing, 385
example, 970 policy file
PIM-SM copying, 110
description, 965 deleting, 115
example, 971 displaying, 111
rendezvous point, 965 renaming, 109
PIM-SSM, 966 policy match conditions, 429
ping policy-based QoS. See QoS
troubleshooting, 63 polling interval, sFlow, 291
platform availability, 35 port
platform dependence, 30 autonegotiation, 130
PoE configuring, 128
budgeted power, 216, 221 duplex setting, 131
capacitance measurement, 224 enabling and disabling, 129
configuration display, 225 flow control, 131
configuring, 220 LFS, 132
default power, 221 link aggregation, 139
deny port, 222 load sharing, 139
denying power, 217 management, 67
devices, 213 monitoring display keys, 234
disconnect precedence, 217, 222 network login, 481
EMS message, 218 numbers and ranges, 48, 128
enabling and disabling power, 220 receive errors, 233
features, 214 SNMP trap, 129
hitless failover support, 213 software-controlled redundant
legacy powered devices, 224 configuring, 160
operator limit, 224 description, 158
port fault state, 218 speed
port labels, 224 configuring, 131
port power limits, 219 displaying, 163
port priority, 217, 222 supported types of, 130
power availability, 214 transmit errors, 232
power budget, 216, 225 utilization, 163
power checking, 215 viewing
powering PoE modules, 215 configuration, 163
required power, 215 information, 163
reserving power, 221 receive errors, 233
resetting ports, 225 statistics, 232
SNMP events, 223 transmit errors, 232
statistics, 225 wildcard combinations, 49, 129
troubleshooting, 215, 216, 222 port lists, 48, 128
upper port power limit, 224 port mode, 733

Index
port priority, STP, 733 probeResetControl, 296

port restart, ESRP, 762 probeSoftwareRev, 295
port restart, netlogin, 521 process
port states, ELSM, 262 control, 108
port weight, ESRP, 748 displaying information, 117
port-based load sharing, 142 error reporting, 1048
port-based load-sharing, 141 management, 117
port-based VLANs, 301–302 restarting, 119
port-mirroring starting, 119
and ELSM, 155 stopping, 118
and load sharing, 153, 154 terminating, 118
and protocol analyzers, 154 profiles, QoS, 441, 443
and sFlow, 153, 154 prompt
description, 152 admin account, 55, 56
displaying, 157 Bootloader, 1022
examples, 156 BootROM, 1022
guidelines, 155 shutdown ports, 56, 252, 255
monitor port, 152 unsaved changes, 56
tagged and untagged frames, 153, 154 user account, 55, 56
traffic filter, 152, 153, 154 propagating labels, 812
troubleshooting, 153, 155 protected VLAN, EAPS, 667
virtual port, 154 protected VLAN, STP, 694
ports, mixed-media modules, 127 protocol analyzers, use with port-mirroring, 154
post-authentication VLAN movement, network protocol filters, 305
login, 501 Protocol Independent Multicast- Dense Mode. See
power checking, PoE modules, 215 PIM-DM
power management Protocol Independent Multicast. See PIM
consumption, 87 Protocol Independent Multicast-Sparse Mode
displaying information, 91 (PIM-SM), 987, 1125
initial system boot-up, 88 Protocol Independent Multicast-Sparse Mode. See
loss of power, 89 PIM-SM
overriding, 90 protocol-based VLANs, 305
re-enabling, 90 proxy ARP
replacement power supply, 89 communicating with devices outside subnet,
Power over Ethernet. 871
See PoE conditions, 870
power supply controller, 87 configuring, 870
powered devices. description, 870
See PoE MAC address in response, 870
primary image, 999 responding to requests, 870
prioritizing entries, FDB, 362 subnets, 871
private AS numbers, 959 public community, SNMP, 94
private community, SNMP, 94 PVST+
privilege levels description, 696, 709
admin, 55 native VLAN, 709
user, 54 VLAN mapping, 709
privileges PW
creating, 57 and LSPs, 824
default, 57
viewing, 58 Q
probe, RMON, 293
probeCapabilities, 295 QoS
probeDateTime, 295 802.1p priority
probeHardwareRev, 295 changing QoS profile mapping, 448

Index
default mapping to QoS profile, 448 priority, 443

overview, 446 qostype priorities, 444
replacing value, 448 queue priority, 443
and ACLs, 440 queues, 438, 440, 443
and duplex, 439 sFlow, 440
and RSVP, 837 Summit X450 specific, 441
applications, 438 traffic groupings
bandwidth utilization, 438 ACL-based, 446
bi-directional rate shaping description, 440, 444
configuring, 464 explicit packet marking, 446
description, 461 source port, 454
maximum bandwidth, 463 VLAN, 454
maximum committed rate, 463 traffic guidelines, 439
minimum bandwidth settings, 463 traffic precedence, 444
BlackDiamond 8800 specific, 441 BD10808 switch, 445
buffer, 441 BD12804 switch, 445
class of service, 446 BlackDiamond 8800 switch, 445
classification priorities, 444 Summit X450 switches, 445
committed rates, 443 troubleshooting, 440, 441, 453, 457, 461
database applications, 439 verifying, 456
default QoS profiles, 442, 444 video applications, 438
description, 437 viewing port settings, 163, 454
DiffServ voice applications, 438
changing mapping to QoS profile, 451 web browsing applications, 439
configuring, 450 weight, 442
default mapping to QoS profile, 451 Quality of Service. See QoS
examining, 451
replacing value, 451 R
viewing mapping to QoS profile, 452
DiffServ model, 793 RADIUS
displaying mapping information, 806 accounting
examples configuring, 549
source port, 454 disabling, 550
VLAN, 454 enabling, 550
EXP bits, 793 password, 550
file server applications, 439 and TACACS+, 68, 547, 555
guidelines, 457 client configuration, 550
ingress hardware queues description, 68, 547
default mapping to priority value, 462 enabling and disabling, 549
description, 462 Merit server configuration (example), 553
ingress QoS profile (IQP), 462 password, 548
maximum bandwidth, 443 per-command authentication, 550
metering, 457, 461 per-command configuration (example), 554
minimum bandwidth, 443 RFC 2138 attributes, 551
monitoring real-time performance, 455 server configuration, 548
OP2, 440 servers, 547
peak rates, 443 TCP port, 550
precedence of traffic, 444 user accounts, creating, 488
priority, 442, 443 RADIUS accounts, network login, 488
profiles RADIUS attributes, network login, 489
default, 442, 444 rapid root failover, 700
description, 440 Rapid Spanning Tree Protocol. See RSTP
naming, 440 rate limiting
parameters, 441, 443 egress traffic, 459, 460

Index
rate shaping, bi-directional. See bi-directional RIP

rate shaping advantages, 912
read-only switch access, 94 and IP multinetting, 876
read-write switch access, 94 configuration example, 915–917
reboot description, 911, 912
MSM, 1005 disabling route advertising, 913
switch, 1005 enabling, 871
receive errors, port, 233 limitations, 912
receiver-mode, 961 poison reverse, 913
redirect page, network login, 507 redistributing routes
redirection, URL. See URL redirection configuring, 914, 934
redundant ports, software-controlled description, 914, 933
configuring, 160 enabling or disabling, 915
description, 158 redistributing to BGP, 960
refresh, ACLs, 377 routing table entries, 912
regions, MSTP, 722 split horizon, 913
related publications, 31 triggered updates, 913
relative route priorities version 2, 913
IPv4, 869 RIPng
IPv6, 900 advantages, 920
Remote Authentication Dial In User Service. See configuration example, 922–923
RADIUS description, 919
remote collector, 289 disabling route advertising, 921
Remote Monitoring. See RMON enabling, 902
remotely assigned label, 795 limitations, 920
renaming a VLAN, 308 poison reverse, 921
rendezvous point, 965 redistributing routes
Rendezvous Points (RPs), 1125 configuring, 921, 944
rescue image, 1044 description, 921, 943
reservation attributes and styles (table), 841 enabling or disabling, 922
reservation error message, 840 routing table entries, 920
reservation message, 839 split horizon, 921
reservation requests, 837 triggered updates, 921
reservation styles, 841 RMON
resilience, 654 agent, 293
Resource ReserVation Protocol. See RSVP alarm actions, 296
responding to ARP requests, 870 Alarms group, 295
restart process, 119 configuring, 296
restart, graceful, 927 description, 293
returning to factory defaults, 1016 Events group, 295
RFC 3618, 1070 features supported, 294
RFCs History group, 294
BGP, 949 management workstation, 293
bridge, 733 output, 297
IPv4 multicast routing, 963 probe, 293
IPv4 unicast routing, 865 probeCapabilities, 295
IPv6 unicast routing, 893, 896, 897 probeDateTime, 295
listing, 1069 probeHardwareRev, 295
OSPF, 925 probeResetControl, 296
RIP, 911 probeSoftwareRev, 295
RIPng, 919 Statistics group, 294
VRRP, 777 trapDestTable, 296
ring topology, 600 route aggregation, 956
route confederations, 953

Index
route flap dampening, 957 message types, 838

route recording, RSVP, 846 objects, 842
route reflectors, 951 path error message, 840
route selection, 959 path message, 839
router interfaces, 866, 894 record route, 843
router types reservation error message, 840
OSPF, 929 reservation message, 839
OSPFv3, 940 reservation requests, 837
routing reservation styles, 841
See IP unicast routing route recording, 846
See IPv6 unicast routing RSVP-TE, 837
Routing Information Protocol, IPv6. See RIPng RSVP-TE, definition of, 791
Routing Information Protocol. See RIP session attribute, 844
routing protocols and virtual routers, 371 shared explicit reservation style, 841
routing table entries traffic engineering extensions, 837
RIP, 912 tunneling, 842
RIPng, 920 wildcard reservation style, 841
routing table IPv6, populating, 899 RSVP-TE, definition of, 791
routing table, populating, 867 rule entry
RSTP ACL, 381
See also STP policy, 428
and STP, 720 rule syntax, ACL, 381
configuring, 732 rule types, 577
designated port rapid behavior, 715
edge safeguard, 712 S
link types
auto, 711 safe defaults mode, 53
broadcast, 711 safe defaults script, 53
configuring, 711 sampling rate, sFlow, 291
description, 711 saving configuration changes, 1015
edge, 711 scoped IPv6 addresses, 897
point-to-point, 711 SCP2, 567, 1002
operation, 713 ScreenPlay
overview, 709 administration, 354
port roles Dashboard, 346
alternate, 710 launching, 345
backup, 710 overview, 343
designated, 710 ports configuration, 348
disabled, 710 setting up, 343
edge, 710 SNMP configuration, 350
root, 710 statistics and monitoring, 351
rapid reconvergence, 716 user accounts, 354
receiving bridge behavior, 716 user sessions, 355
root port rapid behavior, 714 VLAN configuration, 349
timers, 712 secondary image, 999
topology information, propagating, 716 Secure Copy Program 2. See SCP2
RSVP secure MAC
and QoS, 837 configuration, example, 515
definition of, 791, 837 description, 514
explicit route, 843, 845 Secure Shell 2. See SSH2 protocol
fixed filter reservation style, 841 Secure Socket Layer. See SSL
label, 843 security
label request, 843 and safe defaults mode, 525
LSP scaling, 851 egress flooding, 363

Index
security license, 41 Simple Network Management Protocol. See

security name, SNMPv3, 97 SNMP
service contract Simple Network Time Protocol. See SNTP
displaying, 43 slot
service provide, 791 automatic configuration, 125
session refresh, network login, 507 clearing, 126
sessions diagnostics, 235
console, 66 displaying information, 126
deleting, 75 enabling and disabling, 126
maximum number of, 66 manual configuration, 126
shell, 66 mismatch, 126
SSH2, 76, 563 preconfiguring, 126
Telnet, 69 Smart Redundancy
TFTP, 76 configuring, 160
severity levels, EMS, 276 description, 158
sFlow displaying, 163
BlackDiamond 8800 switches, 440 port recovery, 159
configuration example, 292 smart refresh, ACLs, 377
configuring, 288 SNAP protocol, 307
displaying configuration, 293 SNMP
displaying statistics, 293 and safe defaults mode, 93
enabling community strings, 94
on specific ports, 290 configuring, 93
on the switch, 289 settings, displaying, 94
hardware-based sampling, 288 supported MIBs, 93
licensing, 287 system contact, 94
local agent, 289 system location, 94
maximum CPU sample limit, 292 system name, 94
overview, 286 trap receivers, 93
polling interval, 291 using, 91
remote collector, 289 SNMPEngineBoots, 96
resetting values, 292 snmpEngineID, 96
sampling rate, 291 SNMPEngineTime, 96
software-based sampling, 288 SNMPv3
Summit X450 switches, 440 filter profiles and filters, 101
SFTP, 1002 groups, 97
shared explicit reservation style, 841 MIB access control, 98
shared secret notification, 99
RADIUS, 548 overview, 94
RADIUS accounting, 550 security, 96
TACACS+, 556 security name, 97
TACACS+ accounting, 558 tags, notification, 101
shell target address, 100
configuring, 66 target parameters, 100
maximum number of, 66 user name, 97
overview, 66 SNTP
shim header, 791 configuring, 102
described, 793 Daylight Savings Time, 102
illustration, 793 description, 102
shim layer, 793 example, 105
show eaps counters, 672, 682 Greenwich Mean Time offset, 102
show fans, 271 Greenwich Mean Time Offsets (table), 104
showing MPLS information, 803 NTP servers, 102
software functionality, 38

Index
software image. See image displaying information, 572

software licensing, 38 enabling, 570
software module private key, downloading, 571
.xmod file, 1002 private key, pregenerated, 572
activating, 1003 secure web access, 569
description, 1002 using commands, 569
downloading, 1001 stack, 592
overview, 997 stack segment, 594
uninstalling, 1003 stack state, 594
software signature, 999 stackable switch, 592
software version for platforms, 35 stacking link, 592
software-controlled redundant ports stacking port, 592
and link aggregation, 140 stand-alone switch
description, 158 load sharing example, 150
displaying, 163 port number, 49, 129
displaying configuration, 160 standard mode, ESRP domain, 744, 748
troubleshooting, 158, 160 standards, 1069
typical configurations, 159 standby node role, 593
SONET/SDH connection, 134 start process, 119
source IP lockdown startup screen
clearing information, 539 modules shutdown, 56
configuring, 538 switch, 55
displaying information, 539 static IGMP, 968
overview, 538 static MLD, 986
Source Specific Multicast, 966 static MVR, 975
Source-Active (SA), 987 static networks, and BGP, 960
space partitioning, labels, 795 static routes, 868, 899
spanning tree identifier. See StpdID statistics
Spanning Tree Protocol. See STP CPU utilization, 122
spatial reuse, EAPS, 660 port, 232
speed, displaying setting, 163 Statistics, RMON, 294
speed, ports status monitoring, 231
configuring, 131 sticky threshold, ELSM, 267
displaying, 163 stop process, 118
split horizon, RIP, 913 STP
split horizon, RIPng, 921 advanced example, 706
SSH2 client, 566 and ESRP, 775
SSH2 license, 41 and IP multinetting, 877
SSH2 protocol and MVR, 980
ACL policy, 563 and RSTP, 720
authentication key, 561 and VLANs, 693
default port, 563 and VRRP, 782
description, 75, 560 autobind ports, 699
enabling, 561 basic configuration example, 704
maximum number of sessions, 76, 563 bridge priority, 733
sample ACL policies, 564 carrier vlan, 694
TCP port number, 563 compatibility between 802.1D-1998 and
troubleshooting, 561 802.1D-2004, 690
SSL configurable parameters, 733
certificates, downloading, 571 configuration examples, 734
certificates, generating, 570 configuring, 732
certificates, pregenerated, 572 description, 689
description, 569 displaying settings, 163, 739
disabling, 570

Index
domains FAQs, 650

802.1D, 695 LEDs, 619
802.1w, 695 logging in, 603
creating, 693 managing licenses, 604
deleting, 693 Overview, 591
description, 693 topologies, 599
displaying, 740 troubleshooting, 644
mstp, 695 supplicant side requirements, 496
EMISTP switch
example, 706 basic information, 64
rules, 707 monitoring, 231
encapsulation mode reboot, 1005
802.1D, 696 recovery startup screen, 56
description, 696 startup screen, 55
EMISTP, 696 switch management
PVST+, 696 console, 66
forward delay, 733 overview, 65
guidelines, 732 TFTP, 76–77
hello time, 733 user sessions, 66
hitless failover support, 702 switch name, 56
inheriting ports, 700 switch RMON features, 294
manually bind ports, 698 symbols, command syntax, 48
max age, 733 syntax
max hop count, 733 See also CLI
MSTI ID, 733 abbreviated, 46
names, 47 understanding, 45
path cost, 733 syntax helper, 46
port and multiple STPDs, 693 system contact, SNMP, 94
port mode, 733 system diagnostics, 235
port priority, 733 system health check, 245
port states system health checker
blocking, 697 BlackDiamond 10808 switch
disabled, 698 description, 245
displaying, 740 example, 249
forwarding, 698 modes of operation, 246
learning, 697 BlackDiamond 12804 switch
listening, 697 description, 245
protected VLAN, 694 example, 249
PVST+, description, 709 modes of operation, 246
rapid root failover, 700 BlackDiamond 8800 series switch
rules and restrictions, 732 description, 246
StpdID, 697, 733 example, 249
troubleshooting, 732, 1039 modes of operation, 246
StpdID, 697 configuring backplane diagnostics, 247
strings, community, 94 disabling backplane diagnostics, 248, 1056
stub area, OSPF, 929 displaying, 248
stub area, OSPFv3, 941 enabling backplane diagnostics, 247, 1056
subcomponents, EMS, 277 overview
Subnetwork Access Protocol. See SNAP protocol BlackDiamond 10808 switch, 1055
Summit X450e-24p switch, 213 BlackDiamond 12804 switch, 1055
Summit X450e-48p switch, 214 BlackDiamond 8800 series switch, 1056
SummitStack Summit X450 family, 1056
configuration, 607, 611 Summit X450 family
configuration examples, 626 description, 247, 1056

Index
mode of operation, 247 client, 69

system health, monitoring, 245 configuring virtual router, 73
system LEDs, 1033 connecting to another host, 70
system location, SNMP, 94 controlling access, 72
system name, SNMP, 94 default port, 70
system odometer, 1057 default virtual router, 70
system recovery description, 68
configuring, 250 disabling, 75
description, 250 displaying status, 75
displaying, 251 MSM, 61
software, 250 re-enabling, 75
system redundancy sample ACL policies, 73
bulk checkpointing, 80 server, 69
configuring node priority, 78 session
determining the primary node, 78 establishing, 69
dynamic checkpointing, 81 maximum number of, 69
failover, 79 opening, 69
node election, 78 terminating, 75
relaying configurations, 80 viewing, 75
viewing SummitStack, 604
checkpoint statistics, 81 TCP port number, 70
status, 81 using, 68
system temperature, 272 temperature range
system up time, 594 behavior
system virtual routers, 368 modular switch, 1059
Summit X450 family of switches, 1059
T ExtremeXOS, 1059
temperature, displaying
TACACS+ fans, 273
and RADIUS, 68, 547, 555 I/O modules, 272
configuration example, 557 MSM modules, 272
configuring, 556 power controllers, 272
description, 68, 555 power supplies, 273
disabling, 557 Summit X450 family of switches, 272
enabling, 557 Terminal Access Controller Access Control System
password, 556 Plus. See TACACS+
servers, specifying, 555 terminate process, 118
TACACS+ accounting TFTP
configuration example, 559 connecting to another host, 76
configuring, 558 default port, 77
disabling, 559 description, 76
enabling, 559 maximum number of sessions, 76
password, 558 server, 1000
tagging server requirements, 76, 1055
802.1Q, 303 using, 76, 1018
VLAN, 303 TFTP server, troubleshooting, 76
target address, SNMPv3, 100 timeout interval, EDP, 158
target parameters, SNMPv3, 100 timeout, MAC lockdown, 529
TCP MD5, 989 TLS
technical support, contacting, 1063 802.1Q encapsulation, 824
Telnet basic configuration example (figure), 834
ACL policy, 73 characteristics, 828
and safe defaults mode, 72 toggling, ESRP modes of operation, 744, 749
changing port, 73 TOP command, 1055

Index
TOS, 450 BlackDiamond 8800 series switch backup

traceroute, 63 MSM diagnostics, 242, 243
traffic engineering (TE), definition of, 791 BlackDiamond 8800 series switch I/O
traffic filter, port-mirroring, 152, 153, 154 module diagnostics, 239
traffic groupings, and QoS, 444 BlackDiamond 8800 series switch primary
transmit errors, port, 232 MSM diagnostics, 240, 241
trap receivers, SNMP, 93 MSM-1 diagnostics, 239
trapDestTable, 296 MSM-1XL, 239
trial license, 38 MSM-5 diagnostics, 244
triggered updates, RIP, 913 MSM-5R, 244
triggered updates, RIPng, 921 Summit X450 family of switches
Trivial File Transfer Protocol. See TFTP diagnostics, 244
troubleshooting licenses
ACLs, 375 EAPS, 653
AppleTalk, 306 ESRP, 743
ASCII-formatted configuration file, 1016 sFlow, 287
campus mode, 484 upgrading, 38
connectivity, 62 VRRP, 777
copying files, 110 link aggregation, 140, 146
CPU utilization, 121 LLDP, 183, 184
debug mode, EMS, 1048 load sharing, 140, 145, 148
deleting files, 115 MAC-in-MAC, 324
diagnostics memory, 120
Summit X450 family, 1062 module recovery, 260
viewing results, 245 MSM and I/O ports same module, 127
disabling backplane diagnostics, 1056 MSM prompt, 1036
downloads and TFTP, 76, 1055 MSM-1 diagnostics, 1062
EAPS MSM-1XL diagnostics, 1062
control VLAN, 657 passwords, 59
loop protection messages, 670 path MTU discovery, 137
ring ports, 666 ping, 62
enabling backplane diagnostics, 1056 PoE, 215, 216, 217, 221, 222
ESRP, 745, 747, 1040 port configuration, 1037
filenames, 108, 1050 port mirroring, 154
guest VLAN configuration, 500 port-mirroring, 153, 155
hardware table, 1060 guidelines, 152
hitless failover power fluctuation on PoE module, 1060
BlackDiamond 10K, 87 QoS, 441, 450, 453, 454
BlackDiamond 12804, 87 required software, 1038
BlackDiamond 8800, 87 rescue image, 1044
modular switches, all, 86 shutdown state
HQoS, 472 modular switches, 259
IP fragmentation, 137 Summit X450 family, 253
ISP mode, 484 software, 38
jumbo frames, 136 software limits, 1038
Layer 1, 1030 software-controlled redundant ports, 160
Layer 2, 1030 SSH2, 561
Layer 3, 1031 SSL, 57
LEDs SSL commands, 569
BlackDiamond 10808 switch I/O module STP, 732, 1039
diagnostics, 239 system LEDs, 1033
BlackDiamond 12804 switch I/O module TFTP server, 76, 1055
diagnostics, 243 traceroute, 62

Index
unsupported module viewing, 58

BlackDiamond 10808 switch, 1059 USM, SNMPv3 security, 96
BlackDiamond 8800 series switch, 1059 utilization, port, 163
untagged frames on 10Gbps module, 1062
VLANs, 301, 303, 308, 310, 1039 V
vMANs, 135, 316
VRRP, 788, 1041 vendor ID, 488, 504
VRRP and ESRP, 788 Vendor Specific Attribute. See VSA
troubleshooting MPLS, 789 version string, 998
trunks, 303 video applications, and QoS, 438
Tspec object, 837, 839 View-Based Access Control Model, SNMPv3, 98
tunneling, 842 Virtual LANs. See VLANs
IP, 894, 905 virtual link
vMANs, 330 OSPF, 930
Type of Service. See TOS OSPFv3, 941
virtual port, port-mirroring, 154
U virtual private LAN (VPN), definition of, 791
Virtual Router Redundancy Protocol. See VRRP
UDP echo server, 883 virtual routers
Universal Port adding and deleting routing protocols, 371
device profiles, 166 and routing protocols, 371
events that trigger profiles, 169 and VLANs, 300
examples, 172 commands, 369
executing static profiles, 170 configuration domain, 369
overview, 165 configuration example, 373
troubleshooting, 170 configuring routing protocols and VLANs, 372
user and security profiles, 167 creating, 370
variables, 171 default for Telnet, 70
unsupported module error deleting, 370
BlackDiamond 10808 switch, 1059 description, 367
BlackDiamond 8800 series switch, 1059 displaying information, 372
untagged frames, VLANs, 301, 308 system, 368
upgrading the image, 999 user, 369
uplink ports, netlogin, 520 VLAN tagging, 303
uploading VLAN, guest. See guest VLAN
ASCII-formatted configuration, 1017 VLANid, 303
XML-formatted configuration, 1018 VLANs
upstream forwarding, 363 and load sharing, 141, 150
URL redirection, 482 and STP, 693
user account, 54, 57 and virtual routers, 300
user name, local database authentication, 492 and vMANs, 316
user name, SNMPv3, 97 assigning a tag, 303
user sessions, 66 assigning IP address, 309
See also sessions benefits, 299
user virtual routers, 369 configuration examples, 310
User-Based Security Model. See USM configuring, 308
users default tag, 303
access levels, 54 description, 299
adding, 57 disabling, 309
authenticating, 68, 547 disabling route advertising, 913, 921
creating, 57 displaying IPv6 addresses, 312
default, 57 displaying settings, 163, 311
deleting, 58 enabling, 309
passwords, 59 IP fragmentation, 138

Index
IPv4 routing, 871 VRRP

IPv6 address, 310 advertisement interval, 782, 785
IPv6 addresses, 299 and ESRP, 775, 782, 788
IPv6 routing, 901 and IP multinetting, 877
mgmt, 67 and STP, 782
mixing port-based and tagged, 304 configuration parameters (table), 785
names, 47, 307 default gateway, 778
port-based, 301–302 description, 778
precedence, 307 electing the master, 782
protocol filters examples, 786–787
customizing, 306 hitless failover support, 778
deleting, 307 IP address, 785
predefined, 305 licensing, 777
protocol-based, 305, 306 master down interval, 782, 785
QoS profile, 312 master router
renaming, 308 determining, 779
tagged, 303 electing, 781
troubleshooting, 301, 303, 307, 308, 310, multicast address, 782
327, 1039 operation, 783
trunks, 303 ping tracking, 780, 781
types, 300 preempt mode, 785
untagged packets, 301, 303, 308 priority, 779, 781, 785
VLANid, 303 redundancy, 784
vMANs on same port, 316 route table tracking, 779
vMAN ACLs skew time, 782, 785
action modifiers, 393 tracking
actions, 392 description, 779
vMANs example, 780
ACLs, 465 troubleshooting, 788, 1041
and ACLs, 317 virtual IP addresses, 788
and EAPS, 330 virtual router MAC address, 782, 783
and link aggregation, 140 VLAN tracking, 779, 781
and MVR, 982 VRRP virtual router identifier (VRID), 785
and virtual routers, 317 VSA
and VLANs, 316, 317 203
configuring, 327 example, 491
displaying, 330 guidelines, 491
displaying settings, 163 204
EtherTypes, 315 example, 491
example, 335 guidelines, 491
guidelines, 326 205
HQoS, 465 example, 491
jumbo frames, 135, 316 guidelines, 491
licensing, 318 206
multicasting, 318 examples, 488
names, 47 guidelines, 492
tagging ports, 315 209
troubleshooting, 140, 316, 327 example, 491
VLANs on same port, 316 guidelines, 491
voice applications, and QoS, 438 211
VoIP, network measurement, 1066 examples, 490
VPLS guidelines, 490
domains, configuring, 829 definitions
VPLS, definition of, 791 Extreme, 488

Index
NAP, 504
definitions (table), 488, 505
order of use, 489
W
WAN PHY OAM, 134
warranty, 43
web browsing applications, and QoS, 439
web-based authentication, 506
advantages, 483
configuration, example, 510
disabling, 506
disadvantages, 483
enabling, 506
requirements, 482
URL redirection, 482
user login setup, 511
web-based device management. See ScreenPlay
wildcard combinations, port, 49, 129
wildcard reservation style, 841
X
XML, 116
XML configuration mode, 117

Index

ExtremeXOS Concepts Guide 12.0 - 100262-00 Rev 01

Uploaded by

Copyright:

Available Formats

ExtremeXOS Concepts Guide 12.0 - 100262-00 Rev 01

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

ExtremeXOS Concepts Guide 12.0 - 100262-00 Rev 01

Uploaded by

Copyright:

Available Formats

ExtremeXOS Concepts Guide

Software Version 12.0

Extreme Networks, Inc.

2 Extreme XOS 12.0 Concepts Guide

Part 1: Using ExtremeXOS

Chapter 2: Accessing the Switch.................................................................................................... 45

Extreme XOS 12.0 Concepts Guide 3

Chapter 3: Managing the Switch .................................................................................................... 65

4 Extreme XOS 12.0 Concepts Guide

Using the Simple Network Management Protocol .........................................................................91

Chapter 4: Managing the ExtremeXOS Software............................................................................. 107

Chapter 5: Configuring Slots and Ports on a Switch....................................................................... 125

Extreme XOS 12.0 Concepts Guide 5

Enabling Jumbo Frames.....................................................................................................136

Chapter 6: Universal Port............................................................................................................. 165

Chapter 7: CLI Scripting............................................................................................................... 175

6 Extreme XOS 12.0 Concepts Guide

CLI Scripting Examples ...........................................................................................................179

Chapter 8: Link Layer Discovery Protocol...................................................................................... 181

Chapter 9: Connectivity Fault Management—BlackDiamond 10808 and 12800 Series

Chapter 10: Power Over Ethernet.................................................................................................. 213

Extreme XOS 12.0 Concepts Guide 7

Configuring PoE ......................................................................................................................220

Chapter 11: Status Monitoring and Statistics ................................................................................ 231

8 Extreme XOS 12.0 Concepts Guide

Displaying ELSM Information .............................................................................................268

Chapter 12: Virtual LANs ............................................................................................................. 299

Extreme XOS 12.0 Concepts Guide 9

Chapter 13: vMAN and Tunneling ................................................................................................. 313

Chapter 14: Web-Based Device Management................................................................................ 343

10 Extreme XOS 12.0 Concepts Guide

Statistics and Monitoring .........................................................................................................351

Chapter 15: Forwarding Database................................................................................................. 357

Chapter 16: Virtual Routers.......................................................................................................... 367

Chapter 17: Policy Manager ........................................................................................................ 375

Extreme XOS 12.0 Concepts Guide 11

Chapter 18: Access Lists (ACLs)................................................................................................... 379

12 Extreme XOS 12.0 Concepts Guide

Chapter 19: Routing Policies ....................................................................................................... 427

Chapter 20: Quality of Service ..................................................................................................... 437

Extreme XOS 12.0 Concepts Guide 13

Configuring HQoS Ingress and Egress Queues ......................................................................472

Chapter 21: Network Login .......................................................................................................... 481

Chapter 22: Security ................................................................................................................... 523

14 Extreme XOS 12.0 Concepts Guide

Chapter 23: CLEAR-Flow .............................................................................................................. 573

Extreme XOS 12.0 Concepts Guide 15

Chapter 24: Configuring Stacked Switches ................................................................................... 591

16 Extreme XOS 12.0 Concepts Guide

Configuration Examples ...........................................................................................................626

Part 2: Using Switching and Routing Protocols

Extreme XOS 12.0 Concepts Guide 17

Configuring the Link ID of the Shared Port...........................................................................678

Chapter 26: Spanning Tree Protocol............................................................................................. 689

Chapter 27: Extreme Standby Router Protocol ............................................................................... 743

18 Extreme XOS 12.0 Concepts Guide

Chapter 28: Virtual Router Redundancy Protocol........................................................................... 777

Extreme XOS 12.0 Concepts Guide 19

Simple VRRP Network Configuration ...................................................................................783