Linux Pocket Reference For System

Linux Reference for System Administrators
Table of Contents Chapter 1: Basics Chapter 2: Using fdisk Chapter 3: Basic Network trouble-shooting Chapter 4: DNS Chapter 5: NFS Chapter 6: NIS Chapter 7: SAMBA Chapter 8: SQUID Chapter 9: APACHE Chapter 10: SENDMAIL Chapter 11: DHCP Chapter 12: FTP Chapter 13: FIREWALL Chapter 14: KERNEL MODIFICATION AND COMPILATION Chapter 15: Single user and Rescue mode Chapter 16: Advance permissions Wiring / cabling Network Addresses / Classes
Chapter 1: Basics
Let us start with the commonly used commands. D irectory and d isk usage related commands: Following command gives a list of files and directories in a long listing (even hidden files will be displayed), human readable sizes, KB, MB. Also sorted by time (-t) and in reverse order (-r) i.e. Newest in the end of listing. [root@Server /]# ls -l -h -a -t r Following command captures lines showing directories only page per page, to check for files use ^f. (The ^ d or ^f is NOT ctrl-d or ctrl-f). [root@Server /]# ls -l | grep ^d | less To check for free space on your partitions: [root@Server /]# df h To check for the used space by a particular directory: [root@Server /]# du c h /home/xena or [root@Server /]# du s h /home/xena M ounting / un-mounting floppies / cd-rom / directories: To check for currently mounted partitions, use mount command without any arguments: [root@Server /]# mount /dev/hda6 on / type ext3 (rw) none on /proc type proc (rw) usbdevfs on /proc/bus/usb type usbdevfs (rw) /dev/hda1 on /boot type ext3 (rw) /dev/hda2 on /data type ext3 (rw) none on /dev/pts type devpts(rw,gid=5,mode=620) none on /dev/shm type tmpfs (rw) /dev/hda3 on /mnt/hda3 type vfat (rw) You can also use the mounted files system table mtab to check for currently mounted partitions. [root@Server /]# cat /etc/mtab /dev/hda6 / ext3 rw 0 0 none /proc proc rw 0 0 usbdevfs /proc/bus/usb usbdevfs rw 0 0
/dev/hda1 /boot ext3 rw 0 0 /dev/hda2 /data ext3 rw 0 0 none /dev/pts devpts rw,gid=5,mode=620 0 0 none /dev/shm tmpfs rw 0 0 /dev/hda3 /mnt/hda3 vfat rw 0 0 Following command mounts the floppy drive in mount point /mnt/floppy [root@Server /]# mount /dev/fd0 /mnt/floppy Similarly mounting a cd-rom: [root@Server /]# mount /dev/cdrom /mnt/cdrom Following command mounts your VFAT partition, C: drive, in /mnt/windir [root@Server /]# mount -t vfat /dev/hda2 /mnt/windir If you want to look into an .iso image, you can mount it to an empty directory (/mnt/cdimage) using loopback option and then use it as an ordinary directory. [root@Server /]# mount -o loop /downloads/cd1.iso /mnt/cdimage Following command mounts an NFS share /usr/redhat72 from the machine 192.168.1.254 to your local mount point /mnt/cdrom: [root@Server /]# mount t nfs 192.168.1.254:/usr/redhat72 /mnt/cdrom To unmount any mount point use the following command but the mount point must not be busy: [root@Server /]# umount /mnt/cdrom Note: You cannot mount an NTFS partition unless its support is enabled in the Kernel. See chapter Kernel Modification and Compilation. At this time only read-only support for NTFS is available in Red Hat Linux. File copying/moving/renaming/deletion/ compression and decompression: Following command copies all (non-hidden) files, including any subdirectories from /home/xena to /mnt/floppy, overwriting any previously existing files (--reply=yes)
[root@Server /]# cp -vr /home/xena/* /mnt/floppy/ reply=yes The above command will change the ownership of the newly copied files on /mnt/floppy to the uid and gid of the user issuing the copy (cp) command, in this case the username is root. To preserve the file permissions while copying to the new location, use the -p switch as well. For example: [root@Server /]# cp -v -r -p /home/xena/* /mnt/floppy/ reply=yes Following command deletes all files, including subdirectories, forcefully (without asking yes/no) from the directory /tmp. [root@Server /]# rm -fr /tmp/* Following command creates a tar file in /tmp as mytar.tar from the files in current directory [root@Server /]# tar cvf /tmp/mytar.tar * Following command untars the tar file mytar.tar from /tmp in current directory [root@Server /]# tar xvf /tmp/mytar.tar Following command untars the tar file mytar.tar from /tmp in /backup/restore [root@Server /]# tar xvf /tmp/mytar.tar -C /backup/restore Following command adds a file (newfile.pl) to already existing tar file existing.tar . [root@Server /]# tar rvf existing.tar newfile.pl Following command compresses (using maximum compression) the tar file mytar.tar to a 'gzip' file mytar.tar.gz [root@Server /]# gzip -v9 mytar.tar Following command extracts / decompresses the tar file out of the .gz file in the current directory [root@Server /]# gzip -vd mytar.tar.gz
Disk checking and formatting: List of all partitions on all attached hard drives (both SCSI and IDE) [root@Server /]# fdisk -l To resize partitions use: [root@Server /]# parted Following command formats a floppy disk in ext2 format, also checks for and marks, bad sectors: [root@Server /]# mke2fs -c /dev/fd0 Following command formats a floppy disk in dos format [root@Server /]# mkdosfs /dev/fd0 Following command checks for errors on the partition, the partition being checked should be in an un-mounted state during the check. [root@Server /]# e2fsck /dev/hda3 Or [root@Server /]# fsck -t ext2 /dev/hda3 Process management: Following commands shows the list of all running processes with their process ids (PID) from all users: [root@Server /]# ps aux | less You can also use pstree : [root@Server /]# pstree Following command will kill a specific process, say process id 12432: [root@Server /]# kill 12432 Or [root@Server /]# kill s KILL 12432
Or [root@Server /]# kill KILL 12432 Or [root@Server /]# kill 9 12432 Package management: Following command lists all the currently installed packages and captures only those lines which have the word sendmail : [root@Server /]# rpm -qa | grep sendmail Or [root@Server /]# rpm -qa sendmail* To check presence of the package 'bind' in the system, use the following command: [root@Server /]# rpm -q bind Following command lists all the information about the specific installed package. [root@Server /]# rpm -qi sendmail Following command lists all the documentation files and their location of a particular installed package. [root@Server /]# rpm -qd sendmail Following command lists file locations of all of the files of a particular installed package. [root@Server /]# rpm -ql sendmail Following command lists all information of the un-installed package: [root@Server /]# rpm -qpi ymessenger-0.99.19-1.i386.rpm Similarly,you can check the list of files a package contains, before actually installing it: [root@Server /]# rpm -qpl ymessenger-0.99.19-1.i386.rpm Also to check, which package a file belongs to by:
[root@Server /]# rpm -qf /etc/mail/sendmail.cf sendmail-8.12.8-4 If not installed then install it from (mounted) Redhat CD by using the following command: [root@Server /]# rpm -ivh /mnt/cdrom/RedHat/RPMS/bind*.rpm If an older version of the package is installed and you want to upgrade it to a new version, then use (U): [root@Server /]# rpm -Uvh /mnt/cdrom/RedHat/RPMS/bind*.rpm If a version of a package is installed but some of the files got corrupted and want to reinstall the package then use force in your command. [root@Server /]# rpm -ivh -force /mnt/cdrom/RedHat/RPMS/bind*.rpm Following command erases (removes / un-installs) a specific package. [root@Server /]# rpm -e sendmail Sometimes during installation of some rpm package you get some error about missing libraries. To check which package provides you that library, you can use: [root@Server /]# rpm -q -redhatprovides libc.so.6 glibc-2.3.2-11.9 So in the above example you see that if you install glibc-2.3.2-11.9, you will automatically get the required library (libc.so.6), installed. But remember that for the above command to work correctly, you should have rpmdb-redhat-a.b package already installed on your system. [root@Server tmp]# rpm -q rpmdb-redhat rpmdb-redhat-9-0.20060313 System services configuration / device configuration: To check the status of the daemons listening on different ports: [root@Server /]# netstat -at Following command will generate a hardware list of your computer and save it as tar file. [root@Server /]# sysreport
To configure boot up services use setup or ntsysv [root@Server /]# setup Or [root@Server /]# ntsysv Assuming you want to add the ypbind service to run level 3 only; then, [root@Server /]# chkconfig -add ypbind [root@Server /]# chkconfig -level 3 ypbind on To configure various hardware devices (prior to Red hat 8.0), use: [root@Server /]# kbdconfig (for keyboard layout) [root@Server /]# modemtool (for modem) [root@Server /]# mouseconfig (for mouse) [root@Server /]# printconf (for printer configuration, needs X interface) [root@Server /]# timeconfig (for time zone configuration) [root@Server /]# Xconfigurator (for VGA card) [root@Server /]# xvidtune (for Monitor display adjustment) To configure various hardware devices ( Redhat 8.0 and 9.0) , use : [root@Server /]# redhat-config-kbdconfig (for keyboard layout) [root@Server /]# redhat-config-modemtool (for modem) [root@Server /]# redhat-config-mouseconfig (for mouse) [root@Server /]# redhat-config-printconf (for printer configuration, needs X interface) [root@Server /]# redhat-config-timeconfig (for time zone configuration) [root@Server /]# redhat-config-xfree86 (for VGA card) [root@Server /]# xvidtune (for Monitor display adjustment)
To run a shell terminal from X-windows, run the command: [root@Server /]# xterm Or [root@Server /]# konsole User Manageme nt and f ile permissions / ownerships: To add a new group called mailusers use: [root@Server /]# groupadd mailusers To add new user xena and make him a member of the group mailusers, use: [root@Server /]# useradd -s /bin/false -g mailusers xena To change the password if some user, say xena, run the following command as root: [root@Server /]# passwd xena To lock the password of user xena, run the following command as root: [root@Server /]# passwd -l xena To un-lock the password of user xena, run the following command as root: [root@Server /]# passwd -u xena To delete a user say xena, including his home directory, use: [root@Server /]# userdel -r xena To change permissions of an object (file or directory), use chmod: [root@Server /]# chmod 770 myfile.txt (Changes the permissions of the file to rwx for both owner and group) Or [root@Server /]# chmod ug+rwx,o-rwx myfile.txt To change ownership of a directory /project to xena and set the group permissions to the group mailusers, use chown (-R will change ownership of all files and subdirectories under the directory project:
[root@Server /]# chown R xena:mailusers /project For more topics like sudo, umask and setgid, etc, see the chapter, Advance permissions Modules management: To insert a module like you want to add the support for your 3com Ethernet card 3c509, then use: [root@Server /]# modprobe 3c509 io=0x300 irq=5 Or [root@Server /]# insmod 3c509 io=0x300 irq=5 then [root@Server /]# ifconfig eth0 192.168.1.5 netmask 255.255.255.0 up To list out the currently loaded modules type: [root@Server /]# lsmod To remove a module type: [root@Server /]# rmmod 3c509 Routing table and naming: To check the routing table: [root@Server /]# route or [root@Server /]# route -n Note: If the route command sort of hangs / gives delay before showing the routing table and the route -n command shows the routing table in numeric form instantly, this indicates a problem with name resolution. To add a default route or gateway: [root@Server /]# route add default gw Server.mydomain.mydomain.com Or
[root@Server /]# route add default gw 192.168.1.254 To delete a default gateway: [root@Server /]# route del default To add a route to a network [root@Server /]# route add net 192.168.1.0 netmask 255.255.255.0 dev eth0 To delete a network from the routing table [root@Server /]# route del net 192.168.1.0 netmask 255.255.255.0 dev eth0 To add a host route: [root@Server /]# route add 192.168.1.1 eth0 To delete a host route [root@Server /]# route del 192.168.1.1 To trace the path / number of hops to a particular host on the network: [root@Server /]# traceroute wks1.mydomain.com Or [root@Server /]# traceroute www.linux.org To do tracing of the route and ping at the same time, use mtr: [root@Server /]# mtr wks1.mydomain.com To find a host and/or it's IP on the network use the following commands. These will query your default DNS server of your network: [root@Server /]# host wks1.mydomain.com Or [root@Server /]# nslookup wks1.mydomain.com Or
[root@Server /]# dig wks1.mydomain.com You can also use hostname to get the FQDN of a specific host. But the FQDN hostname must be set in either /etc/sysconfig/network file or be set manually using the hostname command. It is a good practice to write FQDN of each computer in it's own /etc/sysconfig/network file. The check that, the command hostname -fqdn must return the FQDN of your machine. See below for more details: [root@Server /]# cat /etc/sysconfig/network NETWORKING=yes HOSTNAME=Server.mydomain.com [root@Server /]# hostname Server.mydomain.com [root@Server /]# hostname --fqdn Server.mydomain.com In order to find out the name of your DNS domain, you should not use domainname command without arguments, as it will return the name of the NIS domain, which is not yet set uptill now. You can use -d switch with domainname command or use dnsdomainname command. The domainname -y will also returns you the name of NIS domain (equal to ypdomainname command). [root@Server /]# domainname (none) [root@Server /]# domainname -d mydomain.com [root@Server /]# dnsdomainname mydomain.com [root@Server /]# domainname -y (none) [root@Server /]# ypdomainname (none) To check the status of a wintendo machine on your network, use: [root@Server /]# nmblookup -A wks1.mydomain.com
Or [root@Server /]# nmblookup -A 192.168.1.1 To check for the open ports on a machine on the network (may be the local host itself): [root@Server /]# nmap wks1.mydomain.com Miscellaneous: To create a boot floppy out of the images provided by the installation CROM: [root@Server /]# dd if=/mnt/cdrom/images/boot.img of=/dev/fd0 To find help on some topic: [root@Server /]# man snmp Or [root@Server /]# man 5 snmp (This is normally referred as snmp(5) in docs) Or [root@Server /]# info snmp Also you can seek for a particual keyword ( say: umask) in the man pages by: [root@Server /]# man -K umask Talking of info command, here are some important shortcut keys while navigating in info pages. Key Functions: N = Takes you to the next page in sequence after the present one. P = Takes you to the previous page in sequence after the present one. U = Takes you up one level in the page hierarchy. Enter = Jumps to the link under the cursor. M = Follow link from menu and asks you which entry to use. F = Follows standard cross reference, which usually has Note: written close to it. L = Displays the last page shown. Equal to pressing back button on a web browser window. Spacebar = Scroll forward one page. Backspace = Scroll backward one page.
To find a file in the /usr directory, use: [root@Server /]# find /usr -name *.doc To find files and directories owned by the group students in the /usr directory, use: [root@Server /]# find /usr -group students Assuming a user XENA left the organization and the files owned by his account name requires search and be deleted for security reasons then use: [root@Server /]# find / -user xena -exec rm '{}' ';' Want to perform some quick mathematical calculations? Then use: [root@Server /]# bc To set up date and time as 20 March 2006 14:02 you should use: [root@Server /]# date 0320140206 To synchronize it with the hardware clock, use: [root@Server /]# hwclock -systohc To check for memory and swap space usage, you can use: [root@Server /]# free Or [root@Server /]# top To check the system load status averages and the uptime: [root@Server /]# uptime To check when the system was booted last time: [root@Server /]# who -b To check the status of a particular service: [root@Server /]# service sendmail status Or
[root@Server /]# /etc/rc.d/init.d/sendmail status To check the status of all services: [root@Server /]# service -status-all To create a boot disk for your system, use the mkbootdisk command, but first check the kernel version of your system by uname -r command or by looking into the /lib/modules direcotry: [root@Server /]# uname -r 2.4.20-8 Or [root@Server /]# ls /lib/modules 2.4.20-8 [root@Server /]# mkbootdisk -device /dev/fd0 2.4.20-8 Sometimes symbolic links to files are broken, eg assuming link file is deleted or some file is over written on it. To create a symbolic link for the /boot/grub/menu.lst files in the /etc directory as grub.conf use: [root@Server /]# ln -s /boot/grub/menu.lst /etc/grub.conf Note: Omitting -s in the ln command, a hard link will be created, which will NOT show up in the ls -l command output as a link file, rather the newly created hard link will be shown as a regular file. Still it will function as a link file. If you want to know the Red Hat version, read the file /etc/redhat-release or /etc/issue [root@Server /]# cat /etc/redhat-release Red Hat Linux release 9 (Shrike) Similarly [root@Server /]# cat /etc/issue Red Hat Linux release 9 (Shrike) Kernel \r on an \m Also uname can be very helpful to determine your processor model, motherboard class/model and machine architecture.
Note: For Intel users the processor may be i386, i486, i586 (P-1) or i686 (P-2 to P-4). For AMD Athlon it shows athlon. Similarly for Intel based machines or AMD based machines or any machine known as IBM-PC/AT compatible will show i386 under uname i command. [root@Server /]# uname -p athlon [root@Server /]# uname -m i686 [root@Server /]# uname -i i386 Important Vi editor commands: Undo: u Redo: ^r Repeat: . Visual mode / highlighter mode/ selection mode (used with arrow keys): v Cut: *c Copy: *y Paste: *p Select all: ggVG Show the name of current file being edited: Ctrl+g Execute shell command: :! Search forward: /linux Search backwards: ?linux Search and replace a text in the entire file: :%s/linux/LINUX/g (This will search and replace the word linux with LINUX in the current file). Let us suppose a situation where you changed path using cd command to some very long/deep level directory and now you want to come back to the previous directory you came from and then from there again you want to come back to the same deep level directory. This creates a lot of typing frustration. The Solution is cd command. [root@Server network-scripts]# pwd /etc/sysconfig/network-scripts [root@Server network-scripts]# cd /lib/modules/ [root@Server modules]# cd /etc/sysconfig/network-scripts
[root@Server network-scripts]# cd /lib/modules To mount flash (USB) disk use: [root@Server modules]# mount /dev/sda1 /mnt/floppy
Chapter 2: Using fdisk

fdisk is a very powerful tool. As a system administrator one should know how to use this tool effectively and efficiently without causing harm to data. This tool will save a lot of time of yours during the installation and also during normal system operation. To check the partition list, use fdisk -l: [root@Server /]# fdisk -l Disk /dev/hda: 255 heads, 63 sectors, 4867 cylinders Units = cylinders of 16065 * 512 bytes Device Boot Start 1 8 1283 3578 3578 3706 3739 3772 End 7 1282 3577 4867 3705 3738 3771 4867 Blocks 56196 10241437+ 18434587+ 10361925 1028128+ 265041 265041 8803588+ Id System
/dev/hda1 /dev/hda2 * /dev/hda3 /dev/hda4 /dev/hda5 /dev/hda6 /dev/hda7 /dev/hda8
83 Linux c Win95 FAT32 (LBA) 83 Linux 5 Extended 83 Linux 83 Linux 82 Linux swap 83 Linux
The * in Boot column shows the active or boot partition, in this case it is /dev/hda2 or 2nd primary partition on FAT32. The Id column contains the partition file system type in hexadecimal format. Also this command will list out all the hard disks attached to the system and will display all the partitions in each hard disk. While creating partitions, it should be noted that at a time there can be a maximum of four primary (with no extended) partitions on one hard drive, numbered 1 to 4. Also there can be only one extended (with max three primary) partition on one hard disk. You can create as many partitions in extended as you want, numbered 5 and above. So in the above example output of the command, fdisk -l, you can see three primary partitions, numbered /dev/hda1 to /dev/hda3 and an extended partition (considered a special type of primary partition) /dev/hda4, defined in hard disk /dev/hda, which is the primary master hard disk on this system. This extended partition contains four logical partitions (as many windows users call / identify them), numbered /dev/hda5 to /dev/hda8.
The IDE devices are numbered using the following scheme: Primary master Primary slave Secondary master Secondary slave /dev/hda /dev/hdb /dev/hdc /dev/hdd
To modify partitions use fdisk in interactive mode. You have to type in the hard disk device id as a parameter. Type the following command and press m to get a list of available commands. [root@Server /]# fdisk /dev/hdc The number of cylinders for this disk is set to 4867. There is nothing wrong with that, but this is larger than 1024, and could in certain setups cause problems with: 1) Software that runs at boot time (e.g:Old versions of LILO) 2) Booting and partitioning software from other OSs (e.g., DOS FDISK, OS/2 FDISK) Command (m for help): m Command a b c d l m n o p q s t u v w x Actions toggle a bootable flag edit bsd disklabel toggle the dos compatibility flag delete a partition list known partition types print this menu add a new partition create a new empty DOS partition table print the partition table quit without saving changes create a new empty Sun disklabel change a partition's system id change display/entry units verify the partition table write table to disk and exit extra functionality (experts only)
Case: Suppose we want to create some partitions on a new hard disk (/dev/hdc), attached to the system as secondary master, of size 2GB. This hard disk also contains some bad
sectors somewhere in the middle and at the end. i.e from cylinder 559 till cylinder 579 and from cylinder 611 till cylinder 1023. (This is told to us by someone who has also been using linux with this hard-disk and has given it to us for experimentation. We in turn want to use this on a old 486 computer lying in the corner of our lab.) We want to create a primary partition of 100MB (primary 1) which will serve as '/boot' later on. This over- comes many boot constraints while installing Linux boot loader: LILO. Then we want to create partitions for win95 and linux. There will be one win95 partition of 100MB (primary 2) and then we will create an extended partition of all leftover space. In this extended partition we will create one partition of 900MB, which will later on serve as '/ and a 64 MB partition which will work as a <swap>.
Configuration / Procedure:
First check the current partition table by pressing p to print the partition table on the screen. Command (m for help):p Disk /dev/hdc: 64 heads, 63 sectors, 1023 cylinders Units = cylinders of 4032 * 512 bytes Device Boot Start End Blocks Id System
This shows that there are no partitions defined on the current hard disk. To add a new partition use n and fdisk will ask you for the type of partition, primary or extended. Type p for primary and give 1 as partition number and +100M as size. Command (m for help):n Command e p Actions extended primary partition (1-4)
p Partition number (1-4): 1 First cylinder (1-1023, default 1): 1 Last cylinder or +sizeM or +sizeK (1-1023, default 1023): +100M Now check the partition table by pressing p again at the fdisk's command prompt. Command (m for help): p Disk /dev/hdc: 64 heads, 63 sectors, 1023 cylinders Units = cylinders of 4032 * 512 bytes
Device Boot /dev/hdc1
Start 1
End 51
Blocks 102815+
Id 83
System Linux
Now we want to add a win95 primary partition (FAT32). Again add a new primary partition by pressing p and press 2 for partition number. Press enter for default starting address and specify +100M for the size. Command (m for help): n Command e p Actions extended primary partition (1-4)
p Partition number (1-4): 2 First cylinder (52-1023, default 52):(press Enter) Using default value 52 Last cylinder or +sizeM or +sizeK (52-1023, default 1023): +100M Command (m for help): p Disk /dev/hdc: 64 heads, 63 sectors, 1023 cylinders Units = cylinders of 4032 * 512 bytes Device Boot /dev/hdc1 /dev/hdc2 Start 1 52 End 51 102 Blocks 102815+ 102816 Id 83 83 System Linux Linux
If you notice here we have created another primary partition of size 100MB and have also confirmed the effects by printing the partition table. But there is something confusing here; under the Id column fdisk is showing the partition as type 83, which is of file system type Linux. But we wanted to make it of type FAT32! For this we will change the system id of this partition by pressing t on the fdisk command prompt and specify that we want to change the system id of partition number 2, to win95 FAT32, which is of type c. Command (m for help): t Partition number (1-4): 2 Hex code (type L to list codes): L
0 1 2 3 4 Empty FAT12 XENIX XENIX FAT16 1c Hidden Win95 FA 70 DiskSecure Mult bb Boot Wizard hid 1e Hidden Win95 FA 75 PC/IX be Solaris boot root 24 NEC DOS 80 Old Minix c1 DRDOS/sec (FATusr 39 Plan 9 81 Minix / old Lin c4 DRDOS/sec (FAT<32M 3c PartitionMagic 82 Linux swap c6 DRDOS/sec (FAT-
5 Extended 40 Venix 80286 83 Linux c7 Syrinx 6 FAT16 41 PPC PReP Boot 84 OS/2 hidden C: da Non-FS data 7 HPFS/NTFS 42 SFS 85 Linux extended db CP/M / CTOS / . 8 AIX 4d QNX4.x 86 NTFS volume set de Dell Utility 9 AIX bootable 4e QNX4.x 2nd part 87 NTFS volume set df BootIt a OS/2 Boot Manag 4f QNX4.x 3rd part 8e Linux LVM e1 DOS access b Win95 FAT32 50 OnTrack DM 93 Amoeba e3 DOS R/O c Win95 FAT32 (LB 51 OnTrack DM6 Aux 94 Amoeba BBT e4 SpeedStor e Win95 FAT16 (LB 52 CP/M 9f BSD/OS eb BeOS fs f Win95 Ext'd (LB 53 OnTrack DM6 Aux a0 IBM Thinkpad hi ee EFI GPT 10 OPUS 54 OnTrackDM6 a5 FreeBSD ef EFI (FAT-12/16/ 11 Hidden FAT12 55 EZ-Drive a6 OpenBSD f0 Linux/PA-RISC b 12 Compaq diagnost 56 Golden Bow a7 NeXTSTEP f1 SpeedStor 14 Hidden FAT16 <3 5c Priam Edisk a8 Darwin UFS f4 SpeedStor 16 Hidden FAT16 61 SpeedStor a9 NetBSD f2 DOS secondary 17 Hidden HPFS/NTF 63 GNU HURD or Sys ab Darwin boot fd Linux raid auto 18 AST SmartSleep 64 Novell Netware b7 BSDI fs fe LANstep 1b Hidden Win95 FA 65 Novell Netware b8 BSDI swap ff BBT
Hex code (type L to list codes): c Changed system type of partition 2 to c (Win95 FAT32 (LBA)) Command (m for help): p Disk /dev/hdc: 64 heads, 63 sectors, 1023 cylinders Units = cylinders of 4032 * 512 bytes Device Boot /dev/hdc1 /dev/hdc2 FAT32 (LBA) Start 1 52 End 51 102 Blocks 102815+ 102816 Id 83 c System Linux Win95
Command (m for help): As you can see, we have verified that the partition id of partition 2 has been successfully changed to Win95 FAT32. Now we also want that our partition 2 should be the boot partition. So we will use a toggle to bootable flag of partition number 2. Command (m for help): a Partition number (1-4): 2 Command (m for help): p Disk /dev/hdc: 64 heads, 63 sectors, 1023 cylinders Units = cylinders of 4032 * 512 bytes
Start 1
End 51
Blocks 102815+
Id 83
System Linux
/dev/hdc2 * FAT32 (LBA)
52
102
102816
Win95
Now we want to create an extended partition of all available leftover space, also within that we will create the partitions which will be used by linux during installation. We can also leave it till here as installation program also contains tools like Disk Druid and fdisk, but lets go on with it for now. Command (m for help): n Command action e extended p primary partition (1-4) e Partition number (1-4): 3 First cylinder (103-1023, default 103):(press Enter) Using default value 103 Last cylinder or +sizeM or +sizeK (103-1023, default 1023): (press Enter) Using default value 1023 Command (m for help): p Disk /dev/hdc: 64 heads, 63 sectors, 1023 cylinders Units = cylinders of 4032 * 512 bytes Device Boot /dev/hdc1 /dev/hdc2 * FAT32 (LBA) /dev/hdc3 Start 1 52 103 End 51 102 Blocks 102815+ 102816 Id 83 c 5 System Linux Win95 Extended
1023 1856736
Command (m for help): n Command l p Actions logical (5 or over) primary partition (1-4)
l First cylinder (103-1023, default 103): (Press Enter) Using default value 103 Last cylinder or +sizeM or +sizeK (103-1023, default 1023): 559
Notice that this time after pressing n for new partition, the options menu is changed. There is logical in place of extended. This is because there can only be one extended partition in one hard disk, and once that is created, this option is no more shown. We are
allowed to create logical partitions in the extended partition. In the above example we have created a cylinder based partition, from cylinder 103 till cylinder 559. We will create the next partition from cylinder 579 till cylinder 611, which will work as a swap partition, after-wards. Command (m for help): n Command Action l logical (5 or over) p primary partition (1-4) l First cylinder (560-1023, default 560): 579 Last cylinder or +sizeM or +sizeK (579-1023, default 1023): 611 Command (m for help): p Disk /dev/hdc: 64 heads, 63 sectors, 1023 cylinders Units = cylinders of 4032 * 512 bytes Device Boot /dev/hdc1 /dev/hdc2 * FAT32 (LBA) /dev/hdc3 /dev/hdc5 /dev/hdc6 Start 1 52 103 103 579 End 51 102 Blocks 102815+ 102816 Id 83 c 5 83 83 System Linux Win95 Extended Linux Linux
1023 1856736 559 921311+ 611 66528
Since this partition has to be a swap partition, we need to change it's system id to 82, which is Linux swap. Command (m for help): t Partition number (1-6): 6 Hex code (type L to list codes):82 Changed system type of partition 6 to 82 (Linux swap) Command (m for help): p Disk /dev/hdc: 64 heads, 63 sectors, 1023 cylinders Units = cylinders of 4032 * 512 bytes
Start 1
End 51
Blocks 102815+
Id 83
System Linux
/dev/hdc2 * FAT32 (LBA) /dev/hdc3 /dev/hdc5 /dev/hdc6 swap
52 103 103 579
102
102816
c 5 83 82
Win95 Extended Linux Linux
1023 1856736 559 921311+ 611 66528
All the partitions are created now so let's save our efforts, i.e writes the partition table by using w on the fdisk prompt. If fdisk was called from a command prompt, then after pressing w, we will be taken back to the shell command prompt. And if we called fdisk during partitioning option of Linux installation, then we will be taken back to the previous screen where you are to select hard disk to run fdisk on. Go back from there and select Disk druid this time to assign mount points and format the partitions during installation. Command (m for help): w The partition table has been altered! Calling ioctl() to re-read partition table. Syncing disks. [root@Server /]# Note: At any instance if you think that you have messed up with things then just exit from fdisk by pressing q instead of w. This will quit fdisk without saving /writing the partition table. You can then re-run it. This normally happens when you delete a partition or assign incorrect system id or incorrect size, etc by mistake. Also deleting a partition is very simple, just press d on the fdisk command prompt, you will be asked for the partition number from 1 to maximum number of partitions you have on the drive. Save your changes by pressing w on the fdisk command prompt.
Partition size calculation:

Now lets calculate the sizes of these cylinder based partitions on the basis of the information available from the print partition table command, p. The statement Units = cylinders of 4032 * 512 bytes, in the output generated by print (p) command of fdisk, means that each cylinder consists of 4032 * 512 bytes, which is equal to 2064384 bytes. We have (51-1) +1 = 51 cylinders in /dev/hdc1 partition, so 51 * 2064384 bytes = 105283584 bytes. Divide this number by 1024 gives 102816 KB (kilo bytes). Further dividing this new number by 1024 again gives 100.40625 MB (mega bytes). This was in fact the size which we specified while creating /dev/hdc1 partition. Since this method of finding out the size of partition is confirmed, now lets find out the size of the partitions /dev/hdc3, /dev/hdc5 and /dev/hdc6, which were also created using cylinder numbers: 1) /dev/hdc3 (extended partition)
(1023-103)+1 = 921 cylinders 921 * 4032 * 512 = 1901297664 bytes 1901297664 / 1024 = 1856736 KB 1856736 /1024 = 1813.21875 MB 1813.21875 / 1024 = 1.770721436 GB, this is nearly equal to 1.8 GB 2)/ dev/hdc5 (559-103)+1 = 457 cylinders 457 * 4032 * 512 = 941359104 bytes 941359104 / 1024 = 919296 KB 919296 / 1024 = 897.75 MB, this is nearly equal to 900 MB. 3)/dev/hdc6 (611-579) + 1 = 33 cylinders 33 * 4032 * 512 = 68124672 bytes 68124672 / 1024 = 66528 KB 66528 / 1204 = 64.96875 MB, this is nearly equal to 65 MB We know that the total numbers of cylinders on this hard disk were 1023, out of which we have used up to 611. So how much space is still left on this hard disk, (this is the space which in fact contains a lot of bad sectors). (1023 - 611) +1 = 413 cylinders 413 * 4032 * 512 = 852590592 bytes 852590592 / 1024 = 832608 KB 832608 / 1024 = 813.09375 MB, this is nearly equal to 813 MB The partitions have been created, now you can use the command mke2fs to format the new Linux partition. You have to use mkswap to format the swap partition and mkdosfs to format the FAT32 partition. [root@Server /]# mke2fs -c /dev/hdc1 mke2fs 1.27 (8-Mar-2002) Filesystem label= OS type: Linux Block size=1024 (log=0) Fragment size=1024 (log=0) 25792 inodes, 102815 blocks 5140 blocks (5.00%) reserved for the super user First data block=1 13 block groups 8192 blocks per group, 8192 fragments per group
1984 inodes per group Superblock backups stored on blocks: 8193, 24577, 40961, 57345, 73729 Checking for bad blocks (read-only test): done Writing inode tables: done Writing superblocks and filesystem accounting information: done This filesystem will be automatically checked every 38 mounts or 180 days, whichever comes first. Use tune2fs -c or -i to override. [root@Server /]# mkdosfs -c -F 32 /dev/hdc2 mkdosfs 2.8 (28 Feb 2001) [root@Server /]# mkswap /dev/hdc6 Setting up swapspace version 1, size = 133052 KiB Further you need to turn on swap space usage by using swapon command. [root@Server /]# swapon -a
Chapter 3: Basic N etwork trouble-shooting

Example network: Suppose our example network is of class C private IP scheme 192.168.1.0 network, which is of-course non-routable. The network has a domain named mydomain.com which is not registered on the Internet and serves as the part of fully qualified domain names of the machines of this network only. There is a modem attached with the server to connect to Internet, so this server also acts as a gateway for this network. This network will be used in all examples of this book, until/unless specified otherwise. See below Figure for detail:
Network: 192.168.1.0 Network Mask: 255.255.255.0 Server is called Server with fully qualified domain name (FQDN) as Server.mydomain.com It's IP is 192.168.1.254 and is also the default gateway for the client computers on this network. Client machines are called wks1, wks2 and so on, with FQDNs as wks1.mydomain.com and wks2.mydomain.com respectively, having IPs 192.168.1.1, 192.168.1.2 and so on.
Basic network configuration and troubleshooting:

Check the network settings in the following order:Note: All commands are to be given as user root. 1) Check the hostname of the computer you are working on, and set it accordingly if not correct. For the time being you can set the hostname by using the hostname command again as: [root@Server /]# hostname Server
[root@Server /]# hostname Server.mydomain.com 2) Check name resolution [root@Server /]# vi /etc/hosts (Client & Server both) In case DNS is being used for name resolution, edit this file and remove any IP and names of computers which are part of this domain / network. For the machine Server, you can leave only the following line in it: 127.0.0.1 localhost.localdomain localhost 192.168.1.254 Server.mydomain.com Server (Interpret this file as IP then machine name then fully qualified domain name then alias or nickname) In case DNS is not being used for name resolution, edit this file and put in IP and names of computers which are part of this domain / network. Remember! In such a case this file needs to be replicated on all the *nix based machines on this network. For the machine Server and wks1, you can have the following lines in this file: 127.0.0.1 localhost.localdomain localhost 192.168.1.254 Server.mydomain.com Server 192.168.1.1 wks1.mydomain.com wks1 3) Check for name resolution order [root@Server /]# vi /etc/host.conf (Client & Server both) In case DNS is being used primarily for name resolution, edit this file and add the following line if not exist: order bind, hosts In case DNS is not being used primarily for name resolution rather local /etc/hosts file is the primary name resolver then edit this file and add the following line if not exist: order hosts, bind Sometimes you will see a multi on line in this file. This means that any particular host in question can have more than one IP addresses in the DNS file. For example in multi-homed computers, which have more then one network interfaces, like firewall machines or routers, etc.
4) Who is the name server for each machine? [root@Server /]# vi /etc/resolv.conf (Client & Server) In case you are on the server machine, and the same machine (i.e. the server) is the DNS server, then edit this file and add the following: nameserver 127.0.0.1 In case you are on the client machine, and some other machine (server) is the DNS server, then edit this file and add the following: nameserver 192.168.1.254 Note: You can have multiple nameserver lines in a /etc/resolv.conf file. Like: nameserver 127.0.0.1 nameserver 192.168.1.254 nameserver 192.168.1.121 5) Check for boot time host name and gateway settings: [root@Server /]# vi /etc/sysconfig/network (Client & Server) This file contains your hostname so in case of server it should contain (the file on the server should look like): NETWORKING=yes HOSTNAME=Server.mydomain.com And the file on the clients must have a GATEWAY mentioned in it: NETWORKING=yes HOSTNAME=wks1.mydomain.com GATEWAY=192.168.1.254 You can also configure the hostname temporarily by using the command: [root@Server /]# hostname Server.mydomain.com 6) Check IP address: [root@Server /]# vi /etc/sysconfig/network-scripts/ifcfgeth0 (Client & Server)
This file contains your IP, which will be configured on machine boot up time. So it should contain a minimum of your IP address, network mask and network address, broadcast address, etc. DEVICE=eth0 BOOTPROTO=static IPADDR=192.168.1.254 NETMASK=255.255.255.0 ONBOOT=yes 7) Now restart the network if required, by following commands or just restart your computer: [root@Server /]# /etc/rc.d/init.d/network restart (Client & Server) Or [root@Server /]# service network restart (Client & Server)
Check the output of ifconfig command after restarting the network. If it doesn't show the new IP which you just gave in the configuration files then this means that your computer is picking up IP from somewhere else. But from where? The answer is most probably from the /etc/sysconfig/networking/devices/ifcfg-eth0 file. Actually on some Linux installations or in some situations, the /etc/sysconfig/networkscripts/ifcfg-eth0 file is a symbolic link to /etc/sysconfig/networking/devices/ifcfg-eth0 file. If this link breaks, then weird things happen. Like, you edit the network-scripts/ifcfgeth0 file but your system picks up the old IP from no where (actually from networking/devices/ifcfg-eth0 file). What to do? Heal the broken hearts. How? Well, change directory to /etc/sysconfig/network-scripts. Copy the ifcfg-eth0 file, which you think is the correct one, to the /etc/sysconfig/networking/devices directory. Or the contrary if you think that the file in networking/devices/ifcfg-eth0 file is the one that should be loaded then delete any ifcfg-eth0 file from the /etc/sysconfig/network-scripts directory. Then create the symbolic link by: [root@Server /]# cd /etc/sysconfig/network-scripts/ [root@Server network-scripts]# ln -s ../networking/devices/ifcfg-eth0 Or
[root@Server /]# ln -s /etc/sysconfig/networking/devices/ifcfg-eth0 (Space) /etc/sysconfig/network-scripts/ifcfg-eth0 Now restart the network again and check the results by ifconfig command. This time it should be ok. Then issue the hostname, ifconfig and route commands to check the changes: [root@Server /]# ifconfig [root@Server /]# route [root@Server /]# hostname If ifconfig doesnt show the desired IP, then try assigning the IP manually for the time being by the command: [root@wks1 /]# ifconfig eth0 192.168.1.254 netmask 255.255.255.0 up If ifconfig command gives any errors like FILE EXISTS, etc, then stop the network first, configure IP by ifconfig command and then again start the network: [root@wks1 /]# service network stop [root@wks1 /]# ifconfig eth0 192.168.1.1 netmask 255.255.255.0 up This time the ifconfig command (without any other parameters), should show you the new IP. The route command should show a default gateway in the routing table only on client computers. It should not show a default gateway for the server, which is not yet connected to any other network. Remember that your loopback interface should also be up for your computer to work correctly. You can check this by ifconfig command without any parameters. [root@wks1 /]# ifconfig And you can turn it up by: [root@wks1 /]# ifup lo Or [root@wks1 /]# ifconfig lo 127.0.0.1 netmask 255.0.0.0 up
At this point you should be able to ping your own computer: [root@wks1 /]# ping 127.0.0.1 [root@wks1 /]# ping 192.168.1.1 If this is correct and the above two commands give you echo reply, then you should also be able to ping gateway IP. [root@wks1 /]# ping 192.168.1.254 If you can ping gateway IP, then it means you should be able to ping any other computer in the network. If other computers are also configured then you should be able to ping them with their respective IPs. Pinging the name e.g. Server.mydomain.com or wks1.mydomain.com will work only if name resolution is working properly, either through hosts files or if DNS is configured on the network server and is running. [root@Server /]# ping Server.mydomain.com [root@Server /]# ping wks1.mydomain.com Note: If ping or any other service like telnet or ssh results into Destination Host Unreachable then check the cable connection and IP addresses on both client and server. Also check for firewall blocking ICMP packets. If it complaints with Unknown host then there is problem with the name resolution. In that case check the hosts file if DNS is not running or not configured. Also check /etc/resolv.conf and other files if DNS is configured. If ping or any other service comes up with Network Unreachable then your computer is not on the same network as the target computer. Further checking during network trouble shooting: 1) Is the network wiring / cable connections / cable configurations correct? You have to be sure that the cabling schemes / wiring is correct and according to some standard. For this you can consult common wiring layouts and cable configurations attached at the end of this book as Appendix A. 2) Is the software package(s) installed? It may happen that during a custom or workstation install you forgot to install the particular packages required by the service. You can use rpm commands to check for the existence of a package. For example you are trying to establish a telnet or secure shell session with another machine and you get a connection refused or Destination host unreachable types of errors. The first error means that the telnet service is not running
on the target computer. It is either not installed at all or is stopped or is being blocked by some firewall. The second error shows problem with your name resolution or physical connection problem, means 3) Does the Firewall permit access to the service? Normally during the installation of Red hat 7.2, 7.3, 8 and 9, you are asked for a firewall. You might have selected High or Medium firewall security option. In this case most of the services will be blocked. The details are: In case of High security the following will happen: No new connections will be allowed targeted to your machines Only DNS and DHCP replies will be allowed by this machine if DNS server (named) and DHCP (dhcpd) are running on this machine Active mode FTP to this machine is not allowed IRC DCC file transfers are not allowed Real Audio and Remote X windows client sessions are not allowed NIS & LDAP replies are not allowed In case of Medium security the following will happen: Access to ports below 1023 will not be allowed NFS and X Font Server and Remote X windows client sessions are not allowed Note: On a Red Hat system, the firewall script generated by firewallconfig Or redhat-config-securitylevel or Gnome Lokkit, is stored as /etc/sysconfig/iptables To temporarily disable a (default) firewall on a redhat 8 and above systems, use : [root@Server /]# iptables -F [root@Server /]# iptables -t nat -F In a RedHat 7.2 system, where ipchains is the default software for firewall implementation, use: [root@Server /]# ipchains -F Also if you want to use iptables instead of ipchains on a RedHat 7.2 system, you have to remove / unload the ipchains module:
[root@Server /]# rmmod ipchains 4) Is the startup script set up correctly to automatically launch the service? Normally you can check this by simply using the setup or ntsysv or chkconfig command or the chkconfig command: [root@Server /]# ntsysv Or [root@Server /]# setup Or [root@Server /]# chkconfig -list | less And if a service (say sendmail) is not configured to start automatically for a run-level (say run-level 3), you can do so by: [root@Server /]# chkconfig -level 3 sendmail on 5) Is the configuration file created for the service? Service might be running but its related configuration files may not be configured properly. eg. For sendmail, the main configuration file is /etc/mail/sendmail.cf. 6) Does the configuration file permit proper access to the service? This is sometimes called service level access control / firewall. In this case some computers / addresses might be restricted not to access the service. For example for sendmail, the access file (normally) is /etc/mail/access, /etc/mail/relay-domains. 7) Are there any other restrictions to the service being shared? Some times standard Linux security features may block access to a particular service which might otherwise be configured and running. E.g. NFS is running in properly configured state but the local file permissions are not allowing users to access the files made available through NFS. Also name resolution, specifically missing reverse lookup PTR entries in the reverse zone file to the corresponding A type entries in the forward zone file of DNS sometimes results in denial of access.
Chapter 4: DNS
BIND (Berkley Internet Naming Daemon) in *nix is a tool to implement DNS (Domain Name Service). The following procedure should be adapted for implementation and troubleshooting of BIND. Packages: For BIND the following packages need to be installed on the server: caching-nameserver-a.b bind-a.b bind-utils-a.b Where a and b represent major and minor version numbers respectively. No special packages on the client is required to be a part of DNS.
Scenario:
In the example network above, we need to implement DNS. The name of the domain will be mydomain.com. See figure. Configuration: The following files need to be present only on the server, before starting name daemon: /etc/named.conf /var/named/localhost.zone (optional) /var/named/0.0.127.in-addr.arpa.zone /var/named/mydomain.com.zone /var/named/1.168.192.in-addr.arpa.zone /var/named/named.ca (this file should exist even if zero byte in size) And amongst them, the following will need configuration: /etc/named.conf /var/named/mydomain.com.zone /var/named/1.168.192.in-addr.arpa.zone 1) First edit the /etc/named.conf file: [root@Server /]# vi /etc/named.conf ## named.conf - configuration for bind # # Generated automatically by bindconf, alchemist et al. controls { inet 127.0.0.1 allow { localhost; } keys { rndckey; }; }; include "/etc/rndc.key"; options { directory "/var/named/"; }; zone "." { type hint; file "named.ca"; }; zone "0.0.127.in-addr.arpa" { type master; file "0.0.127.in-addr.arpa.zone"; allow-update { none; }; };
zone "1.168.192.in-addr.arpa" { type master; file "1.168.192.in-addr.arpa.zone"; allow-update { none; }; }; zone "localhost" { type master; file "localhost.zone"; allow-update { none; }; }; zone "mydomain.com" { type master; file "mydomain.com.zone"; allow-update { none; }; }; Note: Either in the options section (which has global effect), or in the zone section you can further tighten or loosen your security. Like you can use allow-update {192.168.1.0/24 ;}; line to allow DNS table updates by the clients of your example network. Helpful to DHCP servers and windows 2000 clients. Similarly you can use allow-query {192.168.1.0/24 ;}; to restrict the DNS queries to be asked by the clients of your example network only. And you can restrict the list of all of your machines (zone transfer) to the clients of your network only by allow-transfer {192.168.1.0/24 ;}; Note: If you ever happen to mess up with this file, just re-install cachingnameserver package and things should be fine. 2) Now create forward zone /var/named/mydomain.com.zone file for your domain: [root@Server /]# vi /var/named/mydomain.com.zone $TTL 86400 $ORIGIN mydomain.com. @ IN SOA Server.mydomain.com. xena.yahoo.com.uk.( 1 ; serial 28800 ; refresh 7200 ; retry 604800 ; expire 86400 ; ttl ) @ Server IN IN NS A Server.mydomain.com. 192.168.1.254
wks1 wks2 www
IN A IN A IN CNAME
192.168.1.1 192.168.1.2 Server
3) Next create the reverse zone file for your network: [root@Server /]# vi /var/named/1.168.192.in-addr.arpa.zone $TTL 86400 $ORIGIN 1.168.192.in-addr.arpa. @ IN SOA Server.mydomain.com. xena,yahoo.com.uk.( 1 ; serial 28800 ; refresh 7200 ; retry 604800 ; expire 86400 ; ttl ) @ 1 2 254 IN IN IN IN NS PTR PTR PTR Server.mydomain.com. wks1.mydomain.com. wks2.mydomain.com. Server.mydomain.com.
Note: There are no corresponding reverse entries in the reverse zone file for CNAME entries in forward zone file. Also note that you can write the PTR lines as: 1.1.168.192.in-addr.arpa. 2.1.168.192.in-addr.arpa. 254.1.168.192.in-addr.arpa. IN PTR wks1.mydomain.com. IN PTR wks2.mydomain.com. IN PTR Server.mydomain.com.
4) Check existence for /var/named/named.ca file: This file contains the Internet addresses of the root servers (currently 13) in the world so nothing to be messed with in this file. It looks like: [root@Server /]# cat /var/named/named.ca . . . . . . . . 518400 518400 518400 518400 518400 518400 518400 518400 IN IN IN IN IN IN IN IN NS NS NS NS NS NS NS NS K.ROOT-SERVERS.NET. L.ROOT-SERVERS.NET. M.ROOT-SERVERS.NET. I.ROOT-SERVERS.NET. E.ROOT-SERVERS.NET. D.ROOT-SERVERS.NET. A.ROOT-SERVERS.NET. H.ROOT-SERVERS.NET.
. 518400 IN NS C.ROOT-SERVERS.NET. . 518400 IN NS G.ROOT-SERVERS.NET. . 518400 IN NS F.ROOT-SERVERS.NET. . 518400 IN NS B.ROOT-SERVERS.NET. . 518400 IN NS J.ROOT-SERVERS.NET. K.ROOT-SERVERS.NET. 3600000 IN A 193.0.14.129 L.ROOT-SERVERS.NET. 3600000 IN A 198.32.64.12 M.ROOT-SERVERS.NET. 3600000 IN A 202.12.27.33 I.ROOT-SERVERS.NET. 3600000 IN A 192.36.148.17 E.ROOT-SERVERS.NET. 3600000 IN A 192.203.230.10 D.ROOT-SERVERS.NET. 3600000 IN A 128.8.10.90 A.ROOT-SERVERS.NET. 3600000 IN A 198.41.0.4 H.ROOT-SERVERS.NET. 3600000 IN A 128.63.2.53 C.ROOT-SERVERS.NET. 3600000 IN A 192.33.4.12 G.ROOT-SERVERS.NET. 3600000 IN A 192.112.36.4 F.ROOT-SERVERS.NET. 3600000 IN A 192.5.5.241 B.ROOT-SERVERS.NET. 3600000 IN A 128.9.0.107 J.ROOT-SERVERS.NET. 3600000 IN A 192.58.128.30 Actually, initially this file has the same contents as shown above, but in a different layout. This file shown above was derived / created / updated using: [root@Server /]# dig NS . @a.root-servers.net > /var/named/named.ca You should update this file regularly on your system running name server, by issuing the above command on periodic basis. 5) Update your /etc/resolv.conf file where the nameserver is to your machine. Further, this file /etc/resolv.conf, needs to be configured / adjusted on the server and all *nix clients, before starting name daemon. It looks like: [root@Server /]# vi /etc/resolv.conf nameserver 127.0.0.1 nameserver 192.168.1.254 Or domain mydomain.com nameserver 127.0.0.1 nameserver 192.168.1.254 Or search mydomain.com nameserver 127.0.0.1
nameserver 192.168.1.254 And on the client, it might / should look like: domain mydomain.com nameserver 192.168.1.254 Or search mydomain.com nameserver 192.168.1.254 Note: both search and domain keywords will not work at the same time if mentioned together in the same file. Now the configuration finishes here. Start or restart your name daemon by: [root@Server /]# /etc/rc.d/init.d/named start Or [root@Server /]# /etc/rc.d/init.d/named restart Or [root@Server /]# /etc/rc.d/init.d/named reload and check its status by : [root@Server /]# /etc/rc.d/init.d/named status Remember that once you change BIND configuration files, you can / should reload its database by: [root@Server /]# /etc/rc.d/init.d/named reload Log Files: The following log file can be checked on the server for detecting any problems relating to BIND. /var/log/messages Documentation and Help: Extensive Documentation is available in /usr/share/doc/bind-*
Testing : You can use nslookup for your testing (non interactive mode): [root@Server /]# nslookup mydomain.com Note: nslookup is deprecated and may be removed from future releases. Consider using the `dig' or `host' programs instead. Run nslookup with the `-sil[ent]' option to prevent this message from appearing. Server: 192.168.1.254 Address: 192.168.1.254#53 Name: mydomain.com Address: 192.168.1.254 Or start nslookup in interactive mode : [root@Server /]# nslookup Note: nslookup is deprecated and may be removed from future releases. Consider using the `dig' or `host' programs instead. Run nslookup with the `-sil[ent]' option to prevent this message from appearing. > mydomain.com Server: 192.168.1.254 Address: 192.168.1.254#53 Name: mydomain.com Address: 192.168.1.254 > set type=NS > mydomain.com Server: 192.168.1.254 Address: 192.168.1.254#53 mydomain.com nameserver = Server.mydomain.com. > Server (Note here: no NameServer defined inside Server) Server: 192.168.1.254 Address: 192.168.1.254#53
*** Can't find Server: No answer > set type=MX > mydomain.com Server: 192.168.1.254 Address: 192.168.1.254#53 mydomain.com mail exchanger = 10 Server. > set type=ANY > mydomain.com Server: 192.168.1.254 Address: 192.168.1.254#53 Name: mydomain.com Address: 192.168.1.254 mydomain.com origin = Server.mydomain.com. mail addr = root.Server.mydomain.com. serial = 1 refresh = 28800 retry = 7200 expire = 604800 minimum = 86400 mydomain.com nameserver = Server.mydomain.com. mydomain.com mail exchanger = 10 Server.mydomain.com. > set type=A > mydomain.com Server: 192.168.1.254 Address: 192.168.1.254#53 Name: mydomain.com Address: 192.168.1.254 > Server Server: 192.168.1.254 Address: 192.168.1.254#53 Name: Server.mydomain.com Address: 192.168.1.254 > wks1
Server: 192.168.1.254 Address: 192.168.1.254#53 Name: wks1.mydomain.com Address: 192.168.1.1 > wks2 Server: 192.168.1.254 Address: 192.168.1.254#53 Name: wks2.mydomain.com Address: 192.168.1.2 > bigboss (doesn't exist) ;; connection timed out; no servers could be reached > exit You can also run dig: [root@Server /]# dig mydomain.com ; <<>> DiG 9.1.3 <<>> mydomain.com ;; global options: printcmd ;; Got answer: ;; ->>HEADER<<- opcode: QUERY, status: NOERROR, id: 64144 ;; flags: qr aa rd ra; QUERY: 1, ANSWER: 1, AUTHORITY: 1, ADDITIONAL: 0 ;; QUESTION SECTION: ;mydomain.com. IN A ;; ANSWER SECTION: mydomain.com. 86400 IN A 192.168.1.254 ;; AUTHORITY SECTION: mydomain.com. 86400 IN NS Server.mydomain.com. ;; ;; ;; ;; Query time: 17 msec SERVER: 192.168.1.254#53(192.168.1.254) WHEN: Fri Jul 19 14:55:35 2002 MSG SIZE rcvd: 69
[root@Server /]# dig Server.mydomain.com ; <<>> DiG 9.2.1 <<>> Server.mydomain.com ;; global options: printcmd
;; Got answer: ;; ->>HEADER<<- opcode: QUERY, status: NOERROR, id: 43728 ;; flags: qr aa rd ra; QUERY: 1, ANSWER: 1, AUTHORITY: 1, ADDITIONAL: 0 ;; QUESTION SECTION: ;Server.mydomain.com. IN A ;; ANSWER SECTION: Server.mydomain.com. 10800 IN A 192.168.1.254 ;; AUTHORITY SECTION: mydomain.com. 10800 IN NS Server.mydomain.com. ;; ;; ;; ;; Query time: 120 msec SERVER: 192.168.1.254#53(192.168.1.254) WHEN: Mon Jul 14 00:39:37 2003 MSG SIZE rcvd: 63
To check IP of any machine by querying a proper nameserver you can use: [root@Server /]# dig @Server.mydomain.com wks2.mydomain.com A ; <<>> DiG 9.2.1 <<>> @Server.mydomain.com wks2.mydomain.com A ;; global options: printcmd ;; Got answer: ;; ->>HEADER<<- opcode: QUERY, status: NOERROR, id: 34767 ;; flags: qr aa rd ra; QUERY: 1, ANSWER: 1, AUTHORITY: 1, ADDITIONAL: 1 ;; QUESTION SECTION: ;wks2.mydomain.com. IN A ;; ANSWER SECTION: wks2.mydomain.com. 10800 IN A 192.168.1.2 ;; AUTHORITY SECTION: mydomain.com. 10800 IN NS Server.mydomain.com. ;; ADDITIONAL SECTION: Server.mydomain.com. 10800 IN A 192.168.1.254 ;; ;; ;; ;; Query time: 2 msec SERVER: 192.168.1.254#53(Server.mydomain.com) WHEN: Mon Jul 14 00:43:38 2003 MSG SIZE rcvd: 84
Checking for a machine IP which does not exist, will return in: ANSWER:0 [root@Server /]# dig @Server.mydomain.com bigboss.mydomain.com A ; <<>> DiG 9.2.1 <<>> @Server.mydomain.com bigboss.mydomain.com A ;; global options: printcmd ;; Got answer: ;; ->>HEADER<<- opcode: QUERY, status: NXDOMAIN, id: 49812 ;; flags: qr aa rd ra; QUERY: 1, ANSWER: 0, AUTHORITY: 1, ADDITIONAL: 0 ;; QUESTION SECTION: ;bigboss.mydomain.com. IN A ;; AUTHORITY SECTION: mydomain.com. 3600 IN SOA Server.mydomain.com. (space) root.Server.mydomain.com. 1 10800 3600 604800 3600 ;; ;; ;; ;; Query time: 35 msec SERVER: 192.168.1.254#53(Server.mydomain.com) WHEN: Mon Jul 14 00:45:50 2003 MSG SIZE rcvd: 82
Now check for IP to name resolution (reverse name lookups): [root@Server /]# dig @Server.mydomain.com 2.1.168.192.inaddr.arpa PTR ; <<>> DiG 9.2.1 <<>> @Server.mydomain.com 2.1.168.192.inaddr.arpa PTR ;; global options: printcmd ;; Got answer: ;; ->>HEADER<<- opcode: QUERY, status: NOERROR, id: 10703 ;; flags: qr aa rd ra; QUERY: 1, ANSWER: 1, AUTHORITY: 1, ADDITIONAL: 1 ;; QUESTION SECTION: ;2.1.168.192.in-addr.arpa. IN PTR ;; ANSWER SECTION: 2.1.168.192.in-addr.arpa. 10800 IN PTR wks2.mydomain.com. ;; AUTHORITY SECTION: 1.168.192.in-addr.arpa. 10800 IN NS Server.mydomain.com.
;; ADDITIONAL SECTION: Server.mydomain.com. 10800 IN A 192.168.1.254 ;; ;; ;; ;; Query time: 33 msec SERVER: 192.168.1.254#53(Server.mydomain.com) WHEN: Mon Jul 14 00:50:42 2003 MSG SIZE rcvd: 106
This can also be done by the following command: [root@Server /]# dig -x 192.168.1.2 ; <<>> DiG 9.2.1 <<>> -x 192.168.1.2 ;; global options: printcmd ;; Got answer: ;; ->>HEADER<<- opcode: QUERY, status: NOERROR, id: 2549 ;; flags: qr aa rd ra; QUERY: 1, ANSWER: 1, AUTHORITY: 1, ADDITIONAL: 1 ;; QUESTION SECTION: ;2.1.168.192.in-addr.arpa. IN PTR ;; ANSWER SECTION: 2.1.168.192.in-addr.arpa. 86400 IN PTR wks2.mydomain.com. ;; AUTHORITY SECTION: 1.168.192.in-addr.arpa. 86400 IN NS Server.mydomain.com. ;; ADDITIONAL SECTION: Server.mydomain.com. 86400 IN A 192.168.1.254 ;; ;; ;; ;; Query time: 5 msec SERVER: 127.0.0.1#53(127.0.0.1) WHEN: Tue Feb 17 01:49:53 2004 MSG SIZE rcvd: 114
You can get a list of all the machines of a domain by: [root@Server /]# dig axfr mydomain.com ; <<>> DiG 9.2.1 <<>> mydomain.com axfr ;; global options: printcmd mydomain.com. 86400 IN SOA Server.mydomain.com. (space) root.Server.mydomain.com. 1 28800 7200 604800 86400 mydomain.com. 86400 IN NS Server.mydomain.com.
Server.mydomain.com. 86400 IN A 192.168.1.254 wks1.mydomain.com. 86400 IN A 192.168.1.1 wks2.mydomain.com. 86400 IN A 192.168.1.2 mydomain.com. 86400 IN SOA Server.mydomain.com. (space) root.Server.mydomain.com. 1 28800 7200 604800 86400 ;; ;; ;; ;; Query time: 6 msec SERVER: 127.0.0.1#53(127.0.0.1) WHEN: Tue Feb 17 01:50:32 2004 XFR size: 7 records
You can query the MX , SOA and NS records by using : [root@Server /]# dig MX mydomain.com ; <<>> DiG 9.2.1 <<>> MX mydomain.com ;; global options: printcmd ;; Got answer: ;; ->>HEADER<<- opcode: QUERY, status: NOERROR, id: 1644 ;; flags: qr aa rd ra; QUERY: 1, ANSWER: 1, AUTHORITY: 1, ADDITIONAL: 1 ;; QUESTION SECTION: ;mydomain.com. IN MX ;; ANSWER SECTION: mydomain.com. 86400 IN MX 10 Server.mydomain.com. ;; AUTHORITY SECTION: mydomain.com. 86400 IN NS Server.mydomain.com. ;; ADDITIONAL SECTION: Server.mydomain.com. 86400 IN A 192.168.1.254 ;; ;; ;; ;; Query time: 57 msec SERVER: 127.0.0.1#53(127.0.0.1) WHEN: Sat Mar 13 14:50:01 2004 MSG SIZE rcvd: 87
[root@Server /]# dig SOA mydomain.com ; <<>> DiG 9.2.1 <<>> SOA mydomain.com ;; global options: printcmd ;; Got answer: ;; ->>HEADER<<- opcode: QUERY, status: NOERROR, id: 64713
;; flags: qr aa rd ra; QUERY: 1, ANSWER: 1, AUTHORITY: 1, ADDITIONAL: 1 ;; QUESTION SECTION: ;mydomain.com. IN SOA ;; ANSWER SECTION: mydomain.com. 86400 IN SOA Server.mydomain.com. (space) root.Server.mydomain.com. 1 28800 7200 604800 86400 ;; AUTHORITY SECTION: mydomain.com. 86400 IN NS Server.mydomain.com. ;; ADDITIONAL SECTION: Server.mydomain.com. 86400 IN A 192.168.1.254 ;; ;; ;; ;; Query time: 5 msec SERVER: 127.0.0.1#53(127.0.0.1) WHEN: Sat Mar 13 14:51:37 2004 MSG SIZE rcvd: 112
[root@Server /]# dig NS mydomain.com ; <<>> DiG 9.2.1 <<>> NS mydomain.com ;; global options: printcmd ;; Got answer: ;; ->>HEADER<<- opcode: QUERY, status: NOERROR, id: 26538 ;; flags: qr aa rd ra; QUERY: 1, ANSWER: 1, AUTHORITY: 0, ADDITIONAL: 1 ;; QUESTION SECTION: ;mydomain.com. IN NS ;; ANSWER SECTION: mydomain.com. 86400 IN NS Server.mydomain.com. ;; ADDITIONAL SECTION: Server.mydomain.com. 86400 IN A 192.168.1.254 ;; ;; ;; ;; Query time: 22 msec SERVER: 127.0.0.1#53(127.0.0.1) WHEN: Sat Mar 13 14:54:22 2004 MSG SIZE rcvd: 71
Ping Testing :
Ping is basically used for physical connectivity test. It has nothing to do with DNS directly. [root@Server /]# ping Server.mydomain.com
PING Server.mydomain.com (192.168.1.254) from 192.168.1.254 : bytes of data. Warning: time of day goes back, taking countermeasures. 64 bytes from Server.mydomain.com (192.168.1.254): icmp_seq=0 time=685 usec 64 bytes from Server.mydomain.com (192.168.1.254): icmp_seq=1 time=91 usec 64 bytes from Server.mydomain.com (192.168.1.254): icmp_seq=2 time=91 usec --- Server.mydomain.com ping statistics --3 packets transmitted, 3 packets received, 0% packet loss round-trip min/avg/max/mdev = 0.091/0.289/0.685/0.280 ms 56(84) ttl=255 ttl=255 ttl=255
[root@Server /]# ping wks1.mydomain.com

PING wks1.mydomain.com (192.168.1.1) from 192.168.1.254 : of data. Warning: time of day goes back, taking countermeasures. 64 bytes from wks1.mydomain.com (192.168.1.1): icmp_seq=0 time=1.121 msec 64 bytes from wks1.mydomain.com (192.168.1.1): icmp_seq=1 time=516 usec 64 bytes from wks1.mydomain.com (192.168.1.1): icmp_seq=2 time=506 usec --- wks1.mydomain.com ping statistics --3 packets transmitted, 3 packets received, 0% packet loss round-trip min/avg/max/mdev = 0.506/0.714/1.121/0.288 ms 56(84) bytes ttl=255 ttl=255 ttl=255
You should also be able to ping from clients to server and other clients without their local /etc/hosts file having any entries. Note: In windows computers, Network Configuration -> DNS -> Host means the same computer (windows computer) you are sitting at, NOT the server. And DNS means the DNS domain name implemented on the network.
Terminology / abbreviations used in named zone configuration files: SOA: Start Of Authority. The SOA record indicates that this name server is the best source of information for the data within this zone. There can be one and only one SOA record in the zone data file. IN:
IN stands for Internet address. This is one class of address. Others exist but are not in widespread use. Used to assign / specify IP address (A) or nameserver (NS) or pointer (PTR) or Canonical Name (CNAME) specification to machine names. NS: NS stands for NameServer and is used to specify the name of the name server of the zone. mydomain.com. IN NS Server.mydomain.com. Or @ IN NS Server.mydomain.com. MX: MX stands for Mail Exchanger. Suppose we want to use the machine Server.mydomain.com as our mail exchanger server. Then the following configuration needs to be done in the mydomain.com.zone file: @ IN NS Server.mydomain.com. @ IN MX 10 Server.mydomain.com. In the above line having MX in it, the number 10 is the mail delivery priority. It can be any positive number i.e. the smaller the number the higher the priority. Remember that there are no corresponding reverse entries in the reverse zone file for the MX entries in the forward zone files. A: (Address) Indicates / specifies the IP address of the machine. wks1.mydomain.com. IN A 192.168.1.1 PTR: PTR Indicates Pointer. Means what name the IP address points to. The name of the machines in the PTR record must be fully qualified domain name. 1.1.168.192.in-addr.arpa. IN PTR wks1.mydomain.com. Or 1 IN PTR wks1.mydomain.com. Aliases / CNAME: All names provided in the zone files having A record type are in fact canonical names. . E.g. in the following line wks1 is canonical name. wks1 IN A 192.168.1.1
Once we want to use another name for a machine which is already assigned a name then we use aliases. Suppose we want to use our Server as www as well then we will use the following lines in our mydomain.com.zone file: wks1 Server www IN IN IN A A CNAME 192.168.1.1 192.168.1.254 Server
Appending domain names: The domain name is the Origin of all the data in the zone files. Since the origin is appended to names, instead of entering the following line in mydomain.com.zone: wks1.mydomain.com. IN A 192.168.1.1 We will write it as: wks1 IN A 192.168.1.1 So the domain name mydomain.com will automatically be appended to the computer name which does not end with a period (. dot). Also instead of following line in file 1.168.192.in-addr.arpa.zone: 1.1.168.192.in-addr.arpa. IN PTR wks1.mydomain.com. We will write as: 1 IN PTR wks1.mydomain.com. The @ Notation: If the domain name is same as origin, the name can be specified as @ . This is most often seen in SOA records. So the following line in the mydomain.com.zone file, mydomain.com. IN SOA Server root.Server ( Can be written as: @ IN SOA Server root.Server (
TTL:(Time To Live) The TTL on a resource record is the time in seconds any server can cache that record. So if the TTL for a particular resource record is 3600 (seconds), and a server outside your
domain caches that record, it will have to remove the entry from its cache after an hour. If it needs the same data after the hour is up, it'll have to query your name servers again. You can have a separate TTL for SOA resource record and a separate one for your zone data in the A type resource records. Compare these two below: $TTL 86400 $ORIGIN mydomain.com. @ IN SOA Server.mydomain.com. 1 ; serial 28800 ; refresh 7200 ; retry 604800 ; expire 86400 ; minimum ttl ) wks1 3600 IN A 192.168.1.1 wks2 IN A 192.168.1.2 This example is showing that the default TTL for the entire zone is 86400 seconds (or 24 hours). The TTL for the SOA record is also 86400 seconds. The TTL for wks1 is 3600 seconds (or 1 hour). And the TTL for wks2 is not specified. This entire scenario means that any other name server querying our nameserver will keep the queried information in it's cache for 24 hours by default; and will come to us again only when 24 hours have been passed since last query. This is true for the SOA information or information about wks2. But if the queried information was about wks1, then the other nameserver will have to lookup again to our server if 3600 seconds (one hour) has been passed since last query performed for wks1. Refresh: IN case master-slave DNS servers are implemented, this directive (refresh interval) tells the slave: how often to check that its data is up to date? Retry: This is also used in master-slave DNS server implementation. If the slave fails to reach the master nameserver(s) after the refresh period (the host(s) could be down), then it starts trying to connect every retry seconds. Normally, the retry interval is shorter than the refresh interval, but it doesn't have to be. Expire: If the slave fails to contact the master server(s) tills the time specified for session expiring then the slave will expires its data. Expiring the data means the slave stops giving out answers about the data because the data are too old to be useful. Essentially, this field says: at some point, the data are so old that having no data is better than having stale data. The expiration time should always be much larger than the retry and refresh intervals; if
root.Server (
the expire time is smaller than the refresh interval, your slaves will expire their data before trying to load new data.
Chapter 5 : NFS
NFS stands for Network File System. Used for file sharing between *nix machines only. If you want to share files between Windows and UNIX machines then you need to use SAMBA for that. Packages: Following packages needs to be installed on all those machines on which directories need to be shared. nfs-utils-a.b portmap-a.b Case Study:
Fig 1 We need to share the following directories on the server for the entire domain
mydomain.com. /home to be shared with read-write permissions for the entire lan, /usr/project project directory of some developers. /usr2/redhat72 contains all files copied from Redhat 7.2 installation CDs with read-only permissions.
Server side configuration: Following file needs to be configured on the machine on which you want to share the directories. [root@Server /]# /etc/exports The format of this file is: DirectoryPath hostlist(permissions) Where path is the absolute path of the directory you want to share and host list can be one or more hosts that can access the share. Each host is separated by a white space. Hosts can be specified in the following ways: By using a hostname, if the host is part of local domain. e.g. wks1 By using a fully qualified domain name. e.g. wks1.mydomain.com By using a hostname containing a wild card. e.g. *.mydomain.com By using IP address. e.g. 192.168.1.1 By using a network address. e.g. 192.168.1.0/255.255.255.0 or 192.168.1.0/24 And permissions can be (ro) for read only and (rw) for read & write. DO NOT specify any white space between host specifier and permissions. Also w.r.t permissions, by default, if a client is logged in as root on his own machine, then he will NOT be given root permissions on the machine hosting the shares ( he will be squashed !) . Instead, such requests are mapped to user id 65535 (which of course doesn't exist on a normal system). If you want to permit remote root acting as local root on the machine holding shares, then you need to enable a special permission as (no_root_squash).e.g. /etc wks1.mydomain.com(rw,no_root_squash) This line means that if root logs in from wks1.mydomain.com, and he tries to write some files in the /etc directory of the computer holding the shares, then he will be ALLOWED to do all the editing to the files which are writeable by root of local machine. So for now the file looks like this:
[root@Server /]# cat /etc/exports /usr2/redhat72 192.168.1.0/24 (ro,no_root_squash) /home *.mydomain.com(rw) /usr/project *.mydomain.com(rw) Once /etc/exports file is edited, you need to run: [root@Server /]# /etc/rc.d/init.d/portmap restart [root@Server /]# /etc/rc.d/init.d/nfs start Client side configuration: Whatever the name resolution method being used in the domain, it should be working correctly / configured correctly else NFS mount will normally fail and will also give weird messages. If you occasionally need to access the shared directories on the server, then there is nothing to be configured on the client side. You just need to give the following command each time to want to access a share on the server. [root@Server /]# mount -t nfs Server.mydomain.com:/usr2/redhat72 /mnt/cdrom
The above line means that /usr2/redhat72 directory from the server Server.mydomain.com will be mounted on a local directory (mount point) /mnt/cdrom using nfs protocol, or simply nfs file-system. And if you need to mount the share regularly or on every boot, then probably you should make a separate mount point for it on your local machine. e.g. You are working on a project with your team, and the directory /usr/project is on the nfs server Server.mydomain.com, then you should make a mount point in your local work folder as /work/project (note that the directory being used as a mount point should always be empty) and edit your /etc/fstab file and put the following line in it. Server.mydomain.com:/usr/project /work/project nfs defaults 0 0 The script /etc/rc.d/init.d/netfs runs at boot time, automatically mounting the NFS directories specified in the /etc/fstab file. Also the nfslock service provides file locking services for NFS clients. So enable it if the client needs to lock files. Log files: /var/log/messages Documentation and help: www.tldp.org/HOWTO/NFS-HOWTO /usr/share/doc/nfs-utils-* /usr/share/doc/portmap-*
Testing: To check if the NFS service is working properly the tests are performed from the server and client side, both. Testing from server: Check that at least the following three services are shown in the rpcinfo command: portmapper, mountd and nfs. If any one of them is missing then NFS will not work. Give the following command on the server, to see a list of services responding to remote procedure calls. [root@Server /]# rpcinfo -p program vers 100000 2 tcp 100000 2 udp 100024 1 udp 100024 1 tcp 391002 2 tcp 100011 1 udp 100011 2 udp 100011 1 tcp 100011 2 tcp 100003 2 udp 100003 3 udp 100021 1 udp 100021 3 udp 100021 4 udp 100005 1 udp 100005 1 tcp 100005 3 udp 100005 3 tcp proto port 111 portmapper 111 portmapper 1024 status 1024 status 1025 sgi_fam 836 rquotad 836 rquotad 839 rquotad 839 rquotad 2049 nfs 2049 nfs 1026 nlockmgr 1026 nlockmgr 1026 nlockmgr 1027 mountd 1028 mountd 1027 mountd 1028 mountd
Now check if the directories you put in /etc/exports are now available as shares or not by: [root@Server /]# exportfs -v /usr2/redhat72 *.mydomain.com(ro,wdelay,no_root_squash) /home *.mydomain.com(rw,wdelay,root_squash) /usr/project *.mydomain.com(rw,wdelay,root_squash) And also by: [root@Server /]# showmount -e Export list for Server:
/home *.mydomain.com /usr2/redhat72 *.mydomain.com(ro,wdelay,no_root_squash) /usr/project *.mydomain.com(rw,wdelay,root_squash) If you have made NFS shares available to the entire LAN then you should be able to mount the shares on the server itself, in some other directory / mount point: [root@Server /]# mount -t nfs Server.mydomain.com:/usr2/redhat72 /mnt/cdrom Note: If you edit your /etc/exports file later, then you need to run exportfs command again with a -r switch to refresh the share list. i.e. [root@Server /]# exportfs -r
Client side testing and mounting:

You should be able to get rpcinfo from the server, which should be same as the command given on the server: [root@Server /]# rpcinfo -p Server.mydomain.com program vers 100000 2 tcp 100000 2 udp 100024 1 udp 100024 1 tcp 391002 2 tcp 100011 1 udp 100011 2 udp 100011 1 tcp 100011 2 tcp 100003 2 udp 100003 3 udp 100021 1 udp 100021 3 udp 100021 4 udp 100005 1 udp 100005 1 tcp 100005 2 udp 100005 2 tcp 100005 3 udp 100005 3 tcp proto port 111 portmapper 111 portmapper 1024 status 1024 status 1025 sgi_fam 836 rquotad 836 rquotad 839 rquotad 839 rquotad 2049 nfs 2049 nfs 1026 nlockmgr 1026 nlockmgr 1026 nlockmgr 1027 mountd 1028 mountd 1027 mountd 1028 mountd 1027 mountd 1028 mountd
Now check which shares are available on the server by:
[root@Server /]# showmount -e Server.mydomain.com Export list for Server: /home *.mydomain.com /usr2/redhat72 *.mydomain.com /usr/project *.mydomain.com Once shares are visible then, you are ready to mount the files on the local mount point. e.g.: [root@Server /]# mount -t nfs Server.mydomain.com:/usr/project /work/project If the above command gives the prompt back, quietly, then it means that mount was successful. If mount was not successful, then check your NFS server side settings, and most importantly, the IP and hostname of the client computer, where you are trying to mount the server shares. If the IP is not of the same subnet, then your host will not be in the same domain as per defined in DNS table on the server side. So NFS mount won't work. Also if your hostname is not correct, i.e. not defined in the DNS zone file or reverse entry for a computer is missing in the reverse zone file, then again your mount request will be declined. You may get RPC time out type of errors in such cases. Also note the error caught in the /var/log/messages on the NFS server, below: Apr 2 00:40:52 Server rpc.mountd: authenticated mount request from klaptop.mydomain.com:952 for /data (/data) Apr 2 00:40:52 Server rpc.mountd: getfh failed: Operation not permitted This error is encountered when you will try to mount a remote (exported) share over NFS, which is in-fact a directory, mount point. E.g. the above error came by giving the following command on the client side: [root@Server /]# mount -t nfs 192.168.1.254:/data /mnt/nfs In my case, the mount command (on the server side), showed that /data is a mount point and /dev/hda2 is mounted on it. So a directory under /data, like /data/rh9nfs should be exported via NFS instead of /data and that directory in turn can be used / mounted by remote clients. [root@Server /]# mount /dev/hda6 on / type ext3 (rw) none on /proc type proc (rw) usbdevfs on /proc/bus/usb type usbdevfs (rw)
/dev/hda1 on /boot type ext3 (rw) /dev/hda2 on /data type ext3 (rw) none on /dev/pts type devpts (rw,gid=5,mode=620) none on /dev/shm type tmpfs (rw) When some other (exported) directory under the hierarchy of this mount point, is mounted, then it is mounted properly: Apr 2 00:41:01 Server rpc.mountd: request from klaptop.mydomain.com:953 for (/data/rh9nfs) Apr 2 00:42:29 Server rpc.mountd: request from klaptop.mydomain.com:966 for (/data/rh9nfs) authenticated mount /data/rh9nfs authenticated unmount /data/rh9nfs
Chapter 6: NIS
Network Information Service is formerly known as Yellow Pages or YP. NIS is used to provide login authentication to *nix machines. It is NOT used to provide login authentication for windows clients, for that we use SAMBA. Packages: ypserv-a.b (NIS server side) ypbind-a.b (NIS client side) yp-tools-a.b (Client server both) Case Study: In the example network above, we need to implement NIS. NIS domain is always different from the DNS domain and these two should never be the same name i.e. you cannot use the DNS mydomain.com in this case for NIS, the name of the authentication domain for NIS on our network will be logindomain. Configuration: Server side configuration: The following steps need to be configured / adjusted on the server: 1) Set domain name of this domain by using ypdomainname command [root@Server /]# ypdomainname logindomain
2) Start ypserv daemon by: [root@Server /]# /etc/rc.d/init.d/ypserv start 3) Go to /var/yp directory and create a new directory named logindomain [root@Server /]# cd /var/yp [root@Server yp]# mkdir logindomain
4) Now run /usr/lib/yp/ypinit to generate maps here, DO NOT enter the directory logindomain yourself. Press Ctrl-D in response to next host to add: and press Y in response to the question Is this correct? [y/n: y] : [root@Server yp]# /usr/lib/yp/ypinit -m At this point, we have to construct a list of the hosts which will run NIS servers. Server is in the list of NIS server hosts. Please continue to add the names for the other hosts, one per line. When you are done with the list, type a <control D>. next host to add: Server.mydomain.com next host to add: The current list of NIS servers looks like this:
Server.mydomain.com Is this correct? [y/n: y] y We need a few minutes to build the databases... Building /var/yp/logindomain/ypservers... Running /var/yp/Makefile... gmake[1]: Entering directory `/var/yp/logindomain' Updating passwd.byname... Updating passwd.byuid... Updating group.byname... Updating group.bygid... Updating hosts.byname... Updating hosts.byaddr... Updating rpc.byname... Updating rpc.bynumber... Updating services.byname... Updating services.byservicename... Updating netid.byname... Updating protocols.bynumber... Updating protocols.byname... Updating mail.aliases... gmake[1]: Leaving directory `/var/yp/logindomain' Server.mydomain.com has been set up as a NIS master server. Now you can run ypinit -s Server.mydomain.com on all slave servers. 5) It will make the maps of user names and passwords, etc. If it gives any errors during this, repeat the steps from step 1. 6) yppasswdd daemon is used to listen for password change requests from client computers and updates the /etc/passwd and /etc/shadow files on the NIS server. Start it by: [root@Server /]# /etc/rc.d/init.d/yppasswdd start 7) You should also make sure that these services ( ypserv and yppasswdd) are started by default at boot up time. So run setup or ntsysv and enable these services. [root@Server /]# setup [*] ypserv [*] yppasswdd Or [root@Server /]# chkconfig -level 35 ypserv on
[root@Server /]# chkconfig -level 35 yppasswdd on 8) Update /etc/sysconfig/network file and add the line at the end, [root@Server /]# vi /etc/sysconfig/network NETWORKING=yes HOSTNAME=Server.mydomain.com NISDOMAIN=logindomain Client side configuration: 1) Modify the /etc/sysconfig/network and add the line at the end: NISDOMAIN=logindomain [root@Server /]# vi /etc/sysconfig/network NETWORKING=yes HOSTNAME=wks2 GATEWAY=192.168.1.254 NISDOMAIN=logindomain 2) Modify the file /etc/nsswitch.conf and update as follows: [root@Server /]# vi /etc/nsswitch.conf passwd: nis files shadow: nis files group: nis files hosts: dns files nis 3) Run the ypbind service and also put it in the startup by: [root@Server /]# /etc/rc.d/init.d/ypbind start [root@Server /]# setup [*] ypbind Or [root@Server /]# chkconfig -level 35 ypbind on Since this is an authentication service it is better to restart the client machine. Once restarted, on the login screen give in the login name and password of the user already created on the NIS server. Client machine should log the user with an error:
Could not find the home directory of the user, logging on with home directory as /. This is not so important and can be solved. Basically the system tried to login to the home directory on the local machine but since this user did not exist at all on the local computer; the home directory also did not exist. Thus fails to go directly in the home directory. This can be solved by either creating a home directory for this user manually in the /home; or mount the home directory of this user from the NIS server to the local machine using NFS. Assuming our example network, and assuming you want to load home directories from the NIS server permanent and regular basis, you need to add the following line in the /etc/fstab file. 192.168.1.254:/home /home nfs defaults 0 0 Also once the user wants to change his password from a client machine, he needs to use the yppasswd command. Simple passwd command will not work as it is designed to handle login/password stored in local /etc/passwd file only. Documentation: /usr/share/doc/yp-bind-* /usr/share/doc/yp-serv-* /usr/share/doc/yp-tools-* Log files: /var/log/messages Chapter 7: S AMBA SAMBA is derived from SMB, which stands for Session Message Block. It is a protocol used to exchange data between windows and *nix computers. SMB was built on Microsoft's NetBIOS protocol, which can be run over TCP/IP. Packages: samba-a.b (server side) samba-client-a.b (client side) samba-common-a.b (client server both) Case Study: In fact there will be three scenarios: 1. Accessing shares on windows computers from *nix computers 2. Providing access to shares on the SAMBA server to users logging on from
Windows computers, in a work-group model 3. Providing access to shares and logon authentication to users logging on from Windows computers, in a Domain model Case - 1 (workgroup model, shares on windows machine): We have some share (ebooks) available on windows computer (wks1.mydomain.com) and want to access them from a linux machine. Windows side configuration: Nothing special, just enable sharing on the directories / folders you want to share. Optionally set permissions of these shares. Linux side configuration: There is no configuration required. Only the commands smbclient and smbmount will be used to access the windows machine. So the related packages must already be installed on the client side. To see the list of shares available on the windows computer: [root@Server /]# smbclient -L wks1.mydomain.com
Just press enter on the password prompt and you would be given a list of shares available on this computer.
Note: In case of windows 2000 / XP, simple pressing enter will not do the trick. You will be denied access to list of shares in this way. In order to see a list of shares on a windows 2000/XP, you need to specify a valid username already defined and enabled on the windows 2000/XP machine. When asked for password, you have to supply the password of that user. In case of a guest account, which is normally available and enabled on a windows network, you can use guest and a blank password with it: [root@Server /]# smbclient -L wks1 -U guest To connect to the share ebooks on the windows machine, with a windows user xena, use: [root@Server /]# smbclient //wks1/ebooks -U xena If you see the smb:\> prompt, you have successfully logged in. If you dont use the -U <username> switch then the username and password of the currently logged in linux user will be sent to the windows machine, which may get rejected. So it is better to use the username and passwords of the user already created on the windows machine you are connecting to. Once you are logged in, type help for a list of commands. Like FTP client, you can use get or mget and put or mput commands to copy files from and to the windows computer respectively. To exit smbclient, type quit at the smb :\> prompt. If you want, you can mount the remote windows share ebooks on some local mount point on your machine, say /mnt/work. Do do so, use: [root@Server /]# smbmount //wks1/ebooks /mnt/work -o username=xena Or [root@Server /]# mount -t smbfs //wks1/ebooks /mnt/work -o username=xena Case - 2 (workgroup model, shares on linux machine): SAMBA server will be configured on our machine Server.mydomain.com. The workgroup name can be same as the one already being used in the current windows setup. However, we will create a new workgroup named SAMBAWORKGROUP and configure the windows clients to be a part of this workgroup.
Server side configuration: 1) Modify the /etc/samba/smb.conf file and give the following entries: [root@Server /]# vi /etc/samba/smb.conf [global] workgroup = SAMBAWORKGROUP comment = samba server on the local lan placed in the server room hosts allow = 192.168.1. 192.168.2. 127. security = user load printers = yes os level = 64 encrypt passwords = yes smb passwd file = /etc/samba/smbpasswd preferred master = yes domain master = yes wins support = yes
[cdrom] comment = CDROM on SAMBA server path = /mnt/cdrom read only = yes browseable = yes public = yes [floppy] comment = Floppy drive on SAMBA server path = /mnt/floppy read only = no browseable = yes valid users = @students ; note that the group students exists in /etc/groups [linuxproject] comment = project on linux by MS 4 students path = /users/ms4/linuxproject read only = no browseable = yes valid users = Natellie, Smith Note: All paths defined in the smb.conf file must exist. If they dont create them yourself. 2) Test your smb.conf file by running testparm command which is only available for SAMBA. This command should show all the shares without any errors. [root@Server /]# testparm | less 3) If the above command works fine then run the smb services now by: [root@Server /]# /etc/rc.d/init.d/smb start 4) Create users Smith and Natellie( on the local linux machine, if not already created ) and generate SMB passwords for them which will be used to access the shares defined on this computer from the windows client. Use the following commands to generate SMB passwords: [root@Server /]# cat /etc/passwd | mksmbpasswd.sh > /etc/samba/smbpasswd [root@Server /]# smbpasswd -a Smith [root@Server /]# smbpasswd -a Natellie
Note: The first command in step 4 needs to be given once. Client side configuration: On the client side, i.e. on the computer having Microsoft Windows, open the network neighborhood properties, and in the identification tab, type in SAMBAWORKGROUP in the workgroup text box. Also set the other network related properties, e.g. IP addresses, address of DNS and name of Primary WINS server which is the IP address of SAMBA server. Press OK and restart the windows machine. When restarted, types in user name Smith and his password. You will be logged on. Now you should be able to see the SAMBA server in the Network Neighborhood window. Double click it, and you should be able to see the shares: Smith, cdrom, floppy and linuxproject. If you did not log in as a regular linux user, then on double clicking on the SAMBA server will give you a login screen, here again you have a chance to provide a regular linux user account and it's samba password. If still you do not supply a regular login and you just press OK, then you will be presented with the shares on the server without the share Smith displayed. In fact you will be logged in as user nobody, and this will be a lot restrictive access than assumed. To assign a drive letter to a share, you can either right click on the share or select Map network drive. Or you can use the command on the windows DOS shell prompt: C:\> net use F: \\Server.mydomain.com\cdrom Case - 3 (SMB server as domain controller): SAMBA server will be configured on our machine Server.mydomain.com. The workgroup name can be same as the domain name already being used in the current windows setup (provided the NT server for the domain is off line), BUT it should never be of the same name as your DNS domain name. We will create a new workgroup named SAMBADOMAIN on SAMBA machine (the server), and configure the windows clients to be a part of this domain. (Confusing? Well, on a SAMBA machine, in /etc/samba/smb.conf, the word workgroup serves for workgroup as well as domain for windows computers). Server side configuration: 1) Modify the /etc/samba/smb.conf file and give the following entries: [root@Server /]# vi /etc/samba/smb.conf [global] workgroup = SAMBADOMAIN comment = samba server on the local lan placed in the server room hosts allow = 192.168.1. 192.168.2. 127. security = user
load printers = yes encrypt password = yes smb passwd file = /etc/samba/smbpasswd domain logons = yes os level = 64 preferred master = yes domain master = yes wins support = yes [homes] comment = home directories of users read only = no create mask = 0750 browseable = no [netlogon] comment = SAMBA network logon services for Microsoft clients path = /users/netlogon guest ok = yes browseable = no locking = no [cdrom] comment = CDROM on SAMBA server path = /mnt/cdrom read only = yes browseable = yes public = yes [floppy] comment = Floppy drive on SAMBA server path = /mnt/floppy read only = no browseable = yes valid users = @students ; note that the group students exists in /etc/groups [linuxproject] comment = project on linux by MS 4 students path = /users/ms4/linuxproject read only = no
browseable = yes valid users = Natellie, Smith Note: All paths defined in the smb.conf file must exist. If not create them. Also if you are going to use clients of Windows 98 and onwards or Windows NT service pack 3 or higher, then these systems use encrypted passwords by default. We need to provide either encrypted password support in SAMBA server or modify the registry of the windows computers and enable plain text passwords in them. To enable support for encrypted passwords in the SAMBA server, perform the following steps: a. Create a separate password file for Samba. To create one based on your existing /etc/passwd file, at a shell prompt, type the following command: [root@Server /]# cat /etc/passwd | mksmbpasswd.sh > /etc/samba/smbpasswd The mksmbpasswd.sh script is installed in your /usr/bin directory with the samba package. b. Use the following command to change permissions on the Samba password file so that only root has read and write permissions: [root@Server /]# chmod 600 /etc/samba/smbpasswd c. The script does not copy user passwords to the new file. To set each Samba users password, use the command smbpasswd username (replace username with each users username). A Samba user account will not be active until a Samba password is set for it. [root@Server /]# smbpasswd -a Smith d. The next step is to enable encrypted passwords in the Samba configuration file. In the file smb.conf, uncomment the following lines: encrypt password = yes smb passwd file = /etc/samba/smbpasswd 2) Test your smb.conf file by running testparm command which is only available for SAMBA. This command should show all the shares without any errors. [root@Server /]# testparm 3) If the above command works fine then run the smb services now by: [root@Server /]# /etc/rc.d/init.d/smb start
Notes: Windows NT 4: At present, Samba needs to have a *nix account for every entry in the 'smbpasswd' file, so this means that each Windows NT4/2000computer needs a *nix account on the Domain Controller. Machine accounts are differentiated from user accounts by appending a $ to the end of the machine's name. For Windows NT clients, you can create these accounts manually. To make a domain machine account, issue the following commands on the samba Domain Controller as 'root': Create the group "machines" by: [root@Server /]# groupadd machines [root@Server /]# useradd -d /dev/null -g machines -c Machine Account' -s /bin/false -M <NETBIOS_NAME>$ [root@Server /]# smbpasswd -am <NETBIOS_NAME> E.g. To add a machine account of a machine win98 on the samba server, you would use: [root@Server /]# groupadd machines [root@Server /]# useradd -d /dev/null -g machines -c 'Machine Account' s /bin/false -M win98$ [root@Server /]# smbpasswd -am win98 Windows 2000/XP: Windows 2000 is different than NT. As far as NT is concerned, while joining the domain, you are given an option to create a computer account in the domain server. This allows machine accounts to be made on the fly. At the moment the only user that can create computer accounts automatically is root. This means we must first make a samba password for 'root' with: [root@Server /]# smbpasswd -a root It is recommended that you use a different password than real Linux password for root for security reasons. To make a domain machine account, issue the following commands on the Domain Controller as 'root': [root@Server /]# groupadd machines
[root@Server /]# useradd -d /dev/null -g machines -c 'Machine Account' -s /bin/false -M win2k$ [root@Server /]# smbpasswd -am win2k Optionally in order to make the computer accounts on the fly while joining the domain, you need the following entry in the smb.conf file. add user script = /usr/sbin/useradd -d /dev/null -g machines -c 'Machine account' -s /bin/false -M %u Note the absence of $ in this script, SAMBA automatically adds the $ for you when it is configured as domain log on server. Also the group machines were already created manually. This also works for Windows XP. The above solution specifically solves the problem when you are joining the Windows 2000 workstation to a samba domain and you get the following error: "The account used is a computer account. Use your global user account or local user account to access the server". This problem comes with Windows 2000 machines with service pack 2. Another solution is to try changing the windows computer's machine name from Network properties, but this seldom works (depends on the mood of win 2000 machine). Client side configuration: On the client side, i.e. on the computer having Microsoft Windows, open the network neighborhood properties, and in the identification tab, type in SAMBADOMAIN in the DOMAIN textbox. Also set the other network related properties, e.g. IP addresses, address of DNS and name of Primary WINS server which is the IP address of SAMBA server. Press ok and restart the windows machine. Once restarted, type in any user name and password of the user defined on SAMBA server and select the domain SAMBADOMAIN in the drop down list. You will be greeted after a pause of 10 to 30 seconds. Restart and log on to the new domain with the user name Smith, lets say. Now you should be able to see the SAMBA server in the Network Neighborhood window. Double click it, and you should be able to see the shares: Smith cdrom, floppy and linuxproject. On double clicking on the share linuxproject you should be able to see the contents of this share. You will be asked for the password if you logged on as someone else other than Smith or Natellie, this will be the password of one of the users Natellie or Smith (as shown above), which are defined in the machine running SAMBA. Similarly password of one of the users in the group students (defined in SAMBA machine) may be required once accessing the share floppy. To assign a drive letter to a share, you can either right click on the share or select Map to drive. Or you can use the command on the windows DOS shell prompt: C:\> net use F: \\Server.mydomain.com\cdrom
Common problems while joining a Linux SAMBA domain from win2k client: 1) "The account used is a computer account. Use your global user account or local user account to access the server". This problem comes with Windows 2000 machines with service pack 2. Solution is to create a machine account for each win2k computer on the SAMBA computer. Another solution is to try changing the windows computer's machine name from Network properties, but this seldom works (depends on the mood of win 2000 machine). 2) "The credentials supplied conflict with an existing set of credentials" . This is a problem on windows client side.. Windows will not let you connect as two different users to the same share on the remote computer. To over come this problem, log out of the windows computer (restart if it doesn't behave), then again log in first as local administrator and then supply the root login and it's samba generated password while joining the domain. 3) "The domain SMBDOMAIN could not be contacted". Check your SMB configuration. Check for presence of machine accounts (including the group machines) and their SMB passwords. Check the network settings of windows. Restart the windows computer and it should work. Now something about plain text passwords: In the Samba configurations above, we are using encrypted passwords. But if for any reason, for example, incompatibility of samba encrypted passwords and windows encrypted passwords, we can configure both sides to use plain text passwords. On samba side, we will just put a remark on the following two lines in the smb.conf file: #encrypt password = yes #smb passwd file = /etc/samba/smbpasswd And on windows side we have to edit the registry and enable plain text password support. The procedure is different for win 9x, Windows NT and windows 2000. Windows 9x registry settings: Locate the following registry key: HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Services\VxD\ VNETSUP Select Edit -> New and choose to create a new DWORD value.
In Regedit insert a new DWORD value called New Value #1 in the registry. Rename this new bvalue EnablePlainTextPassword and double click this new name. In the dialog box, enter 1 as the value and close regedit program. Restart windows. Windows NT registry settings: Locate the following registry key: HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Services\Rdr\ Prameters Select Edit -> New and choose to create a new DWORD value. In regedit insert a new DWORD value called New Value #1 in the registry. Rename this new bvalue EnablePlainTextPassword and double click this new name. In the dialog box, enter 1 as the value and close regedit program. Restart windows. Windows 2000 registry settings: Method 1: Under windows 2000 this can be done doing: Start -> Settings -> ControlPanel -> Administrative Tools -> Local Security Policy. In the Local Security Settings dialog box, select Local Policies -> Security Options, double click on Send unencrypted password to connect to third party SMB servers and enable it. Reboot the machine. Method 2: Locate the following registry key: HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Services\Lanm anWorkStation\Parameters Select Edit -> New and choose to create a new REG_DWORD value. Regedit will insert a new REG_DWORD value called New Value #1 in the registry. Rename this new bvalue EnablePlainTextPassword and double click this new name. In the dialog box, enter 1 as the value and close regedit program. Restart windows. Log Files: /var/log/messages (file) /var/log/samba (directory)
Documentation and Help: /etc/samba/smb.conf /usr/share/doc/samba-*/docs (excellent documentation) /usr/share/doc/samba-*/docs/DIAGNOSIS.TXT To read more about Using Samba with Windows NT 4.0 and Windows 2000, read ENCRYPTION.TXT, Win95.txt, and WinNT.txt in the directory /usr/share/doc/samba*/docs/textdocs/ Testing: To list the shares available on the local host, use: [root@Server /]# smbclient -L localhost -N To list the shares available on the windows machine, use: [root@Server /]# smbclient -L wks1 -N If no shares are visible, then WINS server may not be responding. To check this use: [root@Server /]# nmblookup -B 192.168.1.255 __SAMBA__ (Notice the two underscores before and after SAMBA) The output of above command should display the IP address of Server.mydomain.com . You should also check that windows machines are properly configured to use WINS by: [root@Server /]# nmblookup -T '*' Client hosts on this network should respond with their IP addresses.
Chapter 8: S QUID
Squid is basically Internet caching server which also does proxying. It is also known as Squid-Proxy server. It caches HTTP and FTP data. Squid listens for user requests for HTTP and FTP on port 3128 by default (can be changed) and contacts the original server if the web page requested is not in the squid's cache. Packages: squid-a.b.STABLE.a.b
Case study: We want our network to share the Internet connection running on the Server.mydomain.com with our clients on our network. SQUID runs on the gateway machine i.e. Server.mydomain.com. Configuration: Normally only very minor changes are required in the Squid file /etc/squid/squid.conf. Main configuration is to create an acl for your network and allow it http access. Also optionally you can make squid listen on port other than default 3128. Update it as follows: [root@Server /]# vi /etc/squid/squid.conf # TAG: http_port http_port 3128 # TAG: cache_dir cache_dir ufs /var/spool/squid 200 16 256 acl mynet src 192.168.1.0/24 http_access allow mynet http_access allow localhost http_access deny all You should also set the visible_hostname tag to the fully qualified domain name of this squid machine. This name should be the one defined in your DNS. Also it must not contain any invalid characters like underscores, etc. Use the same rules which are used to assign host names in DNS. If you put any invalid character here (or
in DNS), you will see ugly messages like mime error, error loading url, etc, etc. visible_hostname Server.mydomain.com You should also set cache_effective_user and cache_effective_group cache_effective_user squid cache_effective_group squid Now create the cache directory (defined in the squid.conf file) above otherwise squid service will fail (very important): [root@Server /]# squid -z Note: If you use some directory for squid's cache other than default /var/spool/squid, then make sure to change the new directory's ownership and permissions for squid:squid, otherwise the squid -z command will fail. Assuming you are using /squid/cache as your cache directory, then:
[root@Server /]# chown squid:squid /squid/cache -R [root@Server /]# chmod 770 /squid/cache -R After the cache directory creation is successful, run the SQUID service by: [root@Server /]# /etc/rc.d/init.d/squid start
Transparent proxying:
SQUID on gateway machine: If you are configuring SQUID to act as a transparent proxy server i.e. you want SQUID to hijack all the requests for HTTP and HTTPS and should never let the users machine contact directly the origin server for web pages and also you dont want to configure each users machine for editing the proxy settings. This can be achieved by using some firewall redirection rules and some settings in SQUID. Note the scenario is that squid is running on the gateway machine, i.e. in this case, Server.mydomain.com, See above Fig 1. In addition to the settings above, the following entries needed to be placed in the squid.conf file. These entries are normally found at the end of squid.conf file. 1) Modify squid.conf file and add / update as below: [root@Server /]# vi /etc/squid/squid.conf # HTTPD-ACCELERATOR OPTIONS httpd_accel_host virtual httpd_accel_port 80 httpd_accel_with_proxy on httpd_accel_uses_host_header on 2) Put the following line in any of your startup scripts or you can make your own script and call it from your startup scripts. Normally rc.local script is recommended for such tasks as it runs after all the scripts and services are initialized/run at system boot up. This line redirects www (port 80) requests to local squid service running at port 3128 in this case. It also assumes that your SQUID server is connected to the hub or switch via eth0 interface and is connected to the Internet via a modem dialup interface ppp0. [root@Server /]# iptables -t nat -A PREROUTING -i eth0 -p tcp --dport 80 -j REDIRECT --to-port 3128
For now apply the iptables command above on the command prompt and restart the squid service by: [root@Server /]# /etc/rc.d/init.d/squid restart Connect to the Internet via your dialup connection. You should see activity in the /var/log/squid/access.log file while the client machines try to connect to the Internet. SQUID on one leg, another Linux mahcine as gateway: In this case the SQUID machine is connected only with one ethernet connection (192.168.1.252) within our network. There is no other interface on this squid machine. The gateway machine (192.168.1.254) is also on Linux and iptables rules are used to redirect port 80 traffic towards the squid machine. See Fig 2. On the SQUID machine, just configure the SQUID for httpd acceleration. No redirection rules required here.
On the gateway machine, give the following command:
[root@Server /]# iptables -t nat -A PREROUTING -i eth0 -s ! 192.168.1.252 -p tcp --dport 80 -j DNAT -to-destination 192.168.1.252:3128 This will be direct all traffic coming from local network, except the SQUID machine, destined for port 80 towards SQUID machine. When SQUID machine needs to bring web content from origin server(s), then this rule will have no effect on it. SQUID on one leg, Cisco router as gateway: If you have a router (say Cisco) as your gateway (192.168.1.254), and your squid machine is again on one leg (192.168.1.252), then you need to define some ACLs on it to get the task done. See Fig 3.
Configure squid machine for httpd acceleration and then configure your Cisco router as following: Define a route map with a name of lets say squid-catch and specify the IP of SQUID machine as the next hop for it.
! route-map squid-catch permit 10 match ip address 110 set ip next-hop 192.168.1.252 ! Now define an access list to trap HTTP requests: ! access-list 110 deny tcp any any neq 80 access-list 110 deny tcp host 192.168.1.252 any access-list 110 permit tcp any any ! Apply this on router's ethernet interface: ! interface ethernet 0 ip policy route-map squid-catch ! Stop porn for God's sake: If you want to restrict users so that they are denied access to certain sites, like porn sites, etc. Then you can make an acl for bad sites put all the possible words which are normally part of url of those bad sites, in a text file with each word at separate line. Create and block the acl as follows in the squid.conf file: acl badsites url_regex -i "/etc/squid/badsites.txt" http_access deny badsites It should be noted that the placement of http_access lines are very important in SQUID's expected behavior. You need to deny badsites BEFORE allowing the http_access to your client src acl. Your configuration with blocked sites would look like this: acl badsites url_regex -i "/etc/squid/badsites.txt" acl mynet src 192.168.1.0/24 http_access deny badsites http_access allow mynet http_access allow localhost http_access deny all Your badsites file should contain all words on separate lines which you want to be search in the URL and then blocked by SQUID. An example badsites.txt file is below:
sex xxx movies This is quite effective approach. You may need to watch your /var/log/squid/access.log for such sites and add those names to this file. Socks : Unfortunately, your clients will not be able to find a SOCKS server on linux machine for their MSN or Yahoo messengers. To support this you have two options. a) Install and configure a SOCKS package. b) Enable IP Masquerading (NAT) on your network (easy and recommended). a) Install and configure SOCKS server: SOCKS stands for SOCK-et-S , its initial internal development name. There is also a Nylon socks project underway on the internet. SOCKS is a networking proxy protocol that enables hosts on one side of a SOCKS server to gain full access to hosts on the other side of the SOCKS server without requiring direct IP-reach ability. SOCKS are often used as a network firewall, redirecting connection requests from hosts on opposite sides of a SOCKS server. The SOCKS server authenticates and authorizes requests, establishes a proxy connection, and relays data between hosts. Quite some information resides on these web pages. Please see to them. www.linuxgazette.com/issue48/tag/27.html www.host.biz/rpm2html/PLF/System_Servers.html www.socks.nec.com/socksfaq.html www.socks.nec.com b) To enable IP masquerading you need to put the following two lines in one of your startup scripts (rc.local) or optionally you can create your own script and call it from /etc/rc.local script. For kernel 2.2 using ipchains, only put the following two lines:[root@Server /]# echo "1" > /proc/sys/net/ipv4/ip_forward [root@Server /]# ipchains -A FORWARD -i ppp0 -s 192.168.1.0/24 -d ! 192.168.1.0/24 -j MASQ
For kernel 2.4 using netfilter (iptables) only put the following two lines :[root@Server /]# echo "1" > /proc/sys/net/ipv4/ip_forward [root@Server /]# iptables -t nat -A POSTROUTING -s 192.168.1.0/24 d ! 192.168.1.0/24 -o ppp0 -j MASQUERADE Common problems: 1) Access denied to SQUID even the src ACL is defined and the URL doesn't contain any word which should get caught in any of the badsites ACL. This happens when there are two (or more) networks which are (mis-)configured in SQUID to be able to access SQUID cache data. For example, the following ACLconfiguration will never work: acl MYNET_A src 192.168.1.0/24 acl MYNET_B src 192.168.0.0/24 http_access allow MYNET_A MYNET_B The above should be written as: acl MYNET_A acl MYNET_B http_access http_access src 192.168.1.0/24 src 192.168.0.0/24 allow MYNET_A allow MYNET_B
Or even this will work: acl MYALL_NETS src 192.168.1.0/24 192.168.0.0/24 http_access allow MYALL_NETS 2) SQUID dead but subsystem locked. This happens because of some configuration problem in squid.conf file, which puts it to stop just after it is started. Check /var/log/messages for any system level or squid related errors. 3) No running copy Normally happens when squid was not able to create the cache directory in the previous steps or is unable to access the cache directories due to ownership / permissions problems. The cache directories must be owned by both user and group squid and should be writeable by squid:squid only. May also be resultant of bad DNS lookups? In bad DNS case, run squid with -D switch.
Log files: /var/log/messages /var/log/squid/access.log /var/log/squid/cache.log Documentation and Help: Extensive help (specially check the FAQ) is available on www.squidcache.org. Also at www.visolve.com. FAQ files in /usr/share/doc/squid-a.b Testing: On the command prompt: [root@Server /]# /etc/rc.d/init.d/squid status You can check the access.log file for user requests for web pages being handles by SQUID, using a separate window or terminal: [root@Server /]# tail -f /var/log/squid/access.log Chapter 9: APACHE Apache is actually "a PATCHy server". It was based on some existing code and a series of "patch files" made to original NCSA server. For many people this name brings the same meaning / feeling / spirit of the Native American Indian tribe of Apache, wellknown for their superior skills in warfare strategy and inexhaustible endurance. Apache is HTTP 1.1 Compliant, open-source web server that is widely used. Packages: Redhat 7.x apache-a.b apacheconf-a.b apache-manual-a.b Redhat 8.0 & 9 httpd-a.b httpd-manual-a.b redhat-config-httpd-a.b (gui)
Case Study: In our example network, we want to implement a local web server under the name www.mydomain.com. I will also introduce ScriptAlias , Alias and UserDir directives. The same server www.mydomain.com will be used to host two virtual websites: www.islam.edu and www.urdulinux.org. All of these websites will be hosted on Server.mydomain.com having IP 192.168.1.254. We will implement it using Name based virtual hosting. Normally you should consider using Name based virtual hosting unless there is a specific need / reason to choose IP based virtual hosting. Configuration Apache configuration file is /etc/httpd/conf/httpd.conf. The file is divided into three main sections. 1. Global Environment 2. Main configuration 3. Virtual Host Configuration
In section 1, only the directive Listen needs to be set. This tells the Apache server, on which IP and port to listen for HTTP requests.
Listen 80 or Listen 192.168.1.254:80 We can also make it listen on any port other than 80. In section 2, set the ServerName and DocumentRoot directives. ServerName would be the name which will be used generally to access your server, i.e. www.mydomain.com. Set DocumentRoot directive to the path where the web pages of your main website reside. Similarly your VirtualHost configurations in section 3 will have the DocumentRoot directive set to the directories containing web content relating to respective sites sites. ServerName www.mydomain.com DocumentRoot /var/www/html Also the default directory options for the document root directory are : <Directory "/var/www/html"> Options Indexes FollowSymLinks AllowOverride None Order allow,deny Allow from all </Directory> Try starting your web server now by: [root@Server /]# /etc/rc.d/init.d/httpd start Open your favorite web browser and type in the URL: http://127.0.0.1 or http://localhost or http://www.mydomain.com You should be able to see a Apache web server Test Page. If you can see the Apache test web page that means you are all set to further configure your server. Now copy your web site related files in this directory /var/www/html make them Apache readable by setting appropriate ownership and permissions and and re-start the web server. You should be able to see your website related main web page.
Directory aliases Suppose you want to provide your users with some helpful documentation on various technologies, stored on your local hard drive. Assuming this local storage is /data/ebooks and you wish that your users should be able to access it as www.mydomain.com/ebooks , It should be kept in mind that your document home directory for www.mydomain.com is /var/www/html. One method is to create a directory named ebooks under /var/www/html and move or copy the contents of /data/ebooks to /var/www/html/ebooks. This is cumbersome and is prone to disk space wastage. The other method is to create a directory alias for this location like: Alias /ebooks/ "/data/ebooks/" <Directory /data/ebooks"> Options Indexes MultiViews AllowOverride None Order allow,deny Allow from all </Directory> Restart your Apache service and you should be able to access your ebooks from the URL: www.mydomain.com/ebooks ScriptAliases Sometimes it is desired that you have another directory for your CGI scripts other than /var/www/cgi-bin. For that you have special directive ScriptAlias. Assume you have /var/www/cgi-admin directory as well, and you want to run your CGI scripts from there too. ScriptAlias /cgi-admin/ "/var/www/cgi-admin/" <Directory "/var/www/cgi-admin"> AllowOverride None Options None Order allow,deny Allow from all </Directory>
UserDirectories This is additional feature which you may want to introduce in your web server configuration. This facility allows users to have their web contents published on your web server directly from their home directories and without requiring any administrative privileges or without the need to edit the httpd.conf file. Using this feature anyone can access the web content published by any user by using the URL: www.mydomain.com/~user. E.g. A user Sam's web content can be viewed by: www.mydomain.com/~sam As per httpd.conf file: The path to the end user account 'public_html' directory must be accessible to the webserver userid. This usually means that ~userid must have permissions of 711, ~userid/public_html must have permissions of 755, and documents contained therein must be world-readable. Otherwise, the client will only receive a "403 Forbidden" message. This means that each user who wants to publish his web content should have his home directory permissions (/home/sam) set as 711 and should create a special directory named public_html (/home/sam/public_html) with permissions 755. Any material / content placed in this directory should have world readable permissions. The web server administrator (you) will have to set the user home directory permissions to 711 and edit the httpd.conf to enable user directory feature. To do so remark the line UserDir disable in the section <IfModule mod_userdir.c> and unremark the line UserDir public_html in the same section. <IfModule mod_userdir.c> # # UserDir is disabled by default since it can confirm the #presence of a username on the system (depending on home #directory permissions). # # UserDir disable # # To enable requests to /~user/ to serve the user's #public_html directory, remove the "UserDir disable" line #above, and #uncomment the following line instead: # UserDir public_html </IfModule> More over you can setup the user public_html directories as read only to avoid any future problems.
# Control access to UserDir directories. The following is #an example for a site where these directories are #restricted to read-only. <Directory /home/*/public_html> AllowOverride FileInfo AuthConfig Limit Options MultiViews Indexes SymLinksIfOwnerMatch IncludesNoExec <Limit GET POST OPTIONS> Order allow,deny Allow from all </Limit> <LimitExcept GET POST OPTIONS> Order deny,allow Deny from all </LimitExcept> </Directory> Your main (basic) configuration of your site www.mydomain.com finishes here. Now lets see what needs to be done for hosting the two example websites discussed above i.e. www.isb.edu and www.mylinux.org. Virtual site hosting needs configuration in your DNS first. For these two sites, we need to define them in our DNS. There is yet another problem. These sites do not fall under our network forward zone data. This means we need to define separate forward zones (define separate domains) in our DNS. Since in these new forward zones, there needs to be only on NS, one A and one CNAME type entries, this is going to do just fine. So first, here is the /etc/named.conf file: [root@Server /]# vi /etc/named.conf ## named.conf - configuration for bind # #Generated automatically by redhat-config-bind,alchemist et #al. # Any changes not supported by redhat-config-bind should be #put in /etc/named.custom # controls { inet 127.0.0.1 allow { localhost; } keys { rndckey; }; }; include "/etc/rndc.key"; options { directory "/var/named/"; }; zone "." {
type file }; zone type file }; zone type file }; zone type file }; zone type file }; zone type file }; zone type file };
hint; "named.ca"; "localhost" { master; "localhost.zone"; "0.0.127.in-addr.arpa" { master; "0.0.127.in-addr.arpa.zone" ; "1.168.192.in-addr.arpa" { master; "1.168.192.in-addr.arpa.zone"; "mydomain.com" { master; "mydomain.com.zone"; "isb.edu" { master; "isb.edu.zone"; "mylinux.org" { master; "mylinux.org.zone";
Now here is the forward zone of my domain mydomain.com. Notice the absence of any mentioning of www.isb.edu or www.mylinux.org. Also notice that there is no separate IP for www, rather www machine is in-fact canonical name for the actual machine named Server. [root@Server /]# vi /var/named/mydomain.com.zone $ORIGIN mydomain.com. $TTL 3h @ IN SOA Server.mydomain.com. root.Server.mydomain.com. ( 1 ; serial 3H ; Refresh after 3 hours 1H ; Retry after 1 hour 1W ; Expire after 1 week 1H ) ; negative caching of TTL of 1 hr @ IN NS Server.mydomain.com. @ IN MX 10 Server.mydomain.com. Server IN A 192.168.1.254 ws1 IN A 192.168.1.1
ws2 IN A 192.168.1.2 www IN CNAME Server Below is what is to be be configured for www.isb.edu : [root@Server /]# vi /var/named/isb.edu.zone $ORIGIN isb.edu. $TTL 3h @ IN SOA Server.mydomain.com. root.Server.mydomain.com. ( 1 ; serial 3H ; Refresh after 3 hours 1H ; Retry after 1 hour 1W ; Expire after 1 week 1H ) ; negative caching of TTL of 1 hr @ IN NS Server.mydomain.com. Server IN A 192.168.1.254 www IN CNAME Server The SOA record of this file may be confusing for some readers. Since there is only one name server in our network which is doing all the work, we need to mention that name server in the SOA record, so that any name queries performed on our name server could tell the querying application / machine, that who is the actual machine resolving all queries. In simple words, the SOA above says that Start Of Authority for the domain isb.edu is the machine Server.mydomain.com. For the domain urdulinux.org we will of-course use the same format as we have used for isb.edu , there is no rocket science involved in it. [root@Server /]# vi /var/named/mylinux.org.zone $ORIGIN mylinux.org. $TTL 3h @ IN SOA Server.mydomain.com. root.Server.mydomain.com. ( 1 ; serial 3H ; Refresh after 3 hours 1H ; Retry after 1 hour 1W ; Expire after 1 week 1H ) ; negative caching of TTL of 1 hr @ IN NS Server.mydomain.com. Server IN A 192.168.1.254 www IN CNAME Server Virtual Host: Now our forward zones properly configured, lets configure section 3 of our Apache web server for Virtual Hosting.
[root@Server /]# vi /etc/httpd/conf/httpd.conf NameVirtualHost 192.168.1.254 <VirtualHost www.mydomain.com> ServerName www.mydomain.com DocumentRoot /var/www/html </VirtualHost> <VirtualHost www.isb.edu> ServerName www.isb.edu DocumentRoot /hostedsites/isb.edu/html </VirtualHost> <VirtualHost www.mylinux.org> Servername www.mylinux.org DocumentRoot /hostedsites/mylinux.org/html </VirtualHost> You can also re-write the VirtualHost section for any virtual site as: <VirtualHost 192.168.1.254> ServerName www.isb.edu DocumentRoot /hostedsites/isb.edu/html </VirtualHost> Notice that we have introduced the virtual host section for www.mydomain.com as well. This is because in case of miss-spelled URLs and sites which are not hosted on this server, Apache goes to the first virtual host defined in the virtual hosts section. See point 1 below. Using an explicit IP address to listen HTTP requests in NameVirtualHost is better. BUT if your web server is on dynamic IP, i.e. it gets it's IP via DHCP or so, then you should not use a fixed IP. In that case you should use an *. Only ServerName and DocumentRoot directives are a must while configuring virtual hosts. There are few more points to keep under consideration while configuring / hosting your web sites. 1) If a URL is typed in by the user in his web browser is mis-spelled or incorrect, then by default, Apache will use the virtual host listed first to display the page i.e. in the above configuration, if the user types in www.linuxdoc.mydomain.com in his web browser, then since this URL / website does not exist in your domain tables, the first virtual host (www.isb.edu) will be used to display the web page, which was not intended. In order to avoid such problems, you can put a sort of dummy virtual host
containing a redirector page or a page telling user that he has not reached where he wished, and that he should type in the address again. 2) Any directives defined outside the virtual host configuration are over written /neglected if the same directives are found inside that specific virtual host. E.g. You have a DocumentRoot directive in the main configuration file as: DocumentRoot /var/www/html But this will not be honored in the virtual host www.isb.edu because that virtual host section has its own DocumentRooot directive as: DocumentRoot /hostedsites/isb.edu/html Providing support of privileged access to some web site, based on user name and password. Sometimes a user needs to be authenticated before displaying the contents of a web page. There are many different ways for doing it but let us the basic method here. Assume your web site mylinux.org is undergoing some refinement and you want it only accessible by its developer Tom using his password bravo. The web pages of the website under discussion are located in the directory: /hostedsites/mylinux.org/html. We need to make a user authentication file (password file) with the help of a utility htpasswd. This file will be placed outside the web pages directory (i.e. in the root-only readable directory /etc/httpd/security) so that cracking is avoided. 1) Create the file (if doesn't exist) and the user "tom" by: [root@Server /]# mkdir /etc/httpd/security [root@Server /]# htpasswd c /etc/httpd/security/.htpassword tom The dot in the beginning of the file makes it hidden so it won't show up in directory listing. Type the password twice to confirm. Also change its ownership so that user Apache is able to read it, by: [root@Server /]# chown apache:apache /etc/httpd/security/.htpassword 2) Create an .htaccess file with the following contents in the same directory which contains your web pages. i.e. /hostedsites/mylinux.org/html, with the command:-
[root@Server /]# vi /hostedsites/mylinux.org/html/.htaccess AuthName "Authorized project developers only" AuthType Basic AuthUserFile /etc/httpd/security/.htpassword require user tom 3) Change the file permissions of the .htaccess file so that only the user Apache is able to change it. [root@Server /]# chown apache:apache /hostedsites/mylinux.org/html/.htaccess 4) Also give only read permissions to users on all files in the web pages directory by:[root@Server /]# chmod o+r /hostedsites/mylinux.org/html/* [root@Server /]# chmod o-wx /hostedsites/mylinux.org/html/* or [root@Server /]# chmod 774 /hostedsites/mylinux.org/html/* 5) Now update the httpd.conf file and set the security options for your DocumentRoot directory in the VirtualHost section of mylinux.org. <VirtualHost www.mylinux.org> Servername www.mylinux.org DocumentRoot /hostedsites/mylinux.org/html <Directory "/hostedsites/mylinux.org/html" AllowOverride All </Directory> </VirtualHost> 6) Now is the time to test your setup. First restart the httpd deamon by [root@Server /]# service httpd restart Open the browser at the server and point to www.mylinux.org and press enter. (You may need to reload the page). A popup screen should come up asking username and password. Supply username tom and password bravo. It should allow access, else error 401 web page will show up, which is "Authorization failed". Log Files /var/log/httpd/access_log /var/log/httpd/error_log /var/log/messages
Documentation and Help: Extensive documentation is available on the Apache Test web page once you simply start your apache web server without configuring it for the first time. Simpler path is /var/www/manual/index.html Also the files in directory /usr/share/doc/httpd, and www.apache.org/docs website. www.geocities.com/mk_azeem
Chapter 10: SENDMAIL

SENDMAIL is world's #1 and most widely used MTA (Mail Transfer Agent). Packages: sendmail-a.b (on server side) sendmail-cf-a.b (on server side) sendmail-doc-a.b (on server side) imap-a.b (on server side, to install both pop3 and imap services) Case Study: We want to establish a mail server in our example network. We will provide a mailing system so that the Linux clients can send and receive inter-office (within the mydomain.com) mails as well as mails coming and going out to rest of the world. While connected to the server via NIS the windows clients will be able to send and receive mails using SMTP, POP and/or IMAP services. In short it is assumed that all the users have their mail accounts on the mail server, so they have to connect to the server, to send and receive mails. They can send mail using our mail server as a relay even if they don't have an account on our mail server. In our example network, our domain name is mydomain.com. Our mail server is Server.mydomain.com on the IP 192.168.1.254. We have SMPT, POP3 and IMAP services running on our mail server. Configuration: It is said in *nix world that You are not a system administrator if you have not configured sendmail once in your life. And you are crazy if you have tried to do it twice. Well, that used to be said once m4 macro utility was not known. Sendmail configuration is possible using one of three methods: 1) By using linuxconf utility this is now obsolete. 2) By using m4 macro configuration utility 3) By editing the sendmail.cf file manually
First of all you should have a proper MX record set up for your mail server in the DNS zone files. Refer to the chapter DNS for this. In short, in addition to NS record you should have lines as following in your forward zone (mydomain.com.zone) file. @ @ IN IN NS MX Server.mydomain.com. 10 Server.mydomain.com.
The number 10 in the example above is the priority of this mail server in case there are more than one mail servers in your setup. The less the number, the higher the priority. The configuration file sendmail.cf used to be found in /etc. In newer versions like Red hat Linux 8.0 and 9, it is found in /etc/mail directory. The first method I am going to use is by m4 macro utility. There is a file present by default in your /etc/mail directory by the name sendmail.mc. Open this file and only change the lines which are shown below: [root@Server /]# vi /etc/mail/sendmail.mc
dnl DAEMON_OPTIONS(`Port=smtp,Addr=127.0.0.1, Name=MTA')dnl CLIENT_OPTIONS(`Family=inet,Address=192.168.1.0')dnl MASQUERADE_AS(`mydomain.com')dnl MASQUERADE_DOMAIN(localhost)dnl MASQUERADE_DOMAIN(localhost.localdomain)dnl The changes above indicate that SMTP daemon will listen on all interfaces of this machine. Notice that I have removed the 127.0.0.1 restriction by commenting the DAEMON_OPTIONS line. Similarly I have introduced my network address 192.168.1.0 in the CLIENT_OPTIONS indicating that only machines of my network can use the SMTP services of this machine. Any mail coming from localhost or localhost.localdomain as well as from any client machine will be translated / re-written as from mydomain.com . Now generate the sendmail.cf file by: [root@Server /]# m4 /etc/mail/sendmail.mc > /etc/mail/sendmail.cf Also open the /etc/mail/local-host-names file and add all possible names of your mail server on a new line in this file. [root@Server /]# vi /etc/mail/local-host-names mydomain.com Server.mydomain.com localhost localhost.localdomain By default your client computers will be denied relaying. That is that by default your client computers / users will not be able to send any mail destined outside mydomain.com. To enable this feature you can use /etc/mail/access.db file. This file is binary and cannot be directly edited. So you have to edit its text version (/etc/mail/access) and then generate the binary file from it. [root@Server /]# vi /etc/mail/access 192.168.1 RELAY mydomain.com RELAY Generate the binary access.db file from this file by : [root@Server /]# makemap hash /etc/mail/access.db < /etc/mail/access Start / restart your sendmail server by: [root@Server /]# service sendmail start
The other method to configure sendmail is by manually editing sendmail.cf . Open it with any text editor and change as following: [root@Server /]# vi /etc/mail/sendmail.cf ################## # local info # ################## # my LDAP cluster # need to set this before any LDAP lookups are done (including classes) #D{sendmailMTACluster}$m Cwlocalhost Server.mydomain.com mydomain.com ; (alternate / possible names for this machine) # file containing names of hosts for which we receive email Fw/etc/mail/local-host-names # my official domain name # ... define this only if sendmail cannot automatically #determine your domain #Dj$w.Foo.COM CP. # "Smart" relay host (may be null) DsServer.mydomain.com # (machine handling all the mail, optional) # class E: names that should be exposed as from this host, #even if we masquerade # class L: names that should be delivered locally, even if #we have a relay # class M: domains that should be converted to $M # class N: domains that should not be converted to $M #CL root C{E}root C{M}localhost C{M}localhost.localdomain #( Masquerade what ?) DMmydomain.com # (Masquerade as what domain? Emails will have this on the right side of @ sign) # my name for error messages DnMAILER-DAEMON
Then somewhere in the middle of the file, under the section OPTIONS, search for SMTP daemon options and modify (disable by putting remark) as follows: ############### # Options # ############### # SMTP daemon options #O DaemonPortOptions=Port=smtp,Addr=127.0.0.1, Name=MTA #or O DaemonPortOptions=Name=MTA # SMTP client options O ClientPortOptions=Family=inet, Address=192.168.1.0 Save the file and start the service by: [root@Server /]# /etc/rc.d/init.d/sendmail start or [root@Server /]# service sendmail restart Configuration on of I MAP and P OP3 services: Kernel 2.2: Edit the file /etc/inetd.conf and find the remarked lines with pop-3 and imap. Un-remark them (enable them by removing the remark so they look like following: [root@Server /]# vi /etc/inetd.conf pop-3 stream tcp nowait root /usr/sbin/tcpd ipop3d imap stream tcp nowait root /usr/sbin/tcpd imapd These services are listed in the file /etc/services with their respective port numbers. For pop3 it is 110 and for imap it is 143. Once done modifying, restart inetd by: [root@Server /]# killall -HUP inetd Kernel 2.4 In the /etc/xinetd.d directory you have separate files dedicated for each service. This is unlike the inetd.conf setup in kernel 2.2, in which all services are listed in one file. So edit the respective files in the /etc/xinetd.d directory and change the
disable=yes line to disable = no, as follows: [root@Server /]# vi /etc/xinetd.d/ipop3 # default: off # description: The POP3 service allows remote users to #access their mail \ # using an POP3 client such as NetscapeCommunicator, mutt,\ # or fetchmail. service pop3 { disable = no socket_type = stream wait = no user = root server = /usr/sbin/ipop3d log_on_success += HOST DURATION log_on_failure += HOST } Similarly modify the imap file as well as: [root@Server /]# vi /etc/xinetd.d/imap # default: off # description:The IMAP service allows remote users to #access their mail using\ # an IMAP client such as Mutt, Pine, fetchmail,or Netscape\ # Communicator. service imap { disable = no socket_type = stream wait = no user = root server = /usr/sbin/imapd log_on_success += HOST DURATION log_on_failure += HOST } Now restart the xinetd services by: [root@Server /]# service xinetd restart
We will see in the testing section that how can we verify if these services are up and working. Client side configuration: Windows clients: Configuring windows clients is not difficult. Open any email program like outlook express or Netscape mail. Create a new account for POP3 based service. Type in the user name and password you have on the mail server. In the outgoing SMTP server type in the IP address of your mail server as 192.168.1.254 on port 25, or if the DNS is working properly, then you can also type in the name of the server as Server.mydomain.com on port 25. Select the POP3 receive mail server and specify your mail server as 192.168.1.254 on port 110 or Server.mydomain.com on port 110. Similar configuration can be used to configure an IMAP account. But port number for IMAP server is 143. Linux clients: Linux clients using X windows system running GNOME or KDE (or any other) can again configure their accounts as they would have in microsoft windows tools. You can use KMAIL, or Evolution. The text clients can use pine or mail or other tools available widely for this purpose. The settings are similar as above. Create a new account for POP3 based service. Type in the user name and password you have on the mail server. In the outgoing SMTP server type in the IP address of your mail server as 192.168.1.254 on port 25, or if the DNS is working properly, then you can also type in the name of the server as Server.mydomain.com on port 25. Select the POP3 receive mail server and specify your mail server as 192.168.1.254 on port 110 or Server.mydomain.com on port 110. Similar configuration can be used to configure an IMAP account. But port number for IMAP server is 143. Log files: /var/log/messages /var/log/maillog Documentation and help /usr/share/doc/sendmail (directory) The website and FAQs on the website www.sendmail.org Testing: After setting up the sendmail.cf file by whatever method, and setting up the POP3 and IMAP services, you should be able to connect to the server using telnet, from server and client machines. Telnet service (on port 23) is not mandatory for these tests.
To test the SMTP server, use: [root@Server /]# telnet 192.168.1.254 25 or [root@Server /]# telnet Server.mydomain.com 25 You should see a message showing the SMTP server is responding. Trying 192.168.1.254... Connected to Server.mydomain.com (192.168.1.254). Escape character is '^]'. 220 Server.home ESMTP Sendmail 8.12.5/8.12.5; Sun, 10 Nov 2002 15:37:47 +0500 Press ^] here and type quit on the telnet> prompt. ^] telnet> quit Connection closed. Now to test the POP3 and IMAP servers, use the same telnet command using port 110 and 143 respectively. [root@Server /]# telnet Server.mydomain.com 110 Trying 192.168.1.254... Connected to Server.mydomain.com (192.168.1.254). Escape character is '^]'. +OK POP3 [192.168.1.254] v2001.78rh server ready ^] telnet> quit Connection closed. [root@Server /]# telnet Server.mydomain.com 143 Trying 192.168.1.254... Connected to Server.mydomain.com (192.168.1.254). Escape character is '^]'. * OK [CAPABILITY IMAP4REV1 LOGIN-REFERRALS STARTTLS AUTH=LOGIN] [192.168.1.254]
IMAP4rev1 2001.315rh at Sun, 10 Nov 2002 15:43:57 +0500 (PKT) ^] telnet> quit Connection closed. Sending mail using Sendmail directly, using telnet: Below is a telnet session which will help you understand how to send email directly using Sendmail. [root@Server /]# telnet Server.mydomain.com 25 Trying 192.168.1.254... Connected to Server.mydomain.com (192.168.1.254). Escape character is '^]'. 220 Server.mydomain.com ESMTP Sendmail 8.12.5/8.12.5; Fri, 18 Jul 2003 01:00:14 +0600 helo mydomain.com 250 Server.mydomain.com Hello Server.mydomain.com [192.168.1.254], pleased to meet you mail from: [email protected] 250 2.1.0 [email protected]... Sender ok rcpt to: [email protected] 250 2.1.5 [email protected]... Recipient ok (will queue) data 354 Enter mail, end with "." on a line by itself This is a test mail being sent from lab environment to my email address ! .2 50 2.0.0 h6HJ0D2i009597 Message accepted for delivery quit 221 2.0.0 Server.mydomain.com closing connection Connection closed by foreign host.
If at any instance the connection is refused then checks: 1) The inetd or xinetd services, they should be enabled 2) Check the sendmail.cf file for any errors 3) Check the firewall settings. There might be firewall rules which are blocking connections to the mail server
While connecting from client machines, using mail programs like outlook or Netscape mail, you might encounter problems like cannot find server Server.mydomain.com. If this is the case then tries the IP address of the mail server in place of the name. If this works then it means that your DNS is not properly configured. Check for MX records in the mydomain.com.zone file. Also check the reverse zone file for your domain. It should be noted here that there are various configuration files / databases in the /etc/mail directory. These are: access List of machines that can use sendmail for outbound mail. domaintable Specifies domain name mapping. local-host-names All possible names of the mail server for which the mail server should receive mails for. mailertable Special instructions for mail routing between various domains. virtusertable Mainly used in virtual hosting. relay-domains Used to permit any domain to relay mail using this mail server, denied by default. Like y.com wants to send mail to z.org using your mail server. Relaying denied error Normally the first and worst problem for an entry-level system administrator which shakes him once at least is thankfully, not much a problem. Re-create /etc/mail/access.db file by the following procedure: [root@Server /]# vi /etc/mail/access localhost.localdomain RELAY localhost RELAY 127.0.0.1 RELAY 192.168.1 RELAY mydomain.com RELAY Note: It is incorrect to write lines like the following for the file above: 192.168.1.0 RELAY (incorrect) *.mydomain.com RELAY (incorrect) After getting the prompt back, update the /etc/mail/access.db file by: [root@Server /]# makemap hash /etc/mail/access.db < /etc/mail/access
Now restart the Sendmail server. If you ever need to allow other domains to send mail to other domains using your mail server then you would configure /etc/mail/relay-domains like this: [root@Server /]# vi /etc/mail/relay-domains 192.168.1 mydomain.com Note: It is incorrect to write lines like the following in the file above: 192.168.1.0 RELAY (incorrect) 192.168.1 RELAY (incorrect) *.mydomain.com RELAY (incorrect) *.mydomain.com (incorrect) After getting the prompt back, restart the Sendmail server.
Chapter 11: DHCP

DHCP stands fro Dynamic Host Configuration Protocol. Hands out IP, Gateway, DNS server, etc, to the client machines. Packages: dhcp-devel-a.b dhcp-a.b Case Study: You have a simple network with a Server (192.168.1.254) trying to cope up with fulfilling all needs of the users. But now users want to have their IP assigned to them automatically whenever they boot and whatever operating system they boot up on their machines. Server side configuration: Normally you have a sample dhcpd file lying in /usr/share/doc/dhcp-3.0pl1 directory of your machine as dhcpd.conf.sample. Copy this file in your /etc directory and rename it as dhcpd.conf . Many people do a mistake here, they rename it to dhcp.conf and expect dhcpd service to work. It won't work like that. Once done, edit it as following: [root@Server /]# vi /etc/dhcpd.conf ddns-update-style interim; ignore client-updates;
subnet 192.168.1.0 netmask 255.255.255.0 { option routers 192.168.1.254; option subnet-mask 255.255.255.0; option nis-domain "logindomain"; option domain-name "mydomain.com"; option domain-name-servers 192.168.1.254; option time-offset -18000; # Eastern Standard Time option netbios-name-servers 192.168.1.254; # SAMBA WINS service range dynamic-bootp 192.168.1.1 192.168.1.200; default-lease-time 21600; max-lease-time 43200; }
Start dhcpd service by : [root@Server /]# service dhcpd start
Client side configuration: a) Windows client: Almost every window user knows that. Go to Network Neighborhood -> Properties -> TCP/IP [your adapter] -> Properties -> Obtain IP address automatically. Also select obtain DNS automatically. That is it. Depending on the version (and mood) of Windows, you may need to restart your client. b) Linux client: Edit /etc/sysconfig/network-scripts/ifcfg-eth0 file and set it up as following: [root@wks1 /]# vi /etc/sysconfig/network-scripts/ifcfg-eth0 DEVICE=eth0 BOOTPROTO=dhcp ONBOOT=yes DHCP_HOSTNAME=wks1 Please note that it is incorrect to write the FQDN here like this: DHCP_HOSTNAME=wks1.mydomain.com Once done, restart the network service of the client. Your client should pick up IP from DHCP server. [root@wks1 /]# service network restart Log files: /var/log/messages
Chapter 12: FTP

File Transfer Protocol. FTP is the easiest way to send and receive files over the network. Imagine a situation when you cannot send large files as attachments to some one. Even if you can, you never know that when that file / message will be sent and whether the other person's mail box was having enough room to accommodate your files attached with the mail message. Amongst many FTP servers, we will discuss only the VSFTPD server in this section. Packages: vsftpd-a.b
Something about F TP ports , A ctive mode and P assive mode FTP : Summary: In Active mode FTP, the FTP server connects to IP and port announced by client, using server's FTP data port (20). In Passive mode FTP, the server assigns / opens port on the server, tell the client about it, and then let the client connect to the new port. More detail: The FTP protocol first opens up a single connection that is called the FTP control session. When we issue commands through this session, other ports are opened to carry the rest of the data related to that specific command. These connections can be done in two ways, either actively or passively. When a connection is done actively, the FTP client sends the server a port and IP address to connect to. After this, the FTP client opens up the port and the server connects to that specified port from its own port 20 (known as FTP-Data) and sends the data over it. Passive FTP works the opposite way. The FTP client tells the server that some specific data is required, upon which the server replies with an IP address to connect to and at which port. The client will, upon receipt of this data, connect to that specific port, from its own port 20(the FTP-data port), and gets the data in question. Case Study: We have a small LAN, which has the server running FTP along DNS to provide file transfer facility to the users. Server side configuration: VSFTPD has the config file as /etc/vsftpd/vsftpd.conf , in Red Hat 9 . In Red Hat 8, it is /etc/vsftpd.conf. This file is quite straight-forward to configure. The default configuration will allow any user to connect to this ftp server as anonymous. If you want to restrict anonymous logins then set the following directives in this file as below: [root@Server /]# vi /etc/vsftpd/vsftpd.conf anonymous_enable=NO To enable local users to login to this FTP server, use: local_enable=YES Set up the welcome banner: ftpd_banner=Welcome to FTP service at mydomain.com.
This will make sure that users cannot change directory outside their home directories: chroot_local_user=YES Now start the vsftpd server by: [root@Server /]# service vsftpd start (RedHat9) Client side configuration: Nothing special is required. You can use any FTP client like: cuteFTP f/windows, gFTP f/Linux, or even ftp command from command line. Below is the command line of ftp session. Here is how we can connect to FTP server: [root@Server /]# ftp Server.mydomain.com Connected to Server.mydomain.com. 220 Welcome to FTP service at mydomain.com 530 Please login with USER and PASS. 530 Please login with USER and PASS. KERBEROS_V4 rejected as an authentication type Name (Server.mydomain.com:root): xena 331 Please specify the password. Password: 230 Login successful. Have fun. Remote system type is UNIX. Using binary mode to transfer files. How to check the directory listing and try to change directories. Notice that chroot environment prevents me from going outside my home directory: ftp>ls 227 Entering Passive Mode (127,0,0,1,52,103) 150 Here comes the directory listing. lrwxrwxrwx 1 500 500 27 Mar 06 04:22 xena -> /home/xena -rw-rw-r-- 1 500 500 4847 Mar 06 04:25 nbbn.sxw 226 Directory send OK. ftp> cd .. 250 Directory successfully changed. ftp> ls 227 Entering Passive Mode (127,0,0,1,192,249) 150 Here comes the directory listing. lrwxrwxrwx 1 500 500 27 Mar 06 04:22 xena -> /home/xena -rw-rw-r-- 1 500 500 4847 Mar 06 04:25 nbbn.sxw 226 Directory send OK.
ftp> cd / 250 Directory successfully changed. ftp> ls 227 Entering Passive Mode (127,0,0,1,53,220) 150 Here comes the directory listing. lrwxrwxrwx 1 500 500 27 Mar 06 04:22 xena -> /home/xena -rw-rw-r-- 1 500 500 4847 Mar 06 Jun 05 04:25 nbbn.sxw 226 Directory send OK. Now I try to upload a file (install.log) from my local directory (/root) . ftp> !ls anaconda-ks.cfg evolution install.log OpenOffice.org1.1.1 Desktop evolution_27Mar06.tar.bz2 install.log.syslog ftp> put install.log local: install.log remote: install.log 227 Entering Passive Mode (127,0,0,1,72,196) 150 Ok to send data. 226 File receive OK. 43909 bytes sent in 0.084 seconds (5.1e+02 Kbytes/s) ftp> ls 227 Entering Passive Mode (127,0,0,1,158,232) 150 Here comes the directory listing. -rw-r--r-- 1 500 500 43909 Mar 06 06:09 install.log lrwxrwxrwx 1 500 500 27 Mar 06 04:22 xena -> /home/xena -rw-rw-r-- 1 500 500 4847 Mar 06 04:25 nbbn.sxw 226 Directory send OK. Now I try to download a file (nbbn.sxw) from my ftp server directory to my local directory (/root) ftp> ls 227 Entering Passive Mode (127,0,0,1,141,229) 150 Here comes the directory listing. -rw-r--r-- 1 500 500 43909 Mar 06 06:09 install.log lrwxrwxrwx 1 500 500 27 Mar 06 04:22 xena -> /home/xena -rw-rw-r-- 1 500 500 4847 Mar 06 04:25 nbbn.sxw 226 Directory send OK. ftp> !ls anaconda-ks.cfg evolution install.log OpenOffice.org1.1.1 Desktop evolution_27Mar06.tar.bz2 install.log.syslog
ftp> get nbbn.sxw local: nbbn.sxw remote: nbbn.sxw 227 Entering Passive Mode (127,0,0,1,94,221) 150 Opening BINARY mode data connection for nbbn.sxw (4847 bytes). 226 File send OK. 4847 bytes received in 0.0097 seconds (4.9e+02 Kbytes/s) ftp> !ls anaconda-ks.cfg evolution install.log nbbn.sxw Desktop evolution_27Mar06.tar.bz2 install.log.syslog OpenOffice.org1.1.1 ftp> Log files: The log file for vsftp is /var/log/vsftpd.log. The two operations I just performed are logged in this file as: [root@Server /]# tail -f /var/log/vsftpd.log Thu Mar 06 12:09:28 2004 1 127.0.0.1 43909 /install.log b _ i r xena ftp 0 * c Thu Mar 06 12:26:56 2004 1 127.0.0.1 4847 /nbbn.sxw b _ o r xena ftp 0 * c
Chapter 13: FIREWALL

Firewall is a machine / mechanism, capable of protecting your network from any outside (or inside) attacks and abuses. It also helps you implement the kind of services you want to allow on your network. In other words, it allows you to decide that which all data packets should go out of your network and which should not, and vice versa. Packages: ipchains-a.b (For kernel 2.2.x) iptables-a.b (For kernel 2.4.x) Case Study: We have a small network, which has the server connected to the outer world by a modem. This means that our server (Server.mydomain.com in this case) will be the gateway for our network. In this scenario the server has two interfaces, eth0 and ppp0. We need to establish a firewall on this gateway machine in order to protect our network from outside attacks. Primarily we want the following:
Nobody should be able to establish a telnet or secure shell connection to our network from Internet. Nobody should be able to initiate (and establish) a connection from Internet with any of your machines other than our web server. Also we want to implement IP Masquerading or NAT, so that our clients can conveniently use their messenger tools like msn messenger and yahoo messenger, etc. Configuration For kernel 2.2: Kernel 2.2 has three paths (chains) for data i.e. INPUT, OUTPUT and FORWARD. List your rules in a script on the gateway machine, i.e. Server.mydomain.com, as these rules are flushed or erased at each computer reboot. You can save your rules in a script say /etc/rc.firewall and call it from /etc/rc.local script. Or you can even put them directly (and simply) in /etc/rc.local Now, to block any incoming telnet requests from outside (not from our network), destined for our network, we will use the following command on our Server.mydomain.com, which is the gateway computer for our network: [root@Server /]# ipchains -A input -i ppp0 -p tcp --dport 22 -j DENY [root@Server /]# ipchains -A input -i ppp0 -p tcp --dport 23 -j DENY To allow any web requests to your web server, use: [root@Server /]# ipchains -A input -i ppp0 -p tcp --dport 80 -j ACCEPT To allow our network clients to access the Internet for IRC, MSN messenger services, etc, follow: 1) Add the following line to our firewall script, which says that rewrite all packet headers to that IP which is assigned to us from our ISP, originating from our lan and are destined anywhere other than our lan: [root@Server /]# ipchains -A forward -i eth0 -o ppp0 -p tcp -j MASQ 2) Enable IP Forwrding in the file /proc/sys/net/ipv4/ip_forward by : [root@Server /]# echo 1 > /proc/sys/net/ipv4/ip_forward
Configuration For kernel 2.4: Kernel 2.4 have five paths (chains) for data i.e. INPUT, OUTPUT, FORWARD, PREROUTING and POSTROUTING. All commands are case sensitive. List your rules in a script on the gateway machine, i.e. Server.mydomain.com, as these rules are flushed or erased at each computer reboot. You can save your rules in a script say /etc/rc.firewall and call it from /etc/rc.local script. Or you can even put them directly (and simply) in /etc/rc.local To block any incoming telnet requests from Internet (not from our network), destined for our network, we will use: [root@Server /]# iptables -A INPUT -i ppp0 -p tcp --dport 23 -j DROP Also [root@Server /]# iptables -t nat -A PREROUTING -i ppp0 -p tcp --dport 23 j DROP To allow any web requests to and from your web server, use the following two commands: [root@Server /]# iptables -A INPUT -i ppp0 -p tcp --dport 80 -j ACCEPT [root@Server /]# iptables -A OUTPUT -o ppp0 -p tcp -d 0/0 -dport 80 -j ACCEPT The following allows connections (initiated by outsider) seeking port 80 from outside on your computer but not any other. [root@Server /]# iptables -A INPUT -i ppp0 -p tcp -syn -dport !80 -j DROP But to deny any connections from any of the outsider to any of your machine on port 80, other than your server, use: [root@Server /]# iptables -A FORWARD -i ppp0 -o eth0 -dport 80 -j DROP To allow our network clients to access the Internet for IRC, MSN messenger services, etc, follow:
1) Add the following line to our firewall script, which says that rewrite all packet headers to that IP which is assigned to us from our ISP, originating from our lan and are destined anywhere other than our lan: [root@Server /]# iptables -t nat -A POSTROUTING -o ppp0 -p tcp -j MASQUERADE 2) Enable IP Forwarding in the file /proc/sys/net/ipv4/ip_forward by: [root@Server /]# echo 1 > /proc/sys/net/ipv4/ip_forward Transparent proxy: If you want to run SQUID as a transparent proxy server and your SQUID machine is the gateway machine in your network, then you need to pickup any request for port 80 from your network before they get to the Internet and redirect to the local proxy server, SQUID, running on port 3128. In order for transparent proxy to work, you need to have transparent proxy support compiled into your kernel. Use the following command on kernel 2.2 machine: [root@Server /]# ipchains -A input -p tcp -i eth0 --dport 80 -j REDIRECT 3128 Use the following command on kernel 2.4 machine: [root@Server /]# iptables -t nat -A PREROUTING -i eth0 -p tcp --dport 80 -j REDIRECT --to-port 3128 Now let us see some explanation of DNAT, SNAT and MASQUERADE targets (including diagrams) MASQUERADE: If you have a network of computers and your dialup machine is linux based and each time you connect to your ISP , you get a different IP address (dynamic public IP), then you need masquerading to allow your client computers web traffic to go out on the internet. It should be noted here that even you don't have SQUID or you don't want to run SQUID and just want to enable windows internet connection sharing type of thing to be implemented, then all you need is the following two commands on your linux dialup machine: [root@Server /]# echo 1 > /proc/sys/net/ipv4/ip_forward [root@Server /]# iptables -t nat -A POSTROUTING -o ppp0 -p tcp -j MASQUERADE Note that MASQUERADEing also works with dynamic private IPs.
SNAT: SNAT is just like MASQUERADE but it is twice fast as compared to MASQUERADE. It is used where we have the same IP each time you connect to your ISP (static public IP). See Fig 1. In this case you don't want time to be wasted by checking each time the IP of the outgoing interface, because you already know it. So just mention that in the iptables command. Assume you get 203.135.10.15 as IP of your modem interface (ppp0), each time when you connect to your ISP, then the internet connection sharing commands will be: [root@Server /]# echo 1 > /proc/sys/net/ipv4/ip_forward [root@Server /]# iptables -t nat -A POSTROUTING -o ppp0 -p tcp -j SNAT -to-source 203.135.10.15 As mentioned earlier, SNAT is twice fast / efficient than MASQUERADE, but works with static IPs only (private and public both). DNAT: In a small network, normally people use SQUID on the same machine which is their dialup machine. Or in other words, they normally run SQUID on the gateway machine. But what if they do not want to run SQUID on the gateway machine? See Figure 2.
Such a case is when you have SQUID on a non-gateway machine on the network with only one network interface card (eth0, with IP: 192.168.1.252, on port 3128) . In such case any web traffic coming from the LAN needs to be directed towards this machine, if transparent proxy setup is desired. This can be achieved by the following commands on the (linux) gateway machine: [root@Server /]# iptables -t nat -A PREROUTING -i eth0 -s ! 192.168.1.252 -p tcp --dport 80 -j DNAT -to-destination 192.168.1.252:3128 DNAT can also be used where you have a setup such as you have your web server on the LAN with only one interface (eth0, with IP: 192.168.1.251). See Figure 3.
You have a permanent (public) IP assigned to your dialup linux machine (202.135.10.18), each time you connect to your ISP. You have your web server registered on the Internet against this IP, BUT you don't have any web server running on this IP.Why? Because as already mentioned you have your webs server in your LAN on a private IP 192.168.1.251. Now you want to redirect any web requests for your web server coming from internet to your server lying inside your LAN. You can do it by: [root@Server /]# iptables -t nat -A PREROUTING -i ppp0 -d 202.135.10.18 -p tcp --dport 80 -j DNAT -to-destination 192.168.1.251:80
Sample firewall configuration / script: A very common request. Especially people ask how to block all those ICMP requests coming in our network from outside (and/or inside)? Plus administrators are too irritated to see people trying to connect to their servers from outside using telnet or ssh, etc. Below is a sample script which you can use for both in networked scenario and standalone scenario. Save this as a script, preferably as /etc/rc.firewall, make it executable and call it from /etc/rc.local [root@Server /]# vi /etc/rc.firewall
#!/bin/bash #following three lines provides Masquerading for our #network. # You don't need these on a stand alone computer. echo Enabling Masquerading ... echo 1 > /proc/sys/net/ipv4/ip_forward iptables -t nat -A POSTROUTING -s 192.168.1.0/24 -d ! 192.168.1.0/24 -o ppp0 -j MASQUERADE # following two lines redirects www (port 80) requests to #local squid # service. Do not use if you don't want to enable #transparent proxying or if you don't know about #transparent proxying. # Needs some configuration in squid.conf file as well. echo Enabling redirection for port 80 assuming Firewall + SQUID on same machine ... iptables REDIRECT iptables REDIRECT -t nat -A --to-port -t nat -A --to-port PREROUTING -i eth0 -p tcp --dport 80 -j 3128 PREROUTING -i eth0 -p udp --dport 80 -j 3128
#iptables -t nat -A PREROUTING -i eth0 -s ! 192.168.1.251 p tcp --dport 80 -j DNAT --to-destination 192.168.1.251:3128 #iptables -t nat -A PREROUTING -i eth0 -s ! 192.168.1.251 p udp --dport 80 -j DNAT --to-destination 192.168.1.251:3128 # Drop connections from internet for ports 0-1023, which are targeting this machine iptables -A INPUT -i ppp0 -p tcp --destination-port 0:1023 -j DROP iptables -A INPUT -i ppp0 -p udp --destination-port 0:1023 -j DROP # lets drop traffic other than 20,21,80,1863 MSN ,5050 yahoo iptables -A FORWARD -i eth0 -o ppp0 -p tcp -m multiport --dport 20,21,110,1863,443,11999,5050,2002,2004 -j ACCEPT iptables -A FORWARD -i eth0 -o ppp0 -p udp -m multiport
--dport 20,21,110,1863,443,11999,5050,2002,2004 -j ACCEPT # the line right next is for FTP ESTABLISHED sessions iptables -A FORWARD -p tcp -m state --state ESTABLISHED,RELATED -j ACCEPT iptables -A FORWARD -p udp -m state --state ESTABLISHED,RELATED -j ACCEPT # 11999 is yahoo games # MSN needs both 1863 and 443, Yahoo needs 5050 (and probably 443 as well) iptables -A FORWARD -i eth0 -o ppp0 -p tcp -m multiport --dport 20,21,110,1863,443,11999,5050,2002,2004 -j ACCEPT #iptables -A FORWARD -i eth0 -o ppp0 -p tcp -d 0/0 --dport 81:65535 -j DROP iptables -A 135,139,445 iptables -A 135,139,445 FORWARD -p tcp -m multiport --destination-port -j DROP FORWARD -p udp -m multiport --destination-port -j DROP
# Also drop connections to this machine from internet for the following multiple ports # my SQUID , postgresql database server and X11 interface iptables -A INPUT -i ppp0 -p tcp -m multiport -destination-port 3128,5432,6000 j DROP iptables -A INPUT -i ppp0 -p udp -m multiport -destination-port 3128,5432,6000 j DROP # And the following DROPS icmp packets coming from internet iptables -A INPUT -i ppp0 -p icmp -j DROP Assign execute permission to the above script by: [root@Server /]# chmod u+x /etc/rc.firewall Edit /etc/rc.local , and add a line to call the above script: [root@Server /]# vi /etc/rc.local #!/bin/sh # #This script will be executed *after* all the other init scripts.
# You can put your own initialization stuff in here if you don't # want to do the full Sys V style init stuff. touch /var/lock/subsys/local /etc/rc.firewall Log files: There are no log files as such, but still you can get most out of /var/log/messages and all log files related to particular services, to check whether desired connections are being made or not. Documentation and Help: www.flounder.net/ipchains/ipchains-howto.html Iptables tutorial 1.1.19 by Oskar Andreason Iptables Basics NHF by prince_kenshi Testing: You can test your Firewall by directly running the related applications like your web browser or try connecting from outside to your computer using telnet, etc. In order to check the current state of your rules you can use: [root@Server /]# ipchains -L (For kernel 2.2) or [root@Server /]# iptables -L (For kernel 2.4) [root@Server /]# iptables -t nat -L (For Kernel 2.4) At any time you can flush your current rule set by: [root@Server /]# ipchains -F (For Kernel 2.2 ) or [root@Server /]# iptables -F (For Kernel 2.4 ) [root@Server /]# iptables -t nat -F (For Kernel 2.4 )
Chapter 14: KERNEL MODIFICATION AND COMPILATION

There are times when you need to recompile your kernel with more (or less) features. Packages: kernel-source-a.b kernel-utils-a.b
glibc-kernheaders-a.b Case Study: Assuming that you want to add NTFS read feature to your kernel. Procedure: Check your kernel version by: [root@Server /]# uname -r Then go to the kernel source directory /usr/src and further into the directory which has the same name / number as your kernel version number, returned to you by the uname command above. On a RedHat 8 system it is 2.4.18-14 [root@Server /]# cd /usr/src/linux-2.4.18-14 Now you can run the kernel configuration by one of three methods: 1) By make config from a command prompt, which will ask you a lot of Yes and No questions. This takes up a lot of time and is rather now obsolete method. [root@Server linux-2.4.18-14]# make config 2) By make menuconfig from the command prompt, which will show a text based menu system and you can select different options and save changes. [root@Server linux-2.4.18-14]# make menuconfig 3) And by make xconfig from a command terminal running in Xwindows. This method is very user friendly. [root@Server linux-2.4.18-14]# make xconfig So for now we will use make menuconfig. Once you are done with your settings of NTFS read feature, you can save the kernel configuration while exiting out of the program. After save and exit, you have to run the following commands in the order in which they appear: [root@Server [root@Server [root@Server [root@Server linux-2.4.18-14]# linux-2.4.18-14]# linux-2.4.18-14]# linux-2.4.18-14]# make make make make dep clean bzImage modules
It may take from few minutes to few hours, depending on the power of your particular machine.
After all of above steps are completed and the prompt returns, you have to copy newly created bzImage to your /boot directory: [root@Server linux-2.4.18-14]# cp arch/i386/boot/bzImage /boot/vmlinuz-ntfs Then issue the commands: [root@Server linux-2.4.18-14]# make modules_install [root@Server linux-2.4.18-14]# cp System.map /boot/System.map-ntfs [root@Server linux-2.4.18-14]# cp .config /boot/config-ntfs The last command is not really necessary. But is is useful to keep the kernel configuration at a separate place as a backup. Then create a boot driver RAM disk file (optional): [root@Server linux-2.4.18-14]# mkinitrd /boot/initrdntfs.img 2.4.18-14 Then you need to update your boot loader configuration as well. LILO: In case you are running LILO, update the /etc/lilo.conf and add a section for your new kernel: [root@Server /]# vi /etc/lilo.conf prompt timeout=50 default=linux boot=/dev/hda map=/boot/map install=/boot/boot.b message=/boot/message linear image=/boot/vmlinuz-2.4.18-14 label=linux initrd=/boot/initrd-2.4.18-14.img read-only append="root=LABEL=/" image=/boot/vmlinuz-ntfs label=linux with NTFS support
initrd=/boot/initrd-ntfs.img read-only append="root=LABEL=/" other=/dev/hda2 optional label=DOS After saving the LILO configuration, run LILO boot map installer: [root@Server /]# /sbin/lilo -v Now you are done, reboot your machine and there will be a new entry in the boot menu Linux with NTFS support, select that and press enter to boot from this kernel. GRUB: In case you have GRUB, then you need to update the /boot/grub/grub.conf file and add a section relating to your new kernel image. [root@Server /]# vi /boot/grub/grub.conf # NOTICE: You have a /boot partition. This means that # all kernel paths are relative to /boot/ default=0 timeout=30 splashimage=(hd0,0)/grub/splash.xpm.gz title Red Hat Linux (2.4.18-14) root (hd0,0) kernel /vmlinuz-2.4.18-14 ro root=/dev/hda3 initrd /initrd-2.4.18-14.img title Red Hat Linux with ntfs support (2.4.18-14) root (hd0,0) kernel /vmlinuz-ntfs ro root=/dev/hda3 initrd /initrd-ntfs.img Now reboot your machine. There is no need to re-run grub program itself to update the boot record with new entries. GRUB reads it's config file at each startup and acts accordingly. For emergencies, you should also make a boot disk before and after kernel compilation: [root@Server /]# mkbootdisk --device /dev/fd0 --verbose 2.4.18-14
Note: If, after kernel compilation, you forgot to run /sbin/lilo and if you did not have an entry for an older kernel in lilo.conf, you have a problem. In many cases, you can boot your Red Hat Linux system from the Red Hat Linux boot disk with your root file system mounted and ready to go. Enter the following command at the boot disk's boot: prompt: linux single root=/dev/hdXX initrd= In the above command, replace the XX in /dev/hdXX with the appropriate letter and number for your root partition. This command, first, starts the boot process in single-user mode, with the root partition set to your root partition. The empty initrd specification bypasses the installation-related image on the boot disk, which will cause you to enter single-user mode immediately. But if your system is SCSI-based, you will not be able to do this. In that case, you will have to access rescue mode using the linux rescue command by using the installation CD # 1, which is the next topic.
Chapter 15: Single user and Rescue mode

Rescue mode: Rescue mode provides the ability to boot a small Linux environment entirely from a diskette, CD-ROM, or using some other method. Normally, you will need to get into rescue mode for one of two reasons: You are unable to boot Linux. (Happens normally when you install some operating system over / after installing linux, like installation of windows over / after installing Linux) You are having hardware or software problems, and you want to get a few important files off your system's hard drive before starting OPERATION UNDER KNIFE. To boot your system in rescue mode, boot from a Red Hat Linux boot disk or Red Hat Linux CD-ROM #1, and enter the following command at the installation boot prompt: boot: linux rescue After booting off a boot disk or Red Hat Linux CD-ROM #1 and providing a valid rescue image, you will see the following message: The rescue environment will now attempt to find your Red Hat Linux installation and mount it under the directory /mnt/sysimage. You can then make any changes required to your system. If you want to proceed with this step choose 'Continue'. If for some reason this process fails you can choose 'Skip' and this step will be skipped and you will go directly to a command shell.
If you select Continue, it will attempt to mount your file system under the directory /mnt/sysimage. If it fails to mount a partition, it will notify you. If you select Skip, your file system will not be mounted. Choose Skip if you think your file system is corrupted. Once you have your system in rescue mode, bash # prompt appears. If you selected Continue to mount your partitions automatically and they were mounted successfully, you are in single-user mode. From here you can do some of your normal tasks like copying files, changing root password, changing network settings, etc. If your file system was mounted properly during rescue boot, and you want to make your system the root partition, use the command: chroot /mnt/sysimage. This is useful if you need to run commands such as rpm that require your root partition to be mounted as /. To exit the chroot environment, type exit, and you will return to the bash# prompt. # chroot /mnt/sysimage However, if your partitions cannot be mounted automatically then first check the list of partitions and try to identify them by: # fdisk -l Then you can mount them by first creating some temporary mount point like /tempusr and mount your partition /dev/hda5 on that mount point (which is your /usr infact). # mkdir /tempusr # mount -t ext2 /dev/hda5 /tempusr Your partition can be of any type, ext2 or ext3, etc. Now if the MBR was over written by some operating system, which you installed after installing Linux, and you were using lilo as boot loader then you can restore / recreate your boat loader by using lilo -r as: # lilo -r /mnt/sysimage (if chroot command is not already issued) or simply # lilo -r / (if chroot command is already given and you are in chroot environment) Added DOS Added Linux-ntfs Added Linux-2.4.18-14
And if you were using GRUB then you would restore your GRUB boot loader by: # grub-install /dev/hda TO synchronize, issue sync command by: # sync And then exit twice to reboot the system. # exit # exit The system will reboot at this point. If using GRUB as boot loader, and you want to recreate your boot loader then you have two options ( in detail) to follow: Method1: Installing GRUB using floppy: You need a GRUB boot floppy to handle the situation. Create one and reboot your computer with it. Once started, GRUB will show the command-line interface. First, set the GRUB's "root device" to the boot directory, like this: grub> root (hd0,0) If you are not sure which partition actually holds these files, use the command `find' like this: grub> find /boot/grub/stage1 This will search for the file name /boot/grub/stage1 and show the devices which contain the file. Once you've set the root device correctly, run the command `setup' grub> setup (hd0) This command will install GRUB on the MBR in the first drive. If you want to install GRUB into the "boot sector" of a partition instead of the MBR; specify a partition into which you want to install GRUB: grub> setup (hd0,0) If you install GRUB into a partition or a drive other than the first one, you must chainload GRUB from another boot loader. Refer to the manual for the respective boot loader to know how to chain-load GRUB. Now you can boot GRUB without a GRUB floppy.
Method2: Using grub-install to instal l GRUB: The usage is basically very easy. You only need to specify one argument to the program, namely, where to install GRUB. The argument can be either of a device file or a GRUB's drive/partition. So, the following will install GRUB into the MBR of the first IDE disk under Linux: # grub-install /dev/hda If it is the first BIOS drive, this is the same as well: # grub-install '(hd0)' (similar to previous command) But all the above examples assume that you use GRUB images under the root directory. If you want GRUB to use images under a directory other than the root directory, you need to specify the option `--root-directory'. The typical usage is that you create a GRUB boot floppy with a file system. Here is an example: # # # # mke2fs /dev/fd0 mount -t ext2 /dev/fd0 /mnt/floppy grub-install root-directory=/mnt/floppy '(fd0)' umount /mnt/floppy
Another example is in case that you have a separate boot partition which is mounted at `/boot'. Since GRUB is a boot loader, it doesn't know anything about mount points at all. Thus, you need to run `grub-install' like this: # grub-install --root-directory=/boot /dev/hda By the way, as noted above, it is quite difficult to guess BIOS drives correctly under a UNIX-like OS. Thus, `grub-install' will prompt you to check if it could really guess the correct mappings, after the installation. Please be careful enough. If the output is wrong, your computer may get stuck at boot-up. Note that `grub-install' is actually just a shell script and the real task is done by the grub shell `grub'. Therefore, you may run `grub' directly to install GRUB, without using `grub-install'. Don't do that, however, unless you are very familiar with the internals of GRUB. Installing a boot loader on a running OS may create unwanted problems. Single-User mode: Normally used when you forget your root password and you have no other way to get in and change it. If you are using GRUB, use the following steps to boot into singleuser mode: If you have a GRUB password configured, type p and enter the password.
Select Red Hat Linux with the version of the kernel that you wish to boot and type e for edit. You will be presented with a list of items in the configuration file for the title you just selected. Select the line that starts with kernel and type e to edit the line. Go to the end of the line and type single or 1 as a separate word (press the [Space bar] and then type single or 1). Press [Enter] to exit edit mode. Back at the GRUB screen, type b to boot into single user mode. If you are using LILO, just specify linux single or linux 1 or linux emergency at the LILO boot prompt (if you are using the graphical LILO, you must press [Ctrl]-[x] to exit the graphical screen and go to the boot: prompt), as: boot: linux single boot: linux emergency In single-user mode boot the computer to runlevel 1. Your local file systems will be mounted, but your network will not be activated. You will have a usable system maintenance shell. In emergency mode, you are booted into the most minimal environment possible. The root file system will be mounted read-only and almost nothing will be set up. The main advantage of emergency mode over single user mode is that your init files are not loaded. If init is corrupted or not working, you can still mount file systems to recover data that could be lost during a re-installation. FSCK: File System ChecK (fsck) is a utility like scandisk or chkdsk utility in DOS. This can be used to check (and repair) errors on partition level (or to create them if you wish!). But this is to be kept in mind that YOU SHOULD NEVER RUN FSCK ON A MOUNTED FILESYSTEM. If you do that, it may (and probably will) destroy your data. If you ever want to run fsck on a file system, first unmount it and then run fsck on it. In case of /boot or /usr or other mount points this may be done by unmounting them. But you will be unable to un-mount root (/) mount point to run fsck on it. For that it is better to restart in single user mode. In single user mode the root (/) file system is mounted read-only. By read only I do not mean that no writing will be allowed on the root partition. Rather the point here is that linux uses Asynchronous File System by default. So in a normal mounted file system, the state of a file might be different in memory as compared to on disk at any given time. Assuming you changed a file and saved it. For you you have saved it but linux has lied to you and is in fact waiting for a proper time to flush the (buffer) changes to the disk. This is change of state of a file in memory and on disk. So if you run fsck on such a file system, the files may get repaired / deleted for inconsistencies, still having their related data in memory (dirty buffers). And you get a (very ugly) and unwanted (corrupted) file system rather repaired one. So what read-only mounted file system means is that writing is allowed as normal, but there will be no data in the buffers. So fsck does what it is supposed to do, that is repair a file system. If you want you can remount a file system read-only at any time by:
[root@Server /]# mount -o remount,ro / Normally you run fsck like: [root@Server /]# fsck /dev/hda1 Here is a sample on how to run fsck on /boot partition which is /dev/hda1 on my system. [root@Server /]# fsck /dev/hda1 fsck 1.32 (09-Nov-2002) e2fsck 1.32 (09-Nov-2002) /dev/hda1 is mounted. WARNING!!! Running e2fsck on a mounted filesystem may cause SEVERE filesystem damage. Do you really want to continue (y/n)? no check aborted. [root@Server /]# umount /boot [root@Server /]# fsck /dev/hda1 fsck 1.32 (09-Nov-2002) e2fsck 1.32 (09-Nov-2002) /boot: clean, 41/26104 files, 12623/104391 blocks [root@Server /]# mount /dev/hda1 /boot Second Hard drive attached d , a bnormal boot up : It is some times seen that if a second (linux installed) hard drive is attached to an already working linux machine, then linux may boot from the other / new hard drive, which is unexpected and of-course un-wanted. What happens is that you have partition labels, not partition numbers in the /etc/inittab file . Once you add another drive which has these partition labels as well, then for some reason, linux picks up that partition as root partition which has label: / and has a lesser partition number in both drives. e.g. On drive /dev/hda , you have root / partition on /dev/hda7 . Once you add another hard drive, say /dev/hdc, and that had linux already installed on it with a root / partition as /dev/hdc2, then on the boot up time, linux will use this root (/dev/hdc2) as root partition and a lot of havoc will be created. You may be asked for root password or to run fsck, etc, etc. And you might be wondering that just minutes ago everything was working perfectly fine. So what you can do to avoid such situation is either:
a) Boot with a special boot time kernel parameter, this may still not get you desired results b) Edit your boot loader files (and /etc/fstab optionally) before attaching the second hard drive or a) Boot with boot time kernel parameter: On LILO ( on my machine) , in the above described scenario, pass an extra parameter : boot: linux root=/dev/hda7 This will over ride the default root partition label mentioned in the /etc/lilo.conf as: append="root=LABEL=/" On GRUB, edit the kernel line : kernel /boot/vmlinuz ro root=LABEL=/ and change it to : kernel /boot/vmlinuz ro root=/dev/hda7 b) Solve this problem permanently by editing your boot loader files as per above discussion. Also since your /etc/fstab file is still having labels for your non-root partitions as well, this might still because problems during booting even if you override root partition location by passing boot time parameter as discussed above. So edit your /etc/fstab file and change the lines having labels to actual partition numbers as per your system. e.g assume you have part of your /etc/fstab as : LABEL=/ / ext3 defaults 1 1 LABEL=/boot /boot ext3 defaults 1 2 Then change it to: /dev/hda7 / ext3 defaults 1 1 /dev/hda1 /boot ext3 defaults 1 2 Installing another version of linux on another partition on and boot problems related to that: This is a very important scenario, and believes me many people just get frustrated during this situation. This kind of situation arises when you already have linux, say
RedHat9 installed on your machine, along with windows and you happen to install another operating system such as Fedora core on another partition. Here is what I have and how I would explain that to you. Below is the output of fdisk -l command on my machine. [root@Server /]# fdisk -l Device Boot Start End Blocks Id System /dev/hda1 1 13 104391 83 Linux /dev/hda2 14 2308 18434587+ 83 Linux /dev/hda3 * 2309 2818 4096575 c W95 FAT32 (LBA) /dev/hda4 2819 4865 16442527+ f W95 Ext'd (LBA) /dev/hda5 2819 3583 6144831 83 Linux /dev/hda6 3584 3599 128488+ 82 Linux swap /dev/hda7 3600 4865 10169113+ 83 Linux The scenario is like this, I had Redhat9 installed on my machine on /dev/hda1 (/boot), /dev/hda7 (/) and /dev/hda6 (swap). I also had windows installed on my machine on /dev/hda3. I already had /dev/hda5 reserved for various experimentation.So I used this partition to install Fedora. So the Fedora was installed on /dev/hda5 (/) and /dev/hda6 (swap), with no separate /boot partition. After installation, Fedora over-wrote my MBR with new GRUB and at the next boot up I had only Fedora core 2 and Others in the GRUB menu. On selecting Others , the computer started booting Windows. This was very undesirable, as I was unable to see (or run) my Redhat9 installation in the menu. When booted up in Fedora, I examined the /etc/grub.conf file. It looked like this: [root@Server /]# /etc/grub.conf # grub.conf generated by anaconda # #Note that you do not have to rerun grub after making #changes to this file # NOTICE: You do not have a /boot partition. This means #that all kernel and initrd paths are relative to /, eg. # root (hd0,4) # kernel /boot/vmlinuz-version ro root=/dev/hda5 # initrd /boot/initrd-version.img #boot=/dev/hda default=0 timeout=10 splashimage=(hd0,4)/boot/grub/splash.xpm.gz title Fedora Core (2.6.5-1.358) root (hd0,4)
kernel /boot/vmlinuz-2.6.5-1.358 ro root=LABEL=/1 rhgb quiet initrd /boot/initrd-2.6.5-1.358.img title Other rootnoverify (hd0,2) chainloader +1 It should be noted here that I installed Fedora using only two partitions (/) and (swap) . It doesn't utilize any /boot partition. This further means that the kernel image is not on a separate partition, but rather withing the (/) partition in the /boot directory. The first line below the title line in the section above shows: root (hd0, 4). This line means that kernel boot files will be on the /dev/hda5 partition (hard drive 0, partition number 4, if counted from zero). Notice that this is in fact our (/) partition. Since boot files are on the (/) partition, that is why the kernel line below it contains the absolute path: kernel /boot/vmlinuz-2.6.5-1.358 ro root=LABEL=/1 rhgb quiet Since the current boot loader is of Fedora, I need to mention my RedHat installation path in this file. I add the following section to this file. title Redhat 9 root (hd0,0) kernel /vmlinuz-2.4.20-8 ro root=/dev/hda7 initrd /initrd-2.4.20-8.img Before trying to understand this piece of code, please recall that my Redhat9 installation uses three partitions, which are: /dev/hda1 (/boot) or (hd0, 0) /dev/hda6 (swap) or (hd0, 5) /dev/hda7 (/) or (hd0, 6) Since my kernel boot files resides on a separate (/boot) partition, which is /dev/hda1 that is why I have placed root (hd0, 0) below title Redhat line. And since the boot partition is separate, all paths are relative to that. So that is why the kernel and initrd lines do not have the /boot mentioned in them. Once I updated my /etc/grub.conf of my Fedora installation and rebooted, I was able to see and run my Redhat9 installation without any problems. Alternative method: There is another method to make this all work. That method is based on the assumption that you want to restore the boot loader of Redhat9 installation and call / access / boot the
newly installed Fedora installation from there. For that you need to follow the following steps. Boot from Redhat9 CD 1, on the install boot prompt, type: linux rescue. The rescue mode will ask you that it will try to find your Redhat9 partitions and mount them. Choose continue here. You will be presented a prompt, with information that your Redhat9 installation has been found and has been mounted under the mount point: (/mnt/sysimage). You need to chroot to this mount point by: bash $ chroot /mnt/sysimage [root@Server /]# This is your Redhat9 installation area. Once there, edit your /etc/grub.conf and introduce the section for your Fedora installation, like: [root@Server /]# vi /etc/grub.conf # grub.conf generated by anaconda # #Note that you do not have to rerun grub after making changes to this file # NOTICE: You have a /boot partition. This means that # all kernel and initrd paths are relative to /boot/, eg. # root (hd0,0) # kernel /vmlinuz-version ro root=/dev/hda7 # initrd /initrd-version.img #boot=/dev/hda default=0 timeout=10 splashimage=(hd0,0)/grub/splash.xpm.gz title Red Hat Linux (2.4.20-8) root (hd0,0) kernel /vmlinuz-2.4.20-8 ro root=/dev/hda7 hdd=idescsi initrd /initrd-2.4.20-8.img title Windows rootnoverify (hd0,2) chainloader +1 title Fedora Core (2.6.5-1.358) root (hd0,4) kernel /boot/vmlinuz-2.6.5-1.358 ro root=/dev/hda5 rhgb quiet initrd /boot/initrd-2.6.5-1.358.img Notice the presence of /boot in the kernel and initrd lines in this new section. And notice the absence of /boot in the kernel and initrd lines in the Redhat9 section.
Once you update your /etc/grub.conf in this Redhat9 installation area. You need to re-install your Redhat9 version of GRUB boot loader on the MBR of your hard drive before rebooting. Do this by: [root@Server / ]# grub-install /dev/hda Once done, you need to exit this chroot environment and reboot normally. For that use: [root@Server / ]# sync [root@Server / ]# exit [root@Server / ]# exit The second exit above will reboot the system. And you will get your system back again with added menu for Fedora.
Chapter 16: Advance permissions

In this topic we would briefly describes setgid, setuid, sticky bits, umask and sudo. Setuid, setgid and sticky bit: Till now you have used three octal numbers for permissions for three types of users, i.e. Owner, group and others. These three octal numbers / digits have values from 000 till 777, representing: no permissions to any one till all permissions for everyone, respectively. There are in fact four octal number positions designed for file and directory permissions. The fourth number happens to be on the left side of any permission set, and if not present, then is assumed to be 0 (zero). This number can be 1 or 2 or 4 maximum and its special meaning are below: 1 2 4 Sticky bit Set group ID on execution Set user ID on execution
Now first s ticky bit: If sticky bit permission is applied on a file like: [root@Server / ]# chmod 1755 myfile.txt Then as per the man page: On older Unix systems, the sticky bit caused executable files to be hoarded in swap space. This feature is not useful on modern VM systems, and the Linux kernel ignores the sticky bit on files. Other kernels may use the sticky bit on files for system-defined purposes. On some systems, only the superuser can set the sticky bit on files. What this means is that sticky bit on files is useless in linux.
Sticky bit on a file can be checked by the ls command. The T indicates that this file has sticky bit on. [root@Server / ]# ls -l -rw-r--r-T 1 root root 4847 Jul 8 12:26 myfile.txt And if sticky bit is implemented on a directory like: [root@Server / ]# chmod 1777 mydirectory Then as per the man page: When the sticky bit is set on a directory, files in that directory may be unlinked or renamed only by root or their owner. Without the sticky bit, anyone able to write to the directory can delete or rename files. The sticky bit is commonly found on directories, such as /tmp, that are world-writable. What this means is that sticky bit on a directory allows users to create files in that directory with a facility that only root or the user who created the file will be able to delete or rename the file. Sticky bit on a directory can be checked by the ls command. The t indicates that this file has sticky bit on. A common example is the /tmp directory on Linux. [root@Server / ]# ls -l drwxrwxrwt 21 root root 4096 Jul 13 17:39 tmp setuid and setgid: Sometimes, unprivileged users must be able to accomplish tasks that require privileges. An example is the passwd program, which allows you to change your password. Changing a user's password requires modifying the password field in the /etc/passwd file. However, you should not give a user access to change this file directly- the user could change everybody else's password as well! Likewise, the mail program requires that you be able to insert a message into the mailbox of another user, yet you should not to give one user unrestricted access to another's mailbox. To get around these problems, linux allows programs to be implemented with special privileges. Processes executing these programs can assume another UID or GID when they're running. A program that changes its UID is called a SUID program (set- UID); a program that changes its GID is called a SGID program (set-GID). A program can be both SUID and SGID at the same time. When a SUID program is run, its effective UID becomes that of the owner of the file, rather than of the user who is running it.
If setuid permission is applied on a file like: [root@Server / ]# chmod 4511 myprogram.sh then myprogram.sh has the ownership and permissions as : -r-s--x--x 1 root admin 16336 Feb 14 2003 myprogram.sh ,then this means that whenever any user, like xena logs on to the system and runs this program then this program is run as not user xena but user root !!!!. The s in the owner's permission set indicates that this file has setuid bit on. If setgid permission is applied on a file like : [root@Server / ]# chmod 2711 myprogram.sh then myprogram.sh has the ownership and permissions as : -rwx--s--x 1 root admin 16336 Feb 14 2003 myprogram.sh ,then this means that any process that executes this SGID program has its effective GID changed to the program's GID (admin in this case). Files created by the process can have their primary group set to this GID as well, depending on the permissions of the directory in which the files are created. The s in the group permission set indicates that this file has setgid bit on. The setgid bit is normally useful on directories. For example you have a directory named newproject and it has ownership as root: students and permissions like: [root@Server / ]# chmod 2770 myprogram.sh [root@Server / ]# ls -l drwxrws--- 2 root students 4096 Jul 13 21:38 newproject This means that any user who creates an object within this directory will effectively create that object with students in the group ownership. In other words, the SGID bit on a directory controls the way that groups are assigned for files created in the directory. If the SGID bit is set, files created in the directory have the same group as the directory if the process creating the file also is in that group. Otherwise, if the SGID bit is not set, or if the process is not in the same group, files created inside the directory have the same group as the user's effective group ID (usually the primary group ID). Setuid and setgid can be security risk you can find files which have setuid and setgid set on them by: [root@Server / ]# find / $-perm -004000 -o -perm -002000 $ -type f -print
This find command starts in the root directory (/) and looks for all files that match mode 002000 (SGID) or mode 004000 (SUID). The -type f option causes the search to be restricted to files. The -print option causes the name of every matching file to be printed. Note that if you are using NFS, you should execute find commands only on your file servers. You should further restrict the find command so that it does not try to search networked disks. Otherwise, use of this command may cause an excessive amount of NFS traffic on your network. To restrict your find command, use the following: [root@Server / ]# find / $ -local -o -prune $ $ -perm 004000 -o -perm -002000 $ type f print Alternatively, if your find command has the -xdev option, you can use it to prevent find from crossing filesystem boundaries. To search the entire filesystem using this option means running the command multiple times, once for each mounted partition. Be sure that you are the superuser when you run find, or you may miss SUID files hidden in protected directories. Umask: The umask (UNIX shorthand for "user file-creation mode mask") is a four-digit octal number that UNIX uses to determine the file permission for newly created files. Every process has its own umask, inherited from its parent process. The umask specifies the permissions you do not want to be given by default to newly created files and directories. Umask works by doing a bitwise AND with the bitwise complement of the umask. Bits that are set in the umask correspond to permissions that are not automatically assigned to newly created files. By default, most UNIX versions specify an octal mode of 666 (any user can read or write the file) when they create new files. Likewise, new programs are created with a mode of 777 (any user can read, write, or execute the program). The most common umask values are 022, 027, and 077. A umask value of 022 lets the owner both read and write all newly created files, but everybody else can only read them: 0666 default file-creation mode (0022) umask 0644 resultant mode A umask value of 077 lets only the file's owner read all newly created files: 0666 default file-creation mode (0077) umask 0600 resultant mode
A simple way to calculate umask values is to remember that the number 2 in the umask turns off write permission, while 7 turns off read, write, and execute permission. A umask value of 002 is commonly used by people who are working on group projects. If you create a file with your umask set to 002, anyone in the file's group will be able to read or modify the file. Everybody else will only be allowed to read it: 0666 default file-creation mode (0002) umask 0664 resultant mode On many UNIX systems, the default umask is 022. This is inherited from the init process, as all processes are descendants of init. Some systems may be configured to use another umask value, or a different value may be set in the startup files. The designers of these systems chose this umask value to foster sharing, an open computing environment, and cooperation among users. Most prototype user accounts shipped with UNIX operating systems specify 022 as the default umask, and many computer centers use this umask when they set up new accounts. Unfortunately, system administrators frequently do not make a point of explaining the umask to novice users, and many users are not aware that most of the files they create are readable by every other user on the system. A recent trend among computing centers has been to set up new accounts with a umask of 077, so a user's files will, by default, be unreadable by anyone else on the system unless the user makes a conscious choice to make them readable.
Common umask settings and their effects: umask User Access Group Access Other 0000 all all all 0002 all all read, execute 0007 all all none 0022 all read, execute read, execute 0027 all read, execute none 0077 all none none
Sudo: Best explanation is the man page itself: sudo allows a permitted user to execute a command as the superuser or another user, as specified in the sudoers file. The real and effective uid and gid are set to match those of the target user as specified in the passwd file (the group vector is also initialized when the target user is not root). By default, sudo requires that users authenticate themselves
with a password (NOTE: by default this is the users password, not the root password). Once a user has been authenticated, a timestamp is updated and the user may then use sudo without a password for a short period of time (5 minutes unless overridden in sudoers). sudo determines who is an authorized user by consulting the file /etc/sudoers. By giving sudo the -v flag a user can update the time stamp without running a command. The password prompt itself will also time out if the users password is not entered within 5 minutes (unless overridden via sudoers). What this says is that sudo is used / can be used, by administrator (root), to grant execute permissions to specific users. For example, users can be granted to execute the mount command, which is by default only executable by root. The benefit of using sudo is that you don't have to mess with the setuid or setgid bits of the target program, e.g. Mount . Rather you just mention the user or group in the /etc/sudoers file and that is it. Users can use sudo command along mount command to get their work done. The /etc/sudoers file is quite easy to understand. But remember you have to edit it NOT with vi. You have to use visudo command. [root@Server / ]# visudo Assume you want to give cdrom mount and unmount ability to the group named students along the ability to shutdown the local machine. You would add the following lines to the /etc/sudoers file. [root@Server / ]# visudo %students ALL=/bin/mount /dev/cdrom, /bin/umount /mnt/cdrom %students localhost=/sbin/shutdown -h now Now the users (members of group students) would use the sudo command to mount various file systems. [root@Server / ]# sudo mount /dev/cdrom /mnt/cdrom [root@Server / ]# Wiring / cabling: Wiring schemes / c able connector schemes: If you hold an RJ-45 connector in your hand in such a way that the eight pins are facing you, the locking clip is away from you or towards back of your hand and the channel to insert cable in this connector is facing downwards, then the pins facing towards you are numbered 1 to 8 from left to right. In a 10 Mbps or 100 Mbps network layout, only pins 1, 2, 3 and 6 are used. Rests are NOT used. It is a miss conception amongst many that if four wires are used then it is a 10 Mbps or Half duplex cable and if it is an eight wire cable then it is 100 Mbps or Full duplex cable. It is NOT like that. 1) Straight through cable
Used to connect dis-similar / dislike devices. PC to Hub/Switch. Router's Ethernet port to Switch. Up linking Switch to switch by using one switch's up link port and the other's non-uplink / normal port. To make a 4 wire (2 pair) straight through cable use: 1 - 1 (White Green - White Green) 2 - 2 (Green - Green) 3 - 3 (White Orange - White Orange) 6 - 6 (Orange - Orange) To make a 8 wire (4 pair) straight through cable use: 1 - 1 (White Green - White Green) 2 - 2 (Green - Green) 3 - 3 (White Orange - White Orange) 4 - 4 (Blue - Blue) 5 - 5 (White Blue - White Blue) 6 - 6 (Orange - Orange) 7 - 7 (White Brown - White Brown) 8 - 8 (Brown - Brown) 2) Cross over cable Cross over cable is used to connect similar devices. Like PC to PC. Router's ethernet port to PC's ethernet port directly. Uplinking switches by using non-uplink / normal ports of both switches. To make a 4 wire (2 pair) cross over cable use: 1 (White Green) 2 (Green) 3 (White Orange) 6 (Orange) 3 (White Green) 6 (Green) 1 (White Orange) 2 (Orange)
To make a 8 wire (4 pair) cross over cable use: 1 (White Green) 2 (Green) 3 (White Orange) 4 (Blue) 5 (White Blue) 6 (Orange) 7 (White Brown) 8 (Brown) 3 (White Green) 6 (Green) 1 (White Orange) 4 (Blue) 5 (White Blue) 2 (Orange) 7 (White Brown) 8 (Brown)
3) Roll over cable It is also known as console cable and is used to connect Router's console port to PC's serial port for Router administration. It is normally available in the form such as it has RJ-45 connectors on both ends and one of the ends is further converted by using a RJ-45 to DB9 converter. Here is the scheme of this cable (RJ-45 to RJ 45): 1-8 2-7 3-6 4-5 5-4 6-3 7-2 8-1 In case you want to make one yourself, get an RJ-45 and a DB9 pin-less connector and connect as below: RJ45 - DB9 1-8 2-6 3-3 4-5 5-5 6 (unused) 7-4 8-7 One more thing I would like to mention here, just in case you forget it. Cisco Routers and Switches, once you connect to them on their console ports, need 9600 8N1 setting in your connecting software, like Minicom (linux) or Hyper terminal (windows) 4) PC serial console cable It is also known as null modem cable and is used to connect PCs over Serial port. Also can be used to log on to a headless Linux PC using another computer using Hyper terminal (windows) or Minicom (linux). Both sides of the cable have DB9 (9pin) connectors (pin-less connectors). (The ones like serial mouse have). Connect as:1-4 2-3
3-2 4-1 5-5 6 (unused) 7-8 8-7 9-9 This is an eight wire cable and you can easily make it of the maximum length 15 meters , using commonly available cat-5 ether-net cable (as it has eight wires too). Network Addresses / Classes Network addresses and classes are also sometimes forgotten by administrators. Here are the schemes: Class A
Class B
Class C
Class D Class E (Multicast) (Research) 224 to 239 240 to 255
1 to 126 First Octet Range Valid Network Numbers 1.0.0.0 to 126.0.0.0
128 to 191
192 to 223
128.1.0.0 to 191.254.0.0
192.0.1.0 to 223.255.254. 0
27 -2 Number of Networks
214 -2
221 -2
224 -2 Number of Hosts per Network Size of Network part of Address Size of Host part of Address 1
216 -2
28 -2
8 Size of Network part, in bits 24 Size of Host part, in bits 255.0.0.0 Default Subnet mask Reserved Network Numbers 0.0.0.0 127.0.0.0
16
24
16
255.255.0.0
255.255.255.0
128.0.0.0 191.255.0.0
192.0.0.0 223.255.255. 0 192.168.0.0 to 192.168.255.2 55
Valid Private IP Addresses
10.0.0.0 to 10.255.255. 255
172.15.0.0 to 172.31.255.2 55

Linux Pocket Reference For System

Uploaded by

Copyright:

Available Formats

Linux Pocket Reference For System

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

Linux Pocket Reference For System

Uploaded by

Copyright:

Available Formats

Linux Reference for System Administrators

Chapter 2: Using fdisk

/dev/hda1 /dev/hda2 * /dev/hda3 /dev/hda4 /dev/hda5 /dev/hda6 /dev/hda7 /dev/hda8

Device Boot /dev/hdc1

Device Boot /dev/hdc1

/dev/hdc2 * FAT32 (LBA)

1023 1856736 559 921311+ 611 66528

Device Boot /dev/hdc1

/dev/hdc2 * FAT32 (LBA) /dev/hdc3 /dev/hdc5 /dev/hdc6 swap

52 103 103 579

Win95 Extended Linux Linux

1023 1856736 559 921311+ 611 66528

Partition size calculation:

Chapter 3: Basic N etwork trouble-shooting

Basic network configuration and troubleshooting:

wks1 wks2 www

192.168.1.1 192.168.1.2 Server

[root@Server /]# ping wks1.mydomain.com

Client side testing and mounting:

Now check which shares are available on the server by:

On the gateway machine, give the following command:

Chapter 10: SENDMAIL

Chapter 11: DHCP

Start dhcpd service by : [root@Server /]# service dhcpd start

Chapter 12: FTP

Chapter 13: FIREWALL

Chapter 14: KERNEL MODIFICATION AND COMPILATION

Chapter 15: Single user and Rescue mode

Chapter 16: Advance permissions

Class D Class E (Multicast) (Research) 224 to 239 240 to 255

1 to 126 First Octet Range Valid Network Numbers 1.0.0.0 to 126.0.0.0

192.0.0.0 223.255.255. 0 192.168.0.0 to 192.168.255.2 55

Valid Private IP Addresses

10.0.0.0 to 10.255.255. 255

You might also like