Mongodb Docs 2010 09 03

.bookmarks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
5
1.1 Development Cycle . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
Creating and Deleting Indexes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
Diagnostic Tools . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
Django and MongoDB . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
Getting Started . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
International Documentation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
Monitoring . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
Older Downloads . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
PyMongo and mod_wsgi . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
Python Tutorial . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
Recommended Production Architectures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
v0.8 Details . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
Building SpiderMonkey . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
Documentation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
Dot Notation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
Dot Notation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Getting the Software . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
Language Support . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
Mongo Administration Guide . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
Working with Mongo Objects and Classes in Ruby . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
MongoDB Language Support . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
Community Info . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
Internals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
TreeNavigation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
Old Pages . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
Storing Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
Indexes in Mongo . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
HowTo . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
Searching and Retrieving . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
Locking . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
Mongo Developers' Guide . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
Locking in Mongo . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
Mongo Database Administration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
Mongo Concepts and Terminology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
MongoDB - A Developer's Tour . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
Updates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
Structuring Data for Mongo . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
Design Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
Document-Oriented Datastore . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
Why so many "Connection Accepted" messages logged? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Why are my datafiles so large? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
Storing Files . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
Introduction - How Mongo Works . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
Optimizing Mongo Performance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
Mongo Usage Basics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
Server-Side Processing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
Home . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
Quickstart . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
Quickstart OS X . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
Quickstart Unix . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
Quickstart Windows . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
Downloads . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
1.0 Changelist . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
1.2.x Release Notes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
1.4 Release Notes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
1.6 Release Notes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
CentOS and Fedora Packages . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
Ubuntu and Debian packages . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22
Version Numbers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
Drivers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
C Language Center . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24
C Tutorial . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
C Sharp Language Center . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31
Driver Syntax Table . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31
Javascript Language Center . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31
node.JS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32
JVM Languages . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32
Python Language Center . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32
PHP Language Center . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32
Installing the PHP Driver . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33
PHP Libraries, Frameworks, and Tools . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33
PHP - Storing Files and Big Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35
Troubleshooting the PHP Driver . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35
Ruby Language Center . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36
Ruby Tutorial . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37
Replica Pairs in Ruby . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42
Replica Sets in Ruby . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43
GridFS in Ruby . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45
Rails - Getting Started . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47
Rails 3 - Getting Started . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49
MongoDB Data Modeling and Rails . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51
Object Mappers for Ruby and MongoDB . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55
Using Mongoid . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57
Ruby External Resources . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59
Frequently Asked Questions - Ruby . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60
Java Language Center . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63
Java Driver Concurrency . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63
Java - Saving Objects Using DBObject . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 64
Java Tutorial . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 64
Java Types . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69
C++ Language Center . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71
C++ BSON Library . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71
C++ Tutorial . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 72
Connecting . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 76
Perl Language Center . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 76
Contributing to the Perl Driver . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 78
Perl Tutorial . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 79
Online API Documentation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 79
Writing Drivers and Tools . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 79
Overview - Writing Drivers and Tools . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 79
bsonspec.org . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 80
Mongo Driver Requirements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 80
Spec, Notes and Suggestions for Mongo Drivers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 84
Feature Checklist for Mongo Drivers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 84
Conventions for Mongo Drivers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 85
Driver Testing Tools . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 85
Mongo Wire Protocol . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 85
BSON . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 90
Mongo Extended JSON . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 92
GridFS Specification . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 94
Implementing Authentication in a Driver . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 96
Notes on Pooling for Mongo Drivers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 97
Driver and Integration Center . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 100
Connecting Drivers to Replica Sets . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 100
Error Handling in Mongo Drivers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 100
Developer Zone . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 101
Tutorial . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 102
Manual . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 107
Connections . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 108
Databases . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 109
Commands . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 109
Mongo Metadata . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 117
Collections . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 118
Capped Collections . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 118
Using a Large Number of Collections . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 119
Data Types and Conventions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 120
Internationalized Strings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 120
Object IDs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 120
Database References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 121
GridFS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 123
Indexes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 124
Using Multikeys to Simulate a Large Number of Indexes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 127
Geospatial Indexing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 128
Indexing as a Background Operation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 131
Multikeys . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 132
Indexing Advice and FAQ . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 133
Inserting . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 136
Legal Key Names . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 137
Schema Design . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 137
Trees in MongoDB . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 139
Optimization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 142
Optimizing Storage of Small Objects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 145
Query Optimizer . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 145
Querying . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 146
Mongo Query Language . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 147
Retrieving a Subset of Fields . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 148
Advanced Queries . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 148
Dot Notation (Reaching into Objects) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 155
Full Text Search in Mongo . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 157
min and max Query Specifiers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 158
OR operations in query expressions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 159
Queries and Cursors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 160
Server-side Code Execution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 162
Sorting and Natural Order . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 165
Aggregation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 165
Removing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 168
Updating . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 169
Atomic Operations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 173
findandmodify Command . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 175
Updating Data in Mongo . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 176
MapReduce . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 178
Data Processing Manual . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 181
mongo - The Interactive Shell . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 182
Overview - The MongoDB Interactive Shell . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 183
dbshell Reference . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 185
Developer FAQ . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 187
Do I Have to Worry About SQL Injection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 188
How does concurrency work . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 189
SQL to Mongo Mapping Chart . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 190
What is the Compare Order for BSON Types . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 193
Admin Zone . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 193
Production Notes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 194
Replication . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 195
Verifying Propagation of Writes with getLastError . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 195
Replica Sets . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 196
About the local database . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 197
Data Center Awareness . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 197
Reconfiguring a replica set when members are down . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 198
Replica Set Design Concepts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 198
Replica Sets Troubleshooting . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 199
Replica Set Tutorial . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 199
Replica Set Configuration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 203
Upgrading to Replica Sets . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 205
Replica Set Admin UI . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 207
Replica Set Commands . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 207
Replica Set FAQ . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 209
Connecting to Replica Sets from Clients . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 209
Replica Sets Limits . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 210
Resyncing a Very Stale Replica Set Member . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 210
Replica Set Internals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 210
Master Slave . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 213
One Slave Two Masters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 216
Replica Pairs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 218
Master Master Replication . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 220
Replication Oplog Length . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 221
Halted Replication . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 222
Sharding . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 223
Sharding Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 224
Configuring Sharding . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 228
A Sample Configuration Session . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 230
Upgrading from a Non-Sharded System . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 232
Sharding Administration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 233
Sharding and Failover . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 234
Sharding Limits . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 235
Sharding Internals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 235
Moving Chunks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 236
Sharding Config Schema . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 236
Sharding Design . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 238
Sharding Use Cases . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 239
Shard Ownership . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 239
Splitting Chunks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 240
Sharding FAQ . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 241
Hosting Center . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 242
Amazon EC2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 242
Joyent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 243
Monitoring and Diagnostics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 243
Checking Server Memory Usage . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 245
Database Profiler . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 245
Http Interface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 247
mongostat . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 249
mongosniff . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 250
Backups . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 251
How to do Snapshotted Queries in the Mongo Database . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 251
Import Export Tools . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 252
Durability and Repair . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 255
Security and Authentication . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 257
Admin UIs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 258
Starting and Stopping Mongo . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 261
Logging . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 263
Command Line Parameters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 263
File Based Configuration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 264
GridFS Tools . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 265
DBA Operations from the Shell . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 265
Architecture and Components . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 267
Troubleshooting . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 267
Excessive Disk Space . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 268
Too Many Open Files . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 269
Contributors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 269
MongoDB kernel code development rules . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 270
Git Commit Rules . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 270
Kernel class rules . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 270
Kernel code style . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 270
Kernel concurrency rules . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 271
Kernel exception architecture . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 272
Kernel string manipulation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 272
Writing Tests . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 272
Project Ideas . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 272
UI . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 273
Source Code . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 274
Building . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 274
Building Boost . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 274
Building for FreeBSD . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 275
Building for Linux . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 275
Building for OS X . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 277
Building for Solaris . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 281
Building for Windows . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 281
Boost 1.41.0 Visual Studio 2010 Binary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 281
Boost and Windows . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 282
Building the Mongo Shell on Windows . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 282
Building with Visual Studio 2008 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 283
Building with Visual Studio 2010 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 284
Building Spider Monkey . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 286
scons . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 287
Database Internals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 288
Caching . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 288
Cursors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 288
Error Codes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 288
Internal Commands . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 288
Replication Internals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 289
Smoke Tests . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 290
Pairing Internals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 291
Contributing to the Documentation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 291
Emacs tips for MongoDB work . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 291
Mongo Documentation Style Guide . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 292
Community . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 294
MongoDB Commercial Services Providers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 295
User Feedback . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 295
Job Board . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 297
About . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 297
Philosophy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 297
Use Cases . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 297
Use Case - Session Objects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 298
Production Deployments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 299
Mongo-Based Applications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 310
Events . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 310
Video & Slides from Recent Events and Presentations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 312
Slide Gallery . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 315
Articles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 316
Benchmarks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 316
FAQ . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 316
Product Comparisons . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 317
Interop Demo (Product Comparisons) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 317
MongoDB, CouchDB, MySQL Compare Grid . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 317
Comparing Mongo DB and Couch DB . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 318
Licensing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 319
Windows . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 320
International Docs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 321
Books . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 321
Doc Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 322
.bookmarks
Recent bookmarks in MongoDB
This page is a container for all the bookmarks in this space. Do not delete or move it or you will lose all your bookmarks.
Bookmarks in MongoDB | Links for MongoDB
The 15 most recent bookmarks in MongoDB

There are no bookmarks to display.
1.1 Development Cycle

Redirection Notice
This page should redirect to [1.2.0 Release Notes].
Creating and Deleting Indexes

Redirection Notice
This page should redirect to Indexes.
Diagnostic Tools
Redirection Notice
This page should redirect to Monitoring and Diagnostics.
Django and MongoDB

Redirection Notice
This page should redirect to Python Language Center.
Getting Started
Redirection Notice
This page should redirect to Quickstart.
International Documentation
Redirection Notice
This page should redirect to International Docs.
Monitoring
Redirection Notice
This page should redirect to Monitoring and Diagnostics.
Older Downloads
Redirection Notice
This page should redirect to Downloads.
PyMongo and mod_wsgi

Redirection Notice
Python Tutorial
Redirection Notice
Recommended Production Architectures

Redirection Notice
This page should redirect to Production Notes.
v0.8 Details
Existing Core Functionality
Basic Mongo database functionality: inserts, deletes, queries, indexing.
Master / Slave Replication
Replica Pairs
Server-side javascript code execution
New to v0.8
Drivers for Java, C++, Python, Ruby.
db shell utility
(Very) basic security
$or
Clean up logging
Performance test baseline
getlasterror
Large capped collections
Bug fixes (compound index keys, etc.)
Import/Export utility
Allow any _id that is unique, and verify uniqueness
Wanted, but may not make it

AMI's
Unlock eval()?
Better disk full handling
better replica pair negotiation logic (for robustness)
Building SpiderMonkey
Redirection Notice
This page should redirect to Building Spider Monkey.
Documentation
Redirection Notice
This page should redirect to Home.
Dot Notation
Redirection Notice
This page should redirect to Dot Notation (Reaching into Objects).
Dot Notation
Redirection Notice
This page should redirect to Dot Notation (Reaching into Objects).
Getting the Software

Placeholder - $$$ TODO
Language Support
Redirection Notice
This page should redirect to Drivers.
Mongo Administration Guide

Redirection Notice
This page should redirect to Admin Zone.
Working with Mongo Objects and Classes in Ruby

Redirection Notice
This page should redirect to Ruby Language Center.
MongoDB Language Support

Redirection Notice
This page should redirect to Language Support.
Community Info
Redirection Notice
This page should redirect to Community.
Internals
Cursors
Tailable Cursors
See p/db/dbclient.h for example of how, on the client side, to support tailable cursors.
Set
Option_CursorTailable = 2
in the queryOptions int field to indicate you want a tailable cursor.
If you get back no results when you query the cursor, keep the cursor live if cursorid is still nonzero. Then, you can issue future getMore
requests for the cursor.
If a getMore request has the resultFlag ResultFlag_CursorNotFound set, the cursor is not longer valid. It should be marked as "dead"
on the client side.
ResultFlag_CursorNotFound = 1
See the Queries and Cursors section of the Mongo Developers' Guide for more information about cursors.
See Also
The Queries and Cursors section of the Mongo Developers' Guide for more information about cursors
TreeNavigation
Old Pages
Storing Data
Redirection Notice
This page should redirect to Inserting.
Indexes in Mongo
Redirection Notice
This page should redirect to Indexes.
HowTo
Redirection Notice
This page should redirect to Developer FAQ.
Searching and Retrieving
Redirection Notice
This page should redirect to Querying.
Locking
Redirection Notice
This page should redirect to Atomic Operations.
Mongo Developers' Guide
Redirection Notice
This page should redirect to Manual.
Locking in Mongo
Redirection Notice
Mongo Database Administration

Redirection Notice
This page should redirect to Admin Zone.
Mongo Concepts and Terminology
Redirection Notice
MongoDB - A Developer's Tour
Redirection Notice
Updates
Redirection Notice
This page should redirect to Updating.
Structuring Data for Mongo
Redirection Notice
This page should redirect to Inserting.
Design Overview
Redirection Notice
This page should redirect to Developer Zone.
Document-Oriented Datastore
Redirection Notice
This page should redirect to Databases.
Why so many "Connection Accepted" messages logged?

Redirection Notice
Why are my datafiles so large?
Redirection Notice
Storing Files
Redirection Notice
This page should redirect to GridFS.
Introduction - How Mongo Works
Redirection Notice
This page should redirect to Developer Zone.
Optimizing Mongo Performance
Redirection Notice
This page should redirect to Optimization.
Mongo Usage Basics
Redirection Notice
This page should redirect to Tutorial.
Server-Side Processing
Redirection Notice
This page should redirect to Server-side Code Execution.
Home
Events
Follow us on Facebook and Twitter to get all of the latest updates!
Coming Soon:
Mongo Boston - September 20

Mongo Berlin - October 4
Mongo Chicago - October 20
Slides and Video now available: MongoSeattle | MongoFR | MongoUK | MongoNYC | MongoSF
More events...
Getting Started
Quickstart | Downloads | Tutorial
Development
Manual
C | C++ | C# & .NET | ColdFusion | Erlang | Factor | Java | Javascript | PHP | Python | Ruby | Perl | More...
Production
Production Notes | Security | Replication | Sharding | Backup
Support
Forum | IRC | Bug tracker | Commercial support | Training | Consulting | Hosting
Community
Blog | Articles | Twitter | [Forum] | Facebook | LinkedIn | Job Board | User groups: NY and SF
Meta
Use Cases | Philosophy | License
Translations
| Deutsch | Español | Français | Italiano | | Português | |
Quickstart
Quickstart OS X
Quickstart Unix
Quickstart Windows
For an even quicker start go to http://try.mongodb.org/.
See Also
SQL to Mongo Mapping Chart

Quickstart OS X
Install MongoDB
The easiest way to install MongoDB is to use a package manager or the pre-built binaries:
Package managers
If you use the Homebrew package manager, run:
$ brew install mongodb
If you use MacPorts you can install with:
$ sudo port install mongodb
This will take a while to install.
32-bit binaries
Note: 64-bit is recommended.
$ curl http://downloads.mongodb.org/osx/mongodb-osx-i386-1.4.4.tgz > mongo.tgz

$ tar xzf mongo.tgz
64-bit binaries
$ curl http://downloads.mongodb.org/osx/mongodb-osx-x86_64-1.4.4.tgz > mongo.tgz

$ tar xzf mongo.tgz
Create a data directory

By default MongoDB will store data in /data/db, but it won't automatically create that directory. To create it, do:
$ mkdir -p /data/db
You can also tell MongoDB to use a different data directory, with the --dbpath option.
Run and connect to the server

First, start the MongoDB server in one terminal:
$ ./mongodb-xxxxxxx/bin/mongod
In a separate terminal, start the shell, which will connect to localhost by default:
$ ./mongodb-xxxxxxx/bin/mongo
> db.foo.save( { a : 1 } )
> db.foo.find()
Congratulations, you've just saved and retrieved your first document with MongoDB!
Learn more
Once you have MongoDB installed and running, head over to the Tutorial.
Quickstart Unix
Install MongoDB
Note: If you are running an old version of Linux and the database doesn't start, or gives a floating point exception, try the "legacy static" version
on the Downloads page instead of the versions listed below.
Package managers
Ubuntu and Debian users can now install nightly snapshots via apt. See Ubuntu and Debian packages for details.
CentOS and Fedora users should head to the CentOS and Fedora Packages page.
32-bit Linux binaries
Note: 64 bit is recommended.
$ curl http://downloads.mongodb.org/linux/mongodb-linux-i686-1.4.4.tgz > mongo.tgz

$ tar xzf mongo.tgz
64-bit Linux binaries
$ curl http://downloads.mongodb.org/linux/mongodb-linux-x86_64-1.4.4.tgz > mongo.tgz

$ tar xzf mongo.tgz
Other Unixes
See the Downloads page for some binaries, and also the Building page for information on building from source.

By default MongoDB will store data in /data/db, but it won't automatically create that directory. To create it, do:
$ sudo mkdir -p /data/db/

$ sudo chown ìd -u` /data/db
You can also tell MongoDB to use a different data directory, with the --dbpath option.

First, start the MongoDB server in one terminal:
$ ./mongodb-xxxxxxx/bin/mongod
In a separate terminal, start the shell, which will connect to localhost by default:
$ ./mongodb-xxxxxxx/bin/mongo
> db.foo.save( { a : 1 } )
> db.foo.find()
Learn more
Once you have MongoDB installed and running, head over to the Tutorial.
Quickstart Windows
Download
32-bit binaries
64-bit binaries
Unzip
Learn more
Download
The easiest (and recommended) way to install MongoDB is to use the pre-built binaries.
32-bit binaries
Download and extract the 32-bit .zip. The "Production" build is recommended.
64-bit binaries
Download and extract the 64-bit .zip.
Note: 64-bit is recommended, although you must have a 64-bit version of Windows to run that version.
Unzip
Unzip the downloaded binary package to the location of your choice. You may want to rename mongo-xxxxxxx to just "mongo" for convenience.

By default MongoDB will store data in C:\data\db, but it won't automatically create that folder, so we do so here:
C:\> mkdir \data

C:\> mkdir \data\db
Or you can do this from the Windows Explorer, of course.

The important binaries for a first run are:
mongod.exe - the database server

mongo.exe - the administrative shell
To run the database, click mongod.exe in Explorer, or run it from a CMD window.
C:\> cd \my_mongo_dir\bin
C:\my_mongo_dir\bin > mongod
Note: It is also possible to run the server as a Windows Service. But we can do that later.
Now, start the administrative shell, either by double-clicking mongo.exe in Explorer, or from the CMD prompt. By default mongo.exe connects to
a mongod server running on localhost and uses the database named test. Run mongo --help to see other options.
C:\> cd \my_mongo_dir\bin
C:\my_mongo_dir\bin> mongo
> // the mongo shell is a javascript shell connected to the db
> 3+3
6
> db
test
> // the first write will create the db:
> db.foo.insert( { a : 1 } )
> db.foo.find()
{ _id : ..., a : 1 }
Learn more
Tutorial
Windows quick links
[Mongo Shell]
Mongo Shell Info
Downloads
See also Packages.
Version OS X 32 OS X 64 Linux 32 Linux 64 Windows Windows Solaris Solaris Source Date

bit bit bit bit 32 bit 64-bit i86pc 64
Production
(Recommended)
1.4.3 os x 10.5+ download download * download * download download download download tgz zip 5/24/2010
os x 10.4 legacy-static legacy-static
nightly os x 10.5+ download download * download * download download download download tgz zip Daily
Previous
Release
nightly os x 10.5+ download download * download * download download download download tgz zip Daily
Dev (unstable)
1.5.x nightly os x 10.5+ download download * download * download download download download tgz zip Daily
Archived list list list list list list list list list
Releases
See Version Numbers

The linux legacy-static builds are only recommended for older systems. If you try to run and get a floating point exception, try the
legacy-static builds. Otherwise you should use the regular ones.
Currently the mongod server must run on little-endian cpu (intel) so if you are using a ppc os x, mongod will not work.
32-bit builds are limited 2gb of data. See http://blog.mongodb.org/post/137788967/32-bit-limitations for more info
See http://buildbot.mongodb.org/waterfall for details of builds and completion times.
Included in Distributions
The MongoDB database server

The MongoDB shell
Backup and restore tools
Import and export tools
GridFS tool
The MongoDB C++ client
Drivers
Information on how to separately download or install the drivers and tools can be found on the Drivers page.
Language Packages Source API Reference
Python bundles github api
PHP pecl github api
Ruby gemcutter github api
Java jar github api
Perl cpan github api
C++ included in database github api
See Drivers for more information and other languages.
Source Code
Source code for MongoDB and all drivers
Packages
MongoDB is included in several different package managers:
For MacPorts, see the mongodb and mongodb-devel packages.

For FreeBSD, see the mongodb and mongodb-devel packages.
For Homebrew, see the mongodb formula.
For ArchLinux, see the mongodb package in the AUR.
For Debian and Ubuntu, see Ubuntu and Debian packages.
For Fedora and CentOS, see CentOS and Fedora Packages.
Documentation
Pre-Exported
You can export yourself: HTML, PDF, or XML.
Logos
MongoDB logos are available for download as attachments on this page.
Powered By MongoDB Badges
We've made badges in beige, brown, blue and green for use on your sites that are powered by MongoDB. They are available below and in
multiple sizes as attachments on this page.
Training
If you're just getting started with MongoDB, consider registering for an upcoming training course.
<#comment><#comment><#comment>
1.0 Changelist
Wrote MongoDB. See documentation.
1.2.x Release Notes

New Features
More indexes per collection

Faster index creation
Map/Reduce
Stored JavaScript functions
Configurable fsync time
Several small features and fixes
DB Upgrade Required
There are some changes that will require doing an upgrade if your previous version is <= 1.0.x. If you're already using a version >= 1.1.x then
these changes aren't required. There are 2 ways to do it:
--upgrade
stop your mongod process
run ./mongod --upgrade
start mongod again
use a slave
start a slave on a different port and data directory
when its synced, shut down the master, and start the new slave on the regular port.
Ask in the forums or IRC for more help.
Replication Changes
There have been minor changes in replication. If you are upgrading a master/slave setup from <= 1.1.2 you have to update the slave first.
mongoimport
mongoimportjson has been removed and is replaced with mongoimport that can do json/csv/tsv
field filter changing
We've changed the semantics of the field filter a little bit. Previously only objects with those fields would be returned. Now the field filter
only changes the output, not which objects are returned. If you need that behavior, you can use $exists
other notes
http://www.mongodb.org/display/DOCS/1.1+Development+Cycle
1.4 Release Notes

We're pleased to announce the 1.4 release of MongoDB. 1.4 is a drop in replacement for 1.2. To upgrade you just need to shutdown mongod,
then restart with the new binaries. (Users upgrading from release 1.0 should review the 1.2 release notes, in particular the instructions for
upgrading the DB format.)
Release 1.4 includes the following improvements over release 1.2:
Core server enhancements
concurrency improvements
indexing memory improvements
background index creation
better detection of regular expressions so the index can be used in more cases
Replication & Sharding
better handling for restarting slaves offline for a while

fast new slaves from snapshots (--fastsync)
configurable slave delay (--slavedelay )
replication handles clock skew on master
$inc replication fixes
sharding alpha 3 - notably 2 phase commit on config servers
Deployment & production
configure "slow threshold" for profiling

ability to do fsync + lock for backing up raw files
option for separate directory per db (--directoryperdb)
http://localhost:28017/_status to get serverStatus via http
REST interface is off by default for security (--rest to enable)
can rotate logs with a db command, logRotate
enhancements to serverStatus command (db.serverStatus()) - counters and replication lag stats
new mongostat tool
Query language improvements
$all with regex

$not
partial matching of array elements $elemMatch
$ operator for updating arrays
$addToSet
$unset
$pull supports object matching
$set with array indices
Geo
2d geospatial search
geo $center and $box searches
1.6 Release Notes

MongoDB 1.6 is a drop-in replacement for 1.4. To upgrade, simply shutdown mongod then restart with the new binaries.
Sharding
Sharding is now production-ready, making MongoDB horizontally scalable, with no single point of failure. A single instance of mongod can now be
upgraded to a distributed cluster with zero downtime when the need arises.
Sharding Tutorial
Sharding Documentation
Upgrading a Single Server to a Cluster
Replica Sets
Replica sets, which provide automated failover among a cluster of n nodes, are also now available.
Plese note that replica pairs are now deprecated; we strongly recommend that replica pair users upgrade to replica sets.
Replica Set Tutorial

Replica Set Documentation
Upgrading Existing Setups to Replica Sets
Other Improvements
The w option (and wtimeout) forces writes to be propagated to n servers before returning success (this works especially well with replica
sets)
$or queries
Improved concurrency
$slice operator for returning subsets of arrays
64 indexes per collection (formerly 40 indexes per collection)
64-bit integers can now be represented in the shell using NumberLong
The findAndModify command now supports upserts. It also allows you to specify fields to return
$showDiskLoc option to see disk location of a document
Support for IPv6 and UNIX domain sockets
Installation
Windows service improvements

The C++ client is a separate tarball from the binaries
1.5.x Release Notes
1.5.8
1.5.7
1.5.6
1.5.5
1.5.4
1.5.3
1.5.2
1.5.1
1.5.0
You can see a full list of all changes on Jira.
Thank you everyone for your support and suggestions!
CentOS and Fedora Packages

10gen now publishes yum-installable RPM packages for CentOS 5.4 (x86 and x86_64) and Fedora 11, 12, and 13 (x64_64 only for the moment).
For each revision in stable, unstable, and snapshot, there are four packages, e.g., mongo-stable, mongo-stable-server, mongo-stable-devel,
mongo-stable-debuginfo, for each of the client, server, headers, and debugging information, respectively.
To use these packages, add one of the following files in /etc/yum.repos.d, and then yum update and yum install your preferred complement of
packages.
For CentOS 5.4 on x86_64:
[10gen]
name=10gen Repository
baseurl=http://downloads.mongodb.org/distros/centos/5.4/os/x86_64/
gpgcheck=0
For CentOS 5.4 on x86
[10gen]
baseurl=http://downloads.mongodb.org/distros/centos/5.4/os/i386/
gpgcheck=0
For Fedora 13:
[10gen]
baseurl=http://downloads.mongodb.org/distros/fedora/13/os/x86_64/
gpgcheck=0
For Fedora 12:
[10gen]
gpgcheck=0
For Fedora 11:
[10gen]
gpgcheck=0
For the moment, these packages aren't signed. (If anybody knows how to automate signing RPMs, please let us know!)
Ubuntu and Debian packages
Please read the notes on the Downloads page. Also, note that these packages are updated daily, and so if you find you can't
download the packages, try updating your apt package lists, e.g., with 'apt-get update' or 'aptitude update'.
10gen publishes apt-gettable packages. Our packages are generally fresher than those in Debian or Ubuntu. We publish 3 distinct packages,
named "mongodb-stable", "mongodb-unstable", "mongodb-snapshot", corresponding to our latest stable release, our latest development release,
and the most recent git checkout at the time of building. Each of these packages conflicts with the others, and with the "mongodb" package in
Debian/Ubuntu.
The packaging is still a work-in-progress, so we invite Debian and Ubuntu users to try them out and let us know how the packaging might be
improved.
To use the packages, add a line to your /etc/apt/sources.list, then 'aptitude update' and one of 'aptitude install mongodb-stable', 'aptitude install
mongodb-unstable' or 'aptitude install mongodb-snapshot'.
For Ubuntu Lucid (10.4) (built using a prerelease installation):
deb http://downloads.mongodb.org/distros/ubuntu 10.4 10gen
For Ubuntu Karmic (9.10):
For Ubuntu Jaunty (9.4):
For Debian Lenny (5.0):
deb http://downloads.mongodb.org/distros/debian 5.0 10gen
These packages are snapshots of our git master branch, and we plan to
update them frequently, so package version numbers will be of the form
YYYYMMDD; when reporting issues with these packages, please include the
package version in your report.
The public gpg key used for signing these packages follows. It should be possible to import the key into apt's public keyring with a command like
this:
sudo apt-key adv --keyserver keyserver.ubuntu.com --recv 7F0CEB10
To configure these packages beyond the defaults, have a look at /etc/mongodb.conf, and/or the initialization script, (/etc/init.d/mongodb on older,
non-Upstart systems, /etc/init/mongodb.conf on Upstart systems). Most MongoDB operational settings are in /etc/mongodb.conf; a few other
settings are in the initialization script. Note that if you customize the userid in the initialization script or the dbpath or logpath settings in
/etc/mongodb.conf, you must ensure that the directories and files you use are writable by the userid you run the server as.
Packages for other distros coming soon!

-----BEGIN PGP PUBLIC KEY BLOCK-----
Version: GnuPG v1.4.10 (Darwin)
mQENBEtsQe8BCACm5G0/ei0IxyjVEp6EEtbEbWK1Q4dKaONtiCODwB8di+L8tlUd
Ra5QYxeyV90C+dqdh34o79enXxT6idHfYYqDdob2/kAPE6vFi4sLmrWIVGCRY++7
RPclZuezPm1sxG1TRAYEsW0VZUE9ofdoQ8xlUZDyn2BSjG8OCT2e4orRg1pHgzw2
n3hnWqJNuJS4jxcRJOxI049THIGUtqBfF8bQoZw8C3Wg/R6pGghUfNjpA6uF9KAH
gnqrC0swZ1/vwIJt9fnvAlzkqLrssYtKH0rMdn5n4g5tJLqY5q/NruHMq2rhoy3r
4MC1w8GTbP7qR83wAyaLJ7xACOKqxB3SrDFJABEBAAG0I1JpY2hhcmQgS3JldXRl
ciA8cmljaGFyZEAxMGdlbi5jb20+iQE4BBMBAgAiBQJLbEHvAhsDBgsJCAcDAgYV
CAIJCgsEFgIDAQIeAQIXgAAKCRCey+xGfwzrEGXbB/4nrmf/2rEnztRelmup3duI
eepzEtw1cv3uHg2oZXGS6S7o5Fsk+amngaWelWKfkSw5La7aH5vL4tKFKUfuaME1
avInDIU/0IEs8jLrdSWq601HowLQcxAhqNPdaGONDtHw56Qhs0Ba8GA6329vLWgZ
ODnXweiNSCDrv3xbIN6IjPyyO5AoUkxmJfD0mVtp3u5Ar7kfIw7ieGGxokaHewNL
Xzqcp9rPiUR6dFw2uRvDdVRrXFUPlgVugaHKytm15JpHmQfyzQiMdYXnIz0oofJO
WM/PY1iw+QJZ2M7PnfbTJeADXIc/EoOAJDRggih533SjhiCaT6FdPMMk6rCZ5cgl
uQENBEtsQe8BCADlNPIJZVSL2i6H9X19YK4CpEqsjIUGISMB1cDT311WFSnhfuMs
GL9xYRb8dlbyeJFFOyHNkIBmH5ekCvGRfS6qJYpcUQZZcWSjEMqBYQV5cw1efd0B
ek64jfvrsLz8+YhKzn+NI8O3nyGvpEEWvOhN4hNjwkDhYbXLvAlsqagbnSMf+Htf
3lgCGYa2gLiNIqNKWCsEVAan/Er6KS39WANGXi6ih0yjReBiU8WR6Qh2ylMi2xKw
yHnTOsbWxP0hqALUa7N4AEGCXS/qn+vUz/hcIbt+eUNy45qoZcTT3dZsWGfJqknh
RFMIuPiej7/WY4Ugzes5NG02ecDkDkpJvrSNABEBAAGJAR8EGAECAAkFAktsQe8C
GwwACgkQnsvsRn8M6xABeggAlNkqbqa12L1bgaCgnGGdCiuXB3F6/VFmSQdUKpts
EuqWH6rSp30r67PupzneX++ouh+9WD5O7gJ0kP3VQJpmXjT/QnN5ANjI4kAtRZUW
qCXlXOxVAeXHL5oiKz0NM23Xc2rNAyfBQY8+SUYrKBalNBq5m68g8oogX8QD5u2F
x+6C+QK9G2EBDD/NWgkKN3GOxpQ5DTdPHI5/fjwYFs1leIaQjjiyJwAifxB/1+w0
VCHe2LDVpRXY5uBTefF2guhVYisKY6n5wNDaQpBmA8w17it5Yp8ge0HMN1A+aZ+6
L6MsuHbG2OYDZgAk8eKhvyd0y/pAhZpNuQ82MMGBmcueSA==
=74Cu
-----END PGP PUBLIC KEY BLOCK-----
Version Numbers
MongoDB uses the odd-numbered versions for development releases.
There are 3 numbers in a MongoDB version: A.B.C
A is the major version. This will rarely change and signify very large changes
B is the release number. This will include many changes including features and things that possible break backwards compatibility. Even
Bs will be stable branches, and odd Bs will be development.
C is the revision number and will be used for bugs and security issues.
For example:
1.0.0 : first GA release

1.0.x : bug fixes to 1.0.x - highly recommended to upgrade, very little risk
1.1.x : development release. this will include new features that are not fully finished, and works in progress. Some things may be different
than 1.0
1.2.x : second GA release. this will be the culmination of the 1.1.x release.
Drivers
MongoDB currently has client support for the following programming languages:
mongodb.org Supported
C
C++
Java
Javascript
Perl
PHP
Python
Ruby
Community Supported
REST
C# and .NET
Clojure
ColdFusion
Blog post: Part 1 | Part 2 | Part 3
http://github.com/virtix/cfmongodb/tree/0.9
Delphi
pebongo - Early stage Delphi driver for MongoDB
Erlang
emongo - An Erlang MongoDB driver that emphasizes speed and stability. "The most emo of drivers."
Erlmongo - an almost complete MongoDB driver implementation in Erlang
Factor
http://github.com/slavapestov/factor/tree/master/extra/mongodb/
Fantom
http://bitbucket.org/liamstask/fantomongo/wiki/Home
F#
http://gist.github.com/218388
Go
gomongo
Groovy
See Java Language Center
Haskell
http://hackage.haskell.org/package/mongoDB
Javascript
Lua
LuaMongo
node.js
Objective C
NuMongoDB
PHP
Asynchronous PHP driver using libevent
PowerShell
Blog post
Python
Ruby
MongoMapper
rmongo - An event-machine-based Ruby driver for MongoDB
jmongo A thin ruby wrapper around the mongo-java-driver for vastly better jruby performance.
em-mongo EventMachine MongoDB Driver (based off of RMongo).
Scala
See JVM Languages
Scheme (PLT)
http://planet.plt-scheme.org/display.ss?package=mongodb.plt&owner=jaymccarthy
docs
Smalltalk
Dolphin Smalltalk
Get Involved, Write a Driver!
Writing Drivers and Tools
C Language Center
C Driver
Download
Build
Notable Projects
C Driver
The MongoDB C Driver is the 10gen-supported driver for MongoDB. It's written in pure C. The goal is to be super strict for ultimate portability, no
dependencies, and very embeddable anywhere.
Tutorial
C Driver README
Source Code
Download
The C driver is hosted at GitHub.com. Check out the latest version with git.
$ git clone git://github.com/mongodb/mongo-c-driver.git
Build
Building with gcc:
$ gcc --std=c99 -Isrc /path/to/mongo-c-driver/src/*.c YOUR_APP.c
Building with scons:
$ scons # this will produce libbson.a and libmongoc.a

$ scons --c99 # this will use c99 mode in gcc (recommended)
$ scons test # this will compile and run the unit tests (optional)
$ scons test --test-server=123.4.5.67 # use remote server for tests
Notable Projects
NuMongodb – An Objective-C wrapper around the MongoDB C driver. It is intended for use with Nu but may be useful in other Objective-C
programming applications.
If you're working on a project that you'd like to have included, let us know.
C Tutorial
Writing Client Code
Connecting
BSON
Inserting
Single
Batch
Querying
Simple Queries
Complex Queries
Sorting
Hints
Explain
Indexing
Updating
Further Reading
This document is an introduction to usage of the MongoDB database from a C program.
First, install Mongo -- see the Quickstart for details.
Next, you may wish to take a look at the Developer's Tour guide for a language independent look at how to use MongoDB. Also, we suggest
some basic familiarity with the mongo shell -- the shell is one's primary database administration tool and is useful for manually inspecting the
contents of a database after your C program runs.
A working C program complete with examples from this tutorial can be found here.
Writing Client Code
For brevity, the examples below are simply inline code.
Connecting
Let's make a tutorial.c file that connects to the database:

#include <stdlib.h>
#include <stdio.h>
#include <string.h>
#include "bson.h"
#include "mongo.h"
int main() {
mongo_connection conn[1]; /* ptr */
mongo_connection_options opts[1];
mongo_conn_return status;
strcpy( opts->host , "127.0.0.1" );

opts->port = 27017;
status = mongo_connect( conn, opts );
switch (status) {
case mongo_conn_success: printf( "connection succeeded\n" ); break;
case mongo_conn_bad_arg: printf( "bad arguments\n" ); return 1;
case mongo_conn_no_socket: printf( "no socket\n" ); return 1;
case mongo_conn_fail: printf( "connection failed\n" ); return 1;
case mongo_conn_not_master: printf( "not master\n" ); return 1;
}
/* CODE WILL GO HERE */
mongo_destroy( conn );
printf( "\nconnection closed\n" );
return 0;
}
If you are using gcc on Linux or OS X, you would compile with something like this, depending on location of your include files:
$ gcc -Isrc --std=c99 /path/to/mongo-c-driver/src/*.c -I /path/to/mongo-c-driver/src/ tutorial.c -o

tutorial
$ ./tutorial
connection succeeded
connection closed
$
BSON
The Mongo database stores data in BSON format. BSON is a binary object format that is JSON-like in terms of the data which can be stored
(some extensions exist, for example, a Date datatype).
To save data in the database we must create bson objects. We use bson_buffer to make bson objects, and bson_iterator to enumerate
bson objects.
Let's now create a BSON "person" object which contains name and age. We might invoke:
bson b[1];
bson_buffer buf[1];
bson_buffer_init( buf )
bson_append_string( buf, "name", "Joe" );
bson_append_int( buf, "age", 33 );
bson_from_buffer( b, buf );
Use the bson_append_new_oid() helper to add an object id to your object. The server will add an _id automatically if it is not included
explicitly.
bson b[1];
bson_buffer buf[1];
bson_buffer_init( buf );
bson_append_new_oid( buf, "_id" );
bson_append_string( buf, "name", "Joe" );
bson_append_int( buf, "age", 33 );
bson_from_buffer( b, buf );
bson_buffer_new_oid( ..., "_id" ) should be at the beginning of the generated object.
When you are done using the object remember to use bson_destroy() to free up the memory allocated by the buffer.
bson_destroy( b )
Inserting
Single
We now save our person object in a persons collection in the database:
mongo_insert( conn, "tutorial.persons", b );
The first parameter to mongo_insert is the pointer to the mongo_connection object. The second parameter is the namespace. tutorial is the
database and persons is the collection name. The third parameter is a pointer to the bson "person" object that we created before.
Batch
We can do batch inserts as well:
static void tutorial_insert_batch( mongo_connection *conn ) {

bson *p, **ps;
bson_buffer *p_buf;
char *names[4];
int ages[] = { 29, 24, 24, 32 };
int i, n = 4;
names[0] = "Eliot"; names[1] = "Mike"; names[2] = "Mathias"; names[3] = "Richard";
ps = ( bson ** )malloc( sizeof( bson * ) * n);
for ( i = 0; i < n; i++ ) {

p = ( bson * )malloc( sizeof( bson ) );
p_buf = ( bson_buffer * )malloc( sizeof( bson_buffer ) );
bson_buffer_init( p_buf );
bson_append_new_oid( p_buf, "_id" );
bson_append_string( p_buf, "name", names[i] );
bson_append_int( p_buf, "age", ages[i] );
bson_from_buffer( p, p_buf );
ps[i] = p;
free( p_buf );
}
mongo_insert_batch( conn, "tutorial.persons", ps, n );
for ( i = 0; i < n; i++ ) {

bson_destroy( ps[i] );
free( ps[i] );
}
}
Querying
Simple Queries
Let's now fetch all objects from the persons collection, and display them.
static void tutorial_empty_query( mongo_connection *conn) {

mongo_cursor *cursor;
bson empty[1];
bson_empty( empty );
cursor = mongo_find( conn, "tutorial.persons", empty, empty, 0, 0, 0 );

while( mongo_cursor_next( cursor ) ) {
bson_print( &cursor->current );
}
mongo_cursor_destroy( cursor );
bson_destroy( empty );
}
empty is the empty BSON object -- we use it to represent {} which indicates an empty query pattern (an empty query is a query for all objects).
We use bson_print() above to print out information about each object retrieved. bson_print() is a diagnostic function which prints an
abbreviated JSON string representation of the object.
mongo_find() returns a mongo_cursor which must be destroyed after use.
Let's now write a function which prints out the name (only) of all persons in the collection whose age is a given value:
static void tutorial_simple_query( mongo_connection *conn ) {

bson query[1];
bson_buffer query_buf[1];
mongo_cursor *cursor;
bson_buffer_init( query_buf );
bson_append_int( query_buf, "age", 24 );
bson_from_buffer( query, query_buf );
cursor = mongo_find( conn, "tutorial.persons", query, NULL, 0, 0, 0 );

while( mongo_cursor_next( cursor ) ) {
bson_iterator it[1];
if ( bson_find( it, &cursor->current, "name" )) {
printf( "name: %s\n", bson_iterator_string( it ) );
}
}
bson_destroy( query );
}
Our query above, written as JSON, is of the form
{ age : <agevalue> }
Queries are BSON objects of a particular format.
In the mongo shell (which uses javascript), we could invoke:
use tutorial;
db.persons.find( { age : 24 } );
Complex Queries
Sometimes we want to do more then a simple query. We may want the results to be sorted in a special way, or what the query to use a certain
index.
Sorting
Let's now make the results from previous query be sorted alphabetically by name. To do this, we change the query statement from:
to:
bson_append_start_object( query_buf, "$query" );
bson_append_finish_object( query_buf );
bson_append_start_object( query_buf, "$orderby" );
bson_append_int( query_buf, "name", 1);
Hints
While the mongo query optimizer often performs very well, explicit "hints" can be used to force mongo to use a specified index, potentially
improving performance in some situations. When you have a collection indexed and are querying on multiple fields (and some of those fields are
indexed), pass the index as a hint to the query:
bson_append_string( query_buf, "name", "Mathias" );
bson_append_start_object( query_buf, "$hint" );
bson_append_int( query_buf, "name", 1);
Explain
A great way to get more information on the performance of your database queries is to use the $explain feature. This will return "explain plan"
type info about a query from the database.
bson_append_bool( query_buf, "$explain", 1);
Indexing
Let's suppose we want to have an index on age so that our queries are fast. We would use:
static void tutorial_index( mongo_connection * conn ) {
bson key[1];
bson_buffer key_buf[1];
bson_buffer_init( key_buf );
bson_append_int( key_buf, "name", 1 );
bson_from_buffer( key, key_buf );
mongo_create_index( conn, "tutorial.persons", key, 0, NULL );
bson_destroy( key );
printf( "simple index created on \"name\"\n" );
bson_buffer_init( key_buf );
bson_append_int( key_buf, "age", 1 );
bson_append_int( key_buf, "name", 1 );
bson_from_buffer( key, key_buf );
mongo_create_index( conn, "tutorial.persons", key, 0, NULL );
bson_destroy( key );
printf( "compound index created on \"age\", \"name\"\n" );

}
Updating
Use the mongo_update() method to perform a database update . For example the following update in the mongo shell :
> use tutorial

> db.persons.update( { name : 'Joe', age : 33 },
... { $inc : { visits : 1 } } )
is equivalent to running the following C function:
static void tutorial_update( mongo_connection *conn ) {

bson cond[1], op[1];
bson_buffer cond_buf[1], op_buf[1];
bson_buffer_init( cond_buf );
bson_append_string( cond_buf, "name", "Joe");
bson_append_int( cond_buf, "age", 33);
bson_from_buffer( cond, cond_buf );
bson_buffer_init( op_buf );
bson_append_start_object( op_buf, "$inc" );
bson_append_int( op_buf, "visits", 1 );
bson_append_finish_object( op_buf );
bson_from_buffer( op, op_buf );
mongo_update(conn, "tutorial.persons", cond, op, 0);
bson_destroy( cond );
bson_destroy( op );
}
Further Reading
This overview just touches on the basics of using Mongo from C++. There are many more capabilities. For further exploration:
See the language-independent Developer's Tour;

Experiment with the mongo shell;
Consider getting involved to make the product (either C driver, tools, or the database itself) better!
C Sharp Language Center

C# Drivers
mongodb-csharp driver
simple-mongodb driver
NoRM
F#
F# Example
Community Articles
A List of C# MongoDB Tools

Experimenting with MongoDB from C#
Using MongoDB from C#
Introduction to MongoDB for .NET
Using Json.NET and Castle Dynamic Proxy with MongoDB
Implementing a Blog Using ASP.NET MVC and MongoDB
Intro Article using a Post and Comments Example
Tools
MongoDB.Emitter Document Wrapper

log4net appender
Support
http://groups.google.com/group/mongodb-csharp
http://groups.google.com/group/mongodb-user
IRC: #mongodb on freenode
See Also
C++ Language Center
Driver Syntax Table

The wiki generally gives examples in JavaScript, so this chart can be used to convert those examples to any language.
JavaScript Python PHP Ruby
[] [] array() []
{} {} new stdClass {}
{x:1} {"x": 1} array('x' => 1) {'x' => 1}
connect("www.example.net") Connection("www.example.net") new Mongo.new("www.example.net")

Mongo("www.example.net")
cursor.next() cursor.next() $cursor->getNext() cursor.next_object()
cursor.hasNext() * $cursor->hasNext() *
collection.findOne() collection.find_one() $collection->findOne() collection.find_one()
db.eval() db.eval() $db->execute() db.eval()
* does not exist in that language

Javascript Language Center
MongoDB can be
Used by clients written in Javascript;

Uses Javascript internally server-side for certain options such as map/reduce;
Has a shell that is based on Javascript for administrative purposes.
[node.JS and V8]
See the node.JS page.
SpiderMonkey
The MongoDB shell extends SpiderMonkey. See the MongoDB shell documentation.
Narwhal
http://github.com/sergi/narwhal-mongodb
MongoDB Server-Side Javascript
Javascript may be executed in the MongoDB server processes for various functions such as query enhancement and map/reduce processing.
See Server-side Code Execution.
node.JS
Node.js is used to write event-driven, scalable network programs in server-side JavaScript. It is similar in purpose to Twisted, EventMachine, etc.
It runs on Google's V8.
Web Frameworks
ExpressJS Mature web framework with MongoDB session support.
3rd Party ORM/ODM
Mongoose - Asynchronous JavaScript Driver with optional support for Modeling.
3rd Party Drivers
node-mongodb - Async Node interface to MongoDB (written in C)

node-mongodb-native - Native async Node interface to MongoDB.
mongo-v8-driver - V8 MongoDB driver (experimental, written in C++).
JVM Languages
moved to Java Language Center
Python Language Center
Redirection Notice
This page should redirect to http://api.mongodb.org/python.
PHP Language Center

Using MongoDB in PHP
To access MongoDB from PHP you will need:
The MongoDB server running - the server is the "mongod" file, not the "mongo" client (note the "d" at the end)
The MongoDB PHP driver installed
Installing the PHP Driver
*NIX
Run:
sudo pecl install mongo
Open your php.ini file and add to it:
extension=mongo.so
It is recommended to add this to the section with the other "extensions", but it will work from anywhere within the php.ini file.
Restart your web server (Apache, nginx, etc.) for the change to take effect.
See the installation docs for configuration information and OS-specific installation instructions.
Windows
Download the correct driver for your environment from http://github.com/mongodb/mongo-php-driver/downloads

VC6 is for Apache (VC9 is for IIS)
Thread safe is for running PHP as an Apache module (typical installation), non-thread safe is for CGI
Unzip and add the php_mongo.dll file to your PHP extensions directory (usually the "ext" folder in your PHP installation.)
Add to your php.ini:
extension=php_mongo.dll
Restart your web server (Apache, IIS, etc.) for the change to take effect
For more information, see the Windows section of the installation docs.
Using the PHP Driver
To get started, see the Tutorial. Also check out the API Documentation.
See Also
PHP Libraries, Frameworks, and Tools for working with Drupal, Cake, Symfony, and more from MongoDB.
Admin UIs
Redirection Notice
This page should redirect to http://www.php.net/manual/en/mongo.installation.php.
PHP Libraries, Frameworks, and Tools

The PHP community has created a huge number of libraries to make working with MongoDB easier and integrate it with existing frameworks.
CakePHP
MongoDB datasource for CakePHP. There's also an introductory blog post on using it with Mongo.
Codeigniter
MongoDB-Codeigniter-Driver
Doctrine
ODM (Object Document Mapper) is an experimental Doctrine MongoDB object mapper. The Doctrine\ODM\Mongo namespace is an experimental
project for a PHP 5.3 MongoDB Object Mapper. It allows you to easily write PHP 5 classes and map them to collections in MongoDB. You just
work with your objects like normal and Doctrine will transparently persist them to Mongo.
This project implements the same "style" of the Doctrine 2 ORM project interface so it will look very familiar to you and it has lots of the same
features and implementations.
Documentation - API, Reference, and Cookbook

Official blog post
Screencast
Blog post on using it with Symfony
Bug tracker
Drupal
MongoDB Integration - Views (query builder) backend, a watchdog implementation (logging), and field storage.
Fat-Free Framework
Fat-Free is a powerful yet lightweight PHP 5.3+ Web development framework designed to help you build dynamic and robust applications - fast!
Kohana Framework
Mango at github
An ActiveRecord-like library for PHP, for the Kohana PHP Framework.
See also PHP Language Center#MongoDb PHP ODM further down.
Lithium
Lithium supports Mongo out-of-the-box.
Tutorial on creating a blog backend.
Log4php
A log4php appender for MongoDB
Memcached
MongoNode
PHP script that replicates MongoDB objects to Memcached.
Symfony 2
Symfony 2 Logger
A centralized logger for Symfony applications. See the blog post.
sfMongoSessionStorage - manages session storage via MongoDB with symfony.
sfStoragePerformancePlugin - This plugin contains some extra storage engines (MongoDB and Memcached) that are currently missing
from the Symfony (>= 1.2) core.
Vork
Vork, the high-performance enterprise framework for PHP natively supports MongoDB as either a primary datasource or
used in conjunction with an RDBMS. Designed for scalability & Green-IT, Vork serves more traffic with fewer servers and
can be configured to operate without any disk-IO.
Vork provides a full MVC stack that outputs semantically-correct XHTML 1.1, complies with Section 508 Accessibility
guidelines & Zend-Framework coding-standards, has SEO-friendly URLs, employs CSS-reset for cross-browser display
consistency and is written in well-documented object-oriented E_STRICT PHP5 code.
An extensive set of tools are built into Vork for ecommerce (cc-processing, SSL, PayPal, AdSense, shipment tracking,
QR-codes), Google Maps, translation & internationalization, Wiki, Amazon Web Services, Social-Networking (Twitter,
Meetup, ShareThis, YouTube, Flickr) and much more.
Zend Framework
Shanty Mongo is a prototype mongodb adapter for the Zend Framework. It's intention is to make working with mongodb documents as
natural and as simple as possible. In particular allowing embeded documents to also have custom document classes.
ZF Cache Backend
A ZF Cache Backend for MongoDB. It support tags and auto-cleaning.
There is a Zend_Nosql_Mongo component proposal.
Stand-Alone Tools
ActiveMongo
ActiveMongo is a really simple ActiveRecord for MongoDB in PHP.
There's a nice introduction to get you started at http://crodas.org/activemongo.php.
MapReduce API
A MapReduce abstraction layer. See the blog post.
MongoDB-MapReduce-PHP at github
MongoDb PHP ODM
MongoDb PHP ODM is a simple object wrapper for the Mongo PHP driver classes which makes using Mongo in your PHP application more like
ORM, but without the suck. It is designed for use with Kohana 3 but will also integrate easily with any PHP application with almost no additional
effort.
Mongodloid
A nice library on top of the PHP driver that allows you to make more natural queries ($query->query('a == 13 AND b >= 8 && c % 3 ==
4');), abstracts away annoying $-syntax, and provides getters and setters.
Project Page
Downloads
Documentation
Morph
A high level PHP library for MongoDB. Morph comprises a suite of objects and object primitives that are designed to make working with MongoDB
in PHP a breeze.
Morph at code.google.com
simplemongophp
Very simple layer for using data objects see blog post
simplemongophp at github
Uniform Server 6-Carbo with MongoDB and phpMoAdmin
The Uniform Server is a lightweight WAMP server solution for running a web server under Windows without having anything to install; just unpack
and run it. Uniform Server 6-Carbo includes the latest versions of Apache2, Perl5, PHP5, MySQL5 and phpMyAdmin. The Uniform Server
MongoDB plugin adds the MongoDB server, phpMoAdmin browser administration interface, the MongoDB PHP driver and a Windows interface to
start and stop both Apache and MongoDB servers. From this interface you can also start either the Mongo-client or phpMoAdmin to administer
MongoDB databases.
Uniform Server 6-Carbo and MongoDB plugin at SourceForge

Uniform Server web site
PHP - Storing Files and Big Data
Redirection Notice
This page should redirect to http://www.php.net/manual/en/class.mongogridfs.php.
Troubleshooting the PHP Driver

Redirection Notice
This page should redirect to http://www.php.net/manual/en/mongo.trouble.php.
Ruby Language Center

This is an overview of the available tools and suggested practices for using Ruby with MongoDB. Those wishing to skip to more detailed
discussion should check out the Ruby Driver Tutorial, Getting started with Rails or Rails 3, and MongoDB Data Modeling and Rails. There are
also a number of good external resources worth checking out.
Ruby Driver
Installing / Upgrading
BSON
Object Mappers
Notable Projects
Ruby Driver
Install the C extension for any performance-critical applications.
The MongoDB Ruby driver is the 10gen-supported driver for MongoDB. It's written in pure Ruby, with a recommended C extension for speed.
The driver is optimized for simplicity. It can be used on its own, but it also serves as the basis for various object-mapping libraries.
Tutorial
Ruby Driver README
API Documentation
Source Code
Installing / Upgrading
The ruby driver is hosted at Rubygems.org. Before installing the driver, make sure you're using the latest version of rubygems (currently 1.3.6):
$ gem update --system
Then install the gems:
$ gem install mongo
To stay on the bleeding edge, check out the latest source from github:
$ git clone git://github.com/mongodb/mongo-ruby-driver.git

$ cd mongo-ruby-driver/
Then, install the driver from there:
$ rake gem:install
BSON
In versions of the Ruby driver prior to 0.20, the code for serializing to BSON existed in the mongo gem. Now, all BSON serialization is handled by
the required bson gem.
gem install bson
For significantly improved performance, install the bson extensions gem:

gem install bson_ext
As long it's in Ruby's load path, bson_ext will be loaded automatically when you require bson.
Note that beginning with version 0.20, the mongo_ext gem is no longer used.
To learn more about the Ruby driver, see the Ruby Tutorial.
Object Mappers
If you need validations, associations, and other high-level data modeling functions, consider using one of the available object mappers. Many of
these exist in the Ruby ecosystem; here we host a list of the most popular ones.
Notable Projects
Tools for working with MongoDB in Ruby are being developed daily. A partial list can eb found in the Projects and Libraries section of our external
resources page.
If you're working on a project that you'd like to have included, let us know.
Ruby Tutorial
This tutorial gives many common examples of using MongoDB with the Ruby driver. If you're looking for information on data modeling, see
MongoDB Data Modeling and Rails. Links to the various object mappers are listed on our object mappers page.
Interested in GridFS? Checkout GridFS in Ruby.
As always, the latest source for the Ruby driver can be found on github.
Installation
A Quick Tour
Using the RubyGem
Making a Connection
Listing All Databases
Dropping a Database
Authentication (Optional)
Getting a List Of Collections
Getting a Collection
Inserting a Document
Finding the First Document In a Collection using find_one()
Adding Multiple Documents
Counting Documents in a Collection
Using a Cursor to get all of the Documents
Getting a Single Document with a Query
Getting a Set of Documents With a Query
Querying with Regular Expressions
Creating An Index
Creating and querying on a geospatial index
Getting a List of Indexes on a Collection
Database Administration
See Also
Installation
The mongo-ruby-driver gem is served through Rubygems.org. To install, make sure you have the latest version of rubygems.
gem update --system
Next, install the mongo rubygem:
gem install mongo
The required bson gem will be installed automatically.
For optimum performance, install the bson_ext gem:

gem install bson_ext
After installing, you may want to look at the examples directory included in the source distribution. These examples walk through some of the
basics of using the Ruby driver.
The full API documentation can be viewed here.
A Quick Tour
Using the RubyGem
All of the code here assumes that you have already executed the following Ruby code:
require 'rubygems' # not necessary for Ruby 1.9

require 'mongo'
Making a Connection
An Mongo::Connection instance represents a connection to MongoDB. You use a Connection instance to obtain an Mongo:DB instance, which
represents a named database. The database doesn't have to exist - if it doesn't, MongoDB will create it for you.
You can optionally specify the MongoDB server address and port when connecting. The following example shows three ways to connect to the
database "mydb" on the local machine:
db = Mongo::Connection.new.db("mydb")
db = Mongo::Connection.new("localhost").db("mydb")
db = Mongo::Connection.new("localhost", 27017).db("mydb")
At this point, the db object will be a connection to a MongoDB server for the specified database. Each DB instance uses a separate socket
connection to the server.
If you're trying to connect to a replica set, see Replica Sets in Ruby.
Listing All Databases
connection = Mongo::Connection.new # (optional host/port args)

connection.database_names.each { |name| puts name }
connection.database_info.each { |info| puts info.inspect}
Dropping a Database
connection.drop_database('database_name')
MongoDB can be run in a secure mode where access to databases is controlled through name and password authentication. When run in this
mode, any client application must provide a name and password before doing any operations. In the Ruby driver, you simply do the following with
the connected mongo object:
auth = db.authenticate(my_user_name, my_password)
If the name and password are valid for the database, auth will be true. Otherwise, it will be false. You should look at the MongoDB log for
further information if available.
Getting a List Of Collections
Each database has zero or more collections. You can retrieve a list of them from the db (and print out any that are there):
db.collection_names.each { |name| puts name }
and assuming that there are two collections, name and address, in the database, you would see
name
address
as the output.
Getting a Collection
You can get a collection to use using the collection method:
coll = db.collection("testCollection")
This is aliased to the [] method:
coll = db["testCollection"]
Once you have this collection object, you can now do things like insert data, query for data, etc.
Once you have the collection object, you can insert documents into the collection. For example, lets make a little document that in JSON would be
represented as
{
"name" : "MongoDB",
"type" : "database",
"count" : 1,
"info" : {
x : 203,
y : 102
}
}
Notice that the above has an "inner" document embedded within it. To do this, we can use a Hash or the driver's OrderedHash (which preserves
key order) to create the document (including the inner document), and then just simply insert it into the collection using the insert() method.
doc = {"name" => "MongoDB", "type" => "database", "count" => 1,

"info" => {"x" => 203, "y" => '102'}}
coll.insert(doc)
Finding the First Document In a Collection using find_one()
To show that the document we inserted in the previous step is there, we can do a simple find_one() operation to get the first document in the
collection. This method returns a single document (rather than the Cursor that the find() operation returns).
my_doc = coll.find_one()
puts my_doc.inspect
and you should see:
{"_id"=>#<BSON::ObjectID:0x118576c ...>, "name"=>"MongoDB", "info"=>{"x"=>203, "y"=>102}, "type"=>

"database", "count"=>1}
Note the _id element has been added automatically by MongoDB to your document.
To demonstrate some more interesting queries, let's add multiple simple documents to the collection. These documents will have the following
form:
{
"i" : value
}
Here's how to insert them:
100.times { |i| coll.insert("i" => i) }
Notice that we can insert documents of different "shapes" into the same collection. These records are in the same collection as the complex
record we inserted above. This aspect is what we mean when we say that MongoDB is "schema-free".
Counting Documents in a Collection
Now that we've inserted 101 documents (the 100 we did in the loop, plus the first one), we can check to see if we have them all using the
count() method.
puts coll.count()
and it should print 101.
Using a Cursor to get all of the Documents
To get all the documents from the collection, we use the find() method. find() returns a Cursor object, which allows us to iterate over the
set of documents that matches our query. The Ruby driver's Cursor implemented Enumerable, which allows us to use Enumerable#each,
{{Enumerable#map}, etc. For instance:
coll.find().each { |row| puts row.inspect }
and that should print all 101 documents in the collection.
Getting a Single Document with a Query
We can create a query hash to pass to the find() method to get a subset of the documents in our collection. For example, if we wanted to find
the document for which the value of the "i" field is 71, we would do the following ;
coll.find("i" => 71).each { |row| puts row.inspect }
and it should just print just one document:
{"_id"=>#<BSON::ObjectID:0x117de90 ...>, "i"=>71}
Getting a Set of Documents With a Query
We can use the query to get a set of documents from our collection. For example, if we wanted to get all documents where "i" > 50, we could
write:
coll.find("i" => {"$gt" => 50}).each { |row| puts row }
which should print the documents where i > 50. We could also get a range, say 20 < i <= 30:
coll.find("i" => {"$gt" => 20, "$lte" => 30}).each { |row| puts row }
Querying with Regular Expressions

Regular expressions can be used to query MongoDB. To find all names that begin with 'a':
coll.find({"name" => /â/})
You can also construct a regular expression dynamically. To match a given search string:
search_string = params['search']
# Constructor syntax
coll.find({"name" => Regexp.new(search_string)})
# Literal syntax
coll.find({"name" => /#{search_string}/})
Although MongoDB isn't vulnerable to anything like SQL-injection, it may be worth checking the search string for anything malicious.
Creating An Index
MongoDB supports indexes, and they are very easy to add on a collection. To create an index, you specify an index name and an array of field
names to be indexed, or a single field name. The following creates an ascending index on the "i" field:
# create_index assumes ascending order; see method docs

# for details
coll.create_index("i")
To specify complex indexes or a descending index you need to use a slightly more complex syntax - the index specifier must be an Array of [field
name, direction] pairs. Directions should be specified as Mongo::ASCENDING or Mongo::DESCENDING:
# explicit "ascending"
coll.create_index([["i", Mongo::ASCENDING]])
Creating and querying on a geospatial index
First, create the index on a field containing long-lat values:
people.create_index([["loc", Mongo::GEO2D]])
Then get a list of the twenty locations nearest to the point 50, 50:
people.find({"loc" => {"$near" => [50, 50]}}, {:limit => 20}).each do |p|
puts p.inspect
end
You can get a list of the indexes on a collection using coll.index_information().
A database can have one of three profiling levels: off (:off), slow queries only (:slow_only), or all (:all). To see the database level:
puts db.profiling_level # => off (the symbol :off printed as a string)

db.profiling_level = :slow_only
Validating a collection will return an interesting hash if all is well or raise an exception if there is a problem.
p db.validate_collection('coll_name')
See Also
Ruby Driver Official Docs

MongoDB Koans A path to MongoDB enlightenment via the Ruby driver.
MongoDB Manual
Replica Pairs in Ruby
Replica Sets will replace replica pairs in MongoDB 1.6. If you are just now setting up an instance, you may want to wait for that
and use master/slave replication in the meantime.
Here follow a few considerations for those using the Ruby driver with MongoDB and replica pairing.
Setup
Connection Failures
Recovery
Testing
Further Reading
Setup
First, make sure that you've correctly paired two mongod instances. If you want to do this on the same machine for testing, make sure you've
created two data directories. The init commands are as follows:
./mongod --pairwith localhost:27018 --dbpath /data/left --port 27017

./mongod --pairwith localhost:27017 --dbpath /data/right --port 27018
When you instantiate a Ruby connection, you'll have to make sure that the driver knows about both instances:
@connection = Connection.paired([['localhost', 27017], ['localhost', 27018]])
Connection Failures
Imagine that our master node goes offline. How will the driver respond?
At first, the driver will try to send operations to what was the master node. These operations will fail, and the driver will raise a ConnectionFailure
exception. It then becomes the client's responsibility to decide how to handle this.
If the client decides to retry, it's not guaranteed that the former slave will have been promoted to master yet, so it's still possible that the driver will
raise another ConnectionFailure. However, once the former slave has become master, typically within a few seconds, subsequent operations
will succeed.
Recovery
Driver users may wish to wrap their database calls with failure recovery code. Here's one possibility:
# Ensure retry upon failure
def rescue_connection_failure(max_retries=5)
success = false
retries = 0
while !success
begin
yield
success = true
rescue Mongo::ConnectionFailure => ex
retries += 1
raise ex if retries >= max_retries
sleep(1)
end
end
end
end
# Wrapping a call to #count()

rescue_connection_failure do
@db.collection('users').count()
end
Of course, the proper way to handle connection failures will always depend on the individual application. We encourage object-mapper and
application developers to publish any promising results.
Testing
The Ruby driver (>= 0.17.2) includes some unit tests for verifying proper replica pair behavior. They reside in tests/replica. You can run them
individually with the following rake tasks:
rake test:pair_count
rake test:pair_insert
rake test:pair_query
Make sure you have a replica pair running locally before trying to run these tests.
Further Reading
Replica Pairs
Pairing Internals
Replica Sets in Ruby

Here follow a few considerations for those using the Ruby driver with MongoDB and replica sets.
Setup
Connection Failures
Recovery
Testing
Further Reading
Setup
First, make sure that you've configured and initialized a replica set.
Connecting to a replica set from the Ruby driver is easy. If you only want to specify a single node, simply pass that node to Connection.new:
@connection = Connection.new('foo.local', 27017)
If you want to pass in multiple seed nodes, use Connection.multi:

@connection = Connection.multi([['n1.mydb.net', 27017],
['n2.mydb.net', 27017], ['n3.mydb.net', 27017]])
In both cases, the driver will attempt to connect to a master node and, when found, will merge any other known members of the replica set into
the seed list.
Connection Failures
Imagine that our master node goes offline. How will the driver respond?
At first, the driver will try to send operations to what was the master node. These operations will fail, and the driver will raise a ConnectionFailure
exception. It then becomes the client's responsibility to decide how to handle this.
If the client decides to retry, it's not guaranteed that another member of the replica set will have been promoted to master right away, so it's still
possible that the driver will raise another ConnectionFailure. However, once a member has been promoted to master, typically within a few
seconds, subsequent operations will succeed.
The driver will essentially cycle through all known seed addresses until a node identifies itself as master.
Recovery
Driver users may wish to wrap their database calls with failure recovery code. Here's one possibility:
# Ensure retry upon failure

def rescue_connection_failure(max_retries=5)
success = false
retries = 0
while !success
begin
yield
success = true
rescue Mongo::ConnectionFailure => ex
retries += 1
raise ex if retries >= max_retries
sleep(1)
end
end
end
end
# Wrapping a call to #count()

rescue_connection_failure do
@db.collection('users').count()
end
Of course, the proper way to handle connection failures will always depend on the individual application. We encourage object-mapper and
application developers to publish any promising results.
Testing
The Ruby driver (>= 1.0.6) includes some unit tests for verifying replica set behavior. They reside in tests/replica_sets. You can run them
individually with the following rake tasks:
rake test:replica_set_count
rake test:replica_set_insert
rake test:pooled_replica_set_insert
rake test:replica_set_query
Make sure you have a replica set running on localhost before trying to run these tests.
Further Reading
Replica Sets
[Replics Set Configuration]
GridFS in Ruby
GridFS, which stands for "Grid File Store," is a specification for storing large files in MongoDB. It works by dividing a file into manageable chunks
and storing each of those chunks as a separate document. GridFS requires two collections to achieve this: one collection stores each file's
metadata (e.g., name, size, etc.) and another stores the chunks themselves. If you're interested in more details, check out the GridFS
Specification.
Prior to version 0.19, the MongoDB Ruby driver implemented GridFS using the GridFS::GridStore class. This class has been deprecated in favor
of two new classes: Grid and GridFileSystem. These classes have a much simpler interface, and the rewrite has resulted in a significant speed
improvement. Reads are over twice as fast, and write speed has been increased fourfold. 0.19 is thus a worthwhile upgrade.
The Grid class

Saving files
File metadata
Safe mode
Deleting files
The GridFileSystem class
Saving files
Deleting files
Metadata and safe mode
Advanced Users
The Grid class
The Grid class represents the core GridFS implementation. Grid gives you a simple file store, keyed on a unique ID. This means that duplicate
filenames aren't a problem. To use the Grid class, first make sure you have a database, and then instantiate a Grid:
@db = Mongo::Connection.new.db('social_site')
@grid = Grid.new(@db)
Saving files
Once you have a Grid object, you can start saving data to it. The data can be either a string or an IO-like object that responds to a #read method:
# Saving string data

id = @grid.put("here's some string / binary data")
# Saving IO data and including the optional filename

image = File.open("me.jpg")
id2 = @grid.put(image, :filename => "me.jpg")
Grid#put returns an object id, which you can use to retrieve the file:
# Get the string we saved

file = @grid.get(id)
# Get the file we saved

image = @grid.get(id2)
File metadata
There are accessors for the various file attributes:

image.filename
# => "me.jpg"
image.content_type
# => "image/jpg"
image.file_length
# => 502357
image.upload_date
# => Mon Mar 01 16:18:30 UTC 2010
# Read all the image's data at once

image.read
# Read the first 100k bytes of the image

image.read(100 * 1024)
When putting a file, you can set many of these attributes and write arbitrary metadata:
# Saving IO data
file = File.open("me.jpg")
id2 = @grid.put(file,
:filename => "my-avatar.jpg"
:content_type => "application/jpg",
:_id => 'a-unique-id-to-use-in-lieu-of-a-random-one',
:chunk_size => 100 * 1024,
:metadata => {'description' => "taken after a game of ultimate"})
Safe mode
A kind of safe mode is built into the GridFS specification. When you save a file, and MD5 hash is created on the server. If you save the file in safe
mode, an MD5 will be created on the client for comparison with the server version. If the two hashes don't match, an exception will be raised.
id2 = @grid.put(image, "my-avatar.jpg", :safe => true)
Deleting files
Deleting a file is as simple as providing the id:
@grid.delete(id2)
The GridFileSystem class
GridFileSystem is a light emulation of a file system and therefore has a couple of unique properties. The first is that filenames are assumed to be
unique. The second, a consequence of the first, is that files are versioned. To see what this means, let's create a GridFileSystem instance:
Saving files
@db = Mongo::Connection.new.db("social_site")
@fs = GridFileSystem.new(@db)
Now suppose we want to save the file 'me.jpg.' This is easily done using a filesystem-like API:
@fs.open("me.jpg", "w") do |f|
f.write image
end
We can then retrieve the file by filename:
image = @fs.open("me.jpg", "r") {|f| f.read }
No problems there. But what if we need to replace the file? That too is straightforward:
image = File.open("me-dancing.jpg")
@fs.open("me.jpg", "w") do |f|
f.write image
end
But a couple things need to be kept in mind. First is that the original 'me.jpg' will be available until the new 'me.jpg' saves. From then on, calls to
the #open method will always return the most recently saved version of a file. But, and this the second point, old versions of the file won't be
deleted. So if you're going to be rewriting files often, you could end up with a lot of old versions piling up. One solution to this is to use the
:delete_old options when writing a file:
image = File.open("me-dancing.jpg")
@fs.open("me.jpg", "w", :delete_old => true) do |f|
f.write image
end
This will delete all but the latest version of the file.
Deleting files
When you delete a file by name, you delete all versions of that file:
@fs.delete("me.jpg")
Metadata and safe mode
All of the options for storing metadata and saving in safe mode are available for the GridFileSystem class:
@fs.open('my-avatar.jpg', w,
:content_type => "application/jpg",
:metadata => {'description' => "taken on 3/1/2010 after a game of ultimate"},
:_id => 'a-unique-id-to-use-instead-of-the-automatically-generated-one',
:safe => true) { |f| f.write image }
Advanced Users
Astute code readers will notice that the Grid and GridFileSystem classes are merely thin wrappers around an underlying GridIO class. This means
that it's easy to customize the GridFS implementation presented here; just use GridIO for all the low-level work, and build the API you need in an
external manager class similar to Grid or GridFileSystem.
Rails - Getting Started

Using Rails 3? See Rails 3 - Getting Started
This tutorial describes how to set up a simple Rails application with MongoDB, using MongoMapper as an object mapper. We assume you're
using Rails versions prior to 3.0.
Configuration
Testing
Coding
Using a Rails Template

All of the configuration steps listed below, and more, are encapsulated in this Rails template (raw version), based on a similar one by Ben Scofield
. You can create your project with the template as follows:
rails project_name -m "http://gist.github.com/219223.txt"
Be sure to replace project_name with the name of your project.
If you want to set up your project manually, read on.
Configuration
1. We need to tell MongoMapper which database we'll be using. Save the following to config/initializers/database.rb:
MongoMapper.database = "db_name-#{Rails.env}"
Replace db_name with whatever name you want to give the database. The Rails.env variable will ensure that a different database is used for
each environment.
2. If you're using Passenger, add this code to config/initializers/database.rb.
if defined?(PhusionPassenger)
PhusionPassenger.on_event(:starting_worker_process) do |forked|
MongoMapper.connection.connect_to_master if forked
end
end
3. Clean out config/database.yml. This file should be blank, as we're not connecting to the database in the traditional way.
4. Remove ActiveRecord from environment.rb.
config.frameworks -= [:active_record]
5. Add MongoMapper to the environment. This can be done by opening config/environment.rb and adding the line:
config.gem 'mongo_mapper'
Once you've done this, you can install the gem in the project by running:
rake gems:install
rake gems:unpack
Testing
It's important to keep in mind that with MongoDB, we cannot wrap test cases in transactions. One possible work-around is to invoke a teardown
method after each test case to clear out the database.
To automate this, I've found it effective to modify ActiveSupport::TestCase with the code below.
# Drop all columns after each test case.
def teardown
MongoMapper.database.collections.each do |coll|
coll.remove
end
end
# Make sure that each test case has a teardown

# method to clear the db after each test.
def inherited(base)
base.define_method teardown do
super
end
end
This way, all test classes will automatically invoke the teardown method. In the example above, the teardown method clears each collection. We
might also choose to drop each collection or drop the database as a whole, but this would be considerably more expensive and is only necessary
if our tests manipulate indexes.
Usually, this code is added in test/test_helper.rb. See the aforementioned rails template for specifics.
Coding
If you've followed the foregoing steps (or if you've created your Rails with the provided template), then you're ready to start coding. For help on
that, you can read about modeling your domain in Rails.
Rails 3 - Getting Started

It's not difficult to use MongoDB with Rails 3. Most of it comes down to making sure that you're not loading ActiveRecord and understanding how
to use Bundler, the new Ruby dependency manager.
Install the Rails 3 Pre-release

Configure your application
Bundle and Initialize
Bundling
Initializing
Running Tests
ActiveModel Compatibility
Conclusion
See also
Install the Rails 3 Pre-release
If you haven't done so already, install the Rails 3 pre-release. This requires installing the dependencies manually and then installing the Rails 3
pre-release gem:
# Use sudo if your setup requires it

gem install tzinfo builder i18n memcache-client rack \
rake rack-test rack-mount erubis mail text-format \
thor bundler
gem install rails --prerelease
Configure your application
The important thing here is to avoid loading ActiveRecord. One way to do this is with the --skip-activerecord switch. So you'd create your
app skeleton like so:
rails my_app --skip-activerecord
Alternatively, if you've already created your app (or just want to know what this actually does), have a look at config/application.rb and
change the first lines from this:
require "rails/all"
to this:
require "action_controller/railtie"
require "action_mailer/railtie"
require "active_resource/railtie"
require "rails/test_unit/railtie"
It's also important to make sure that the reference to active_record in the generator block is commented out:
# Configure generators values. Many other options are available, be sure to check the documentation.
# config.generators do |g|
# g.orm :active_record
# g.template_engine :erb
# g.test_framework :test_unit, :fixture => true
# end
As of this this writing, it's commented out by default, so you probably won't have to change anything here.
Bundle and Initialize
The final step involves bundling any gems you'll need and then creating an initializer for connecting to the database.
Bundling
Edit Gemfile, located in the Rails root directory. By default, our Gemfile will only load Rails:
gem "rails", "3.0.0.beta"
Normally, using MongoDB will simply mean adding whichever OM framework you want to work with, as these will require the "mongo" gem by
default.
# Edit this Gemfile to bundle your application's dependencies.
source 'http://gemcutter.org'

gem "mongo_mapper"
However, there's currently an issue with loading mongo_ext, as the current gemspec isn't compatible with the way Bundler works. We'll be fixing
that soon; just pay attention to this issue.
In the meantime, you can use the following work-around:
# Edit this Gemfile to bundle your application's dependencies.
require 'rubygems'
require 'mongo'
source 'http://gemcutter.org'

gem "mongo_mapper"
Requiring rubygems and mongo before running the gem command will ensure that mongo_ext is loaded. If you'd rather not load rubygems, just
make sure that both mongo and mongo_ext are in your load path when you require mongo.
Once you've configured your Gemfile, run the bundle installer:
bundle install
Initializing
Last item is to create an initializer to connect to MongoDB. Create a Ruby file in config/initializers. You can give it any name you want;
here we'll call it config/initializers/mongo.rb:
MongoMapper.connection = Mongo::Connection.new('localhost', 27017)

MongoMapper.database = "#myapp-#{Rails.env}"
if defined?(PhusionPassenger)
PhusionPassenger.on_event(:starting_worker_process) do |forked|
MongoMapper.connection.connect_to_master if forked
end
end
Running Tests
A slight modification is required to get rake test working (thanks to John P. Wood). Create a file lib/tasks/mongo.rake containing the
following:
namespace :db do
namespace :test do
task :prepare do
# Stub out for MongoDB
end
end
end
Now the various rake test tasks will run properly. See John's post for more details.
ActiveModel Compatibility
ActiveModel is a series of interfaces designed to make any object-mapping library compatible with the various helper methods across the Rails
stack. To see the status of ActiveModel integration on the various object mappers, see our object mappers page.
Briefly, Mongoid supports ActiveModel via a prerelease branch. MongoMapper will be adding support in the near future. In the meantime, use the
MongoMapper Rails 3 Branch.
Conclusion
That should be all. You can now start creating models based on whichever OM you've installed.
Note that this document is a work in progress. If you have any helpful comments, please add them below.
See also
Rails 3 App skeleton with MongoMapper

Rails 3 Release Notes
MongoDB Data Modeling and Rails

This tutorial discusses the development of a web application on Rails and MongoDB. MongoMapper will serve as our object mapper. The goal is
to provide some insight into the design choices required for building on MongoDB. To that end, we'll be constructing a simple but non-trivial social
news application. The source code for newsmonger is available on github for those wishing to dive right in.
Modeling Stories
Caching to Avoid N+1
A Note on Denormalization
Fields as arrays
Atomic Updates
Modeling Comments
Linear, Embedded Comments
Nested, Embedded Comments
Comment collections
Unfinished business
Assuming you've configured your application to work with MongoMapper, let's start thinking about the data model.
Modeling Stories
A news application relies on stories at its core, so we'll start with a Story model:
class Story
include MongoMapper::Document
key :title, String

key :url, String
key :slug, String
key :voters, Array
key :votes, Integer, :default => 0
key :relevance, Integer, :default => 0
# Cached values.
key :comment_count, Integer, :default => 0
key :username, String
# Note this: ids are of class ObjectId.

key :user_id, ObjectId
timestamps!
# Relationships.
belongs_to :user
# Validations.
validates_presence_of :title, :url, :user_id
end
Obviously, a story needs a title, url, and user_id, and should belong to a user. These are self-explanatory.
Caching to Avoid N+1
When we display our list of stories, we'll need to show the name of the user who posted the story. If we were using a relational database, we
could perform a join on users and stores, and get all our objects in a single query. But MongoDB does not support joins and so, at times, requires
bit of denormalization. Here, this means caching the 'username' attribute.
A Note on Denormalization
Relational purists may be feeling uneasy already, as if we were violating some universal law. But let's bear in mind that MongoDB collections are
not equivalent to relational tables; each serves a unique design objective. A normalized table provides an atomic, isolated chunk of data. A
document, however, more closely represents an object as a whole. In the case of a social news site, it can be argued that a username is intrinsic
to the story being posted.
What about updates to the username? It's true that such updates will be expensive; happily, in this case, they'll be rare. The read savings
achieved in denormalizing will surely outweigh the costs of the occasional update. Alas, this is not hard and fast rule: ultimately, developers must
evaluate their applications for the appropriate level of normalization.
Fields as arrays
With a relational database, even trivial relationships are blown out into multiple tables. Consider the votes a story receives. We need a way of
recording which users have voted on which stories. The standard way of handling this would involve creating a table, 'votes', with each row
referencing user_id and story_id.
With a document database, it makes more sense to store those votes as an array of user ids, as we do here with the 'voters' key.
For fast lookups, we can create an index on this field. In the MongoDB shell:
db.stories.ensureIndex('voters');
Or, using MongoMapper, we can specify the index in config/initializers/database.rb:
Story.ensure_index(:voters)
To find all the stories voted on by a given user:
Story.all(:conditions => {:voters => @user.id})
Atomic Updates
Storing the voters array in the Story class also allows us to take advantage of atomic updates. What this means here is that, when a user
votes on a story, we can
1. ensure that the voter hasn't voted yet, and, if not,

2. increment the number of votes and
3. add the new voter to the array.
MongoDB's query and update features allows us to perform all three actions in a single operation. Here's what that would look like from the shell:
// Assume that story_id and user_id represent real story and user ids.
db.stories.update({_id: story_id, voters: {'$ne': user_id}},
{'$inc': {votes: 1}, '$push': {voters: user_id}});
What this says is "get me a story with the given id whose voters array does not contain the given user id and, if you find such a story, perform
two atomic updates: first, increment votes by 1 and then push the user id onto the voters array."
This operation highly efficient; it's also reliable. The one caveat is that, because update operations are "fire and forget," you won't get a response
from the server. But in most cases, this should be a non-issue.
A MongoMapper implementation of the same update would look like this:
def self.upvote(story_id, user_id)

collection.update({'_id' => story_id, 'voters' => {'$ne' => user_id}},
{'$inc' => {'votes' => 1}, '$push' => {'voters' => user_id}})
end
Modeling Comments
In a relational database, comments are usually given their own table, related by foreign key to some parent table. This approach is occasionally
necessary in MongoDB; however, it's always best to try to embed first, as this will achieve greater query efficiency.
Linear, Embedded Comments
Linear, non-threaded comments should be embedded. Here are the most basic MongoMapper classes to implement such a structure:
class Story
many :comments
end
class Comment
include MongoMapper::EmbeddedDocument
key :body, String
belongs_to :story
end
If we were using the Ruby driver alone, we could save our structure like so:
@stories = @db.collection('stories')
@document = {:title => "MongoDB on Rails",
:comments => [{:body => "Revelatory! Loved it!",
:username => "Matz"
}
]
}
@stories.save(@document)
Essentially, comments are represented as an array of objects within a story document. This simple structure should be used for any one-to-many
relationship where the many items are linear.
Nested, Embedded Comments
But what if we're building threaded comments? An admittedly more complicated problem, two solutions will be presented here. The first is to
represent the tree structure in the nesting of the comments themselves. This might be achieved using the Ruby driver as follows:
@stories = @db.collection('stories')
@document = {:title => "MongoDB on Rails",
:comments => [{:body => "Revelatory! Loved it!",
:username => "Matz",
:comments => [{:body => "Agreed.",
:username => "rubydev29"
}
]
}
]
}
@stories.save(@document)
Representing this structure using MongoMapper would be tricky, requiring a number of custom mods.
But this structure has a number of benefits. The nesting is captured in the document itself (this is, in fact, how Business Insider represents
comments). And this schema is highly performant, since we can get the story, and all of its comments, in a single query, with no application-side
processing for constructing the tree.
One drawback is that alternative views of the comment tree require some significant reorganizing.
Comment collections
We can also represent comments as their own collection. Relative to the other options, this incurs a small performance penalty while granting us
the greatest flexibility. The tree structure can be represented by storing the unique path for each leaf (see Mathias's original post on the idea).
Here are the relevant sections of this model:
class Comment
key :body, String

key :depth, Integer, :default => 0
key :path, String, :default => ""
# Note: we're intentionally storing parent_id as a string

key :parent_id, String
key :story_id, ObjectId
timestamps!
# Relationships.
belongs_to :story
# Callbacks.
after_create :set_path
private
# Store the comment's path.

def set_path
unless self.parent_id.blank?
parent = Comment.find(self.parent_id)
self.story_id = parent.story_id
self.depth = parent.depth + 1
self.path = parent.path + ":" + parent.id
end
save
end
The path ends up being a string of object ids. This makes it easier to display our comments nested, with each level in order of karma or votes. If
we specify an index on story_id, path, and votes, the database can handle half the work of getting our comments in nested, sorted order.
The rest of the work can be accomplished with a couple grouping methods, which can be found in the newsmonger source code.
It goes without saying that modeling comments in their own collection also facilitates various site-wide aggregations, including displaying the
latest, grouping by user, etc.
Unfinished business
Document-oriented data modeling is still young. The fact is, many more applications will need to be built on the document model before we can
say anything definitive about best practices. So the foregoing should be taken as suggestions, only. As you discover new patterns, we encourage
you to document them, and feel free to let us know about what works (and what doesn't).
Developers working on object mappers and the like are encouraged to implement the best document patterns in their code, and to be wary of
recreating relational database models in their apps.
Object Mappers for Ruby and MongoDB

Although it's possible to use the Ruby driver by itself, sometimes you want validations, associations, and many of the other conveniences
provided by ActiveRecord. Here, then, is a list of the most popular object mappers available for working with Ruby and MongoDB.
Recommendations
Libraries
MongoMapper
Mongoid
Mongomatic
MongoDoc
MongoModel
Candy
MongoRecord
Recommendations
First we advise that you get to know how the database itself works. This is best accomplished by playing with the shell and experimenting with the
Ruby driver (or any of the other drivers, for that matter)*.
Once you understand how MongoDB works, you'll be in a good position to choose the object mapper that best suits your needs. So long as you
pick an OM that's used in production and is actively developed, you really can't make a bad choice.
Libraries
MongoMapper
John Nunemaker's OM. Used in production and actively-developed. ActiveModel support forthcoming.
Installation:
gem install mongo_mapper
Source:
mongo_mapper on github
Documentation:
MongoMapper on google groups

#mongomapper on freenode.
Articles:
Getting Started with MongoMapper

MongoMapper and Rails
More MongoMapper Awesomeness.
Mongoid
If you're using Mongoid, please read the Using Mongoid page on this site.
Durran Jordan's OM. Used in production and actively-developed. Supports ActiveModel and Rails 3.
Installation:
gem install mongoid
Source:
mongoid on github
Documentation:
Docs at mongoid.org
Mongomatic
A simple object mapper emphasizing MongoDB idioms.
Installation:
gem install mongomatic
Source:
mongomatic on github
MongoDoc
MongoDoc is a simple, fast ODM for MongoDB. The project will eventually be merged into Mongoid.
Installation:
gem install mongodoc
Source:
mongodoc on github
MongoModel
Sam Pohlenz's OM. Actively-developed.
Notes:
An OM with emphaisis on ActiveModel compatibility.
Installation:
gem install mongomodel
Source:
mongomodel on github
Candy
Stephen Eley's OM. Actively-developed.
Notes:
From the README:

Candy's goal is to provide the simplest possible object persistence for the MongoDB database. By "simple" we mean "nearly invisible." Candy
doesn't try to mirror ActiveRecord or DataMapper. Instead, we play to MongoDB's unusual strengths – extremely fast writes and a set of
field-specific update operators – and do away with the cumbersome, unnecessary methods of last-generation workflows.
Installation:
gem install candy
Source:
candy on github
MongoRecord
10gen's original OM.
Notes:
MongoRecord is an ActiveRecord-like OM, and the first of its kind developed for MongoDB. Favored by a contingent of developers for its
simplicity, MongoRecord currently receives a lot of love from Nate Wiger.
Installation:
gem install mongo_record
Source:
mongo-record on github
Using Mongoid
Mongoid is a mature ODM for MongoDB. Much work has gone into the project, and it sports an active user community and excellent
documentation. That said, we've seen a few of Mongoid's design decisions cause problems for users in production. This page is an attempt to
make current and future users aware of these issues.
Most of the issues mentioned here will be resolved with the release of Mongoid 2.0.
These principles should be applied to every MongoDB object mapper, regardless of implementation language. We still recommend Mongoid, but
only if the following points are taken into account.
Object ids should be stored as object ids, not as strings.

Index creation shouldn't be defined in the model and triggered on load.
Embedded documents and related documents are both appropriate solutions.
Automated sorting by _id ( #last method).
Safe mode should not be enabled for every operation by default.
Final thoughts
Object ids should be stored as object ids, not as strings.

Problem:
Mongoid currently stores object ids as string, by default. There are a couple of problems with this. The first is that the object id type is the reigning
convention; if you ever need to switch languages or ODMs, you'll find yourself in trouble, since all other projects will expect an object id proper
and not a string. Since it's impossible to change the _id attribute, the only way to convert a collection to use object ids proper is to rewrite the
collection. For many users, this would be incredibly inconvenient.
The second problem is that storing an object id as a string requires 16 extra bytes of storage per object id. This will be duplicated in any index
where it appears. Thus, storing object ids as strings needlessly increases storage size.
A final problem with storing object ids as strings is that upserts will always insert an object id proper. If you ever want to use upserts, you must
store object ids and object ids proper. Otherwise, your collection will contain object ids of multiple types.
Solution
In the Mongoid config, make sure you set
use_object_ids: true
All object mappers should use real, non-string object ids by default.
Index creation shouldn't be defined in the model and triggered on load.
Problem
Mongoid (and MongoMapper) allows developers to define indexes on the model layer. Because of this, we've now seen numerous examples of
developers accidentally triggering new index builds on production boxes. This can cause unwanted downtime and worse, depending on how the
developers react.
We've also noticed that this index creation API doesn't encourage compound indexes, which are often the best choice for certain queries. If you're
not familiar with compound indexes, read our indexing advice
Solution
The best solution for the moment is to avoid defining indexes in the model. We'd recommend creating some kind of rake task that updates
indexes so that you never inadvertently trigger an index creation. You can easily accomplish this with the Ruby driver API. Read the docs on
create_index
It's also important to know how and when to create compound indexes. Again, check our this indexing advice.
Embedded documents and related documents are both appropriate solutions.
Problem
The Mongoid API and documentation can lead newer users to believe that embedded documents are almost always the way to represent
one-to-many relations. However, there are plenty of situations where storing the related documents in a separate collection is the better choice.
Too much embedding can lead to extra large documents (> 100kb), which can be hard to work with. In addition, these large documents can be
inefficient when updating on the server, transferring across the network, and/or serializing to and from BSON.
Solution
Know that related documents can be a good choice just as frequently as embedded documents. Consult these schema design resources:
Schema design docs

MongoDB Schema Design slides
Schema Design in MongoDB Video from MongoNY
Automated sorting by _id (#last method).
Problem
We've seen methods that invoke automated sorts by _id to preserve order. This is fine if the right indexes are defined, but can start to result in
slow queries as the collection grows.
Solution
If you have large collections, be sure that you have the proper indexes defined so that sorts on _id don't slow down your system.
Safe mode should not be enabled for every operation by default.
Problem
Mongoid uses :safe mode by default for every insert and update. This isn't always necessary. Safe mode should be used when an error is
expected or when you want to use replication acknowledgment.
Solution
We suggest disabling safe mode in the Mongoid config, especially if performance is critical.
persist_in_safe_mode: false
Then, decide which operations actually need it. If you have a unique on a collection, or if you're doing updates that you think could fail, then safe
mode is a good idea. If you're logging, doing analytics, or performing a bulk insert, it may not be necessary.
The API should allow users to set safe mode on individual insert and update operations.
Final thoughts
This isn't an indictment of Mongoid or of any other object mapper. We think Mongoid is a great project, and we hope to see it continue to flourish.
But we've seen enough pain caused by some of its design decisions to warrant this page. Our only hope is that the recommendations here help
users have a better experience with Mongoid and MongoDB.
Ruby External Resources

There are a number of good resources appearing all over the web for learning about MongoDB and Ruby. A useful selection is listed below. If you
know of others, do let us know.
Screencasts
Presentations
Articles
Projects
Libraries
Screencasts
Introduction to MongoDB - Part I

An introduction to MongoDB via the MongoDB shell.
Introduction to MongoDB - Part II

In this screencast, Joon You teaches how to use the Ruby driver to build a simple Sinatra app.
Introduction to MongoDB - Part III

For the final screencast in the series, Joon You introduces MongoMapper and Rails.
RailsCasts: MongoDB & MongoMapper

Ryan Bates' RailsCast introducing MongoDB and MongoMapper.
Presentations
Introduction to MongoDB (Video)

Mike Dirolf's introduction to MongoDB at Pivotal Labs, SF.
MongoDB: A Ruby Document Store that doesn't rhyme with 'Ouch' (Slides)
Wynn Netherland's introduction to MongoDB with some comparisons to CouchDB.
MongoDB (is) for Rubyists (Slides)

Kyle Banker's presentation on why MongoDB is for Rubyists (and all human-oriented programmers).
Introduction to Mongoid and MongoDB (Video)

Durran Jordan discusses Mongoid, MongoDB, and how HashRocket uses these tools in production.
Articles
Why I Think Mongo is to Databases What Rails was to Frameworks

What if a key-value store mated with a relational database system?
John Nunemaker's articles on MongoDB.
A series of articles on aggregation with MongoDB and Ruby:

1. Part I: Introduction of Aggregation in MongoDB
2. Part II: MongoDB Grouping Elaborated
3. Part III: Introduction to Map-Reduce in MongoDB
Does the MongoDB Driver Support Feature X?

An explanation of how the MongoDB drivers usually automatically support new database features.
Projects
Mongo Queue
An extensible thread safe job/message queueing system that uses mongodb as the persistent storage engine.
Resque-mongo
A port of the Github's Resque to MongoDB.
Mongo Admin
A Rails plugin for browsing and managing MongoDB data. See the live demo.
Sinatra Resource
Resource Oriented Architecture (REST) for Sinatra and MongoMapper.
Shorty
A URL-shortener written with Sinatra and the MongoDB Ruby driver.
NewsMonger
A simple social news application demonstrating MongoMapper and Rails.
Data Catalog API

From Sunlight Labs, a non-trivial application using MongoMapper and Sinatra.
Watchtower
An example application using Mustache, MongoDB, and Sinatra.
Shapado
A question and answer site similar to Stack Overflow. Live version at shapado.com.
Libraries
ActiveExpando
An extension to ActiveRecord to allow the storage of arbitrary attributes in MongoDB.
ActsAsTree (MongoMapper)
ActsAsTree implementation for MongoMapper.
Machinist adapter (MongoMapper)

Machinist adapter using MongoMapper.
Mongo-Delegate
A delegation library for experimenting with production data without altering it. A quite useful pattern.
Remarkable Matchers (MongoMapper)

Testing / Matchers library using MongoMapper.
OpenIdAuthentication, supporting MongoDB as the datastore

Brandon Keepers' fork of OpenIdAuthentication supporting MongoDB.
MongoTree (MongoRecord)
MongoTree adds parent / child relationships to MongoRecord.
Merb_MongoMapper
a plugin for the Merb framework for supporting MongoMapper models.
Mongolytics (MongoMapper)
A web analytics tool.
Rack-GridFS
A Rack middleware component that creates HTTP endpoints for files stored in GridFS.
Frequently Asked Questions - Ruby

This is a list of frequently asked questions about using Ruby with MongoDB. If you have a question you'd like to have answered here, please add
it in the comments.
Can I run [insert command name here] from the Ruby driver?
Does the Ruby driver support an EXPLAIN command?
I see that BSON supports a symbol type. Does this mean that I can store Ruby symbols in MongoDB?
Why can't I access random elements within a cursor?
Why can't I save an instance of TimeWithZone?
I keep getting CURSOR_NOT_FOUND exceptions. What's happening?
I periodically see connection failures between the driver and MongoDB. Why can't the driver retry the operation automatically?
Can I run [insert command name here] from the Ruby driver?
Yes. You can run any of the available database commands from the driver using the DB#command method. The only trick is to use an
OrderedHash when specifying the command. For example, here's how you'd run an asynchronous fsync from the driver:
# This command is run on the admin database.

@db = Mongo::Connection.new.db('admin')
# Build the command.

cmd = OrderedHash.new
cmd['fsync'] = 1
cmd['async'] = true
# Run it.
@db.command(cmd)
It's important to keep in mind that some commands, like fsync, must be run on the admin database, while other commands can be run on any
database. If you're having trouble, check the command reference to make sure you're using the command correctly.
Does the Ruby driver support an EXPLAIN command?
Yes. explain is, technically speaking, an option sent to a query that tells MongoDB to return an explain plan rather than the query's results. You
can use explain by constructing a query and calling explain at the end:
@collection = @db['users']
result = @collection.find({:name => "jones"}).explain
The resulting explain plan might look something like this:
{"cursor"=>"BtreeCursor name_1",
"startKey"=>{"name"=>"Jones"},
"endKey"=>{"name"=>"Jones"},
"nscanned"=>1.0,
"n"=>1,
"millis"=>0,
"oldPlan"=>{"cursor"=>"BtreeCursor name_1",
"endKey"=>{"name"=>"Jones"}
},
"allPlans"=>[{"cursor"=>"BtreeCursor name_1",
"endKey"=>{"name"=>"Jones"}}]
}
Because this collection has an index on the "name" field, the query uses that index, only having to scan a single record. "n" is the number of
records the query will return. "millis" is the time the query takes, in milliseconds. "oldPlan" indicates that the query optimizer has already seen this
kind of query and has, therefore, saved an efficient query plan. "allPlans" shows all the plans considered for this query.
I see that BSON supports a symbol type. Does this mean that I can store Ruby symbols in MongoDB?
You can store Ruby symbols in MongoDB, but only as values. BSON specifies that document keys must be strings. So, for instance, you can do
this:
@collection = @db['test']
boat_id = @collection.save({:vehicle => :boat})

car_id = @collection.save({"vehicle" => "car"})
@collection.find_one('_id' => boat_id)

{"_id" => ObjectID('4bb372a8238d3b5c8c000001'), "vehicle" => :boat}
@collection.find_one('_id' => car_id)

{"_id" => ObjectID('4bb372a8238d3b5c8c000002'), "vehicle" => "car"}
Notice that the symbol values are returned as expected, but that symbol keys are treated as strings.
Why can't I access random elements within a cursor?
MongoDB cursors are designed for sequentially iterating over a result set, and all the drivers, including the Ruby driver, stick closely to this
directive. Internally, a Ruby cursor fetches results in batches by running a MongoDB getmore operation. The results are buffered for efficient
iteration on the client-side.
What this means is that a cursor is nothing more than a device for returning a result set on a query that's been initiated on the server. Cursors are
not containers for result sets. If we allow a cursor to be randomly accessed, then we run into issues regarding the freshness of the data. For
instance, if I iterate over a cursor and then want to retrieve the cursor's first element, should a stored copy be returned, or should the cursor re-run
the query? If we returned a stored copy, it may not be fresh. And if the the query is re-run, then we're technically dealing with a new cursor.
To avoid those issues, we're saying that anyone who needs flexible access to the results of a query should store those results in an array and
then access the data as needed.
Why can't I save an instance of TimeWithZone?
MongoDB stores times in UTC as the number of milliseconds since the epoch. This means that the Ruby driver serializes Ruby Time objects only.
While it would certainly be possible to serialize a TimeWithZone, this isn't preferable since the driver would still deserialize to a Time object.
All that said, if necessary, it'd be easy to write a thin wrapper over the driver that would store an extra time zone attribute and handle the
serialization/deserialization of TimeWithZone transparently.
I keep getting CURSOR_NOT_FOUND exceptions. What's happening?
The most likely culprit here is that the cursor is timing out on the server. Whenever you issue a query, a cursor is created on the server. Cursor
naturally time out after ten minutes, which means that if you happen to be iterating over a cursor for more than ten minutes, you risk a
CURSOR_NOT_FOUND exception.
There are two solutions to this problem. You can either:
1. Limit your query. Use some combination of limit and skip to reduce the total number of query results. This will, obviously, bring down the
time it takes to iterate.
2. Turn off the cursor timeout. To do that, invoke find with a block, and pass :timeout => true:
@collection.find({}, :timeout => false) do |cursor|

cursor.each do |document
# Process documents here
end
end
I periodically see connection failures between the driver and MongoDB. Why can't the driver retry the operation automatically?
A connection failure can indicate any number of failure scenarios. Has the server crashed? Are we experiencing a temporary network partition? Is
there a bug in our ssh tunnel?
Without further investigation, it's impossible to know exactly what has caused the connection failure. Furthermore, when we do see a connection
failure, it's impossible to know how many operations prior to the failure succeeded. Imagine, for instance, that we're using safe mode and we send
an $inc operation to the server. It's entirely possible that the server has received the $inc but failed on the call to getLastError. In that case,
retrying the operation would result in a double-increment.
Because of the indeterminacy involved, the MongoDB drivers will not retry operations on connection failure. How connection failures should be
handled is entirely dependent on the application. Therefore, we leave it to the application developers to make the best decision in this case.
The drivers will reconnect on the subsequent operation.
Java Language Center

Basics
Tutorial
API Documentation
Downloads
Specific Topics
Concurrency
Saving Objects
Data Types
3rd Party
POJO Mappers
Morphia - Type-Safe Wrapper with DAO/Datastore abstractions

pojo to MongoDB
mungbean
daybreak PoJo mapping for Java & MongoDB using Java 5 annotations.
Code Generation
Sculptor - mongodb-based DSL -> Java (code generator)
GuicyData - DSL -> Java generator with Guice integration
Blog Entries
Misc
log4j -> mongodb appender

(Experimental, Type4) JDBC driver
Other JVM-based Languages
Clojure
Groovy
Groovy Tutorial for MongoDB
MongoDB made more Groovy
GMongo, a Groovy wrapper to the mongodb Java driver
GMongo 0.5 Released
Scala
Lift-MongoDB - Wrapper, Mapper, and Record back-end implementation. Part of the Lift Web Framework .
mongo-scala-driver is a thin wrapper around mongo-java-driver to make working with MongoDB more Scala-like.
Wiki
Mailing list
Casbah Casbah is a Scala oriented series of wrappers and extensions to the MongoDB Java driver to provide a more
scala-friendly interface to MongoDB. Implements the Scala 2.8 collection interfaces to improve interaction, and a fluid query
syntax which closely matches the MongoDB interface. Support for ORM-style Object mapping is coming soon, as well.
Tutorial
Mailing List
GitHub Project Page
JavaScript
MongoDB-Rhino - A toolset to provide full integration between the Rhino JavaScript engine for the JVM and MongoDB. Uses the
MongoDB Java driver.
JRuby
jmongo A thin ruby wrapper around the mongo-java-driver for vastly better jruby performance.
If there is a project missing here, just add a comment or email the list and we'll add it.
Presentations
Using MongoDB with Scala - Brendan McAdams' Presentation at the New York Scala Enthusiasts (August 2010)
Java Development - Brendan McAdams' Presentation from MongoNYC (May 2010)
Java Development - James Williams' Presentation from MongoSF (April 2010)
Building a Mongo DSL in Scala at Hot Potato - Lincoln Hochberg's Presentation from MongoSF (April 2010)
Java Driver Concurrency
The Java MongoDB driver is thread safe. If you are using in a web serving environment, for example, you should create a single Mongo instance,
and you can use it in every request. The Mongo object maintains an internal pool of connections to the database (default pool size of 10).
However, if you want to ensure complete consistency in a "session" (maybe an http request), you probably want the driver to use the same socket
for that session (which isn't necessarily the case since Mongo instances have built-in connection pooling). This is only necessary for a write heavy
environment, where you might read data that you wrote.
To do that, you would do something like:
DB db...;
db.requestStart();
code....
db.requestDone();
Java - Saving Objects Using DBObject

The Java driver provides a DBObject interface to save custom objects to the database.
For example, suppose one had a class called Tweet that they wanted to save:
public class Tweet implements DBObject {

/* ... */
}
Then you can say:
Tweet myTweet = new Tweet();

myTweet.put("user", userId);
myTweet.put("message", msg);
myTweet.put("date", new Date());
collection.insert(myTweet);
When a document is retrieved from the database, it is automatically converted to a DBObject. To convert it to an instance of your class, use
DBCollection.setObjectClass():
collection.setObjectClass(Tweet);
Tweet myTweet = (Tweet)collection.findOne();
Java Tutorial
Introduction
A Quick Tour
Making A Connection
Getting A List Of Collections
Getting A Collection
Finding the First Document In A Collection using findOne()
Counting Documents in A Collection
Using a Cursor to Get All the Documents
Getting A Single Document with A Query
Getting A Set of Documents With a Query
Creating An Index
Quick Tour of the Administrative Functions
Getting A List of Databases
Dropping A Database
Introduction
This page is a brief overview of working with the MongoDB Java Driver.
For more information about the Java API, please refer to the online API Documentation for Java Driver
A Quick Tour
Using the Java driver is very simple. First, be sure to include the driver jar mongo.jar in your classpath. The following code snippets come from
the examples/QuickTour.java example code found in the driver.
Making A Connection
To make a connection to a MongoDB, you need to have at the minimum, the name of a database to connect to. The database doesn't have to
exist - if it doesn't, MongoDB will create it for you.
Additionally, you can specify the server address and port when connecting. The following example shows three ways to connect to the database
mydb on the local machine :
import com.mongodb.Mongo;
import com.mongodb.DB;
import com.mongodb.DBCollection;
import com.mongodb.BasicDBObject;
import com.mongodb.DBObject;
import com.mongodb.DBCursor;
Mongo m = new Mongo();

Mongo m = new Mongo( "localhost" );
Mongo m = new Mongo( "localhost" , 27017 );
DB db = m.getDB( "mydb" );
At this point, the db object will be a connection to a MongoDB server for the specified database. With it, you can do further operations.
Note: The Mongo object instance actually represents a pool of connections to the database; you will only need one object of class Mongo even
with multiple threads. See the concurrency doc page for more information.
MongoDB can be run in a secure mode where access to databases is controlled through name and password authentication. When run in this
mode, any client application must provide a name and password before doing any operations. In the Java driver, you simply do the following with
the connected mongo object :
boolean auth = db.authenticate(myUserName, myPassword);
If the name and password are valid for the database, auth will be true. Otherwise, it will be false. You should look at the MongoDB log for
further information if available.
Most users run MongoDB without authentication in a trusted environment.
Getting A List Of Collections
Each database has zero or more collections. You can retrieve a list of them from the db (and print out any that are there) :
Set<String> colls = db.getCollectionNames();
for (String s : colls) {

System.out.println(s);
}
and assuming that there are two collections, name and address, in the database, you would see
name
address
as the output.
Getting A Collection
To get a collection to use, just specify the name of the collection to the getCollection(String collectionName) method:
DBCollection coll = db.getCollection("testCollection")
Once you have this collection object, you can now do things like insert data, query for data, etc
Once you have the collection object, you can insert documents into the collection. For example, lets make a little document that in JSON would be
represented as
{
"name" : "MongoDB",
"type" : "database",
"count" : 1,
"info" : {
x : 203,
y : 102
}
}
Notice that the above has an "inner" document embedded within it. To do this, we can use the BasicDBObject class to create the document
(including the inner document), and then just simply insert it into the collection using the insert() method.
BasicDBObject doc = new BasicDBObject();
doc.put("name", "MongoDB");
doc.put("type", "database");
doc.put("count", 1);
BasicDBObject info = new BasicDBObject();
info.put("x", 203);
info.put("y", 102);
doc.put("info", info);
coll.insert(doc);
Finding the First Document In A Collection using findOne()
To show that the document we inserted in the previous step is there, we can do a simple findOne() operation to get the first document in the
collection. This method returns a single document (rather than the DBCursor that the find() operation returns), and it's useful for things where
there only is one document, or you are only interested in the first. You don't have to deal with the cursor.
DBObject myDoc = coll.findOne();

System.out.println(myDoc);
and you should see
{ "_id" : "49902cde5162504500b45c2c" , "name" : "MongoDB" , "type" : "database" , "count" : 1 , "info"

: { "x" : 203 , "y" : 102} , "_ns" : "testCollection"}
Note the _id and _ns elements have been added automatically by MongoDB to your document. Remember, MongoDB reserves element names
that start with _ for internal use.
In order to do more interesting things with queries, let's add multiple simple documents to the collection. These documents will just be
{
"i" : value
}
and we can do this fairly efficiently in a loop
for (int i=0; i < 100; i++) {

coll.insert(new BasicDBObject().append("i", i));
}
Notice that we can insert documents of different "shapes" into the same collection. This aspect is what we mean when we say that MongoDB is
"schema-free"
Counting Documents in A Collection
Now that we've inserted 101 documents (the 100 we did in the loop, plus the first one), we can check to see if we have them all using the
getCount() method.
System.out.println(coll.getCount());
and it should print 101.
Using a Cursor to Get All the Documents
In order to get all the documents in the collection, we will use the find() method. The find() method returns a DBCursor object which allows
us to iterate over the set of documents that matched our query. So to query all of the documents and print them out :
DBCursor cur = coll.find();
while(cur.hasNext()) {
System.out.println(cur.next());
}
and that should print all 101 documents in the collection.
Getting A Single Document with A Query
We can create a query to pass to the find() method to get a subset of the documents in our collection. For example, if we wanted to find the
document for which the value of the "i" field is 71, we would do the following ;
BasicDBObject query = new BasicDBObject();
query.put("i", 71);
cur = coll.find(query);
}
and it should just print just one document
{ "_id" : "49903677516250c1008d624e" , "i" : 71 , "_ns" : "testCollection"}

Getting A Set of Documents With a Query
We can use the query to get a set of documents from our collection. For example, if we wanted to get all documents where "i" > 50, we could write
:
query = new BasicDBObject();
query.put("i", new BasicDBObject("$gt", 50)); // e.g. find all where i > 50
}
which should print the documents where i > 50. We could also get a range, say 20 < i <= 30 :
query = new BasicDBObject();
query.put("i", new BasicDBObject("$gt", 20).append("$lte", 30)); // i.e. 20 < i <= 30
}
Creating An Index
MongoDB supports indexes, and they are very easy to add on a collection. To create an index, you just specify the field that should be indexed,
and specify if you want the index to be ascending (1) or descending (-1). The following creates an ascending index on the "i" field :
coll.createIndex(new BasicDBObject("i", 1)); // create index on "i", ascending
You can get a list of the indexes on a collection :
List<DBObject> list = coll.getIndexInfo();
for (DBObject o : list) {

System.out.println(o);
}
and you should see something like
{ "name" : "i_1" , "ns" : "mydb.testCollection" , "key" : { "i" : 1} , "_ns" : "system.indexes"}
Quick Tour of the Administrative Functions
Getting A List of Databases
You can get a list of the available databases:
for (String s : m.getDatabaseNames()) {

System.out.println(s);
}
Dropping A Database
You can drop a database by name using the Mongo object:
m.dropDatabase("my_new_db");
Java Types
Object Ids
Regular Expressions
Dates/Times
Database References
Binary Data
Embedded Documents
Arrays
Object Ids
com.mongodb.ObjectId is used to autogenerate unique ids.
ObjectId id = new ObjectId();

ObjectId copy = new ObjectId(id);
Regular Expressions
The Java driver uses java.util.regex.Pattern for regular expressions.
Pattern john = Pattern.compile("joh?n", CASE_INSENSITIVE);

BasicDBObject query = new BasicDBObject("name", john);
// finds all people with "name" matching /joh?n/i

DBCursor cursor = collection.find(query);
Dates/Times
The java.util.Date class is used for dates.
Date now = new Date();

BasicDBObject time = new BasicDBObject("ts", now);
collection.save(time);
Database References
com.mongodb.DBRef can be used to save database references.

DBRef addressRef = new DBRef(db, "foo.bar", address_id);
DBObject address = addressRef.fetch();
DBObject person = BasicDBObjectBuilder.start()

.add("name", "Fred")
.add("address", addressRef)
.get();
collection.save(person);
DBObject fred = collection.findOne();

DBRef addressObj = (DBRef)fred.get("address");
addressObj.fetch()
Binary Data
An array of bytes (byte[]) can be used for binary data.
Embedded Documents
Suppose we have a document that, in JavaScript, looks like:
{
"x" : {
"y" : 3
}
}
The equivalent in Java is:
BasicDBObject y = new BasicDBObject("y", 3);

BasicDBObject x = new BasicDBObject("x", y);
Arrays
Anything that extends List in Java will be saved as an array.
So, if you are trying to represent the JavaScript:
{
"x" : [
1,
2,
{"foo" : "bar"},
4
]
}
you could do:
ArrayList x = new ArrayList();

x.add(1);
x.add(2);
x.add(new BasicDBObject("foo", "bar"));
x.add(4);
BasicDBObject doc = new BasicDBObject("x", x);

C++ Language Center
A C++ driver is available for communicating with the MongoDB. As the database is written in C++, the driver actually uses some core MongoDB
code -- this is the same driver that the database uses itself for replication.
The driver has been compiled successfully on Linux, OS X, Windows, and Solaris.
API Documentation
MongoDB C++ Client Tutorial
Using BSON from C++
HOWTO
Connecting
Tailable Cursors
Mongo Database and C++ Driver Source Code (at github). See the client subdirectory for client driver related files.
Download
Additional Notes
The Building documentation covers compiling the entire database, but some of the notes there may be helpful for compiling client
applications too.
There is also a pure C driver for MongoDB. For true C++ apps we recommend using the C++ driver.
C++ BSON Library

Overview
Examples
API Docs
Short Class Names
Overview
The MongoDB C++ driver library includes a bson package that implements the BSON specification (see http://www.bsonspec.org/). This library
can be used standalone for object serialization and deserialization even when one is not using MongoDB at all.
Include bson/bson.h or db/jsobj.h in your application (not both). bson.h is new and may not work in some situations, was is good for light
header-only usage of BSON (see the bsondemo.cpp example).
Key classes:
BSONObj a BSON object
BSONElement a single element in a bson object. This is a key and a value.
BSONObjBuilder used to make BSON objects
BSONObjIterator to enumerate BSON objects
BSONObjBuilder b;
b.append("name", "Joe");
b.append("age", 33);
BSONObj p = b.obj();
Or more concisely:
BSONObj p = BSONObjBuilder().append("name", "Joe").append("age", 33).obj();
We can also create objects with a stream-oriented syntax:

BSONObjBuilder b;
b << "name" << "Joe" << "age" << 33;
The macro BSON lets us be even more compact:
BSONObj p = BSON( "name" << "Joe" << "age" << 33 );
Use the GENOID helper to add an object id to your object. The server will add an _id automatically if it is not included explicitly.
BSONObj p = BSON( GENOID << "name" << "Joe" << "age" << 33 );
// result is: { _id : ..., name : "Joe", age : 33 }
GENOID should be at the beginning of the generated object. We can do something similar with the non-stream builder syntax:
BSONObj p =
BSONObjBuilder().genOID().append("name","Joe").append("age",33).obj();
Examples
http://github.com/mongodb/mongo/blob/master/bson/bsondemo/bsondemo.cpp
API Docs
http://api.mongodb.org/cplusplus/
Short Class Names
Add
using namespace bson;
to your code to use the following shorter more C++ style names for the BSON classes:
// from bsonelement.h
namespace bson {
typedef mongo::BSONElement be;
typedef mongo::BSONObj bo;
typedef mongo::BSONObjBuilder bob;
}
(Or one could use bson::bo fully qualified for example).
Also available is bo::iterator as a synonym for BSONObjIterator.
C++ Tutorial
Installing the Driver Library and Headers
Unix
From Source
Windows
Compiling
Writing Client Code
Connecting
BSON
Inserting
Querying
Indexing
Sorting
Updating
Further Reading
This document is an introduction to usage of the MongoDB database from a C++ program.
First, install Mongo -- see the Quickstart for details.
Next, you may wish to take a look at the Developer's Tour guide for a language independent look at how to use MongoDB. Also, we suggest
some basic familiarity with the mongo shell -- the shell is one's primary database administration tool and is useful for manually inspecting the
contents of a database after your C++ program runs.
Installing the Driver Library and Headers
A good source for general information about setting up a MongoDB development environment on various operating systems is the building page.
Unix
From Source
For Unix, the Mongo driver library is libmongoclient.a.
When installing from source, use scons --full install to install the libraries. --full tells the install target to include the library and header
files; by default library and header files are installed in /usr/local. You can use --prefix to change the install path: scons --prefix
/opt/mongo --full install. You can also specify --sharedclient to build a shared library instead of a statically linked library.
The normal db distribution used to include the C++ driver, but now you have to build from source.
Windows
For more information on Boost setup see the Building for Windows page.
Compiling
The C++ drivers requires the pcre and boost libraries (with headers) to compile. Be sure they are in your include and lib paths. You can usually
install them from your OS's package manager if you don't already have them.
Writing Client Code
Note: for brevity, the examples below are simply inline code. In a real application one will define classes for each database object typically.
Connecting
Let's make a tutorial.cpp file that connects to the database (see client/examples/tutorial.cpp for full text of the examples below):
#include <iostream>
#include "client/dbclient.h"
using namespace mongo;
void run() {
DBClientConnection c;
c.connect("localhost");
}
int main() {
try {
run();
cout << "connected ok" << endl;
} catch( DBException &e ) {
cout << "caught " << e.what() << endl;
}
return 0;
}
If you are using gcc on Linux or OS X, you would compile with something like this, depending on location of your include files and libraries:
$ g++ tutorial.cpp -lmongoclient -lboost_thread-mt -lboost_filesystem -lboost_program_options -o
tutorial
$ ./tutorial
connected ok
$
Depending on your boost version you might need to link against the boost_system library as well: -lboost_system. Also, you
may need to append "-mt" to boost_filesystem and boost_program_options. And, of course, you may need to use -I and -L to
specify the locations of your mongo and boost headers and libraries.
BSON
The Mongo database stores data in BSON format. BSON is a binary object format that is JSON-like in terms of the data which can be stored
(some extensions exist, for example, a Date datatype).
To save data in the database we must create objects of class BSONObj. The components of a BSONObj are represented as BSONElement
objects. We use BSONObjBuilder to make BSON objects, and BSONObjIterator to enumerate BSON objects.
BSONObjBuilder b;
b.append("name", "Joe");
b.append("age", 33);
Or more concisely:
BSONObj p = BSONObjBuilder().append("name", "Joe").append("age", 33).obj();
We can also create objects with a stream-oriented syntax:
BSONObjBuilder b;
b << "name" << "Joe" << "age" << 33;
The macro BSON lets us be even more compact:
Use the GENOID helper to add an object id to your object. The server will add an _id automatically if it is not included explicitly.
BSONObj p = BSON( GENOID << "name" << "Joe" << "age" << 33 );
// result is: { _id : ..., name : "Joe", age : 33 }
GENOID should be at the beginning of the generated object. We can do something similar with the non-stream builder syntax:
BSONObj p =
BSONObjBuilder().genOID().append("name","Joe").append("age",33).obj();
Inserting
We now save our person object in a persons collection in the database:
c.insert("tutorial.persons", p);
The first parameter to insert is the namespace. tutorial is the database and persons is the collection name.
Querying
Let's now fetch all objects from the persons collection, and display them. We'll also show here how to use count().
cout << "count:" << c.count("tutorial.persons") << endl;
auto_ptr<DBClientCursor> cursor =
c.query("tutorial.persons", emptyObj);
while( cursor->more() )
cout << cursor->next().toString() << endl;
emptyObj is the empty BSON object -- we use it to represent {} which indicates an empty query pattern (an empty query is a query for all objects).
We use BSONObj::toString() above to print out information about each object retrieved. BSONObj::toString is a diagnostic function which prints
an abbreviated JSON string representation of the object. For full JSON output, use BSONObj::jsonString.
Let's now write a function which prints out the name (only) of all persons in the collection whose age is a given value:
void printIfAge(DBClientConnection&c, int age) {

auto_ptr<DBClientCursor> cursor =
c.query("tutorial.persons", QUERY( "age" << age ) );
while( cursor->more() ) {
BSONObj p = cursor->next();
cout << p.getStringField("name") << endl;
}
}
getStringField() is a helper that assumes the "name" field is of type string. To manipulate an element in a more generic fashion we can retrieve
the particular BSONElement from the enclosing object:
BSONElement name = p["name"];

// or:
//BSONElement name = p.getField("name");
See the api docs, and jsobj.h, for more information.
Our query above, written as JSON, is of the form
{ age : <agevalue> }
Queries are BSON objects of a particular format -- in fact, we could have used the BSON() macro above instead of QUERY(). See class Query in
dbclient.h for more information on Query objects, and the Sorting section below.
In the mongo shell (which uses javascript), we could invoke:
use tutorial;
db.persons.find( { age : 33 } );
Indexing
Let's suppose we want to have an index on age so that our queries are fast. We would use:
c.ensureIndex("tutorial.persons", fromjson("{age:1}"));
The ensureIndex method checks if the index exists; if it does not, it is created. ensureIndex is intelligent and does not repeat transmissions to the
server; thus it is safe to call it many times in your code, for example, adjacent to every insert operation.
In the above example we use a new function, fromjson. fromjson converts a JSON string to a BSONObj. This is sometimes a convenient way to
specify BSON. Alternatively we could have written:
c.ensureIndex("tutorial.persons", BSON( "age" << 1 ));
Sorting
Let's now make the results from printIfAge sorted alphabetically by name. To do this, we change the query statement from:
auto_ptr<DBClientCursor> cursor = c.query("tutorial.persons", QUERY( "age" << age ) );
to
to auto_ptr<DBClientCursor> cursor = c.query("tutorial.persons", QUERY( "age" << age ).sort("name")

);
Here we have used Query::sort() to add a modifier to our query expression for sorting.
Updating
Use the update() method to perform a database update . For example the following update in the mongo shell :
> use tutorial

> db.persons.update( { name : 'Joe', age : 33 },
... { $inc : { visits : 1 } } )
is equivalent to the following C++ code:
db.update( "tutorial.persons" ,
BSON( "name" << "Joe" << "age" << 33 ),
BSON( "$inc" << BSON( "visits" << 1 ) ) );
Further Reading
This overview just touches on the basics of using Mongo from C++. There are many more capabilities. For further exploration:
See the language-independent Developer's Tour;

Experiment with the mongo shell;
Review the doxygen API docs;
See connecting pooling information in the API docs;
See GridFS file storage information in the API docs;
See the HOWTO pages under the C++ Language Center
Consider getting involved to make the product (either C++ driver, tools, or the database itself) better!
Connecting
The C++ driver includes several classes for managing collections under the parent class DBClientInterface.
In general, you will want to instantiate either a DBClientConnection object, or a DBClientPaired object. DBClientConnection is our normal
connection class for a connection to a single MongoDB database server (or shard manager). We use DBClientPaired to connect to database
replica pairs.
See http://api.mongodb.org/cplusplus/ for details on each of the above classes.
Note : replica pairs will soon be replaced by Replica Sets; a new / adjusted interface will be available then.
Perl Language Center

Installing
CPAN
Manual (Non-CPAN) Installation
Big-Endian Systems
Next Steps
MongoDB Perl Tools
Entities::Backend::MongoDB
MojoX::Session::Store::MongoDB
MongoDB::Admin
MongoX
Installing
Start a MongoDB server instance (mongod) before installing so that the tests will pass. The mongod cannot be running as a
slave for the tests to pass.
Some tests may be skipped, depending on the version of the database you are running.
CPAN
$ sudo cpan MongoDB
The Perl driver is available through CPAN as the package MongoDB. It should build cleanly on *NIX and Windows (via Strawberry Perl).
Manual (Non-CPAN) Installation
If you would like to try the latest code or are contributing to the Perl driver, it is available at Github. There is also documentation generated after
every commit.
You can see if it's a good time to grab the bleeding edge code by seeing if the build is green.
To build the driver, run:
$ perl Makefile.PL
$ make
$ make test # make sure mongod is running, first
$ sudo make install
Please note that the tests will not pass without a mongod process running.
Big-Endian Systems
The driver will work on big-endian machines, but the database will not. The tests assume that mongod will be running on localhost unless
%ENV{MONGOD} is set. So, to run the tests, start the database on a little-endian machine (at, say, "example.com") and then run the tests with:
MONGOD=example.com make test
A few tests that require a database server on "localhost" will be skipped.
Next Steps
There is a tutorial and API documentation on CPAN.
If you're interested in contributing to the Perl driver, check out Contributing to the Perl Driver.
MongoDB Perl Tools
Entities::Backend::MongoDB
Entities::Backend::MongoDB is a backend for the Entities user management and authorization system stores all entities and relations
between them in a MongoDB database, using the MongoDB module. This is a powerful, fast backend that gives you all the features of MongoDB.
MojoX::Session::Store::MongoDB
MojoX::Session::Store::MongoDB is a store for MojoX::Session that stores a session in a MongoDB database. Created by Ask Bjørn
Hansen.
MongoDB::Admin
MongoDB::Admin is a collection of MongoDB administrative functions. Created by David Burley.
MongoX
MongoX - DSL sugar for MongoDB
Contributing to the Perl Driver

The easiest way to contribute is to file bugs and feature requests on Jira.
If you would like to help code the driver, read on...
Finding Something to Help With
Fixing Bugs
You can choose a bug on Jira and fix it. Make a comment that you're working on it, to avoid overlap.
Writing Tests
The driver could use a lot more tests. We would be grateful for any and all tests people would like to write.
Adding Features
If you think a feature is missing from the driver, you're probably right. Check on IRC or the mailing list, then go ahead and create a Jira case and
add the feature. The Perl driver was a bit neglected for a while (although it's now getting a lot of TLC) so it's missing a lot of things that the other
drivers have. You can look through their APIs for ideas.
Contribution Guildlines
The best way to make changes is to create an account on [Github], fork the driver, make your improvements, and submit a merge request.
To make sure your changes are approved and speed things along:
Write tests. Lots of tests.

Document your code.
Write POD, when applicable.
Bonus (for C programmers, particularly):
Make sure your change works on Perl 5.8, 5.10, Windows, Mac, Linux, etc.
Code Layout
The important files:

| perl_mongo.c # serialization/deserialization
| mongo_link.c # connecting to, sending to, and receiving from the database
- lib
- MongoDB
| Connection.pm # connection, queries, inserts... everything comes through here
| Database.pm
| Collection.pm
| Cursor.pm
| OID.pm
| GridFS.pm
- GridFS
| File.pm
- xs
| Mongo.xs
| Connection.xs
| Cursor.xs
| OID.xs
Perl Tutorial
Redirection Notice
This page should redirect to http://search.cpan.org/dist/MongoDB/lib/MongoDB/Tutorial.pod.
Online API Documentation

MongoDB API and driver documentation is available online. It is updated daily.
Java Driver API Documentation

C++ Driver API Documentation
Python Driver API Documentation
Ruby Driver API Documentation
PHP Driver API Documentation
See Also
Mongo Query Language

mongosniff
--objcheck command line parameter
Overview - Writing Drivers and Tools

This section contains information for developers that are working with the low-level protocols of Mongo - people who are writing drivers and
higher-level tools.
Documents of particular interest :
BSON http://bsonspec.org Description of the BSON binary document format. Fundamental to how Mongo and it's client software works.
Mongo Wire Protocol Specification for the basic socket communications protocol used between Mongo and clients.
Mongo Driver Requirements Description of what functionality is expected from a Mongo Driver
GridFS Specification Specification of GridFS - a convention for storing large objects in Mongo
Mongo Extended JSON Description of the extended JSON protocol for the REST-ful interface (ongoing development)
Additionally we recommend driver authors take a look at existing driver source code as an example.
bsonspec.org
Mongo Driver Requirements

This is a high-level list of features that a driver for MongoDB might provide. We attempt to group those features by priority. This list should be
taken with a grain of salt, and probably used more for inspiration than as law that must be adhered to. A great way to learn about implementing a
driver is by reading the source code of any of the existing drivers, especially the ones listed as "mongodb.org supported".
High priority
BSON serialization/deserialization
full cursor support (e.g. support OP_GET_MORE operation)
close exhausted cursors via OP_KILL_CURSORS
support for running database commands
handle query errors
convert all strings to UTF-8 (part of proper support for BSON)
hint, explain, count, $where
database profiling: set/get profiling level, get profiling info
advanced connection management (replica pairs, slave okay)
automatic reconnection
Medium priority
validate a collection in a database

buffer pooling
Tailable cursor support
A driver should be able to connect to a single server. By default this must be localhost:27017, and must also allow the server to be specified
by hostname and port.
Mongo m = new Mongo(); // go to localhost, default port
Mongo m = new Mongo(String host, int port);
How the driver does this is up to the driver - make it idiomatic. However, a driver should make it explicit and clear what is going on.
Pair Mode Connection
A driver must be able to support "Pair Mode" configurations, where two mongod servers are specified, and configured for hot-failover.
The driver should determine which of the pair is the current master, and send all operations to that server. In the event of an error, either socket
error or a "not a master" error, the driver must restart the determination process. It must not assume the other server in the pair is now the master.
ServerPair sp = new ServerPair(INETAddr...);

Mongo m = new Mongo(sp)
A driver may optionally allow a driver to connect deliberately to the "non-master" in the pair, for debugging, admin or operational purposes.
ServerPair sp = new ServerPair(INETAddr...);

sp.setTarget(ServerPair.SHADOW_MASTER);
Mongo m = new Mongo(sp);
1. Cluster Mode Connect to master in master-slave cluster
ServerCluster sc = new ServerCluster(INETAddr...); // again, give one and discover?

Mongo m = new Mongo(sc);
## Connect to slave in read-only mode in master-slave cluster

ServerCluster sc = new ServerCluster(INETAddr...); // again, give one and discover?
sc.setTarget(...)
Mongo m = new Mongo(sc);
or maybe make it like *Default/Simple* w/ a flag?
Other than that, we need a way to get a DB object :
DB db = m.getDB(name);
And a list of db names (useful for tools...) :
List<String> getDBNameList();
Database Object
Simple operations on a database object :

/**
* get name of database
*/
String dbName = db.getName();
/**
* Get a list of all the collection names in this database
*/
List<String> cols = db.getCollectionNames();
/**
* get a collection object. Can optionally create it if it
* doesn't exist, or just be strict. (XJDM has strictness as an option)
*/
Collection coll = db.getCollection(string);
/**
* Create a collection w/ optional options. Can fault
* if the collection exists, or can just return it if it already does
*/
Collection coll = db.createCollection( string);
Collection coll = db.createCollection( string, options);
/**
* Drop a collection by its name or by collection object.
* Driver could invalidate any outstanding Collection objects
* for that collection, or just hope for the best.
*/
boolean b = db.dropCollection(name);
boolean b = db.dropCollection(Collection);
/**
* Execute a command on the database, returning the
* BSON doc with the results
*/
Document d = db.executeCommand(command);
/**
* Close the [logical] database
*/
void db.close();
/**
* Erase / drop an entire database
*/
bool dropDatabase(dbname)
These methods have to do with database metadata: profiling levels and collection validation. Each admin object is associated with a database.
These methods could either be built into the Database class or provided in a separate Admin class whose instances are only available from a
database instance.
/* get an admin object from a database object. */
Admin admin = db.getAdmin();
/**
* Get profiling level. Returns one of the strings "off", "slowOnly", or
* "all". Note that the database returns an integer. This method could
* return an int or an enum instead --- in Ruby, for example, we return
* symbols.
*/
String profilingLevel = admin.getProfilingLevel();
/**
* Set profiling level. Takes whatever getProfilingLevel() returns.
*/
admin.setProfilingLevel("off");
/**
* Retrieves the database's profiling info.
*/
Document profilingInfo = admin.getProfilingInfo();
/**
* Returns true if collection is valid; raises an exception if not.
*/
boolean admin.validateCollection(collectionName);
Collection Basic Ops
/**
* full query capabilities - limit, skip, returned fields, sort, etc
*/
Cursor find(...);
void insert(...) // insert one or more objects into the collection, local variants on args
void remove(query) // remove objects that match the query
void modify(selector, modifier) // modify all objects that match selector w/ modifier object
void replace(selector, object) // replace first object that match selector w/ specified
object
void repsert(selector, object) // replace first object that matches, or insert **upsert w/
modifier makes no logical sense*
long getCount();
long getCount(query);
Index Operations
void createIndex( index_info)

void dropIndex(name)
void dropIndexes()
List<info> getIndexInformation()
Misc Operations
document explain(query)
options getOptions();
string getName();
void close();
Cursor Object
document getNextDocument()
iterator getIterator() // again, local to language
bool hasMore()
void close()
Spec, Notes and Suggestions for Mongo Drivers

Assume that the BSON objects returned from the database may be up to 4MB. This size may change over time but for now the limit is 4MB per
object. We recommend you test your driver with 4MB objects.
See Also
Driver Requirements
BSON
The main Database Internals page
Feature Checklist for Mongo Drivers
Functionality Checklist
This section lists tasks the driver author might handle.
Essential
BSON serialization/deserialization
Basic operations: query, save, update, remove, ensureIndex, findOne, limit, sort
Fetch more data from a cursor when necessary (dbGetMore)
Sending of KillCursors operation when use of a cursor has completed (ideally for efficiently these are sent in batches)
Convert all strings to utf8
Authentication
Recommended
automatic doc_id generation (important when using replication)

Database $cmd support and helpers
Detect { $err: ... } response from a db query and handle appropriately --see Error Handling in Mongo Drivers
Automatically connect to proper server, and failover, when connecting to a Replica Set
ensureIndex commands should be cached to prevent excessive communication with the database. (Or, the driver user should be
informed that ensureIndex is not a lightweight operation for the particular driver.)
Support for objects up to 4MB in size
More Recommended
lasterror helper functions

count() helper function
$where clause
eval()
File chunking
hint fields
explain helper
Automatic _id index creation (maybe the db should just do this???)
More Optional
addUser, logout helpers

Allow client user to specify Option_SlaveOk for a query
Tailable cursor support
In/out buffer pooling (if implementing in a garbage collected languages)
More Optional
[connection pooling]
Automatic reconnect on connection failure
DBRef Support:
Ability to generate easily
Automatic traversal
See Also
The Driver and Integration Center for information about the latest drivers
The [top page] for this section
The starting point for all Home
Conventions for Mongo Drivers
Interface Conventions
It is desirable to keep driver interfaces consistent when possible. Of course, idioms vary by language, and when they do adaptation is appropriate.
However, when the idiom is the same, keeping the interfaces consistent across drivers is desirable.
Terminology
In general, use these terms when naming identifiers. Adapt the names to the normal "punctuation" style of your language -- foo_bar in C might
be fooBar in Java.
database - what does this mean?

collection
index
Driver Testing Tools

Object IDs
driverOIDTest for testing toString
> db.runCommand( { "driverOIDTest" : new ObjectId() } )

{
"oid" : ObjectId("4b8991f221752a6e61a88267"),
"str" : "4b8991f221752a6e61a88267",
"ok" : 1
}
Mongo Wire Protocol

Introduction
Messages Types and Formats
Standard Message Header
Request Opcodes
Client Request Messages
OP_UPDATE
OP_INSERT
OP_QUERY
OP_GETMORE
OP_DELETE
OP_KILL_CURSORS
OP_MSG
Database Response Messages
OP_REPLY
Introduction
The Mongo Wire Protocol is a simple socket-based, request-response style protocol. Clients communicate with the database server through a
regular TCP/IP socket.
Default Socket Port

The default port is 27017, but this is configurable and will vary.
Clients should connect to the database with a regular TCP/IP socket. Currently, there is no connection handshake.
To describe the message structure, a C-like struct is used. The types used in this document (cstring, int32, etc.) are the
same as those defined in the BSON specification. The standard message header is typed as MsgHeader. Integer constants are
in capitals (e.g. ZERO for the integer value of 0).
In the case where more than one of something is possible (like in a OP_INSERT or OP_KILL_CURSORS), we again use the
notation from the BSON specification (e.g. int64*). This simply indicates that one or more of the specified type can be written
to the socket, one after another.
Byte Ordering
Note that like BSON documents, all data in the mongo wire protocol is little-endian.
Messages Types and Formats
TableOfContents
There are two types of messages, client requests and database responses, each having a slightly different structure.
Standard Message Header
In general, each message consists of a standard message header followed by request-specific data. The standard message header is structured
as follows :
struct MsgHeader {
int32 messageLength; // total message size, including this
int32 requestID; // identifier for this message
int32 responseTo; // requestID from the original request
// (used in reponses from db)
int32 opCode; // request type - see table below
}
messageLength : This is the total size of the message in bytes. This total includes the 4 bytes that holds the message length.
requestID : This is a client or database-generated identifier that uniquely identifies this message. For the case of client-generated messages
(e.g. CONTRIB:OP_QUERY and CONTRIB:OP_GET_MORE), it will be returned in the
responseTo field of the CONTRIB:OP_REPLY message. Along with the reponseTo field in responses, clients can use this to associate query
responses with the originating query.
responseTo : In the case of a message from the database, this will be the requestID taken from the CONTRIB:OP_QUERY or
CONTRIB:OP_GET_MORE messages from the client. Along with the requestID field in queries, clients can use this to associate query
responses with the originating query.
opCode : Type of message. See the table below in the next section.
Request Opcodes
TableOfContents
The following are the currently supported opcodes :
Opcode Name opCode value Comment
OP_REPLY 1 Reply to a client request. responseTo is set
OP_MSG 1000 generic msg command followed by a string
OP_UPDATE 2001 update document
OP_INSERT 2002 insert new document
RESERVED 2003 formerly used for OP_GET_BY_OID
OP_QUERY 2004 query a collection
OP_GET_MORE 2005 Get more data from a query. See Cursors

OP_DELETE 2006 Delete documents
OP_KILL_CURSORS 2007 Tell database client is done with a cursor
Client Request Messages
TableOfContents
Clients can send all messages except for CONTRIB:OP_REPLY. This is reserved for use by the database.
Note that only the CONTRIB:OP_QUERY and CONTRIB:OP_GET_MORE messages result in a response from the database. There will be no
response sent for any other message.
You can determine if a message was successful with a $$$ TODO get last error command.
OP_UPDATE
The OP_UPDATE message is used to update a document in a collection. The format of a OP_UPDATE message is
struct OP_UPDATE {
MsgHeader header; // standard message header
int32 ZERO; // 0 - reserved for future use
cstring fullCollectionName; // "dbname.collectionname"
int32 flags; // bit vector. see below
document selector; // the query to select the document
document update; // specification of the update to perform
}
fullCollectionName : The full collection name. The full collection name is the concatenation of the database name with the collection name,
using a "." for the concatenation. For example, for the database "foo" and the collection "bar", the full collection name is "foo.bar".
flags :
bit num name description
0 Upsert If set, the database will insert the supplied object into the collection if no matching document is found.
1 MultiUpdate If set, the database will update all matching objects in the collection. Otherwise only updates first matching doc.
2-31 Reserved Must be set to 0.
selector : BSON document that specifies the query for selection of the document to update.
update : BSON document that specifies the update to be performed. For information on specifying updates see the documentation on Updating.
There is no response to an OP_UPDATE message.
OP_INSERT
The OP_INSERT message is used to insert one or more documents into a collection. The format of the OP_INSERT message is
struct {
document* documents; // one or more documents to insert into the collection
}
documents : One or more documents to insert into the collection. If there are more than one, they are written to the socket in sequence, one after
another.
There is no response to an OP_INSERT message.
OP_QUERY
The OP_QUERY message is used to query the database for documents in a collection. The format of the OP_QUERY message is :
struct OP_QUERY {
int32 flags; // bit vector of query options. See below for details.
int32 numberToSkip; // number of documents to skip
int32 numberToReturn; // number of documents to return
// in the first OP_REPLY batch
document query; // query object. See below for details.
[ document returnFieldSelector; ] // Optional. Selector indicating the fields
// to return. See below for details.
}
flags :
bit name description

num
0 Reserved Must be set to 0.
1 TailableCursor Tailable means cursor is not closed when the last data is retrieved. Rather, the cursor marks the final object's
position. You can resume using the cursor later, from where it was located, if more data were received. Like any
"latent cursor", the cursor may become invalid at some point (CursorNotFound) – for example if the final object it
references were deleted.
2 SlaveOk Allow query of replica slave. Normally these return an error except for namespace "local".
3 OplogReplay Internal replication use only - driver should not set
4 NoCursorTimeout The server normally times out idle cursors after an inactivity period (10 minutes) to prevent excess memory use. Set
this option to prevent that.
5 AwaitData Use with TailableCursor. If we are at the end of the data, block for a while rather than returning no data. After a
timeout period, we do return as normal.
6 Exhaust Stream the data down full blast in multiple "more" packages, on the assumption that the client will fully read all data
queried. Faster when you are pulling a lot of data and know you want to pull it all down. Note: the client is not
allowed to not read all the data unless it closes the connection.
numberToSkip : Sets the number of documents to omit - starting from the first document in the resulting dataset - when returning the result of
the query.
numberToReturn : Limits the number of documents in the first CONTRIB:OP_REPLY message to the query. However, the database will still
establish a cursor and return the cursorID to the client if there are more results than numberToReturn. If the client driver offers 'limit'
functionality (like the SQL LIMIT keyword), then it is up to the client driver to ensure that no more than the specified number of document are
returned to the calling application. If numberToReturn is 0, the db will used the default return size. If the number is negative, then the database
will return that number and close the cursor. No futher results for that query can be fetched. If numberToReturn is 1 the server will treat it as -1
(closing the cursor automatically).
query : BSON document that represents the query. The query will contain one or more elements, all of which must match for a document to be
included in the result set. Possible elements include $query, $orderby, $hint, $explain, and $snapshot.
returnFieldsSelector : OPTIONAL BSON document that limits the fields in the returned documents. The returnFieldsSelector contains one
or more elements, each of which is the name of a field that should be returned, and and the integer value 1. In JSON notation, a
returnFieldsSelector to limit to the fields "a", "b" and "c" would be :
{ a : 1, b : 1, c : 1}
The database will respond to an OP_QUERY message with an CONTRIB:OP_REPLY message.
OP_GETMORE
The OP_GETMORE message is used to query the database for documents in a collection. The format of the OP_GETMORE message is :
struct {
int32 numberToReturn; // number of documents to return
int64 cursorID; // cursorID from the OP_REPLY
}
numberToReturn : Limits the number of documents in the first CONTRIB:OP_REPLY message to the query. However, the database will still
establish a cursor and return the cursorID to the client if there are more results than numberToReturn. If the client driver offers 'limit'
functionality (like the SQL LIMIT keyword), then it is up to the client driver to ensure that no more than the specified number of document are
returned to the calling application. If numberToReturn is 0, the db will used the default return size.
cursorID : Cursor identifier that came in the CONTRIB:OP_REPLY. This must be the value that came from the database.
The database will respond to an OP_GETMORE message with an CONTRIB:OP_REPLY message.
OP_DELETE
The OP_DELETE message is used to remove one or more messages from a collection. The format of the OP_DELETE message is :
struct {
int32 flags; // bit vector - see below for details.
document selector; // query object. See below for details.
}
flags :

num
0 SingleRemove If set, the database will remove only the first matching document in the collection. Otherwise all matching documents
will be removed.
selector : BSON document that represent the query used to select the documents to be removed. The selector will contain one or more
elements, all of which must match for a document to be removed from the collection. Please see $$$ TODO QUERY for more information.
There is no reponse to an OP_DELETE message.
OP_KILL_CURSORS
The OP_KILL_CURSORS message is used to close an active cursor in the database. This is necessary to ensure that database resources are
reclaimed at the end of the query. The format of the OP_KILL_CURSORS message is :
struct {
int32 numberOfCursorIDs; // number of cursorIDs in message
int64* cursorIDs; // sequence of cursorIDs to close
}
numberOfCursorIDs : The number of cursors that are in the message.
cursorIDs : "array" of cursor IDs to be closed. If there are more than one, they are written to the socket in sequence, one after another.
Note that if a cursor is read until exhausted (read until OP_QUERY or OP_GETMORE returns zero for the cursor id), there is no need to kill the
cursor.
OP_MSG
Deprecated. OP_MSG sends a diagnostic message to the database. The database sends back a fixed resonse. The format is
struct {
cstring message; // message for the database
}
Drivers do not need to implement OP_MSG.
Database Response Messages
TableOfContents
OP_REPLY
The OP_REPLY message is sent by the database in response to an CONTRIB:OP_QUERY or CONTRIB:OP_GET_MORE

message. The format of an OP_REPLY message is:
struct {
int32 responseFlags; // bit vector - see details below
int64 cursorID; // cursor id if client needs to do get more's
int32 startingFrom; // where in the cursor this reply is starting
int32 numberReturned; // number of documents in the reply
document* documents; // documents
}
responseFlags :

num
0 CursorNotFound Set when getMore is called but the cursor id is not valid at the server. Returned with zero results.
1 QueryFailure Set when query failed. Results consist of one document containing an "$err" field describing the failure.
2 ShardConfigStale Drivers should ignore this. Only mongos will ever see this set, in which case, it needs to update config from the
server.
3 AwaitCapable Set when the server supports the AwaitData Query option. If it doesn't, a client should sleep a little between
getMore's of a Tailable cursor. Mongod version 1.6 supports AwaitData and thus always sets AwaitCapable.
4-31 Reserved Ignore
cursorID : The cursorID that this OP_REPLY is a part of. In the event that the result set of the query fits into one OP_REPLY message,
cursorID will be 0. This cursorID must be used in any CONTRIB:OP_GET_MORE messages used to get more data, and also must be closed
by the client when no longer needed via a CONTRIB:OP_KILL_CURSORS message.
BSON
bsonspec.org
BSON and MongoDB
Language-Specific Examples
C
C++
Java
PHP
Python
Ruby
MongoDB Document Types
bsonspec.org
BSON is a bin-ary-en-coded seri-al-iz-a-tion of JSON-like doc-u-ments. BSON is designed to be lightweight, traversable, and efficient. BSON, like
JSON, supports the embedding of objects and arrays within other objects and arrays. See bsonspec.org for the spec and more information in
general.
BSON and MongoDB
MongoDB uses BSON as the data storage and network transfer format for "documents".
BSON at first seems BLOB-like, but there exists an important difference: the Mongo database understands BSON internals. This means that
MongoDB can "reach inside " BSON objects, even nested ones. Among other things, this allows MongoDB to build indexes and match objects
against query expressions on both top-level and nested BSON keys.
See also: the BSON blog post.
Language-Specific Examples
We often map from a language's "dictionary" type – which may be its native objects – to BSON. The mapping is particularly natural in dynamically
typed languages:
JavaScript: {"foo" : "bar"}

Perl: {"foo" => "bar"}
PHP: array("foo" => "bar")
Python: {"foo" : "bar"}
Ruby: {"foo" => "bar"}
Java: DBObject obj = new BasicDBObject("foo", "bar");
bson b;
bson_buffer buf;
bson_buffer_init( &buf )
bson_append_string( &buf, "name", "Joe" );
bson_append_int( &buf, "age", 33 );
bson_from_buffer( &b, &buf );
bson_print( &b );
See http://github.com/mongodb/mongo-c-driver/blob/master/src/bson.h for more information.
C++

cout << p.toString() << endl;
cout << p["age"].number() << endl;
See the BSON section of the C++ Tutorial for more information.
Java
BasicDBObject doc = new BasicDBObject();

doc.put("name", "MongoDB");
doc.put("type", "database");
doc.put("count", 1);
BasicDBObject info = new BasicDBObject();
info.put("x", 203);
info.put("y", 102);
doc.put("info", info);
coll.insert(doc);
PHP
The PHP driver includes bson_encode and bson_decode functions. bson_encode takes any PHP type and serializes it, returning a string of
bytes:
$bson = bson_encode(null);
$bson = bson_encode(true);
$bson = bson_encode(4);
$bson = bson_encode("hello, world");
$bson = bson_encode(array("foo" => "bar"));
$bson = bson_encode(new MongoDate());
Mongo-specific objects (MongoId, MongoDate, MongoRegex, MongoCode) will be encoded in their respective BSON formats. For other objects,
it will create a BSON representation with the key/value pairs you would get by running for ($object as $key => $value).
bson_decode takes a string representing a BSON object and parses it into an associative array.
Python
>>> from pymongo.bson import BSON

>>> bson_string = BSON.from_dict({"hello": "world"})
>>> bson_string
'\x16\x00\x00\x00\x02hello\x00\x06\x00\x00\x00world\x00\x00'
>>> bson_string.to_dict()
{u'hello': u'world'}
PyMongo also supports "ordered dictionaries" through the pymongo.son module. The BSON class can handle SON instances using the same
methods you would use for regular dictionaries.
Ruby
There are now two gems that handle BSON-encoding: bson and bson_ext. These gems can be used to work with BSON independently of the
MongoDB Ruby driver.
irb
>> require 'rubygems'
=> true
>> require 'bson'
=> true
>> doc = {:hello => "world"}
>> bson = BSON.serialize(doc).to_s
=> "\026\000\000\000\002hello\000\006\000\000\000world\000\000"
>> BSON.deserialize(bson.unpack("C*"))
=> {"hello" => "world"}
The BSON class also supports ordered hashes. Simply construct your documents using the OrderedHash class, also found in the MongoDB Ruby
Driver.
MongoDB Document Types
MongoDB uses BSON documents for three things:
1. Data storage (user documents). These are the regular JSON-like objects that the database stores for us. These BSON documents are
sent to the database via the INSERT operation. User documents have limitations on the "element name" space due to the usage of
special characters in the JSON-like query language.
a. A user document element name cannot begin with "$".
b. A user document element name cannot have a "." in the name.
c. The element name "_id" is reserved for use as a primary key id, but you can store anything that is unique in that field.
The database expects that drivers will prevent users from creating documents that violate these constraints.
2. Query "Selector" Documents : Query documents (or selectors) are BSON documents that are used in QUERY, DELETE and UPDATE
operations. They are used by these operations to match against documents. Selector objects have no limitations on the "element name"
space, as they must be able to supply special "marker" elements, like "$where" and the special "command" operations.
3. "Modifier" Documents : Documents that contain 'modifier actions' that modify user documents in the case of an update (see Updating).
Mongo Extended JSON

Mongo's REST interface supports storage and retrieval of JSON documents. Special representations are used for BSON types that do not have
obvious JSON mappings, and multiple representations are allowed for some such types. The REST interface supports three different modes for
document output { Strict, JS, TenGen }, which serve to control the representations used. Mongo can of course understand all of these
representations in REST input.
Strict mode produces output conforming to the JSON spec http://www.json.org.

JS mode uses some Javascript types to represent certain BSON types.
TenGen mode uses some Javascript types and some 10gen specific types to represent certain BSON types.
The following BSON types are represented using special conventions:
Type Strict JS TenGen Explanation
data_binary <bindata> is
the base64
{ "$binary" { "$binary" : { "$binary" : representation
: "<bindata>", "<bindata>", of a binary
"<bindata>" "$type" : "<t>" } "$type" : "<t>" } string. <t> is
, "$type" : the
"<t>" } hexadecimal
representation
of a single byte
indicating the
data type.
data_date <date> is the

JSON
{ "$date" : Date( <date> ) Date( <date> ) representation
<date> } of a 64 bit
unsigned
integer for
milliseconds
since epoch.
data_regex <sRegex> is a
string of valid
{ "$regex" /<jRegex>/<jOptions> /<jRegex>/<jOptions> JSON
: characters.
"<sRegex>", <jRegex> is a
"$options" : string that may
"<sOptions>" contain valid
} JSON
characters and
unescaped '"'
characters, but
may not
contain
unescaped '/'
characters.
<sOptions> is
a string
containing
letters of the
alphabet.
<jOptions> is a
string that may
contain only
the characters
'g', 'i', and 'm'.
Because the
JS and
TenGen
representations
support a
limited range of
options, any
nonconforming
options will be
dropped when
converting to
this
representation.
data_oid <id> is a 24
character
{ "$oid" : { "$oid" : "<id>" } ObjectId( "<id>" ) hexadecimal
"<id>" } string. Note
that these
representations
require a
data_oid value
to have an
associated
field name
"_id".
data_ref <name> is a
string of valid
{ "$ref" : { "$ref" : "<name>" Dbref( "<name>", JSON
"<name>", , "$id" : "<id>" } "<id>" ) characters.
"$id" : <id> is a 24
"<id>" } character
hexadecimal
string.
GridFS Specification
Introduction
Specification
Storage Collections
files
chunks
Indexes
Introduction
GridFS is a storage specification for large objects in MongoDB. It works by splitting large object into small chunks, usually 256k in size. Each
chunk is stored as a separate document in a chunks collection. Metadata about the file, including the filename, content type, and any optional
information needed by the developer, is stored as a document in a files collection.
So for any given file stored using GridFS, there will exist one document in files collection and one or more documents in the chunks collection.
If you're just interested in using GridFS, see the docs on storing files. If you'd like to understand the GridFS implementation, read on.
Specification
Storage Collections
GridFS uses two collections to store data:
files contains the object metadata

chunks contains the binary chunks with some additional accounting information
In order to make more than one GridFS namespace possible for a single database, the files and chunks collections are named with a prefix. By
default the prefix is fs., so any default GridFS store will consist of collections named fs.files and fs.chunks. The drivers make it possible to
change this prefix, so you might, for instance, have another GridFS namespace specifically for photos where the collections would be
photos.files and photos.chunks.
Here's an example of the standard GridFS interface in Java:

/*
* default root collection usage - must be supported
*/
GridFS myFS = new GridFS(myDatabase); // returns a default GridFS (e.g. "fs" root
collection)
myFS.storeFile(new File("/tmp/largething.mpg")); // saves the file into the "fs" GridFS store
/*
* specified root collection usage - optional
*/
GridFS myContracts = new GridFS(myDatabase, "contracts"); // returns a GridFS where

"contracts" is root
myFS.retrieveFile("smithco", new File("/tmp/smithco_20090105.pdf")); // retrieves object whose
filename is "smithco"
Note that the above API is for demonstration purposes only - this spec does not (at this time) recommend any API. See individual driver
documentation for API specifics.
files
Documents in the files collection require the following fields:
{
"_id" : <unspecified>, // unique ID for this file
"length" : data_number, // size of the file in bytes
"chunkSize" : data_number, // size of each of the chunks. Default is 256k
"uploadDate" : data_date, // date when object first stored
"md5" : data_string // result of running the "filemd5" command on this file's
chunks
}
Any other desired fields may be added to the files document; common ones include the following:
{
"filename" : data_string, // human name for the file
"contentType" : data_string, // valid mime type for the object
"aliases" : data_array of data_string, // optional array of alias strings
"metadata" : data_object, // anything the user wants to store
}
Note that the _id field can be of any type, per the discretion of the spec implementor.
chunks
The structure of documents from the chunks collection is as follows:
{
"_id" : <unspecified>, // object id of the chunk in the _chunks collection
"files_id" : <unspecified>, // _id of the corresponding files collection entry
"n" : chunk_number, // chunks are numbered in order, starting with 0
"data" : data_binary, // the chunk's payload as a BSON binary type
}
Notes:
The _id is whatever type you choose. As with any MongoDB document, the default will be a BSON object id.
The files_id is a foreign key containing the _id field for the relevant files collection entry
Indexes
GridFS implementations should create a unique, compound index in the chunks collection for files_id and n. Here's how you'd do that from
the shell:
db.fs.chunks.ensureIndex({files_id:1, n:1}, {unique: true});
This way, a chunk can be retrieved efficiently using it's files_id and n values. Note that GridFS implementations should use findOne
operations to get chunks individually, and should not leave open a cursor to query for all chunks. So to get the first chunk, we could do:
db.fs.chunks.findOne({files_id: myFileID, n: 0});
Implementing Authentication in a Driver

The current version of Mongo supports only very basic authentication. One authenticates a username and password in the context of a particular
database. Once authenticated, the user has full read and write access to the database in question.
The admin database is special. In addition to several commands that are administrative being possible only on admin, authentication on admin
gives one read and write access to all databases on the server. Effectively, admin access means root access to the db.
Note on a single socket we may authenticate for any number of databases, and as different users. This authentication persists for the life of the
database connection (barring a logout command).
The Authentication Process
Authentication is a two step process. First the driver runs a getnonce command to get a nonce for use in the subsequent authentication. We
can view a sample getnonce invocation from dbshell:
> db.$cmd.findOne({getnonce:1})
{ "nonce":"7268c504683936e1" , "ok":1
The nonce returned is a hex String.
The next step is to run an authenticate command for the database on which to authenticate. The authenticate command has the form:
{ authenticate : 1, user : username, nonce : nonce, key : digest }
where
username is a username in the database's system.users collection;

nonce is the nonce returned from a previous getnonce command;
digest is the hex encoding of a MD5 message digest which is the MD5 hash of the concatenation of (nonce, username, password_digest
), where password_digest is the user's password value in the pwd field of the associated user object in the database's system.users
collection. pwd is the hex encoding of MD5( username + ":mongo:" + password_text).
Authenticate will return an object containing
{ ok : 1 }
when successful.
Details of why an authentication command failed may be found in the Mongo server's log files.
The following code from the Mongo Javascript driver provides an example implementation:
DB.prototype.addUser = function( username , pass ){
var c = this.getCollection( "system.users" );
var u = c.findOne( { user : username } ) || { user : username };

u.pwd = hex_md5( username + ":mongo:" + pass );
print( tojson( u ) );
c.save( u );
}
DB.prototype.auth = function( username , pass ){

var n = this.runCommand( { getnonce : 1 } );
var a = this.runCommand(
{
authenticate : 1 ,
user : username ,
nonce : n.nonce ,
key : hex_md5( n.nonce + username + hex_md5( username + ":mongo:" + pass ) )
}
);
return a.ok;
}
Logout
Drivers may optionally implement the logout command which deauthorizes usage for the specified database for this connection. Note other
databases may still be authorized.
Alternatively, close the socket to deauthorize.
> db.$cmd.findOne({logout:1})
{
"ok" : 1.0
}
Replica Pairs and Authentication
For drivers that support replica pairs, extra care with replication is required.
When switching from one server in a pair to another (on a failover situation), you must reauthenticate. Clients will likely want to cache
authentication from the user so that the client can reauthenticate with the new server when appropriate.
Be careful also with operations such as Logout - if you log out from only half a pair, that could be an issue.
Authenticating with a server in slave mode is allowed.
See Also
Security and Authentication
Notes on Pooling for Mongo Drivers

Note that with the db write operations can be sent asynchronously or synchronously (the latter indicating a getlasterror request after the write).
When asynchronous, one must be careful to continue using the same connection (socket). This ensures that the next operation will not begin until
after the write completes.
Pooling and Authentication
An individual socket connection to the database has associated authentication state. Thus, if you pool connections, you probably want a separate
pool for each authentication case (db + username).
Pseudo-code
The following pseudo-code illustrates our recommended approach to implementing connection pooling in a driver's connection class. This handles
authentication, grouping operations from a single "request" onto the same socket, and a couple of other gotchas:
class Connection:
init(pool_size, addresses, auto_start_requests):
this.pool_size = pool_size
this.addresses = addresses
this.auto_start_requests = auto_start_requests
this.thread_map = {}
this.locks = Lock[pool_size]
this.sockets = Socket[pool_size]
this.socket_auth = String[pool_size][]
this.auth = {}
this.find_master()
find_master():
for address in this.addresses:
if address.is_master():
this.master = address
pick_and_acquire_socket():
choices = random permutation of [0, ..., this.pool_size - 1]
choices.sort(order: ascending,
value: size of preimage of choice under this.thread_map)
for choice in choices:

if this.locks[choice].non_blocking_acquire():
return choice
sock = choices[0]
this.locks[sock].blocking_acquire()
return sock
get_socket():
if this.thread_map[current_thread] >= 0:
sock_number = this.thread_map[current_thread]
this.locks[sock_number].blocking_acquire()
else:
sock_number = this.pick_and_lock_socket()
if this.auto_start_requests or current_thread in this.thread_map:
this.thread_map[current_thread] = sock_number
if not this.sockets[sock_number]:
this.sockets[sock_number] = Socket(this.master)
return sock_number
send_message_without_response(message):
sock_number = this.get_socket()
this.check_auth()
this.sockets[sock_number].send(message)
this.locks[sock_number].release()
send_message_with_response(message):
sock_number = this.get_socket()
this.check_auth()
this.sockets[sock_number].send(message)
result = this.sockets[sock_number].receive()
this.locks[sock_number].release()
return result
# start_request is only needed if auto_start_requests is False

start_request():
this.thread_map[current_thread] = -1
end_request():
delete this.thread_map[current_thread]
authenticate(database, username, password):

# TODO should probably make sure that these credentials are valid,
# otherwise errors are going to be delayed until first op.
this.auth[database] = (username, password)
logout(database):
delete this.auth[database]
check_auth(sock_number):
for db in this.socket_auth[sock_number]:
if db not in this.auth.keys():
this.sockets[sock_number].send(logout_message)
this.socket_auth[sock_number].remove(db)
for db in this.auth.keys():
if db not in this.socket_auth[sock_number]:
this.sockets[sock_number].send(authenticate_message)
this.socket_auth[sock_number].append(db)
# somewhere we need to do error checking - if you get not master then everything
# in this.sockets gets closed and set to null and we call find_master() again.
# we also need to reset the socket_auth information - nothing is authorized yet
# on the new master.
See Also
Driver and Integration Center
Connecting Drivers to Replica Sets

Ideally a MongoDB driver can connect to a cluster of servers which represent a replica set, and automatically find the right set member with which
to communicate. Failover should be automatic too. The general steps are:
1. The user, when opening the connection, specifies host[:port] for one or more members of the set. Not all members need be specified --
in fact the exact members of the set might change over time. This list for the connect call is the seed list.
2. The driver then connects to all servers on the seed list, perhaps in parallel to minimize connect time. Send an ismaster command to each
server.
3. When the server is in replSet mode, it will return a hosts field with all members of the set that are potentially eligible to serve data. The
client should cache this information. Ideally this refreshes too, as the set's config could change over time.
4. Choose a server with which to communicate.
a. If ismaster == true, that server is primary for the set. This server can be used for writes and immediately consistent reads.
b. If secondary == true, that server is not primary, but is available for eventually consistent reads. In this case, you can use the
primary field to see which server the master should be.
5. If an error occurs with the current connection, find the new primary and resume use there.
For example, if we run the ismaster command on a non-primary server, we might get something like:
> db.runCommand("ismaster")
{
"ismaster" : false,
"secondary" : true,
"hosts" : [
"ny1.acme.com",
"ny2.acme.com",
"sf1.acme.com"
],
"passives" : [
"ny3.acme.com",
"sf3.acme.com"
],
"arbiters" : [
"sf2.acme.com",
]
"primary" : "ny2.acme.com",
"ok" : true
}
There are three servers with priority > 0 (ny1, ny2, and sf1), two passive servers (ny3 and sf3), and an arbiter (sf2). The primary should be ny2,
but the driver should call ismaster on that server before it assumes it is.
Error Handling in Mongo Drivers

If an error occurs on a query (or getMore operation), Mongo returns an error object instead of user data.
The error object has a first field guaranteed to have the reserved key $err. For example:
{ $err : "some error message" }
The $err value can be of any type but is usually a string.

Drivers typically check for this return code explicitly and take action rather than returning the object to the user. The query results flags include a
set bit when $err is returned.
/* db response format
Query or GetMore: // see struct QueryResult

int resultFlags;
int64 cursorID;
int startingFrom;
int nReturned;
list of marshalled JSObjects;
*/
struct QueryResult : public MsgData {

enum {
ResultFlag_CursorNotFound = 1, /* returned, with zero results, when getMore is called but the
cursor id is not valid at the server. */
ResultFlag_ErrSet = 2 /* { $err : ... } is being returned */
};
...
};
See Also
Developer Zone
Tutorial
Shell
Manual
Databases
Collections
Indexes
Data Types and Conventions
GridFS
Inserting
Updating
Querying
Removing
Optimization
Developer FAQ
If you have a comment or question about anything, please contact us through IRC (freenode.net#mongodb) or the mailing list, rather than leaving
a comment at the bottom of a page. It is easier for us to respond to you through those channels.
Introduction
MongoDB is a collection-oriented, schema-free document database.
By collection-oriented, we mean that data is grouped into sets that are called 'collections'. Each collection has a unique name in the database,
and can contain an unlimited number of documents. Collections are analogous to tables in a RDBMS, except that they don't have any defined
schema.
By schema-free, we mean that the database doesn't need to know anything about the structure of the documents that you store in a collection. In
fact, you can store documents with different structure in the same collection if you so choose.
By document, we mean that we store data that is a structured collection of key-value pairs, where keys are strings, and values are any of a rich
set of data types, including arrays and documents. We call this data format "BSON" for "Binary Serialized dOcument Notation."
MongoDB Operational Overview

MongoDB is a server process that runs on Linux, Windows and OS X. It can be run both as a 32 or 64-bit application. We recommend running in
64-bit mode, since Mongo is limited to a total data size of about 2GB for all databases in 32-bit mode.
The MongoDB process listens on port 27017 by default (note that this can be set at start time - please see Command Line Parameters for more
information).
Clients connect to the MongoDB process, optionally authenticate themselves if security is turned on, and perform a sequence of actions, such as
inserts, queries and updates.
MongoDB stores its data in files (default location is /data/db/), and uses memory mapped files for data management for efficiency.
MongoDB can also be configured for automatic data replication , as well as automatic fail-over .
For more information on MongoDB administration, please see Mongo Administration Guide.
MongoDB Functionality
As a developer, MongoDB drivers offer a rich range of operations:
Queries: Search for documents based on either query objects or SQL-like "where predicates". Queries can be sorted, have limited return
sizes, can skip parts of the return document set, and can also return partial documents.
Inserts and Updates : Insert new documents, update existing documents.
Index Management : Create indexes on one or more keys in a document, including substructure, deleted indexes, etc
General commands : Any MongoDB operation can be managed via DB Commands over the regular socket.
Tutorial
Getting the Database
Getting A Database Connection
Inserting Data into A Collection
Accessing Data From a Query
Specifying What the Query Returns
findOne() - Syntactic Sugar
Limiting the Result Set via limit()
More Help
What Next
Getting the Database
First, run through the Quickstart guide for your platform to get up and running.
Getting A Database Connection
Let's now try manipulating the database with the database shell . (We could perform similar operations from any programming language using an
appropriate driver. The shell is convenient for interactive and administrative use.)
Start the MongoDB JavaScript shell with:
$ bin/mongo
By default the shell connects to database "test" on localhost. You then see:
MongoDB shell version: <whatever>

url: test
connecting to: test
type "help" for help
>
"connecting to:" tells you the name of the database the shell is using. To switch databases, type:
> use mydb

switched to db mydb
To see a list of handy commands, type help.

Tip for Developers with Experience in Other Databases
You may notice, in the examples below, that we never create a database or collection. MongoDB does not require that you do
so. As soon as you insert something, MongoDB creates the underlying collection and database. If you query a collection that
does not exist, MongoDB treats it as an empty collection.
Switching to a database with the use command won't immediately create the database - the database is created lazily the first
time data is inserted. This means that if you use a database for the first time it won't show up in the list provided by ` show dbs`
until data is inserted.
Inserting Data into A Collection
Let's create a test collection and insert some data into it. We will create two objects, j and t, and then save them in the collection things.
In the following examples, '>' indicates commands typed at the shell prompt.
> j = { name : "mongo" };

{"name" : "mongo"}
> t = { x : 3 };
{ "x" : 3 }
> db.things.save(j);
> db.things.save(t);
> db.things.find();
{ "_id" : ObjectId("4c2209f9f3924d31102bd84a"), "name" : "mongo" }
{ "_id" : ObjectId("4c2209fef3924d31102bd84b"), "x" : 3 }
>
A few things to note :
We did not predefine the collection. The database creates it automatically on the first insert.
The documents we store can have any "structure" - in fact in this example, the documents have no common data elements at all. In
practice, one usually stores documents of the same structure within collections. However, this flexibility means that schema migration and
augmentation are very easy in practice - rarely will you need to write scripts which perform "alter table" type operations.
Upon being inserted into the database, objects are assigned an object ID (if they do not already have one) in the field _id.
When you run the above example, your ObjectID values will be different.
Let's add some more records to this collection:
> for (var i = 1; i <= 20; i++) db.things.save({x : 4, j : i});

> db.things.find();
{ "_id" : ObjectId("4c220a42f3924d31102bd856"), "x" : 4, "j" : 1 }
{ "_id" : ObjectId("4c220a42f3924d31102bd85a"), "x" : 4, "j" : 5 }
{ "_id" : ObjectId("4c220a42f3924d31102bd85b"), "x" : 4, "j" : 6 }
{ "_id" : ObjectId("4c220a42f3924d31102bd85c"), "x" : 4, "j" : 7 }
{ "_id" : ObjectId("4c220a42f3924d31102bd85d"), "x" : 4, "j" : 8 }
{ "_id" : ObjectId("4c220a42f3924d31102bd85e"), "x" : 4, "j" : 9 }
{ "_id" : ObjectId("4c220a42f3924d31102bd85f"), "x" : 4, "j" : 10 }
has more
Note that not all documents were shown - the shell limits the number to 20 when automatically iterating a cursor. Since we already had 2
documents in the collection, we only see the first 18 of the newly-inserted documents.
If we want to return the next set of results, there's the it shortcut. Continuing from the code above:

has more
> it
Technically, find() returns a cursor object. But in the cases above, we haven't assigned that cursor to a variable. So, the shell automatically
iterates over the cursor, giving us an initial result set, and allowing us to continue iterating with the it command.
But we can also work with the cursor directly; just how that's done is discussed in the next section.
Accessing Data From a Query
Before we discuss queries in any depth, lets talk about how to work with the results of a query - a cursor object. We'll use the simple find()
query method, which returns everything in a collection, and talk about how to create specific queries later on.
In order to see all the elements in the collection when using the mongo shell, we need to explicitly use the cursor returned from the find()
operation.
Lets repeat the same query, but this time use the cursor that find() returns, and iterate over it in a while loop :
> var cursor = db.things.find();

> while (cursor.hasNext()) printjson(cursor.next());
The above example shows cursor-style iteration. The hasNext() function tells if there are any more documents to return, and the next()
function returns the next document. We also used the built-in tojson() method to render the document in a pretty JSON-style format.
When working in the JavaScript shell, we can also use the functional features of the language, and just call forEach on the cursor. Repeating
the example above, but using forEach() directly on the cursor rather than the while loop:
> db.things.find().forEach(printjson);
In the case of a forEach() we must define a function that is called for each document in the cursor.
In the mongo shell, you can also treat cursors like an array :
> var cursor = db.things.find();

> printjson(cursor[4]);
When using a cursor this way, note that all values up to the highest accessed (cursor[4] above) are loaded into RAM at the same time. This is
inappropriate for large result sets, as you will run out of memory. Cursors should be used as an iterator with any query which returns a large
number of elements.
In addition to array-style access to a cursor, you may also convert the cursor to a true array:
> var arr = db.things.find().toArray();

> arr[5];
Please note that these array features are specific to mongo - The Interactive Shell, and not offered by all drivers.
MongoDB cursors are not snapshots - operations performed by you or other users on the collection being queried between the first and last call to
next() of your cursor may or may not be returned by the cursor. Use explicit locking to perform a snapshotted query.
Specifying What the Query Returns
Now that we know how to work with the cursor objects that are returned from queries, lets now focus on how to tailor queries to return specific
things.
In general, the way to do this is to create "query documents", which are documents that indicate the pattern of keys and values that are to be
matched.
These are easier to demonstrate than explain. In the following examples, we'll give example SQL queries, and demonstrate how to represent the
same query using MongoDB via the mongo shell. This way of specifying queries is fundamental to MongoDB, so you'll find the same general
facility in any driver or language.
SELECT * FROM things WHERE name="mongo"
> db.things.find({name:"mongo"}).forEach(printjson);
SELECT * FROM things WHERE x=4
> db.things.find({x:4}).forEach(printjson);
The query expression is an document itself. A query document of the form { a:A, b:B, ... } means "where a==A and b==B and ...". More
information on query capabilities may be found in the Queries and Cursors section of the Mongo Developers' Guide.
MongoDB also lets you return "partial documents" - documents that have only a subset of the elements of the document stored in the database.
To do this, you add a second argument to the find() query, supplying a document that lists the elements to be returned.
To illustrate, lets repeat the last example find({x:4}) with an additional argument that limits the returned document to just the "j" elements:
SELECT j FROM things WHERE x=4
> db.things.find({x:4}, {j:true}).forEach(printjson);

{ "_id" : ObjectId("4c220a42f3924d31102bd856"), "j" : 1 }
{ "_id" : ObjectId("4c220a42f3924d31102bd85a"), "j" : 5 }
{ "_id" : ObjectId("4c220a42f3924d31102bd85b"), "j" : 6 }
{ "_id" : ObjectId("4c220a42f3924d31102bd85c"), "j" : 7 }
{ "_id" : ObjectId("4c220a42f3924d31102bd85d"), "j" : 8 }
{ "_id" : ObjectId("4c220a42f3924d31102bd85e"), "j" : 9 }
{ "_id" : ObjectId("4c220a42f3924d31102bd85f"), "j" : 10 }
Note that the "_id" field is always returned.
findOne() - Syntactic Sugar
For convenience, the mongo shell (and other drivers) lets you avoid the programming overhead of dealing with the cursor, and just lets you
retrieve one document via the findOne() function. findOne() takes all the same parameters of the find() function, but instead of returning a
cursor, it will return either the first document returned from the database, or null if no document is found that matches the specified query.
As an example, lets retrieve the one document with name=='mongo'. There are many ways to do it, including just calling next() on the cursor
(after checking for null, of course), or treating the cursor as an array and accessing the 0th element.
However, the findOne() method is both convenient and efficient:
> printjson(db.things.findOne({name:"mongo"}));
This is more efficient because the client requests a single object from the database, so less work is done by the database and the network. This is
the equivalent of find({name:"mongo"}).limit(1).
Limiting the Result Set via limit()
You may limit the size of a query's result set by specifing a maximum number of results to be returned via the limit() method.
This is highly recommended for performance reasons, as it limits the work the database does, and limits the amount of data returned over the
network. For example:
> db.things.find().limit(3);
More Help
In addition to the general "help" command, you can call help on db and db.whatever to see a summary of methods available.
If you are curious about what a function is doing, you can type it without the {{()}}s and the shell will print the source, for example:
> printjson
function (x) {
print(tojson(x));
}
mongo is a full JavaScript shell, so any JavaScript function, syntax, or class can be used in the shell. In addition, MongoDB defines some of its
own classes and globals (e.g., db). You can see the full API at http://api.mongodb.org/js/.
What Next
After completing this tutorial the next step to learning MongoDB is to dive into the manual for more details.
Manual
This is the MongoDB manual. Except where otherwise noted, all examples are in JavaScript for use with the mongo shell. There is a table
available giving the equivalent syntax for each of the drivers.
Connections
Databases
Commands
Clone Database
fsync Command
Index-Related Commands
Last Error Commands
Windows Service
Viewing and Terminating Current Operation
Validate Command
getLastError
List of Database Commands
Mongo Metadata
Collections
Capped Collections
Using a Large Number of Collections
Internationalized Strings
Object IDs
Database References
GridFS
Indexes
Using Multikeys to Simulate a Large Number of Indexes
Geospatial Indexing
Indexing as a Background Operation
Multikeys
Indexing Advice and FAQ
Inserting
Legal Key Names
Schema Design
Trees in MongoDB
Optimization
Optimizing Storage of Small Objects
Query Optimizer
Querying
Retrieving a Subset of Fields
Advanced Queries
Dot Notation (Reaching into Objects)
Full Text Search in Mongo
min and max Query Specifiers
OR operations in query expressions
Queries and Cursors
Tailable Cursors
Server-side Code Execution
Sorting and Natural Order
Aggregation
Removing
Updating
Atomic Operations
findandmodify Command
Updating Data in Mongo
MapReduce
Data Processing Manual
Connections
MongoDB is a database server: it runs in the foreground or background and waits for connections from the user. Thus, when you start MongoDB,
you will see something like:
~/$ ./mongod
#
# some logging output
#
Tue Mar 9 11:15:43 waiting for connections on port 27017
Tue Mar 9 11:15:43 web admin interface listening on port 28017
It will stop printing output at this point but it hasn't frozen, it is merely waiting for connections on port 27017. Once you connect and start sending
commands, it will continue to log what it's doing. You can use any of the MongoDB drivers or Mongo shell to connect to the database.
You cannot connect to MongoDB by going to http://localhost:27017 in your web browser. The database cannot be accessed via HTTP on port
27017.
Standard Connection String Format
The uri scheme described on this page is not yet supported by all of the drivers. Refer to a specific driver's documentation to
see how much (if any) of the standard connection uri is supported. All drivers support an alternative method of specifying
connections if this format is not supported.
mongodb://[username:password@]host1[:port1][,host2[:port2],...[,hostN[:portN]]][/database]
mongodb:// is a required prefix to identify that this is a string in the standard connection format.
username:password@ are optional. If given, the driver will attempt to login to a database after connecting to a database server.
host1 is the only required part of the URI. It identifies a server address to connect to.
:portX is optional and defaults to :27017 if not provided.
/database is the name of the database to login to and thus is only relevant if the username:password@ syntax is used. If not
specified the "admin" database will be used by default.
As many hosts as necessary may be specified (for connecting to replica pairs/sets).
Examples
Connect to a database server running locally on the default port:
mongodb://localhost
Connect and login to the admin database as user "fred" with password "foobar":
mongodb://fred:foobar@localhost
Connect and login to the "baz" database as user "fred" with password "foobar":
mongodb://fred:foobar@localhost/baz
Connect to a replica pair, with one server on example1.com and another server on example2.com:
mongodb://example1.com:27017,example2.com:27017
Connect to a replica set with three servers running on localhost (on ports 27017, 27018, and 27019):
mongodb://localhost,localhost:27018,localhost:27019
Connection Pooling
The server will use one thread per TCP connection, therefore it is highly recomended that your application use some sort of connection pooling.
Luckily, most drivers handle this for you behind the scenes. One notable exception is setups where your app spawns a new process for each
request, such as CGI and some configurations of PHP.
Databases
Each MongoDB server can support multiple databases. Each database is independent, and the data for each database is stored separately, for
security and ease of management.
A database consists of one or more collections, the documents (objects) in those collections, and an optional set of security credentials for
controlling access.
Commands
Clone Database
fsync Command
Last Error Commands
Windows Service
Validate Command
getLastError
Mongo Metadata
Commands
Introduction
The Mongo database has a concept of a database command. Database commands are ways to ask the database to perform special operations,
or to request information about its current operational status.
Introduction
Privileged Commands
Getting Help Info for a Command
More Command Documentation
A command is sent to the database as a query to a special collection namespace called $cmd. The database will return a single document with
the command results - use findOne() for that if your driver has it.
The general command syntax is:
db.$cmd.findOne( { <commandname>: <value> [, options] } );
The shell provides a helper function for this:
db.runCommand( { <commandname>: <value> [, options] } );
For example, to check our database's current profile level setting, we can invoke:
> db.runCommand({profile:-1});
{
"was" : 0.0 ,
"ok" : 1.0
}
For many db commands, some drivers implement wrapper methods are implemented to make usage easier. For example, the mongo shell offers
> db.getProfilingLevel()
0.0
Let's look at what this method is doing:
> print( db.getProfilingLevel )

function () {
var res = this._dbCommand({profile:-1});
return res ? res.was : null;
}
> print( db._dbCommand )

function (cmdObj) {
return this.$cmd.findOne(cmdObj);
}
Many commands have helper functions - see your driver's documentation for more information.
Privileged Commands
Certain operations are for the database administrator only. These privileged operations may only be performed on the special database named
admin.
> use admin;

> db.runCommand("shutdown"); // shut down the database
If the db variable is not set to 'admin', you can use _adminCommand to switch to the right database automatically (and just for that operation):
> db._adminCommand("shutdown");
(For this particular command there is also a shell helper function, db.shutdownServer.)
Getting Help Info for a Command
Use commandHelp in shell to get help info for a command:

> db.commandHelp("datasize")
help for: datasize example: { datasize:"blog.posts", keyPattern:{x:1}, min:{x:10}, max:{x:55} }
NOTE: This command may take awhile to run
(Help is not yet available for some commands.)
More Command Documentation
Clone Database
fsync Command
Last Error Commands
Windows Service
Validate Command
getLastError
Commands Quick Reference Card
Clone Database
MongoDB includes commands for copying a database from one server to another.
// copy an entire database from one name on one server to another

// name on another server. omit <from_hostname> to copy from one
// name to another on the same server.
db.copyDatabase(<from_dbname>, <to_dbname>, <from_hostname>);
// if you must authenticate with the source database
db.copyDatabase(<from_dbname>, <to_dbname>, <from_hostname>, <username>, <password>);
// in "command" syntax (runnable from any driver):
db.runCommand( { copydb : 1, fromdb : ..., todb : ..., fromhost : ... } );
// command syntax for authenticating with the source:
n = db.runCommand( { copydbgetnonce : 1, fromhost: ... } );
db.runCommand( { copydb : 1, fromhost: ..., fromdb: ..., todb: ..., username: ..., nonce: n.nonce,
key: <hash of username, nonce, password > } );
// clone the current database (implied by 'db') from another host

var fromhost = ...;
print("about to get a copy of database " + db + " from " + fromhost);
db.cloneDatabase(fromhost);
// in "command" syntax (runnable from any driver):
db.runCommand( { clone : fromhost } );
fsync Command
fsync Command
Lock, Snapshot and Unlock
Caveats
Snapshotting Slaves
See Also
Version 1.3.1 and higher
The fsync command allows us to flush all pending writes to datafiles. More importantly, it also provides a lock option that makes backups easier.
fsync Command
The fsync command forces the database to flush all datafiles:
> use admin

> db.runCommand({fsync:1});
By default the command returns after synchronizing. To return immediately use:
> db.runCommand({fsync:1,async:true});
To fsync on a regular basis, use the --syncdelay command line option (see mongod --help output). By default a full flush is forced every 60
seconds.
Lock, Snapshot and Unlock
The fsync command supports a lock option that allows one to safely snapshot the database's datafiles. While locked, all write operations are
blocked, although read operations are still allowed. After snapshotting, use the unlock command to unlock the database and allow locks again.
Example:
> use admin

switched to db admin
> db.runCommand({fsync:1,lock:1})
{
"info" : "now locked against writes",
"ok" : 1
}
> db.currentOp()
{
"inprog" : [
],
"fsyncLock" : 1
}
>// do some work here: for example, snapshot datafiles...

>// runProgram("/path/to/my-filesystem-snapshotting-script.sh")
> db.$cmd.sys.unlock.findOne();
{ "ok" : 1, "info" : "unlock requested" }
> // unlock is now requested. it may take a moment to take effect.
> db.currentOp()
{ "inprog" : [ ] }
Caveats
While the database can be read while locked for snapshotting, if a write is attempted, this will block readers due to the database's use of a
read/write lock. This should be addressed in the future : http://jira.mongodb.org/browse/SERVER-1423
Snapshotting Slaves
The above procedure works on replica slaves. The slave will not apply operations while locked. However, see the above caveats section.
See Also
Backups
Create Index
ensureIndex() is the helper function for this. Its implementation creates an index by adding its info to the system.indexes table.
> db.myCollection.ensureIndex(<keypattern>);
> // same as:
> db.system.indexes.insert({ name: "name", ns: "namespaceToIndex",
key: <keypattern> });
Note: Once you've inserted the index, all subsequent document inserts for the given collection will be indexed, as will all pre-existing documents
in the collection. If you have a large collection, this can take a significant amount of time and will block other operations. However, beginning with
version 1.3.2, you can specify that indexing happen in the background. See the background indexing docs for details.
You can query system.indexes to see all indexes for a table foo:
>db.system.indexes.find( { ns: "foo" } );
In some drivers, ensureIndex() remembers if it has recently been called, and foregoes the insert operation in that case. Even if this is not the
case, ensureIndex() is a cheap operation, so it may be invoked often to ensure that an index exists.
Dropping an Index
From the shell:
db.mycollection.dropIndex(<name_or_pattern>)
db.mycollection.dropIndexes()
// example:
t.dropIndex( { name : 1 } );
From a driver (raw command object form; many drivers have helpers):
{ deleteIndexes: <collection_name>, index: <index_name> }

// "*" for <index_name> will drop all indexes except _id
Index Namespace
Each index has a namespace of its own for the btree buckets. The namespace is:
<collectionnamespace>.$<indexname>
This is an internal namespace that cannot be queried directly.
Last Error Commands
Since MongoDB doesn't wait for a response by default when writing to the database, a couple commands exist for ensuring that these operations
have succeeded. These commands can be invoked automatically with many of the drivers when saving and updating in "safe" mode. But what's
really happening is that a special command called getlasterror is being invoked. Here, we explain how this works.
getlasterror
Drivers
Use Cases
Mongo Shell Behavior
fsync option
With Replication
getPrevError
getlasterror
The getlasterror command checks for an error on the last database operation for this connection. Since it's a command, there are a few ways
to invoke it:
> db.$cmd.findOne({getlasterror:1})
Or
> db.runCommand("getlasterror")
Or you can use the helper:
> db.getLastError()
For more about commands, see the command documentation.

Drivers
The drivers support getlasterror in the command form and many also offer a "safe" mode for operations. If you're using Python, for example,
you automatically call getlasterror on insert as follows:
collection.save({"name": "MongoDB"}, safe=True)
If the save doesn't succeed, an exception will be raised. For more on "safe" mode, see your driver's documentation.
Use Cases
getlasterror is primarily useful for write operations (although it is set after a command or query too). Write operations by default do not have a
return code: this saves the client from waiting for client/server turnarounds during write operations. One can always call getLastError if one wants
a return code.
If you're writing data to MongoDB on multiple connections, then it can sometimes be important to call getlasterror on one connection to be
certain that the data has been committed to the database. For instance, if you're writing to connection #1 and want those writes to be reflected in
reads from connection #2, you can assure this by calling getlasterror after writing to connection #1.
Note: The special mongo wire protocol killCursors operation does not support getlasterror. (This is really only of significant to driver developers .)
Mongo Shell Behavior
The database shell performs a resetError() before each read/eval/print loop command evaluation - and automatically prints the error, if one
occurred, after each evaluation. Thus, after an error, at the shell prompt db.getLastError() will return null. However, if called before returning
to the prompt, the result is as one would expect:
> try { db.foo.findOne() } catch(e) { print("preverr:" + tojson(db.getPrevError())); print("lasterr:"

+ tojson(db.getLastError()));}
preverr:{"err" : "unauthorized" , "nPrev" : 1 , "ok" : 1}
lasterr:"unauthorized"
fsync option
Include the fsync option to force the database to fsync all files before returning (v1.3+):
> db.runCommand({getlasterror:1,fsync:true})
{ "err" : null, "n" : 0, "fsyncFiles" : 2, "ok" : 1 }
With Replication
See blocking for replication.
getPrevError
Note: getPrevError may be deprecated in the future.
When performing bulk write operations, resetError() and getPrevError() can be an efficient way to check for success of the operation.
For example if we are inserting 1,000 objects in a collection, checking the return code 1,000 times over the network is slow. Instead one might do
something like this:
db.resetError();
for( loop 1000 times... )
db.foo.save(something...);
if( db.getPrevError().err )
print("didn't work!");
Windows Service
On windows mongod.exe has native support for installing and running as a windows service.
Service Related Commands

The service related commands are:
mongod --install
mongod --service
mongod --remove
mongod --reinstall
You may also option pass the following to --install and --reinstall
--serviceName {arg}
--serviceUser {arg}
--servicePassword {arg}
The --install and --remove options install and remove the mongo daemon as a windows service respectively. The --service option starts the
service. --reinstall will attempt to remove the service, and then install it. If the service is not already installed, --reinstall will still work.
Both --remove and --reinstall will stop the service if it is currently running.
To change the name of the service use --serviceName. To make mongo execute as a local or domain user, as opposed to the Local System
account, use --serviceUser and --servicePassword.
Whatever other arguments you pass to mongod.exe on the command line alongside --install are the arguments that the service is configured to
execute mongod.exe with. Take for example the following command line:
mongod --bind_ip 127.0.0.1 --logpath d:\mongo\logs --logappend --dbpath d:\mongo\data --directoryperdb

--install
Will cause a service to be created called Mongo that will execute the following command:
mongod --bind_ip 127.0.0.1 --logpath d:\mongo\logs --logappend --dbpath d:\mongo\data --directoryperdb
View Current Operation(s) in Progress

Terminate (Kill) an Operation in Progress
View Current Operation(s) in Progress
> db.currentOp();
> // same as: db.$cmd.sys.inprog.findOne()
{ inprog: [ { "opid" : 18 , "op" : "query" , "ns" : "mydb.votes" ,
"query" : "{ score : 1.0 }" , "inLock" : 1 }
]
}
Fields:
opid - an incrementing operation number. Use with killOp().

op - the operation type (query, update, etc.)
ns - namespace for the operation (database + collection name)
query - the query spec, if operation is a query
NOTE: currentOp's output format varies from version 1.0 and version 1.1 of MongoDB. The format above is for 1.1 and higher.
You can also do
db.$cmd.sys.inprog.find()
or this version which prints all connections
db.$cmd.sys.inprog.find( { $all : 1 } )
Terminate (Kill) an Operation in Progress
// <= v1.2
> db.killOp()
> // same as: db.$cmd.sys.killop.findOne()
{"info" : "no op in progress/not locked"}
// v>= 1.3
> db.killOp(1234/*opid*/)
> // same as: db.$cmd.sys.killop.findOne({op:1234})
Validate Command
Use this command to check that a collection is valid (not corrupt) and to get various statistics.
This command scans the entire collection and its indexes and will be very slow on large datasets.
From the mongo shell:
> db.foo.validate()
{"ns" : "test.foo" , "result" : "
validate
details: 08D03C9C ofs:963c9c
firstExtent:0:156800 ns:test.foo
lastExtent:0:156800 ns:test.foo
# extents:1
datasize?:144 nrecords?:3 lastExtentSize:2816
padding:1
first extent:
loc:0:156800 xnext:null xprev:null
ns:test.foo
size:2816 firstRecord:0:1568b0 lastRecord:0:156930
3 objects found, nobj:3
192 bytes data w/headers
144 bytes data wout/headers
deletedList: 0000000100000000000
deleted: n: 1 size: 2448
nIndexes:1
test.foo.$x_1 keys:3
" , "ok" : 1 , "valid" : true , "lastExtentSize" : 2816}
From a driver one might invoke the driver's equivalent of:
> db.$cmd.findOne({validate:"foo" } );
validate takes an optional scandata parameter which skips the scan of the base collection (but still scans indexes).
> db.$cmd.findOne({validate:"foo", scandata:true});
getLastError
Redirection Notice
This page should redirect to Last Error Commands in about 3 seconds.
Most drivers, and the db shell, support a getlasterror capability. This lets one check the error code on the last operation.
Database commands, as well as queries, have a direct return code.
getlasterror is primarily useful for write operations (although it is set after a command or query too). Write operations by default do not have a
return code: this saves the client from waiting for client/server turnarounds during write operations. One can always call getLastError if one wants
a return code.
> db.runCommand("getlasterror")
> db.getLastError()
Note: The special mongo wire protocol killCursors operation does not support getlasterror. (This is really only of significant to driver developers .)
getPrevError
Note: getPrevError may be deprecated in the future.
When performing bulk write operations, resetError() and getPrevError() can be an efficient way to check for success of the operation.
For example if we are inserting 1,000 objects in a collection, checking the return code 1,000 times over the network is slow. Instead one might do
something like this:
db.resetError();
for( loop 1000 times... )
db.foo.save(something...);
if( db.getPrevError().err )
print("didn't work!");
Last Error in the Shell
The database shell performs a resetError() before each read/eval/print loop command evaluation - and automatically prints the error, if one
occurred, after each evaluation. Thus, after an error, at the shell prompt db.getLastError() will return null. However, if called before returning
to the prompt, the result is as one would expect:
> try { db.foo.findOne() } catch(e) { print("preverr:" + tojson(db.getPrevError())); print("lasterr:"

+ tojson(db.getLastError()));}
preverr:{"err" : "unauthorized" , "nPrev" : 1 , "ok" : 1}
lasterr:"unauthorized"
FSync with GetLastError
Include the fsync option to force the database to fsync all files before returning (v1.3+):
> db.runCommand({getlasterror:1,fsync:true})
{ "err" : null, "n" : 0, "fsyncFiles" : 2, "ok" : 1 }

List of MongoDB Commands
See the Commands page for details on how to invoke a command.
Also: with v1.5+, run mongod with --rest enabled, and then go to http://localhost:28017/_commands
Commands Quick Reference Card
Mongo Metadata
The <dbname>.system.* namespaces in MongoDB are special and contain database system information. System collections include:
system.namespaces lists all namespaces.

system.indexes lists all indexes.
Additional namespace / index metadata exists in the database.ns files, and is opaque.
system.profile stores database profiling information.
system.users lists users who may access the database.
local.sources stores replica slave configuration data and state.
Information on the structure of a stored object is stored within the object itself. See BSON .
There are several restrictions on manipulation of objects in the system collections. Inserting in system.indexes adds an index, but otherwise
that table is immutable (the special drop index command updates it for you). system.users is modifiable. system.profile is droppable.
Note: $ is a reserved character. Do not use it in namespace names or within field names. Internal collections for indexes use the $ character in
their names. These collection store b-tree bucket data and are not in BSON format (thus direct querying is not possible).
Collections
MongoDB collections are essentially named groupings of documents. You can think of them as roughly equivalent to relational database tables.
Details
A MongoDB collection is a collection of BSON documents. These documents are usually have the same structure, but this is not a requirement
since MongoDB is a schema-free database. You may store a heterogeneous set of documents within a collection, as you do not need predefine
the collection's "columns" or fields.
A collection is created when the first document is inserted.
Collection names should begin with letters or an underscore and may include numbers; $ is reserved. Collections can be organized in
namespaces; these are named groups of collections defined using a dot notation. For example, you could define collections blog.posts and
blog.authors, both reside under "blog". Note that this is simply an organizational mechanism for the user -- the collection namespace is flat from
the database's perspective.
Programmatically, we access these collections using the dot notation. For example, using the mongo shell:
if( db.blog.posts.findOne() )
print("blog.posts exists and is not empty.");
The maximum size of a collection name is 128 characters (including the name of the db and indexes). It is probably best to keep it under 80/90
chars.
See also:
Capped Collections
Capped Collections
Capped collections are fixed sized collections that have a very high performance auto-LRU age-out feature (age out is based on insertion order).
In addition, capped collections automatically, with high performance, maintain insertion order for the objects in the collection; this is very powerful
for certain use cases such as logging.
Creating a Fixed Size (capped) Collection
Unlike a standard collection, you must explicitly create a capped collection, specifying a collection size in bytes. The collection's data space is
then preallocated. Note that the size specified includes database headers.
db.createCollection("mycoll", {capped:true, size:100000})
Usage and Restrictions
You may insert new objects in the capped collection.

You may update the existing objects in the collection. However, the objects must not grow in size. If they do, the update will fail. (There
are some possible workarounds which involve pre-padding objects; contact us in the support forums for more information, if help is
needed.)
The database does not allow deleting objects from a capped collection. Use the drop() method to remove all rows from the collection.
Note: After the drop you must explicitly recreate the collection.
Maximum size for a capped collection is currently 1e9 bytes on a thirty-two bit machine. The maximum size of a capped collection on a
sixty-four bit machine is constrained only by system resources.
Behavior
Once the space is fully utilized, newly added objects will replace the oldest objects in the collection.
If you perform a find() on the collection with no ordering specified, the objects will always be returned in insertion order. Reverse order
is always retrievable with find().sort({$natural:-1}).
Applications
Logging. Capped collections provide a high-performance means for storing logging documents in the database. Inserting objects in an
unindexed capped collection will be close to the speed of logging to a filesystem. Additionally, with the built-in LRU mechanism, you are
not at risk of using excessive disk space for the logging.
Caching. If you wish to cache a small number of objects in the database, perhaps cached computations of information, the capped tables
provide a convenient mechanism for this. Note that for this application you will likely use an index on the capped table as there will be
more reads than writes.
Auto Archiving. If you know you want data to automatically "roll out" over time as it ages, a capped collection can be an easier way to
support than writing manual archival cron scripts.
Recommendations
For maximum performance, do not create indexes on a capped collection. If the collection will be written to much more than it is read
from, it is better to have no indexes. Note that you may create indexes on a capped collection; however, you are then moving from "log
speed" inserts to "database speed" inserts -- that is, it will still be quite fast by database standards.
Use natural ordering to retrieve the most recently inserted elements from the collection efficiently. This is (somewhat) analogous to tail on
a log file.
Capping the Number of Objects
You may also cap the number of objects in the collection. Once the limit is reached, items roll out on a least recently inserted basis.
To cap on number of objects, specify a max: parameter on the createCollection() call.
Note: When specifying a cap on the number of objects, you must also cap on size. Be sure to leave enough room for your chosen number of
objects or items will roll out faster than expected. You can use the validate() utility method to see how much space an existing collection uses,
and from that estimate your size needs.
db.createCollection("mycoll", {capped:true, size:100000, max:100});

db.mycoll.validate();
Tip: When programming, a handy way to store the most recently generated version of an object can be a collection capped with max=1.
Preallocating space for a normal collection
The createCollection command may be used for non capped collections as well. For example:
db.createCollection("mycoll", {size:10000000});
db.createCollection("mycoll", {size:10000000, autoIndexId:false});
Explicitly creating a non capped collection via createCollection allows parameters of the new collection to be specified. For example,
specification of a collection size causes the corresponding amount of disk space to be preallocated for use by the collection. The autoIndexId
field may be set to true or false to explicitly enable or disable automatic creation of a unique key index on the _id object field. By default, such an
index is created for non capped collections but is not created for capped collections.
An index is not automatically created on _id for capped collections by default
See Also
The Sorting and Natural Order section of this Guide

A technique one can use with MongoDB in certain situations is to have several collections to store information instead of a single collection. By
doing this, certain repeating data no longer needs to be stored in every object, and an index on that key may be eliminated. More importantly for
performance (depending on the problem), the data is then clustered by the grouping specified.
For example, suppose we are logging objects/documents to the database, and want to have M logs: perhaps a dev log, a debug log, an ops log,
etc. We could store them all in one collection 'logs' containing objects like:
{ log : 'dev', ts : ..., info : ... }

However, if the number of logs is not too high, it might be better to have a collection per log. We could have a 'logs.dev' collection, a 'logs.debug'
collection, 'logs.ops', etc.:
// logs.dev:
{ ts : ..., info : ... }
Of course, this only makes sense if we do not need to query for items from multiple logs at the same time.
Generally, having a large number of collections has no significant performance penalty, and results in very good performance.
Limits
By default MongoDB has a limit of approximately 24,000 namespaces per database. Each collection counts as a namespace, as does each
index. Thus if every collection had one index, we can create up to 12,000 collections. Use the --nssize parameter to set a higher limit.
Be aware that there is a certain minimum overhead per collection -- a few KB. Further, any index will require at least 8KB of data space as the
b-tree page size is 8KB.
--nssize
If more collections are required, run mongod with the --nssize parameter specified. This will make the <database>.ns file larger and support more
collections. Note that --nssize sets the size used for newly created .ns files -- if you have an existing database and wish to resize, after running
the db with --nssize, run the db.repairDatabase() command from the shell to adjust the size.
Maximum .ns file size is 2GB.
MongoDB (BSON) Data Types
Mongo uses special data types in addition to the basic JSON types of string, integer, boolean, double, null, array, and object. These types include
date, object id, binary data, regular expression, and code. Each driver implements these types in language-specific ways, see your driver's
documentation for details.
See BSON for a full list of database types.
Internationalization
See Internationalized strings
Database References
See Database References and Schema Design
MongoDB supports UTF-8 for strings in stored objects and queries. (Specifically, BSON strings are UTF-8.)
Generally, drivers for each programming language convert from the language's string format of choice to UTF-8 when serializing and deserializing
BSON. For example, the Java driver converts Java Unicode strings to UTF-8 on serialization.
In most cases this means you can effectively store most international characters in MongoDB strings. A few notes:
MongoDB regex queries support UTF-8 in the regex string.

Currently, sort() on a string uses strcmp: sort order will be reasonable but not fully international correct. Future versions of MongoDB
may support full UTF-8 sort ordering.
Object IDs
Documents in MongoDB are required to have a key, _id, which uniquely identifies them.
Document IDs: _id

The BSON ObjectId Datatype
BSON ObjectID Specification
Document Timestamps
Sequence Numbers
Document IDs: _id
Every MongoDB document has an _id field as its first attribute. This value usually a BSON ObjectId. Such an id must be unique for each
member of a collection; this is enforced if the collection has an index on _id, which is the case by default.
If a user tries to insert a document without providing an id, the database will automatically generate an _object id and store it the _id field.
Users are welcome to use their own conventions for creating ids; the _id value may be of any type so long as it is a unique.
The BSON ObjectId Datatype
Although _id values can be of any type, a special BSON datatype is provided for object ids. This type is a 12-byte binary value designed to have
a reasonably high probability of being unique when allocated. All of the officially-supported MongoDB drivers use this type by default for _id
values. Also, the Mongo database itself uses this type when assigning _id values on inserts where no _id value is present.
In the MongoDB shell, ObjectId() may be used to create ObjectIds. ObjectId(string) creates an object ID from the specified hex string.
> x={ name: "joe" }

{ name : "joe" }
> db.people.save(x)
{ name : "joe" , _id : ObjectId( "47cc67093475061e3d95369d" ) }
> x
{ name : "joe" , _id : ObjectId( "47cc67093475061e3d95369d" ) }
> db.people.findOne( { _id: ObjectId( "47cc67093475061e3d95369d" ) } )
{ _id : ObjectId( "47cc67093475061e3d95369d" ) , name : "joe" }
> db.people.findOne( { _id: new ObjectId( "47cc67093475061e3d95369d" ) } )
{ _id : ObjectId( "47cc67093475061e3d95369d" ) , name : "joe" }
BSON ObjectID Specification
A BSON ObjectID is a 12-byte value consisting of a 4-byte timestamp (seconds since epoch), a 3-byte machine id, a 2-byte process id, and a
3-byte counter. Note that the timestamp and counter fields must be stored big endian unlike the rest of BSON. This is because they are compared
byte-by-byte and we want to ensure a mostly increasing order. Here's the schema:
0 1 2 3 4 5 6 7 8 9 10 11
time machine pid inc
Document Timestamps
One useful consequence of this specification is that it provides documents with a creation timestamp for free. All of the drivers implement methods
for extracting these timestamps; see the relevant api docs for details.
Sequence Numbers
Traditional databases often use monotonically increasing sequence numbers for primary keys. In MongoDB, the preferred approach is to use
Object IDs instead. Object IDs are more synergistic with sharding and distribution.
However, sometimes you may want a sequence number. The Insert if Not Present section of the Atomic Operations page shows an example of
how to do this.
Database References
Simple Manual References
DBRef
DBRef in Different Languages / Drivers
C#
C++
Java
Javascript (mongo shell)
PHP
Python
Ruby
See Also
As MongoDB is non-relational (no joins), references ("foreign keys") between documents are generally resolved client-side by additional queries
to the server. Two conventions are common for references in MongoDB: first simple manual references, and second, the DBRef standard, which
many drivers support explicitly.
Note: Often embedding of objects eliminates the need for references, but sometimes references are still appropriate.
Simple Manual References
Generally, manually coded references work just fine. We simply store the value that is present in _id in some other document in the database.
For example:
> p = db.postings.findOne();
{
"_id" : ObjectId("4b866f08234ae01d21d89604"),
"author" : "jim",
"title" : "Brewing Methods"
}
> // get more info on author
> db.users.findOne( { _id : p.author } )
{ "_id" : "jim", "email" : "[email protected]" }
DBRef
DBRef is a more formal specification for creating references between documents. DBRefs (generally) include a collection name as well as an
oject id. Most developers only use DBRefs if the collection can change from one document to the next. If your referenced collection will always
be the same, the manual references outlined above are more efficient.
A DBRef is a reference from one document (object) to another within a database. A database reference is a standard embedded (JSON/BSON)
object: we are defining a convention, not a special type. By having a standard way to represent, drivers and data frameworks can add helper
methods which manipulate the references in standard ways.
DBRef's have the advantage of allowing optional automatic client-side dereferencing with some drivers, although more features may be added
later. In many cases, you can just get away with storing the _id as a reference then dereferencing manually as detailed in the "Simple Manual
References" section above.
Syntax for a DBRef reference value is
{ $ref : <collname>, $id : <idvalue>[, $db : <dbname>] }
where <collname> is the collection name referenced (without the database name), and <idvalue> is the value of the _id field for the object
referenced. $db is optional (currently unsupported by many of the drivers) and allows for references to documents in other databases (specified
by <dbname>).
The ordering for DBRefs does matter, fields must be in the order specified above.
The old BSON DBRef datatype is deprecated.
DBRef in Different Languages / Drivers
C#
Use the DBRef class. It takes the collection name and _id as parameters to the constructor. Then you can use the FollowReference method on
the Database class to get the referenced document.
C++
The C++ driver does not yet provide a facility for automatically traversing DBRefs. However one can do it manually of course.
Java
Java supports DB references using the DBRef class.
Javascript (mongo shell)
Example:
> x = { name : 'Biology' }
{ "name" : "Biology" }
> db.courses.save(x)
> x
{ "name" : "Biology", "_id" : ObjectId("4b0552b0f0da7d1eb6f126a1") }
> stu = { name : 'Joe', classes : [ new DBRef('courses', x._id) ] }
// or we could write:
// stu = { name : 'Joe', classes : [ {$ref:'courses',$id:x._id} ] }
> db.students.save(stu)
> stu
{
"name" : "Joe",
"classes" : [
{
"$ref" : "courses",
"$id" : ObjectId("4b0552b0f0da7d1eb6f126a1")
}
],
"_id" : ObjectId("4b0552e4f0da7d1eb6f126a2")
}
> stu.classes[0]
{ "$ref" : "courses", "$id" : ObjectId("4b0552b0f0da7d1eb6f126a1") }
> stu.classes[0].fetch()
{ "_id" : ObjectId("4b0552b0f0da7d1eb6f126a1"), "name" : "Biology" }
>
PHP
PHP supports DB references with the MongoDBRef class, as well as creation and deferencing methods at the database (MongoDB::createDBRef
and MongoDB::getDBRef) and collection (MongoCollection::createDBRef and MongoCollection::getDBRef) levels.
Python
To create a DB reference in python use the pymongo.dbref.DBRef class. You can also use the dereference method on Database instances to
make dereferencing easier.
Python also supports auto-ref and auto-deref - check out the auto_reference example.
Ruby
Ruby also supports DB references using the DBRef class and a dereference method on DB instances. For example:
@db = Connection.new.db("blog")
@user = @db["users"].save({:name => "Smith"})
@post = @db["posts"].save({:title => "Hello World", :user_id => @user.id})
@ref = DBRef.new("users", @post.user_id)
assert_equal @user, @db.dereference(@ref)
See Also
Schema Design
GridFS
GridFS is a specification for storing large files in MongoDB. All of the officially supported driver implement the GridFS spec.
Rationale
Implementation
Language Support
Command Line Tools
See also
Rationale
The database supports native storage of binary data within BSON objects. However, BSON objects in MongoDB are limited to 4MB in size. The
GridFS spec provides a mechanism for transparently dividing a large file among multiple documents. This allows us to efficiently store large
objects, and in the case of especially large files, such as videos, permits range operations (e.g., fetching only the first N bytes of a file).
Implementation
To facilitate this, a standard is specified for the chunking of files. Each file has a metadata object in a files collection, and one or more chunk
objects in a chunks collection. Details of how this is stored can be found in the GridFS Specification; however, you do not really need to read that,
instead, just look at the GridFS API in each language's client driver or mongofiles tool.
Language Support
Most drivers include GridFS implementations; for languages not listed below, check the driver's API documentation. (If a language does not
include support, see the GridFS specification -- implementing a handler is usually quite easy.)
Command Line Tools
Command line tools are available to write and read GridFS files from and to the local filesystem.
See also
C++
A PHP GridFS Blog Article
Indexes
Indexes enhance query performance, often dramatically. It's important to think about the kinds of queries your application will need so that you
can define relevant indexes. Once that's done, actually creating the indexes in MongoDB is relatively easy.
Indexes in MongoDB are conceptually similar to those in RDBMSes like MySQL. You will want an index in MongoDB in the same sort of situations
where you would have wanted an index in MySQL.
Basics
Default Indexes
Embedded Keys
Documents as Keys
Arrays
Compound Keys Indexes
Unique Indexes
Missing Keys
Duplicate Values
Background Index Building
Dropping Indexes
ReIndex
Additional Notes on Indexes
Index Performance
Using sort() without an Index
Geospatial
Webinar
Basics
An index is a data structure that collects information about the values of the specified fields in the documents of a collection. This data structure is
used by Mongo's query optimizer to quickly sort through and order the documents in a collection. Formally speaking, these indexes are
implemented as "B-Tree" indexes.
In the shell, you can create an index by calling the ensureIndex() function, and providing a document that specifies one or more keys to index.
Referring back to our examples database from Mongo Usage Basics, we can index on the 'j' field as follows:
db.things.ensureIndex({j:1});
The ensureIndex() function only creates the index if it does not exist.
Once a collection is indexed on a key, random access on query expressions which match the specified key are fast. Without the index, MongoDB
has to go through each document checking the value of specified key in the query:
db.things.find({j : 2}); // fast - uses index
db.things.find({x : 3}); // slow - has to check all because 'x' isn't indexed
You can run db.things.getIndexes(); to see the existing indexes on the collection.
Default Indexes
An index is always created on _id. This index is special and cannot be deleted. The _id index enforces uniqueness for its keys. For Capped
Collections no index is created.
Embedded Keys
With MongoDB you can even index on a key inside of an embedded document. For example:
db.things.ensureIndex({"address.city": 1})
Documents as Keys
Indexed fields may be of any type, including documents:
db.factories.insert( { name: "xyz", metro: { city: "New York", state: "NY" } } );

db.factories.ensureIndex( { metro : 1 } );
// this query can use the above index:
db.factories.find( { metro: { city: "New York", state: "NY" } } );
An alternative to documents as keys it to create a compound index such as:
db.factories.ensureIndex( { "metro.city" : 1, "metro.state" : 1 } );

// these queries can use the above index:
db.factories.find( { "metro.city" : "New York", "metro.state" : "NY" } );
db.factories.find( { "metro.city" : "New York" } );
db.factories.find().sort( { "metro.city" : 1, "metro.state" : 1 } );
db.factories.find().sort( { "metro.city" : 1 } )
There are pros and cons to the two approaches. When using the entire (sub-)document as a key, compare order is predefined and is ascending
key order in the order the keys occur in the BSON document. With compound indexes reaching in, you can mix ascending and descending keys,
and the query optimizer will then be able to use the index for queries on solely the first key(s) in the index too.
Arrays
When a document's stored value for a index key field is an array, MongoDB indexes each element of the array. See the Multikeys page for more
information.
Compound Keys Indexes
In addition to single-key basic indexes, MongoDB also supports multi-key "compound" indexes. Just like basic indexes, you use the
ensureIndex() function in the shell to create the index, but instead of specifying only a single key, you can specify several :
db.things.ensureIndex({j:1, name:-1});
When creating an index, the number associated with a key specifies the direction of the index, so it should always be 1 (ascending) or -1
(descending). Direction doesn't matter for single key indexes or for random access retrieval but is important if you are doing sorts or range queries
on compound indexes.
If you have a compound index on multiple fields, you can use it to query on the beginning subset of fields. So if you have an index on
a,b,c
you can use it query on

a
a,b
a,b,c
New in 1.6+
Now you can also use the compound index to service any combination of equality (and some inequality) queries from the
constitute fields.
Unique Indexes
MongoDB supports unique indexes, which guarantee that no documents are inserted whose values for the indexed keys match those of an
existing document. To create an index that guarantees that no two documents have the same values for both firstname and lastname you
would do:
db.things.ensureIndex({firstname: 1, lastname: 1}, {unique: true});
Missing Keys
When a document is saved to a collection with unique indexes, any missing indexed keys will be inserted with null values. Thus, it won't be
possible to insert multiple documents missing the same indexed key.
db.things.ensureIndex({firstname: 1}, {unique: true});

db.things.save({lastname: "Smith"});
// Next operation will fail because of the unique index on firstname.

db.things.save({lastname: "Jones"});
Duplicate Values
A unique index cannot be created on a key that has duplicate values. If you would like to create the index anyway, keeping the first document the
database indexes and deleting all subsequent documents that have duplicate values, add the dropDups option.
db.things.ensureIndex({firstname : 1}, {unique : true, dropDups : true})
Background Index Building
By default, building an index blocks other database operations. v1.3.2 and higher has a background index build option .
Dropping Indexes
To delete all indexes on the specified collection:
db.collection.dropIndexes();
To delete a single index:
db.collection.dropIndex({x: 1, y: -1})
Running directly as a command without helper:

// note: command was "deleteIndexes", not "dropIndexes", before MongoDB v1.3.2
// remove index with key pattern {y:1} from collection foo
db.runCommand({dropIndexes:'foo', index : {y:1}})
// remove all indexes:
db.runCommand({dropIndexes:'foo', index : '*'})
ReIndex
The reIndex command will rebuild all indexes for a collection.
db.myCollection.reIndex()
// same as:
db.runCommand( { reIndex : 'myCollection' } )
Usually this is unnecessary. You may wish to do this if the size of your collection has changed dramatically or the disk space used by indexes
seems oddly large.
Repair database recreates all indexes in the database.
Additional Notes on Indexes
MongoDB indexes (and string equality tests in general) are case sensitive.
When you update an object, if the object fits in its previous allocation area, only those indexes whose keys have changed are updated.
This improves performance. Note that if the object has grown and must move, all index keys must then update, which is slower.
Index information is kept in the system.indexes collection, run db.system.indexes.find() to see example data.
Index Performance
Indexes make retrieval by a key, including ordered sequential retrieval, very fast. Updates by key are faster too as MongoDB can find the
document to update very quickly.
However, keep in mind that each index created adds a certain amount of overhead for inserts and deletes. In addition to writing data to the base
collection, keys must then be added to the B-Tree indexes. Thus, indexes are best for collections where the number of reads is much greater than
the number of writes. For collections which are write-intensive, indexes, in some cases, may be counterproductive. Most collections are
read-intensive, so indexes are a good thing in most situations.
Using sort() without an Index
You may use sort() to return data in order without an index if the data set to be returned is small (less than four megabytes). For these cases it
is best to use limit() and sort() together.
Geospatial
See Geospatial Indexing page.
Webinar
Indexing with MongoDB webinar video (http://bit.ly/bQK1Op) and slides (http://bit.ly/9qeMjC). More detailed slides (http://bit.ly/dAHrQk).

One way to work with data that has a high degree of options for queryability is to use the multikey indexing feature where the keys are objects.
For example:
> x = {
> ... _id : "abc",
> ... cost : 33,
> ... attribs : [
> ... { color : 'red' },
> ... { shape : 'rect' },
> ... { color : 'blue' },
> ... { avail : true } ]
> ... };
> db.foo.insert(x);
> db.foo.ensureIndex({attribs:1});
> db.foo.find( { attribs : {color:'blue'} } ); // uses index
> db.foo.find( { attribs : {avail:false} } ); // uses index
In addition to being able to have an unlimited number of attributes types, we can also add new types dynamically.
This is mainly useful for simply attribute lookups; the above pattern is not necessary helpful for sorting or certain other query types.
See Also
Discussion thread MongoDB for a chemical property search engine for a more complex real world example.
Geospatial Indexing
Creating the Index
Querying
Compound Indexes
geoNear Command
Bounds Queries
The Earth is Round but Maps are Flat
New Spherical Model
Sharded Environments
Implementation
v1.3.3+
MongoDB supports two-dimensional geospatial indexes. It is designed with location-based queries in mind, such as "find me the closest N items
to my location." It can also efficiently filter on additional criteria, such as "find me the closest N museums to my location."
In order to use the index, you need to have a field in your object that is either a sub-object or array where the first 2 elements are x,y coordinates
(or y,x - just be consistent; it might be advisible to use order-preserving dictionaries/hashes in your client code, to ensure consistency). Some
examples:
{ loc : [ 50 , 30 ] }
{ loc : { x : 50 , y : 30 } }
{ loc : { foo : 50 , y : 30 } }
{ loc : { lat : 40.739037, long: 73.992964 } }
Creating the Index
db.places.ensureIndex( { loc : "2d" } )
By default, the index assumes you are indexing latitude/longitude and is thus configured for a [-180..180] value range.
If you are indexing something else, you can specify some options:
db.places.ensureIndex( { loc : "2d" } , { min : -500 , max : 500 } )
that will scale the index to store values between -500 and 500. Currently geo indexing is limited to indexing squares with no "wrapping" at the
outer boundaries. You cannot insert values on the boundaries, for example, using the code above, the point (-500, -500) could not to be
inserted.
you can only have 1 geo2d index per collection right now
Querying
The index can be used for exact matches:
db.places.find( { loc : [50,50] } )
Of course, that is not very interesting. More important is a query to find points near another point, but not necessarily matching exactly:
db.places.find( { loc : { $near : [50,50] } } )
The above query finds the closest points to (50,50) and returns them sorted by distance (there is no need for an additional sort parameter). Use
limit() to specify a maximum number of points to return (a default limit of 100 applies if unspecified):
db.places.find( { loc : { $near : [50,50] } } ).limit(20)
Compound Indexes
MongoDB geospatial indexes optionally support specification of secondary key values. If you are commonly going to be querying on both a
location and other attributes at the same time, add the other attributes to the index. The other attributes are annotated within the index to make
filtering faster. For example:
db.places.ensureIndex( { location : "2d" , category : 1 } );

db.places.find( { location : { $near : [50,50] }, category : 'coffee' } );
geoNear Command
While the find() syntax above is typically preferred, MongoDB also has a geoNear command which performs a similar function. The geoNear
command has the added benefit of returning the distance of each item from the specified point in the results, as well as some diagnostics for
troubleshooting.
> db.runCommand( { geoNear : "places" , near : [50,50], num : 10 } );
> db.runCommand({geoNear:"asdf", near:[50,50]})
{
"ns" : "test.places",
"near" : "1100110000001111110000001111110000001111110000001111",
"results" : [
{
"dis" : 69.29646421910687,
"obj" : {
"_id" : ObjectId("4b8bd6b93b83c574d8760280"),
"y" : [
1,
1
],
"category" : "Coffee"
}
},
{
"dis" : 69.29646421910687,
"obj" : {
"_id" : ObjectId("4b8bd6b03b83c574d876027f"),
"y" : [
1,
1
]
}
}
],
"stats" : {
"time" : 0,
"btreelocs" : 1,
"btreelocs" : 1,
"nscanned" : 2,
"nscanned" : 2,
"objectsLoaded" : 2,
"objectsLoaded" : 2,
"avgDistance" : 69.29646421910687
},
"ok" : 1
}
The above command will return the 10 closest items to (50,50). (The loc field is automatically determined by checking for a 2d index on the
collection.)
If you want to add an additional filter, you can do so:
> db.runCommand( { geoNear : "places" , near : [ 50 , 50 ], num : 10,

... query : { type : "museum" } } );
query can be any regular mongo query.
Bounds Queries
v1.3.4
$within can be used instead of $near to find items within a shape. At the moment, $box (rectangles) and $center (circles) are supported.
To query for all points within a rectangle, you must specify the lower-left and upper-right corners:
> box = [[40, 40], [60, 60]]

> db.places.find({"loc" : {"$within" : {"$box" : box}}})
A circle is specified by a center point and radius:
> center = [50, 50]

> radius = 10
> db.places.find({"loc" : {"$within" : {"$center" : [center, radius]}}})
The Earth is Round but Maps are Flat
The current implementation assumes an idealized model of a flat earth, meaning that an arcdegree of latitude (y) and longitude (x) represent the
same distance everywhere. This is only true at the equator where they are both about equal to 69 miles or 111km. However, at the 10gen offices
at { x : -74 , y : 40.74 } one arcdegree of longitude is about 52 miles or 83 km (latitude is unchanged). This means that something 1 mile to the
north would seem closer than something 1 mile to the east.
New Spherical Model
In 1.7.0 we added support for correctly using spherical distances by adding "Sphere" to the name of the query. So far, only $centerSphere queries
are supported but support for $nearSphere is planned. There are a few caveats that you must be aware of when using spherical distances:
1. The code assumes that you are using decimal degrees in (X,Y) / (longitude, latitude) order. This is the same order used for the GeoJSON
spec.
2. All distances use radians. This allows you to easily multiply by the radius of the earth (about 6371 km or 3959 miles) to get the distance in
your choice of units. Conversely, divide by the radius of the earth when doing queries.
3. We don't currently handle wrapping at the poles or at the transition from -180° to +180° longitude, however we detect when a search
would wrap and raise an error.
Support for geospatial in sharded collections is coming; please watch this ticket: http://jira.mongodb.org/browse/SHARDING-83.
In the meantime sharded clusters can use geospatial indexes for unsharded collections within the cluster.
Implementation
The current implementation encodes geographic hash codes atop standard MongoDB b-trees. Results of $near queries are exact. The problem
with geohashing is that prefix lookups don't give you exact results, especially around bit flip areas. MongoDB solves this by doing a grid by grid
search after the initial prefix scan. This guarantees performance remains very high while providing correct results.
Indexing as a Background Operation
Slaves and replica secondaries build all indexes in the foreground in certain releases (including the latest). Thus even when
using background:true on the primary, the slave/secondary will be unavailable to service queries while the index builds there.
By default the ensureIndex() operation is blocking, and will stop other operations on the database from proceeding until completed. However,
in v1.3.2+, a background indexing option is available.
To build an index in the background, add background:true to your index options. Examples:
> db.things.ensureIndex({x:1}, {background:true});

> db.things.ensureIndex({name:1}, {background:true, unique:true,
... dropDups:true});
With background mode enabled, other operations, including writes, will not be obstructed during index creation. The index is not used for queries
until the build is complete.
Although the operation is 'background' in the sense that other operations may run concurrently, the command will not return to the shell prompt
until completely finished. To do other operations at the same time, open a separate mongo shell instance.
Please note that background mode building uses an incremental approach to building the index which is slower than the default foreground mode:
time to build the index will be greater.
While the build progresses, it is possible to see that the operation is still in progress with the db.currentOp() command (will be shown as an
insert to system.indexes). You may also use db.killOp() to terminate the build process.
While the build progresses, the index is visible in system.indexes, but it is not used for queries until building completes.
Notes
Only one index build at a time is permitted per collection.

Some administrative operations, such as repairDatabase, are disallowed while a background indexing job is in progress.
v1.4 and higher (for production usage)
Multikeys
MongoDB provides an interesting "multikey" feature that can automatically index arrays of an object's values. A good example is tagging.
Suppose you have an article tagged with some category names:
$ dbshell
> db.articles.save( { name: "Warm Weather", author: "Steve",
tags: ['weather', 'hot', 'record', 'april'] } )
> db.articles.find()
{"name" : "Warm Weather" , "author" : "Steve" ,
"tags" : ["weather","hot","record","april"] , "_id" : "497ce4051ca9ca6d3efca323"}
We can easily perform a query looking for a particular value in the tags array:
> db.articles.find( { tags: 'april' } )

Further, we can index on the tags array. Creating an index on an array element indexes results in the database indexing each element of the
array:
> db.articles.ensureIndex( { tags : 1 } )

true
> db.articles.find( { tags: 'april' } )
> db.articles.find( { tags: 'april' } ).explain()
{"cursor" : "BtreeCursor tags_1" , "startKey" : {"tags" : "april"} ,
"endKey" : {"tags" : "april"} , "nscanned" : 1 , "n" : 1 , "millis" : 0 }
Embedded object fields in an array
Additionally the same technique can be used for fields in embedded objects:
> db.posts.find( { "comments.author" : "julie" } )

{"title" : "How the west was won" ,
"comments" : [{"text" : "great!" , "author" : "sam"},
{ "text" : "ok" , "author" : "julie"}],
"_id" : "497ce79f1ca9ca6d3efca325"}
Querying on all values in a given set
By using the $all query option, a set of values may be supplied each of which must be present in a matching object field. For example:
> db.articles.find( { tags: { $all: [ 'april', 'record' ] } } )

> db.articles.find( { tags: { $all: [ 'april', 'june' ] } } )
> // no matches
See Also
The Multikeys section of the Full Text Search in Mongo document for information about this feature.

We get a lot of questions about indexing. Here we provide answers to a number of these. There are a couple of points to keep in mind, though.
First, indexes in MongoDB work quite similarly to indexes in MySQL, and thus many of the techniques for building efficient indexes in MySQL
apply to MongoDB.
Second, and even more importantly, know that advice on indexing can only take you so far. The best indexes for your application should always
be based on a number of important factors, including the kinds of queries you expect, the ratio of reads to writes, and even the amount of free
memory on your system. This means that the best strategy for designing indexes will always be to profile a variety of index configurations with
data sets similar to the ones you'll be running in production, and see which perform best. There's no substitute for good empirical analyses.
Indexing Strategies
Create indexes to match your queries.
One index per query.
Make sure your indexes can fit in RAM.
Be careful about single-key indexes with low selectivity.
Use explain.
Understanding explain's output.
Pay attention to the read/write ratio of your application.
Indexing Properties
1. The sort column must be the last column used in the index.
2. The range query must also be the last column in an index. This is an axiom of 1 above.
3. Only use a range query or sort on one column.
4. Conserve indexes by re-ordering columns used on equality (non-range) queries.
5. MongoDB's $ne or $nin operator's aren't efficient with indexes.
FAQ
I've started building an index, and the database has stopped responding. What's going on? What do I do?
I'm using $ne or $nin in a query, and while it uses the index, it's still slow. What's happening?
Indexing Strategies
Here are some general principles for building smart indexes.
Create indexes to match your queries.
If you only query on a single key, then a single-key index will do. For instance, maybe you're searching for a blog post's slug:
db.posts.find({ slug : 'state-of-mongodb-2010' })
In this case, a unique index on a single key is best:
db.ensureIndex({ slug: 1 }, {unique: true});
However, it's common to query on multiple keys and to sort the results. For these situations, compound indexes are best. Here's an example for
querying the latest comments with a 'mongodb' tag:
db.comments.find({ tags : 'mongodb'}).sort({ created_at : -1 });
And here's the proper index:
db.comments.ensureIndex({tags : 1, created_at : -1});
Note that if we wanted to sort by created_at ascending, this index would be less effective.
One index per query.
It's sometimes thought that queries on multiple keys can use multiple indexes; this is not the case with MongoDB. If you have a query that selects
on multiple keys, and you want that query to use an index efficiently, then a compound-key index is necessary.
Make sure your indexes can fit in RAM.

The shell provides a command for returning the total index size on a given collection:
db.comments.totalIndexSize();
65443
If your queries seem sluggish, you should verify that your indexes are small enough to fit in RAM. For instance, if you're running on 4GB RAM and
you have 3GB of indexes, then your indexes probably aren't fitting in RAM. You may need to add RAM and/or verify that all the indexes you've
created are actually being used.
Be careful about single-key indexes with low selectivity.
Suppose you have a field called 'status' where the possible values are 'new' and 'processeed'. If you add an index on 'status' then you've created
a low-selectivity index, meaning that the index isn't going to be very helpful in locating records and might just be be taking up space.
A better strategy, depending on your queries, of course, would be to create a compound index that includes the low-selectivity field. For instance,
you could have a compound-key index on 'status' and 'created_at.'
Another option, again depending on your use case, might be to use separate collections, one for each status. As with all the advice here,
experimentation and benchmarks will help you choose the best approach.
Use explain.
MongoDB includes an explain command for determining how your queries are being processed and, in particular, whether they're using an
index. explain can be used from of the drivers and also from the shell:
db.comments.find({ tags : 'mongodb'}).sort({ created_at : -1 }).explain();
This will return lots of useful information, including the number of items scanned, the time the query takes to process in milliseconds, which
indexes the query optimizer tried, and the index ultimately used.
If you've never used explain, now's the time to start.
Understanding explain's output.
There are three main fields to look for when examining the explain command's output:
cursor: the value for cursor can be either BasicCursor or BtreeCursor. The second of these indicates that the given query is using
an index.
nscanned: he number of documents scanned.
n: the number of documents returned by the query. You want the value of n to be close to the value of nscanned. What you want to
avoid is doing a collection scan, that is, where every document in the collection is accessed. This is the case when nscanned is equal to
the number of documents in the collection.
millis: the number of milliseconds require to complete the query. This value is useful for comparing indexing strategies, indexed vs.
non-indexed queries, etc.
Pay attention to the read/write ratio of your application.
This is important because, whenever you add an index, you add overhead to all insert, update, and delete operations on the given collection. If
your application is read-heavy, as are most web applications, the additional indexes are usually a good thing. But if your application is
write-heavy, then be careful when creating new indexes, since each additional index with impose a small write-performance penalty.
In general, don't be cavalier about adding indexes. Indexes should be added to complement your queries. Always have a good reason for
adding a new index, and make sure you've benchmarked alternative strategies.
Indexing Properties
Here are a few properties of compound indexes worth keeping in mind (Thanks to Doug Green and Karoly Negyesi for their help on this).
These examples assume a compound index of three fields: a, b, c. So our index creation would look like this:
db.foo.ensureIndex({a: 1, b: 1, c: 1})
Here's some advice on using an index like this:

This information is no longer strictly correct in 1.6.0+; compound indexes can now be used to service queries where range or
filter fields are used within the compound index, not just fields used from left to right. Please run explain to see how the
compound index is used.
1. The sort column must be the last column used in the index.
Good:
find(a=1).sort(a)
find(a=1).sort(b)
find(a=1, b=2).sort(c)
Bad:
find(a=1).sort(c)
even though c is the last column used in the index, a is that last column used, so you can only sort on a or b.
2. The range query must also be the last column in an index. This is an axiom of 1 above.
Good:
find(a=1,b>2)
find(a>1 and a<10)
find(a>1 and a<10).sort(a)
Bad:
find(a>1, b=2)
3. Only use a range query or sort on one column.
Good:
find(a=1,b=2).sort(c)
find(a=1,b>2)
find(a=1,b>2 and b<4)
find(a=1,b>2).sort(b)
Bad:
find(a>1,b>2)
find(a=1,b>2).sort(c)
4. Conserve indexes by re-ordering columns used on equality (non-range) queries.
Imagine you have the following two queries:
find(a=1,b=1,d=1)
find(a=1,b=1,c=1,d=1)
A single index defined on a, b, c, and d can be used for both queries.

If, however, you need to sort on the final value, you might need two indexes
5. MongoDB's $ne or $nin operator's aren't efficient with indexes.
When excluding just a few documents, it's better to retrieve extra rows from MongoDB and do the exclusion on the client side.
FAQ
I've started building an index, and the database has stopped responding. What's going on? What do I do?
Building an index can be an IO-intensive operation, especially you have a large collection. This is true on any database system that supports
secondary indexes, including MySQL. If you'll need to build an index on a large collection in the future, you'll probably want to consider building
the index in the background, a feature available beginning with 1.3.2. See the docs on background indexing for more info.
As for the long-building index, you only have a few options. You can either wait for the index to finish building or kill the current operation (see
killOp()). If you choose the latter, the partial index will be deleted.
I'm using $ne or $nin in a query, and while it uses the index, it's still slow. What's happening?
The problem with $ne and $nin is that much of an index will match queries like these. If you need to use $nin, it's often best to make sure that an
additional, more selective criterion is part of the query.
Inserting
When we insert data into MongoDB, that data will always be in document-form. Documents are data structure analogous to JSON, Python
dictionaries, and Ruby hashes, to take just a few examples. Here, we discuss more about document-orientation and describe how to insert data
into MongoDB.
Document-Orientation
JSON
Mongo-Friendly Schema
Store Example
Document-Orientation
Document-oriented databases store "documents" but by document we mean a structured document – the term perhaps coming from the phrase
"XML document". However other structured forms of data, such as JSON or even nested dictionaries in various languages, have similar
properties.
The documents stored in Mongo DB are JSON-like. JSON is a good way to store object-style data from programs in a manner that is
language-independent and standards based.
To be efficient, MongoDB uses a format called BSON which is a binary representation of this data. BSON is faster to scan for specific fields than
JSON. Also BSON adds some additional types such as a data data type and a byte-array (bindata) datatype. BSON maps readily to and from
JSON and also to various data structures in many programming languages.
Client drivers serialize data to BSON, then transmit the data over the wire to the db. Data is stored on disk in BSON format. Thus, on a retrieval,
the database does very little translation to send an object out, allowing high efficiency. The client driver unserialized a received BSON object to its
native language format.
JSON
For example the following "document" can be stored in Mongo DB:
{ author: 'joe',
created : new Date('03/28/2009'),
title : 'Yet another blog post',
text : 'Here is the text...',
tags : [ 'example', 'joe' ],
comments : [ { author: 'jim', comment: 'I disagree' },
{ author: 'nancy', comment: 'Good post' }
]
}
This document is a blog post, so we can store in a "posts" collection using the shell:
> doc = { author : 'joe', created : new Date('03/28/2009'), ... }

> db.posts.insert(doc);
MongoDB understands the internals of BSON objects -- not only can it store them, it can query on internal fields and index keys based upon
them. For example the query
> db.posts.find( { "comments.author" : "jim" } )
is possible and means "find any blog post where at least one comment subjobject has author == 'jim'".
Mongo-Friendly Schema
Mongo can be used in many ways, and one's first instincts when using it are probably going to be similar to how one would write an application
with a relational database. While this work pretty well, it doesn't harness the real power of Mongo. Mongo is designed for and works best with a
rich object model.
Store Example
If you're building a simple online store that sells products with a relation database, you might have a schema like:
item
title
price
sku
item_features
sku
feature_name
feature_value
You would probably normalize it like this because different items would have different features, and you wouldn't want a table with all possible
features. You could model this the same way in mongo, but it would be much more efficient to do
item : {
"title" : <title> ,
"price" : <price> ,
"sku" : <sku> ,
"features" : {
"optical zoom" : <value> ,
...
}
}
This does a few nice things
you can load an entire item with one query

all the data for an item is on the same place on disk, thus only one seek is required to load
Now, at first glance there might seem to be some issues, but we've got them covered.
you might want to insert or update a single feature. mongo lets you operate on embedded files like:
db.items.update( { sku : 123 } , { "$set" : { "features.zoom" : "5" } } )
Does adding a feature require moving the entire object on disk? No. mongo has a padding heuristic that adapts to your data so it will
leave some empty space for the object to grow. This will prevent indexes from being changed, etc.
Legal Key Names

Key names in inserted documents are limited as follows:
The '$' character must not be the first character in the key name.
The '.' character must not appear anywhere in the key name.
Schema Design
Introduction
Embed vs. Reference
Use Cases
Index Selection
How Many Collections?
See Also
Introduction
With Mongo, you do less "normalization" than you would perform designing a relational schema because there are no server-side "joins".
Generally, you will want one database collection for each of your top level objects.
You do not want a collection for every "class" - instead, embed objects. For example, in the diagram below, we have two collections, students and
courses. The student documents embed address documents and the "score" documents, which have references to the courses.
Compare this with a relational schema, where you would almost certainly put the scores in a separate table, and have a foreign-key relationship
back to the students.
Embed vs. Reference
The key question in Mongo schema design is "does this object merit its own collection, or rather should it embed in objects in other collections?"
In relational databases, each sub-item of interest typically becomes a separate table (unless denormalizing for performance). In Mongo, this is not
recommended - embedding objects is much more efficient. Data is then colocated on disk; client-server turnarounds to the database are
eliminated. So in general the question to ask is, "why would I not want to embed this object?"
So why are references slow? Let's consider our students example. If we have a student object and perform:
print( student.address.city );
This operation will always be fast as address is an embedded object, and is always in RAM if student is in RAM. However for
print( student.scores[0].for_course.name );
if this is the first access to scores[0], the shell or your driver must execute the query
// pseudocode for driver or framework, not user code
student.scores[0].for_course = db.courses.findOne({_id:_course_id_to_find_});
Thus, each reference traversal is a query to the database. Typically, the collection in question is indexed on _id. The query will then be
reasonably fast. However, even if all data is in RAM, there is a certain latency given the client/server communication from appserver to database.
In general, expect 1ms of time for such a query on a ram cache hit. Thus if we were iterating 1,000 students, looking up one reference per student
would be quite slow - over 1 second to perform even if cached. However, if we only need to look up a single item, the time is on the order of 1ms,
and completely acceptable for a web page load. (Note that if already in db cache, pulling the 1,000 students might actually take much less than 1
second, as the results return from the database in large batches.)
Some general rules on when to embed, and when to reference:
"First class" objects, that are at top level, typically have their own collection.
Line item detail objects typically are embedded.
Objects which follow an object modelling "contains" relationship should generally be embedded.
Many to many relationships are generally by reference.
Collections with only a few objects may safely exist as separate collections, as the whole collection is quickly cached in application server
memory.
Embedded objects are harder to reference than "top level" objects in collections, as you cannot have a DBRef to an embedded object (at
least not yet).
It is more difficult to get a system-level view for embedded objects. For example, it would be easier to query the top 100 scores across all
students if Scores were not embedded.
If the amount of data to embed is huge (many megabytes), you may reach the limit on size of a single object.
If performance is an issue, embed.
Use Cases
Let's consider a few use cases now.
1. Customer / Order / Order Line-Item
orders should be a collection. customers a collection. line-items should be an array of line-items embedded in the order object.
1. Blogging system.
posts should be a collection. post author might be a separate collection, or simply a field within posts if only an email address.
comments should be embedded objects within a post for performance.
Index Selection
A second aspect of schema design is index selection. As a general rule, where you want an index in a relational database, you want an index in
Mongo.
The _id field is automatically indexed.

Fields upon which keys are looked up should be indexed.
Sort fields generally should be indexed.
The MongoDB profiling facility provides useful information for where an index should be added that is missing.
Note that adding an index slows writes to a collection, but not reads. Use lots of indexes for collections with a high read : write ratio (assuming
one does not mind the storage overage). For collections with more writes than reads, indexes are very expensive.
How Many Collections?
As Mongo collections are polymorphic, one could have a collection objects and put everything in it! This approach is taken by some object
databases. For performance reasons, we do not recommend this approach. Data within a Mongo collection tends to be contiguous on disk. Thus,
table scans of the collection are possible, and efficient. Collections are very important for high throughput batch processing.
See Also
Schema Design talk from MongoNY

DBRef
Trees in MongoDB
Next: Advanced Queries
Trees in MongoDB
Patterns
Full Tree in Single Document
Parent Links
Child Links
Array of Ancestors
Materialized Paths (Full Path in Each Node)
acts_as_nested_set
See Also
The best way to store a tree usually depends on the operations you want to perform; see below for some different options. In practice, most
developers find that one of the "Full Tree in Single Document", "Parent Links", and "Array of Ancestors" patterns works best.
Patterns
Full Tree in Single Document

{
comments: [
{by: "mathias", text: "...", replies: []}
{by: "eliot", text: "...", replies: [
{by: "mike", text: "...", replies: []}
]}
]
}
Pros:
Single document to fetch per page

One location on disk for whole tree
You can see full structure easily
Cons:
Hard to search
Hard to get back partial results
Can get unwieldy if you need a huge tree (there is a 4MB per doc limit)
Parent Links
Storing all nodes in a single collection, with each node having the id of its parent, is a simple solution. The biggest problem with this approach is
getting an entire subtree requires several query turnarounds to the database (or use of db.eval).
> t = db.tree1;
> t.find()
{ "_id" : 1 }
{ "_id" : 2, "parent" : 1 }
{ "_id" : 3, "parent" : 1 }
{ "_id" : 4, "parent" : 2 }
{ "_id" : 5, "parent" : 4 }
{ "_id" : 6, "parent" : 4 }
> // find children of node 4

> t.ensureIndex({parent:1})
> t.find( {parent : 4 } )
{ "_id" : 5, "parent" : 4 }
{ "_id" : 6, "parent" : 4 }
Child Links
Another option is storing the ids of all of a node's children within each node's document. This approach is fairly limiting, although ok if no
operations on entire subtrees are necessary. It may also be good for storing graphs where a node has multiple parents.
> t = db.tree2
> t.find()
{ "_id" : 1, "children" : [ 2, 3 ] }
{ "_id" : 2 }
{ "_id" : 3, "children" : [ 4 ] }
{ "_id" : 4 }
> // find immediate children of node 3

> t.findOne({_id:3}).children
[ 4 ]
> // find immediate parent of node 3

> t.ensureIndex({children:1})
> t.find({children:3})
{ "_id" : 1, "children" : [ 2, 3 ] }
Array of Ancestors
Here we store all the ancestors of a node in an array. This makes a query like "get all descendents of x" fast and easy.
> t = db.mytree;
> t.find()
{ "_id" : "a" }
{ "_id" : "b", "ancestors" : [ "a" ], "parent" : "a" }
{ "_id" : "c", "ancestors" : [ "a", "b" ], "parent" : "b" }
{ "_id" : "d", "ancestors" : [ "a", "b" ], "parent" : "b" }
{ "_id" : "e", "ancestors" : [ "a" ], "parent" : "a" }
{ "_id" : "f", "ancestors" : [ "a", "e" ], "parent" : "e" }
{ "_id" : "g", "ancestors" : [ "a", "b", "d" ], "parent" : "d" }
> t.ensureIndex( { ancestors : 1 } )
> // find all descendents of b:

> t.find( { ancestors : 'b' })
{ "_id" : "c", "ancestors" : [ "a", "b" ], "parent" : "b" }
{ "_id" : "d", "ancestors" : [ "a", "b" ], "parent" : "b" }
{ "_id" : "g", "ancestors" : [ "a", "b", "d" ], "parent" : "d" }
> // get all ancestors of f:

> anc = db.mytree.findOne({_id:'f'}).ancestors
[ "a", "e" ]
> db.mytree.find( { _id : { $in : anc } } )
{ "_id" : "a" }
{ "_id" : "e", "ancestors" : [ "a" ], "parent" : "a" }
ensureIndex and MongoDB's multikey feature makes the above queries efficient.
In addition to the ancestors array, we also stored the direct parent in the node to make it easy to find the node's immediate parent when that is
necessary.
Materialized Paths (Full Path in Each Node)
Materialized paths make certain query options on trees easy. We store the full path to the location of a document in the tree within each node.
Usually the "array of ancestors" approach above works just as well, and is easier as one doesn't have to deal with string building, regular
expressions, and escaping of characters. (Theoretically, materialized paths will be slightly faster.)
The best way to do this with MongoDB is to store the path as a string and then use regex queries. Simple regex expressions beginning with "^"
can be efficiently executed. As the path is a string, you will need to pick a delimiter character -- we use ',' below. For example:
> t = db.tree
test.tree
> // get entire tree -- we use sort() to make the order nice
> t.find().sort({path:1})
{ "_id" : "a", "path" : "a," }
{ "_id" : "b", "path" : "a,b," }
{ "_id" : "c", "path" : "a,b,c," }
{ "_id" : "d", "path" : "a,b,d," }
{ "_id" : "g", "path" : "a,b,g," }
{ "_id" : "e", "path" : "a,e," }
{ "_id" : "f", "path" : "a,e,f," }
{ "_id" : "g", "path" : "a,b,g," }
> t.ensureIndex( {path:1} )
> // find the node 'b' and all its descendents:

> t.find( { path : /â,b,/ } )
{ "_id" : "b", "path" : "a,b," }
{ "_id" : "c", "path" : "a,b,c," }
{ "_id" : "d", "path" : "a,b,d," }
{ "_id" : "g", "path" : "a,b,g," }
// or if its path not already known:

> b = t.findOne( { _id : "b" } )
{ "_id" : "b", "path" : "a,b," }
> t.find( { path : new RegExp("^" + b.path) } )
{ "_id" : "b", "path" : "a,b," }
{ "_id" : "c", "path" : "a,b,c," }
{ "_id" : "d", "path" : "a,b,d," }
{ "_id" : "g", "path" : "a,b,g," }
Ruby example: http://github.com/banker/newsmonger/blob/master/app/models/comment.rb
acts_as_nested_set
See http://api.rubyonrails.org/classes/ActiveRecord/Acts/NestedSet/ClassMethods.html
This pattern is best for datasets that rarely change as modifications can require changes to many documents.
See Also
Sean Cribbs blog post (source of several ideas on this page).
Optimization
Optimizing A Simple Example
Optimization #1: Create an index
Optimization #2: Limit results
Optimization #3: Select only relevant fields
Using the Profiler
Optimizing Statements that Use count()
Increment Operations
Circular Fixed Size Collections
Server Side Code Execution
Explain
Hint
See Also
Optimizing A Simple Example
This section describes proper techniques for optimizing database performance.
Let's consider an example. Suppose our task is to display the front page of a blog - we wish to display headlines of the 10 most recent posts. Let's
assume the posts have a timestamp field ts.
The simplest thing we could write might be:

articles = db.posts.find().sort({ts:-1}); // get blog posts in reverse time order
for (var i=0; i< 10; i++) {

print(articles[i].getSummary());
}
Optimization #1: Create an index
Our first optimization should be to create an index on the key that is being used for the sorting:
db.posts.ensureIndex({ts:1});
With an index, the database is able to sort based on index information, rather than having to check each document in the collection directly. This
is much faster.
Optimization #2: Limit results
MongoDB cursors return results in groups of documents that we'll call 'chunks'. The chunk returned might contain more than 10 objects - in some
cases, much more. These extra objects are a waste of network transmission and resources both on the app server and the database.
As we know how many results we want, and that we do not want all the results, we can use the limit() method for our second optimization.
articles = db.posts.find().sort({ts:-1}).limit(10); // 10 results maximum
Now, we'll only get 10 results returned to client.
Optimization #3: Select only relevant fields
The blog post object may be very large, with the post text and comments embedded. Much better performance will be achieved by selecting only
the fields we need:
articles = db.posts.find({}, {ts:1,title:1,author:1,abstract:1}).sort({ts:-1}).limit(10);

articles.forEach( function(post) { print(post.getSummary()); } );
The above code assumes that the getSummary() method only references the fields listed in the find() method.
Note if you fetch only select fields, you have a partial object. An object in that form cannot be updated back to the database:
a_post = db.posts.findOne({}, Post.summaryFields);

a_post.x = 3;
db.posts.save(a_post); // error, exception thrown
Using the Profiler
MongoDB includes a database profiler which shows performance characteristics of each operation against the database. Using the profiler you
can find queries (and write operations) which are slower than they should be; use this information, for example, to determine when an index is
needed. See the [Performance Tuning] section of the Mongo Developers' Guide for more information.
Optimizing Statements that Use count()
To speed operations that rely on count(), create an index on the field involved in the count query expression.
db.posts.ensureIndex({author:1});
db.posts.find({author:"george"}).count();
Increment Operations
MongoDB supports simple object field increment operations; basically, this is an operation indicating "increment this field in this document at the
server". This can be much faster than fetching the document, updating the field, and then saving it back to the server and are particularly useful
for implementing real time counters. See the Updates section of the Mongo Developers' Guide for more information.
Circular Fixed Size Collections
MongoDB provides a special circular collection type that is pre-allocated at a specific size. These collections keep the items within well-ordered
even without an index, and provide very high-speed writes and reads to the collection. Originally designed for keeping log files - log events are
stored in the database in a circular fixed size collection - there are many uses for this feature. See the Capped Collections section of the Mongo
Developers' Guide for more information.
Server Side Code Execution
Occasionally, for maximal performance, you may wish to perform an operation in process on the database server to eliminate client/server
network turnarounds. These operations are covered in the Server-Side Processing section of the Mongo Developers' Guide.
Explain
A great way to get more information on the performance of your database queries is to use the $explain feature. This will display "explain plan"
type info about a query from the database.
When using the mongo - The Interactive Shell, you can find out this "explain plan" via the explain() function called on a cursor. The result will
be a document that contains the "explain plan".
db.collection.find(query).explain();
provides information such as the following:
{
"cursor" : "BasicCursor",
"indexBounds" : [ ],
"nscanned" : 57594,
"nscannedObjects" : 57594,
"n" : 3 ,
"millis" : 108
}
This will tell you the type of cursor used (BtreeCursor is another type – which will include a lower & upper bound), the number of records the
DB had to examine as part of this query, the number of records returned by the query, and the time in milliseconds the query took to execute.
Hint
While the mongo query optimizer often performs very well, explicit "hints" can be used to force mongo to use a specified index, potentially
improving performance in some situations. When you have a collection indexed and are querying on multiple fields (and some of those fields are
indexed), pass the indexe as a hint to the query. You can do this in two different ways. You may either set it per query, or set it for the entire
collection.
To set the hint for a particular query, call the hint() function on the cursor before accessing any data, and specify a document with the key to be
used in the query:
db.collection.find({user:u, foo:d}).hint({user:1});
Be sure to Index
For the above hints to work, you need to have run ensureIndex() to index the collection on the user field.
To force the query optimizer to not use indexes (do a table scan), use:
> db.collection.find().hint({$natural:1})
See Also

Query Optimizer
currentOp()

MongoDB records have a certain amount of overhead per object (BSON document) in a collection. This overhead is normally insignificant, but if
your objects are tiny (just a few bytes, maybe one or two fields) it would not be. Below are some suggestions on how to optimize storage
efficiently in such situations.
Using the _id Field Explicitly
Mongo automatically adds an object ID to each document and sets it to a unique value. Additionally this field in indexed. For tiny objects this
takes up significant space.
The best way to optimize for this is to use _id explicitly. Take one of your fields which is unique for the collection and store its values in _id. By
doing so, you have explicitly provided IDs. This will effectively eliminate the creation of a separate _id field. If your previously separate field was
indexed, this eliminates an extra index too.
Using Small Field Names
Consider a record
{ last_name : "Smith", best_score: 3.9 }
The strings "last_name" and "best_score" will be stored in each object's BSON. Using shorter strings would save space:
{ lname : "Smith", score : 3.9 }
Would save 9 bytes per document. This of course reduces expressiveness to the programmer and is not recommended unless you have a
collection where this is of significant concern.
Field names are not stored in indexes as indexes have a predefined structure. Thus, shortening field names will not help the size of indexes. In
general it is not necessary to use short field names.
Combining Objects
Fundamentally, there is a certain amount of overhead per document in MongoDB. One technique is combining objects. In some cases you may
be able to embed objects in other objects, perhaps as arrays of objects. If your objects are tiny this may work well, but will only make sense for
certain use cases.
Query Optimizer
The MongoDB query optimizer generates query plans for each query submitted by a client. These plans are executed to return results. Thus,
MongoDB supports ad hoc queries much like say, MySQL.
The database uses an interesting approach to query optimization though. Traditional approaches (which tend to be cost-based and statistical) are
not used, as these approaches have a couple of problems.
First, the optimizer might consistently pick a bad query plan. For example, there might be correlations in the data of which the optimizer is
unaware. In a situation like this, the developer might use a query hint.
Also with the traditional approach, query plans can change in production with negative results. No one thinks rolling out new code without testing
is a good idea. Yet often in a production system a query plan can change as the statistics in the database change on the underlying data. The
query plan in effect may be a plan that never was invoked in QA. If it is slower than it should be, the application could experience an outage.
The Mongo query optimizer is different. It is not cost based -- it does not model the cost of various queries. Instead, the optimizer simply tries
different query plans and learn which ones work well. Of course, when the system tries a really bad plan, it may take an extremely long time to
run. To solve this, when testing new plans, MongoDB executes multiple query plans in parallel. As soon as one finishes, it terminates the other
executions, and the system has learned which plan is good. This works particularly well given the system is non-relational, which makes the
space of possible query plans much smaller (as there are no joins).
Sometimes a plan which was working well can work poorly -- for example if the data in the database has changed, or if the parameter values to
the query are different. In this case, if the query seems to be taking longer than usual, the database will once again run the query in parallel to try
different plans.
This approach adds a little overhead, but has the advantage of being much better at worst-case performance.
See Also
MongoDB hint() and explain() operators
Querying
One of MongoDB's best capabilities is its support for dynamic (ad hoc) queries. Systems that support dynamic queries don't require any special
indexing to find data; users can find data using any criteria. For relational databases, dynamic queries are the norm. If you're moving to MongoDB
from a relational databases, you'll find that many SQL queries translate easily to MongoDB's document-based query language.
Query Expression Objects

Query Options
Field Selection
Sorting
Skip and Limit
slaveOk
Cursors
More info
Quick Reference Card
See Also
Query Expression Objects
MongoDB supports a number of query objects for fetching data. Queries are expressed as BSON documents which indicate a query pattern. For
example, suppose we're using the MongoDB shell and want to return every document in the users collection. Our query would look like this:
db.users.find({})
In this case, our selector is an empty document, which matches every document in the collection. Here's a more selective example:
db.users.find({'last_name': 'Smith'})
Here our selector will match every document where the last_name attribute is 'Smith.'
MongoDB support a wide array of possible document selectors. For more examples, see the MongoDB Tutorial or the section on Advanced
Queries. If you're working with MongoDB from a language driver, see the driver docs:
Query Options
Field Selection
In addition to the query expression, MongoDB queries can take some additional arguments. For example, it's possible to request only certain
fields be returned. If we just wanted the social security numbers of users with the last name of 'Smith,' then from the shell we could issue this
query:
// retrieve ssn field for documents where last_name == 'Smith':

db.users.find({last_name: 'Smith'}, {'ssn': 1});
// retrieve all fields *except* the thumbnail field, for all documents:
db.users.find({}, {thumbnail:0});
Note the _id field is always returned even when not explicitly requested.
Sorting
MongoDB queries can return sorted results. To return all documents and sort by last name in ascending order, we'd query like so:
db.users.find({}).sort({last_name: 1});
Skip and Limit
MongoDB also supports skip and limit for easy paging. Here we skip the first 20 last names, and limit our result set to 10:
db.users.find().skip(20).limit(10);
db.users.find({}, {}, 10, 20); // same as above, but less clear
slaveOk
When querying a replica pair or replica set, drivers route their requests to the master mongod by default; to perform a query against an
(arbitrarily-selected) slave, the query can be run with the slaveOk option. Here's how to do so in the shell:
db.getMongo().setSlaveOk(); // enable querying a slave

db.users.find(...)
Note: some language drivers permit specifying the slaveOk option on each find(), others make this a connection-wide setting. See your
language's driver for details.
Cursors
Database queries, performed with the find() method, technically work by returning a cursor. Cursors are then used to iteratively retrieve all the
documents returned by the query. For example, we can iterate over a cursor in the mongo shell like this:
> var cur = db.example.find();

> cur.forEach( function(x) { print(tojson(x))});
{"n" : 1 , "_id" : "497ce96f395f2f052a494fd4"}
{"n" : 2 , "_id" : "497ce971395f2f052a494fd5"}
{"n" : 3 , "_id" : "497ce973395f2f052a494fd6"}
>
More info
This was just an introduction to querying in Mongo. For the full details please look in at the pages in the "Querying" sub-section to the right of your
screen.
Quick Reference Card
Download the Query and Update Modifier Quick Reference Card
See Also
Queries and Cursors

Advanced Queries
Query Optimizer

Queries in MongoDB are expressed as JSON (BSON). Usually we think of query object as the equivalent of a SQL "WHERE" clause:
> db.users.find( { x : 3, y : "abc" } ).sort({x:1}); // select * from users where x=3 and y='abc'
order by x asc;
However, the MongoDB server actually looks at all the query parameters (ordering, limit, etc.) as a single object. In the above example from the
mongo shell, the shell is adding some syntactic sugar for us. Many of the drivers do this too. For example the above query could also be written:
> db.users.find( { $query : { x : 3, y : "abc" }, $orderby : { x : 1 } } );
The possible specifies in the query object are:
$query - the evaluation or "where" expression

$orderby - sort order desired
$hint - hint to query optimizer
$explain - if true, return explain plan results instead of query results
$snapshot - if true, "snapshot mode"

By default on a find operation, the entire object is returned. However we may also request that only certain fields be returned. This is somewhat
analogous to the list of column specifiers in a SQL SELECT statement (projection). Regardless of what field specifiers are included, the _id field
is always returned.
// select z from things where x="john"

db.things.find( { x : "john" }, { z : 1 } );
Field Negation
We can say "all fields except x" – for example to remove specific fields that you know will be large:
// get all posts about 'tennis' but without the comments field
db.posts.find( { tags : 'tennis' }, { comments : 0 } );
Dot Notation
You can retrieve partial sub-objects via Dot Notation.
> t.find({})
{ "_id" : ObjectId("4c23f0486dad1c3a68457d20"), "x" : { "y" : 1, "z" : [ 1, 2, 3 ] } }
> t.find({}, {'x.y':1})
{ "_id" : ObjectId("4c23f0486dad1c3a68457d20"), "x" : { "y" : 1 } }
Retrieving a Subrange of Array Elements
You can use the $slice operator to retrieve a subrange of elements in an array.
New in MongoDB 1.5.1
db.posts.find({}, {comments:{$slice: 5}}) // first 5 comments

db.posts.find({}, {comments:{$slice: -5}}) // last 5 comments
db.posts.find({}, {comments:{$slice: [20, 10]}}) // skip 20, limit 10
db.posts.find({}, {comments:{$slice: [-20, 10]}}) // 20 from end, limit 10
See Also
example slice1
Advanced Queries
Introduction
$slice operator
Conditional Operators : <, <=, >, >=
Conditional Operator : $ne
Conditional Operator : $in
Conditional Operator : $nin
Conditional Operator : $mod
Conditional Operator: $all
Conditional Operator : $size
Conditional Operator: $exists
Conditional Operator: $type
$or
Regular Expressions
Value in an Array
Conditional Operator: $elemMatch
Value in an Embedded Object
Meta operator: $not
Javascript Expressions and $where
sort()
limit()
skip()
snapshot()
count()
group()
Special operators
See Also
Introduction
MongoDB offers a rich query environment with lots of features. This page lists some of those features.
Queries in MongoDB are represented as JSON-style objects, very much like the documents we actually store in the database. For example:
// i.e., select * from things where x=3 and y="foo"

db.things.find( { x : 3, y : "foo" } );
Note that any of the operators on this page can be combined in the same query document. For example, to find all document where j is not equal
to 3 and k is greater than 10, you'd query like so:
db.things.find({j: {$ne: 3}, k: {$gt: 10} });
By default on a find operation, the entire document/object is returned. However we may also request that only certain fields are returned. Note
that the _id field is always returned automatically.
// select z from things where x=3

db.things.find( { x : 3 }, { z : 1 } );
You can also remove specific fields that you know will be large:
// get all posts about mongodb without comments

db.posts.find( { tags : 'mongodb' }, { comments : 0 } );
$slice operator
You can use the $slice operator to retrieve a subset of elements in an array.
db.posts.find({}, {comments:{$slice: 5}}) // first 5 comments

db.posts.find({}, {comments:{$slice: -5}}) // last 5 comments
db.posts.find({}, {comments:{$slice: [20, 10]}}) // skip 20, limit 10
db.posts.find({}, {comments:{$slice: [-20, 10]}}) // 20 from end, limit 10
More examples at example slice1
Conditional Operators : <, <=, >, >=
Use these special forms for greater than and less than comparisons in queries, since they have to be represented in the query document:
db.collection.find({ "field" : { $gt: value } } ); // greater than : field > value
db.collection.find({ "field" : { $lt: value } } ); // less than : field < value
db.collection.find({ "field" : { $gte: value } } ); // greater than or equal to : field >= value
db.collection.find({ "field" : { $lte: value } } ); // less than or equal to : field <= value
For example:
db.things.find({j : {$lt: 3}});

db.things.find({j : {$gte: 4}});
You can also combine these operators to specify ranges:
db.collection.find({ "field" : { $gt: value1, $lt: value2 } } ); // value1 < field < value
Conditional Operator : $ne
Use $ne for "not equals".
db.things.find( { x : { $ne : 3 } } );
Conditional Operator : $in
The $in operator is analogous to the SQL IN modifier, allowing you to specify an array of possible matches.
db.collection.find( { "field" : { $in : array } } );
Let's consider a couple of examples. From our things collection, we could choose to get a subset of documents based upon the value of the 'j'
key:
db.things.find({j:{$in: [2,4,6]}});
Suppose the collection updates is a list of social network style news items; we want to see the 10 most recent updates from our friends. We
might invoke:
db.updates.ensureIndex( { ts : 1 } ); // ts == timestamp
var myFriends = myUserObject.friends; // let's assume this gives us an array of DBRef's of my friends
var latestUpdatesForMe = db.updates.find( { user : { $in : myFriends } } ).sort( { ts : -1 }
).limit(10);
Conditional Operator : $nin
The $nin operator is similar to $in except that it selects objects for which the specified field does not have any value in the specified array. For
example
db.things.find({j:{$nin: [2,4,6]}});
would match {j:1,b:2} but not {j:2,c:9}.
Conditional Operator : $mod
The $mod operator allows you to do fast modulo queries to replace a common case for where clauses. For example, the following $where query:
db.things.find( "this.a % 10 == 1")
can be replaced by:

db.things.find( { a : { $mod : [ 10 , 1 ] } } )
Conditional Operator: $all
The $all operator is similar to $in, but instead of matching any value in the specified array all values in the array must be matched. For
example, the object
{ a: [ 1, 2, 3 ] }
would be matched by
db.things.find( { a: { $all: [ 2, 3 ] } } );
but not
db.things.find( { a: { $all: [ 2, 3, 4 ] } } );
An array can have more elements than those specified by the $all criteria. $all specifies a minimum set of elements that must be matched.
Conditional Operator : $size
The $size operator matches any array with the specified number of elements. The following example would match the object { a:["foo"]},
since that array has just one element:
db.things.find( { a : { $size: 1 } } );
You cannot use $size to find a range of sizes (for example: arrays with more than 1 element). If you need to query for a range, create an extra
size field that you increment when you add elements.
Conditional Operator: $exists
Check for existence (or lack thereof) of a field.
db.things.find( { a : { $exists : true } } ); // return object if a is present

db.things.find( { a : { $exists : false } } ); // return if a is missing
Currently $exists is not able to use an index. Indexes on other fields are still used.
Conditional Operator: $type
The $type operator matches values based on their BSON type.
db.things.find( { a : { $type : 2 } } ); // matches if a is a string

db.things.find( { a : { $type : 16 } } ); // matches if a is an int
Possible types are:
Type Name Type Number
Double 1
String 2
Object 3
Array 4
Binary data 5
Object id 7
Boolean 8
Date 9
Null 10
Regular expression 11
JavaScript code 13
Symbol 14
JavaScript code with scope 15
32-bit integer 16
Timestamp 17
64-bit integer 18
Min key 255
Max key 127
For more information on types and BSON in general, see http://www.bsonspec.org.
$or
The $or operator lets you use a boolean or expression to do queries. You give $or a list of expressions, any of which can satisfy the query.
Simple:
db.foo.find( { $or : [ { a : 1 } , { b : 2 } ] } )
With another field
db.foo.find( { name : "bob" , $or : [ { a : 1 } , { b : 2 } ] } )
The $or operator retrieves matches for each or clause individually and eliminates duplicates when returning results. A number of $or
optimizations are planned for 1.8. See this thread for details.
Regular Expressions
You may use regexes in database query expressions:
db.customers.find( { name : /acme.*corp/i } );
For simple prefix queries (also called rooted regexps) like /^prefix/, the database will use an index when available and appropriate (much like
most SQL databases that use indexes for a LIKE 'prefix%' expression). This only works if you don't have i (case-insensitivity) in the flags.
While /â/, /â.*/, and /â.*$/ are equivalent and will all use an index in the same way, the later two require scanning the
whole string so they will be slower. The first format can stop scanning after the prefix is matched.
MongoDB uses PCRE for regular expressions. Valid flags are:
i - Case insensitive. Letters in the pattern match both upper and lower
case letters.
m - Multiline. By default, Mongo treats the subject string as consisting of a single line of characters (even if it actually contains newlines).
The "start of line" metacharacter (^) matches only at the start of the string, while the "end of line" metacharacter ($) matches only at the
end of the string, or before a terminating newline.
When m it is set, the "start of line" and "end of line" constructs match immediately following or immediately before internal newlines in the
subject string, respectively, as well as at the very start and end. If there are no newlines in a subject string, or no occurrences of ^ or $ in
a pattern, setting m has no effect.
x - Extended. If set, whitespace data characters in the pattern are totally ignored except when escaped or inside a character class.
Whitespace does not include the VT character (code 11). In addition, characters between an unescaped # outside a character class and
the next newline, inclusive, are also ignored.
This option makes it possible to include comments inside complicated patterns. Note, however, that this applies only to data characters.
Whitespace characters may never appear within special character sequences in a pattern, for example within the sequence (?( which
introduces a conditional subpattern.
Value in an Array
To look for the value "red" in an array field colors:
db.things.find( { colors : "red" } );
That is, when "colors" is inspected, if it is an array, each value in the array is checked. This technique may be mixed with the embedded object
technique below.
Conditional Operator: $elemMatch
Version 1.3.1 and higher.
Use $elemMatch to check if an element in an array matches the specified match expression.
> t.find( { x : { $elemMatch : { a : 1, b : { $gt : 1 } } } } )

{ "_id" : ObjectId("4b5783300334000000000aa9"),
"x" : [ { "a" : 1, "b" : 3 }, 7, { "b" : 99 }, { "a" : 11 } ]
}
Note that a single array element must match all the criteria specified; thus, the following query is semantically different in that each criteria can
match a different element in the x array:
> t.find( { "x.a" : 1, "x.b" : { $gt : 1 } } )
See the dot notation page for more.
Value in an Embedded Object
For example, to look author.name=="joe" in a postings collection with embedded author objects:
db.postings.find( { "author.name" : "joe" } );
See the dot notation page for more.
Meta operator: $not
Version 1.3.3 and higher.
The $not meta operator can be used to negate the check performed by a standard operator. For example:
db.customers.find( { name : { $not : /acme.*corp/i } } );
db.things.find( { a : { $not : { $mod : [ 10 , 1 ] } } } );
Javascript Expressions and $where
In addition to the structured query syntax shown so far, you may specify query expressions as Javascript. To do so, pass a string containing a
Javascript expression to find(), or assign such a string to the query object member $where. The database will evaluate this expression for
each object scanned. When the result is true, the object is returned in the query results.
For example, the following statements all do the same thing:
db.myCollection.find( { a : { $gt: 3 } } );
db.myCollection.find( { $where: "this.a > 3" } );
db.myCollection.find("this.a > 3");
f = function() { return this.a > 3; } db.myCollection.find(f);
Javascript executes more slowly than the native operators listed on this page, but is very flexible. See the server-side processing page for more
information.
sort()
sort() is analogous to the ORDER BY statement in SQL - it requests that items be returned in a particular order. We pass sort() a key
pattern which indicates the desired order for the result.
db.myCollection.find().sort( { ts : -1 } ); // sort by ts, descending order
sort() may be combined with the limit() function. In fact, if you do not have a relevant index for the specified key pattern, limit() is
recommended as there is a limit on the size of sorted results when an index is not used. Without a limit(), or index, a full in-memory sort must
be done but by using a limit() it reduces the memory and increases the speed of the operation by using an optimized sorting algorithm.
limit()
limit() is analogous to the LIMIT statement in MySQL: it specifies a maximum number of results to return. For best performance, use limit()
whenever possible. Otherwise, the database may return more objects than are required for processing.
db.students.find().limit(10).forEach( function(student) { print(student.name + "<p>"); } );
In the shell (and most drivers), a limit of 0 is equivalent to setting no limit at all.
skip()
The skip() expression allows one to specify at which object the database should begin returning results. This is often useful for implementing
"paging". Here's an example of how it might be used in a JavaScript application:
function printStudents(pageNumber, nPerPage) {

print("Page: " + pageNumber);
db.students.find().skip((pageNumber-1)*nPerPage).limit(nPerPage).forEach( function(student) {
print(student.name + "<p>"); } );
}
snapshot()
Indicates use of snapshot mode for the query. Snapshot mode assures no duplicates are returned, or objects missed, which were present at both
the start and end of the query's execution (even if the object were updated). If an object is new during the query, or deleted during the query, it
may or may not be returned, even with snapshot mode.
Note that short query responses (less than 1MB) are always effectively snapshotted.
Currently, snapshot mode may not be used with sorting or explicit hints.
count()
The count() method returns the number of objects matching the query specified. It is specially optimized to perform the count in the MongoDB
server, rather than on the client side for speed and efficiency:
nstudents = db.students.find({'address.state' : 'CA'}).count();
Note that you can achieve the same result with the following, but the following is slow and inefficient as it requires all documents to be put into
memory on the client, and then counted. Don't do this:
nstudents = db.students.find({'address.state' : 'CA'}).toArray().length; // VERY BAD: slow and uses
excess memory
On a query using skip() and limit(), count ignores these parameters by default. Use count(true) to have it consider the skip and limit values in the
calculation.
n = db.students.find().skip(20).limit(10).count(true);
group()
The group() method is analogous to GROUP BY in SQL. group() is more flexible, actually, allowing the specification of arbitrary reduction
operations. See the Aggregation section of the Mongo Developers' Guide for more information.
Special operators
only return the index key
db.foo.find()._addSpecial("$returnKey" , true )
limit the number of items to scan
t.find()._addSpecial( "$maxScan" , 50 )
See Also
* Optimizing Queries (including explain() and hint())

Dot Notation vs. Subobjects
Array Element by Position
Matching with $elemMatch
MongoDB is designed for store JSON-style objects. The database understands the structure of these objects and can reach into them to evaluate
query expressions.
Let's suppose we have some objects of the form:
> db.persons.findOne()
{ name: "Joe", address: { city: "San Francisco", state: "CA" } ,
likes: [ 'scuba', 'math', 'literature' ] }
Querying on a top-level field is straightforward enough using Mongo's JSON-style query objects:
> db.persons.find( { name : "Joe" } )
But what about when we need to reach into embedded objects and arrays? This involves a bit different way of thinking about queries than one
would do in a traditional relational DBMS. To reach into embedded objects, we use a "dot notation":
> db.persons.find( { "address.state" : "CA" } )
Reaching into arrays is implicit: if the field being queried is an array, the database automatically assumes the caller intends to look for a value
within the array:
> db.persons.find( { likes : "math" } )
We can mix these styles too, as in this more complex example:

> db.blogposts.findOne()
{ title : "My First Post", author: "Jane",
comments : [{ by: "Abe", text: "First" },
{ by : "Ada", text : "Good post" } ]
}
> db.blogposts.find( { "comments.by" : "Ada" } )
We can also create indexes of keys on these fields:
db.persons.ensureIndex( { "address.state" : 1 } );
db.blogposts.ensureIndex( { "comments.by" : 1 } );
Dot Notation vs. Subobjects
Suppose there is an author id, as well as name. To store the author field, we can use an object:
> db.blog.save({ title : "My First Post", author: {name : "Jane", id : 1}})
If we want to find any authors named Jane, we use the notation above:
> db.blog.findOne({"author.name" : "Jane"})
To match only objects with these exact keys and values, we use an object:
db.blog.findOne({"author" : {"name" : "Jane", "id" : 1}})
Note that
db.blog.findOne({"author" : {"name" : "Jane"}})
will not match, as subobjects have to match exactly (it would match an object with one field: { "name" : "Jane"}). Note that the embedded
document must also have the same key order, so:
db.blog.findOne({"author" : {"id" : 1, "name" : "Jane"}})
will not match, either. This can make subobject matching unwieldy in languages whose default document representation is unordered.
Array Element by Position
Array elements also may be accessed by specific array position:
// i.e. comments[0].by == "Abe"

> db.blogposts.find( { "comments.0.by" : "Abe" } )
(The above examples use the mongo shell's Javascript syntax. The same operations can be done in any language for which Mongo has a driver
available.)
Matching with $elemMatch
Using the $elemMatch query operator (mongod >= 1.3.1), you can match an entire document within an array. This is best illustrated with an
example. Suppose you have the following two documents in your collection:
// Document 1
{ "foo" : [
{
"shape" : "square",
"color" : "purple",
"thick" : false
},
{
"shape" : "circle",
"color" : "red",
"thick" : true
}
] }
// Document 2
{ "foo" : [
{
"shape" : "square",
"color" : "red",
"thick" : true
},
{
"shape" : "circle",
"color" : "purple",
"thick" : false
}
] }
You want to query for a purple square, and so you write the following:
db.foo.find({"foo.shape": "square", "foo.color": "purple"})
The problem with this query is that it will match the second in addition to matching the first document. In other words, the standard query syntax
won't restrict itself to a single document within the foo array. As mentioned above, subobjects have to match exactly, so
db.foo.find({foo: {"shape": "square", "color": "purple"} } )
won't help either, since there's a third attribute specifying thickness.
To match an entire document within the foo array, you need to use $elemMatch. To properly query for a purple square, you'd use $elemMatch like
so:
db.foo.find({foo: {"$elemMatch": {shape: "square", color: "purple"}}})
The query will return the first document, which contains the purple square you're looking for.

Introduction
Multikeys (Indexing Values in an Array)
Text Search
Comparison to Full Text Search Engines
Real World Examples
Introduction
Mongo provides some functionality that is useful for text search and tagging.
Multikeys (Indexing Values in an Array)
The Mongo multikey feature can automatically index arrays of values. Tagging is a good example of where this feature is useful. Suppose you
have an article object/document which is tagged with some category names:
obj = {
name: "Apollo",
text: "Some text about Apollo moon landings",
tags: [ "moon", "apollo", "spaceflight" ]
}
and that this object is stored in db.articles. The command
db.articles.ensureIndex( { tags: 1 } );
will index all the tags on the document, and create index entries for "moon", "apollo" and "spaceflight" for that document.
You may then query on these items in the usual way:
> print(db.articles.findOne( { tags: "apollo" } ).name);

Apollo
The database creates an index entry for each item in the array. Note an array with many elements (hundreds or thousands) can make inserts very
expensive. (Although for the example above, alternate implementations are equally expensive.)
Text Search
It is fairly easy to implement basic full text search using multikeys. What we recommend is having a field that has all of the keywords in it,
something like:
{ title : "this is fun" ,

_keywords : [ "this" , "is" , "fun" ]
}
Your code must split the title above into the keywords before saving. Note that this code (which is not part of Mongo DB) could do stemming, etc.
too. (Perhaps someone in the community would like to write a standard module that does this...)
Comparison to Full Text Search Engines
MongoDB has interesting functionality that makes certain search functions easy. That said, it is not a dedicated full text search engine.
For example, dedicated engines provide the following capabilities:
built-in text stemming

ranking of queries matching various numbers of terms (can be done with MongoDB, but requires user supplied code to do so)
bulk index building
Bulk index building makes building indexes fast, but has the downside of not being realtime. MongoDB is particularly well suited for problems
where the search should be done in realtime. Traditional tools are often not good for this use case.
Real World Examples
The Business Insider web site uses MongoDB for its blog search function in production.
Mark Watson's opinions on Java, Ruby, Lisp, AI, and the Semantic Web - A recipe example in Ruby.
min and max Query Specifiers

The min() and max() functions may be used in conjunction with an index to constrain query matches to those having index keys between the
min and max keys specified. The min() and max() functions may be used individually or in conjunction. The index to be used may be specified
with a hint() or one may be inferred from pattern of the keys passed to min() and/or max().
db.f.find().min({name:"barry"}}.max({name:"larry"}).hint({name:1});
db.f.find().min({name:"barry"}}.max({name:"larry"});
db.f.find().min({last_name:"smith",first_name:"john"}};
If you're using the standard query syntax, you must distinguish between the $min and $max keys and the query selector itself. See here:
db.f.find({$min: {name:"barry"}, $max: {name:"larry"}, $query:{}});
The min() value is included in the range and the max() value is excluded.
Normally, it is much preferred to use $gte and $lt rather than to use min and max, as min and max require a corresponding index. Min and max
are primarily useful for compound keys: it is difficult to express the last_name/first_name example above without this feature (it can be done using
$where).
min and max exist primarily to support the mongos (sharding) process.

Query objects in Mongo by default AND expressions together. Before 1.5.3 MongoDB did not include an "$or" operator for such queries, however
there are ways to express such queries.
$in
The $in operator indicates a "where value in ..." expression. For expressions of the form x == a OR x == b, this can be represented as
{ x : { $in : [ a, b ] } }
$where
We can provide arbitrary Javascript expressiosn to the server via the $where operator. This provides a means to perform OR operations. For
example in the mongo shell one might invoke:
db.mycollection.find( { $where : function() { return this.a == 3 || this.b == 4; } } );
The following syntax is briefer and also works; however, if additional structured query components are present, you will need the $where form:
db.mycollection.find( function() { return this.a == 3 || this.b == 4; } );
$or
The $or operator lets you use a boolean or expression to do queries. You give $or a list of expressions, any of which can satisfy the query.
Simple:
db.foo.find( { $or : [ { a : 1 } , { b : 2 } ] } )
With another field
db.foo.find( { name : "bob" , $or : [ { a : 1 } , { b : 2 } ] } )
The $or operator retrieves matches for each or clause individually and eliminates duplicates when returning results.
See Also
Advanced Queries
Queries and Cursors

Queries to MongoDB return a cursor, which can be iterated to retrieve results. The exact way to query will vary with language driver. Details below
focus on queries from the MongoDB shell (i.e. the mongo process).
The shell find() method returns a cursor object which we can then iterate to retrieve specific documents from the result. We use hasNext()
and next() methods for this purpose.
for( var c = db.parts.find(); c.hasNext(); ) {

print( c.next());
}
Additionally in the shell, forEach() may be used with a cursor:
db.users.find().forEach( function(u) { print("user: " + u.name); } );
Array Mode in the Shell
Note that in some languages, like JavaScript, the driver supports an "array mode". Please check your driver documentation for specifics.
In the db shell, to use the cursor in array mode, use array index [] operations and the length property.
Array mode will load all data into RAM up to the highest index requested. Thus it should not be used for any query which can return very large
amounts of data: you will run out of memory on the client.
You may also call toArray() on a cursor. toArray() will load all objects queries into RAM.
Getting a Single Item
The shell findOne() method fetches a single item. Null is returned if no item is found.
findOne() is equivalent in functionality to:
function findOne(coll, query) {

var cursor = coll.find(query).limit(1);
return cursor.hasNext() ? cursor.next() : null;
}
Tip: If you only need one row back and multiple match, findOne() is efficient, as it performs the limit() operation, which limits the objects
returned from the database to one.
Querying Embedded Objects
To find an exact match of an entire embedded object, simply query for that object:
db.order.find( { shipping: { carrier: "usps" } } );
The above query will work if { carrier: "usps" } is an exact match for the entire contained shipping object. If you wish to match any sub-object with
shipping.carrier == "usps", use this syntax:
db.order.find( { "shipping.carrier" : "usps" } );
See the dot notation docs for more information.
Greater Than / Less Than

db.myCollection.find( { a : { $gt : 3 } } );
db.myCollection.find( { a : { $gte :3 } } );
db.myCollection.find( { a : { $lt :3 } } );
db.myCollection.find( { a : { $lte :3 } } ); // a <= 3
Latent Cursors and Snapshotting
A latent cursor has (in addition to an initial access) a latent access that occurs after an intervening write operation on the database collection (i.e.,
an insert, update, or delete). Under most circumstances, the database supports these operations.
Conceptually, a cursor has a current position. If you delete the item at the current position, the cursor automatically skips its current position
forward to the next item.
Mongo DB cursors do not provide a snapshot: if other write operations occur during the life of your cursor, it is unspecified if your application will
see the results of those operations or not. See the snapshot docs for more information.
Auditing allocated cursors
Information on allocated cursors may be obtained using the {cursorInfo:1} command.
db.runCommand({cursorInfo:1})
See Also
Advanced Queries
Multikeys in the HowTo
Tailable Cursors
MongoDB has a feature known as tailable cursors which are similar to the Unix "tail -f" command.
Tailable means the cursor is not closed once all data is retrieved. Rather, the cursor marks the last known object's position and you can resume
using the cursor later, from where that object was located, provided more data is available.
Like all "latent cursors", the cursor may become invalid at any point -- for example if the last object returned is deleted. Thus, you should be
prepared to requery if the cursor is dead. You can determine if a cursor is dead by checking its id. An id of zero indicates a dead cursor.
In addition, the cursor may be dead upon creation if the initial query returns no matches. In this case a requery is required to create a persistent
tailable cursor.
Tailable cursors are only allowed on capped collections and can only return objects in natural order.
If the field you wish to "tail" is indexed, simply requerying for "field > value" is already quite efficient. Tailable will be slightly
faster in situations such as that. However, if the field is not indexed, tailable provides a huge improvement in performance.
Mongo replication uses this feature to follow the end of the master server's replication op log collection -- the tailable feature eliminates the need
to create an index for the oplog at the master, which would slow log writes.
C++ example:
#include "client/dbclient.h"
#include "util/goodies.h"
using namespace mongo;
void tail(DBClientBase& conn, const char *ns) {

BSONElement lastId = minKey.firstElement();
Query query = Query();
auto_ptr<DBClientCursor> c =
conn.query(ns, query, 0, 0, 0, QueryOption_CursorTailable);
while( 1 ) {
if( !c->more() ) {
if( c->isDead() ) {
break; // we need to requery
}
// all data (so far) exhausted, wait for more

sleepsecs(1);
continue;
}
BSONObj o = c->next();
lastId = o["_id"];
cout << o.toString() << endl;
}
}
Server-side Code Execution

$where Clauses and Functions in Queries
Restrictions
Map/Reduce
Using db.eval()
Examples
Storing functions server-side
Notes on Concurrency
Running .js files via a mongo shell instance on the server
Mongo supports the execution of code inside the database process.
$where Clauses and Functions in Queries
In addition to the regular document-style query specification for find() operations, you can also express the query either as a string containing a
SQL-style WHERE predicate clause, or a full JavaScript function.
When using this mode of query, the database will call your function, or evaluate your predicate clause, for each object in the collection.
In the case of the string, you must represent the object as "this" (see example below). In the case of a full JavaScript function, you use the normal
JavaScript function syntax.
The following four statements in mongo - The Interactive Shell are equivalent:
db.myCollection.find( { a : { $gt: 3 } } );
db.myCollection.find( { $where: "this.a > 3" });
db.myCollection.find( "this.a > 3" );
db.myCollection.find( { $where: function() { return this.a > 3;}});
The first statement is the preferred form. It will be at least slightly faster to execute because the query optimizer can easily interpret that query and
choose an index to use.
You may mix data-style find conditions and a function. This can be advantageous for performance because the data-style expression will be
evaluated first, and if not matched, no further evaluation is required. Additionally, the database can then consider using an index for that
condition's field. To mix forms, pass your evaluation function as the $where field of the query object. For example:
db.myCollection.find( { active: true, $where: function() { return obj.credits - obj.debits < 0; } } );
You may mix data-style find conditions and a function. This can be advantageous for performance because the data-style expression will be
evaluated first, and if not matched, no further evaluation is required. Additionally, the database can then consider using an index for that
condition's field. For example:
db.myCollection.find( { active: true, $where: "this.credits - this.debits < 0" } );
Restrictions
Do not write to the collection being inspected from the $where expression.
Map/Reduce
MongoDB supports Javascript-based map/reduce operations on the server. See the map/reduce documentation for more information.
Using db.eval()
Use map/reduce instead of db.eval() for long running jobs. db.eval blocks other operations!
db.eval() is used to evaluate a function (written in JavaScript) at the database server.

This is useful if you need to touch a lot of data lightly. In that scenario, network transfer of the data could be a bottleneck.
db.eval() returns the return value of the function that was invoked at the server. If invocation fails an exception is thrown.
For a trivial example, we can get the server to add 3 to 3:
> db.eval( function() { return 3+3; } );

6
>
Let's consider an example where we wish to erase a given field, foo, in every single document in a collection. A naive client-side approach would
be something like
function my_erase() {
db.things.find().forEach( function(obj) {
delete obj.foo;
db.things.save(obj);
} );
}
my_erase();
Calling my_erase() on the client will require the entire contents of the collection to be transmitted from server to client and back again.
Instead, we can pass the function to eval(), and it will be called in the runtime environment of the server. On the server, the db variable is set to
the current database:
db.eval(my_erase);
Examples
> myfunc = function(x){ return x; };
> db.eval( myfunc, {k:"asdf"} );

{ k : "asdf" }
> db.eval( myfunc, "asdf" );

"asdf"
> db.eval( function(x){ return x; }, 2 );

2.0
If an error occurs on the evaluation (say, a null pointer exception at the server), an exception will be thrown of the form:
{ dbEvalException: { errno : -3.0 , errmsg : "invoke failed" , ok : 0.0 } }
Example of using eval() to do equivalent of the Mongo count() function:
function mycount(collection) {
return db.eval( function(){return db[collection].find({},{_id:ObjId()}).length();} );
}
Example of using db.eval() for doing an atomic increment, plus some calculations:
function inc( name , howMuch ){

return db.eval(
function(){
var t = db.things.findOne( { name : name } );
t = t || { name : name , num : 0 , total : 0 , avg : 0 };
t.num++;
t.total += howMuch;
t.avg = t.total / t.num;
db.things.save( t );
return t;
}
);
}
db.things.remove( {} );
print( tojson( inc( "eliot" , 2 )) );
print( tojson( inc( "eliot" , 3 )) );
Storing functions server-side
in version 1.1.1 and above
There is a special system collection called system.js that can store JavaScript function to be re-used. To store a function, you would do:
db.system.js.save( { _id : "foo" , value : function( x , y ){ return x + y; } } );
_id is the name of the function, and is unique per database.
Once you do that, you can use foo from any JavaScript context (db.eval, $where, map/reduce)
See http://github.com/mongodb/mongo/tree/master/jstests/storefunc.js for a full example
Notes on Concurrency
eval() blocks the entire mongod process while running. Thus, its operations are atomic but prevent other operations from processing.
When more concurrency is needed consider using map/reduce instead of eval().

Running .js files via a mongo shell instance on the server
This is a good technique for performing batch administrative work. Run mongo on the server, connecting via the localhost interface. The
connection is then very fast and low latency. This is friendlier than db.eval() as db.eval() blocks other operations.

"Natural order" is defined as the database's native ordering of objects in a collection.
When executing a find() with no parameters, the database returns objects in forward natural order.
For standard tables, natural order is not particularly useful because, although the order is often close to insertion order, it is not guaranteed to be.
However, for Capped Collections, natural order is guaranteed to be the insertion order. This can be very useful.
In general, the natural order feature is a very efficient way to store and retrieve data in insertion order (much faster than say, indexing on a
timestamp field). But remember, the collection must be capped for this to work.
In addition to forward natural order, items may be retrieved in reverse natural order. For example, to return the 50 most recently inserted items
(ordered most recent to less recent) from a capped collection, you would invoke:
> c=db.cappedCollection.find().sort({$natural:-1}).limit(50)
Sorting can also be done on arbitrary keys in any collection. For example, this sorts by 'name' ascending, then 'age' descending:
> c=db.collection.find().sort({name : 1, age : -1})
See Also
The Capped Collections section of this Guide

Advanced Queries
Aggregation
Mongo includes utility functions which provide server-side count, distinct, and group by operations. More advanced aggregate functions
can be crafted using MapReduce.
Count
Distinct
Group
Examples
Using Group from Various Languages
Map/Reduce
See Also
Count
count() returns the number of objects in a collection or matching a query. If a document selector is provided, only the number of matching
documents will be returned.
size() is like count() but takes into consideration any limit() or skip() specified for the query.
db.collection.count(selector);
For example:
print( "# of objects: " + db.mycollection.count() );

print( db.mycollection.count( {active:true} );
count is faster if an index exists for the condition in the selector. For example, to make the count on active fast, invoke
db.mycollection.ensureIndex( {active:1} );
Distinct
The distinct command returns returns a list of distinct values for the given key across a collection.
Command is of the form:
{ distinct : <collection_name>, key : <key>[, query : <query>] }
although many drivers have a helper function for distinct.
> db.addresses.insert({"zip-code": 10010})

> // shell helper:

> db.addresses.distinct("zip-code");
[ 10010, 99701 ]
> // running as a command manually:

> db.runCommand( { distinct: 'addresses', key: 'zip-code' } )
{ "values" : [ 10010, 99701 ], "ok" : 1 }
distinct may also reference a nested key:
> db.comments.save({"user": {"points": 25}})

> db.comments.distinct("user.points");
[ 25, 31 ]
You can add an optional query parameter to distinct as well
> db.address.distinct( "zip-code" , { age : 30 } )
Note: the distinct command results are returned as a single BSON object. If the results could be large (> 4 megabytes), use map/reduce instead.
Group
Note: currently one must use map/reduce instead of group() in sharded MongoDB configurations.
group returns an array of grouped items. The command is similar to SQL's group by. The SQL statement
select a,b,sum(c) csum from coll where active=1 group by a,b
corresponds to the following in MongoDB:
db.coll.group(
{key: { a: true, b:true },
cond: { active:1 },
reduce: function(obj,prev) { prev.csum += obj.c; },
initial: { csum: 0 }
});
Note: the result is returned as a single BSON object and for this reason must be fairly small – less than 10,000 keys, else you will get an
exception. For larger grouping operations without limits, please use map/reduce .
group takes a single object parameter containing the following fields:
key: Fields to group by.

reduce: The reduce function aggregates (reduces) the objects iterated. Typical operations of a reduce function include summing and
counting. reduce takes two arguments: the current document being iterated over and the aggregation counter object. In the example
above, these arguments are named obj and prev.
initial: initial value of the aggregation counter object.
keyf: An optional function returning a "key object" to be used as the grouping key. Use this instead of key to specify a key that is not an
existing member of the object (or, to access embedded members). Set in lieu of key.
cond: An optional condition that must be true for a row to be considered. This is essentially a find() query expression object. If null, the
reduce function will run against all rows in the collection.
finalize: An optional function to be run on each item in the result set just before the item is returned. Can either modify the item (e.g.,
add an average field given a count and a total) or return a replacement object (returning a new object with just _id and average fields).
See jstests/group3.js for examples.
To order the grouped data, simply sort it client-side upon return. The following example is an implementation of count() using group().
function gcount(collection, condition) {

var res =
db[collection].group(
{ key: {},
initial: {count: 0},
reduce: function(obj,prev){ prev.count++;},
cond: condition } );
// group() returns an array of grouped items. here, there will be a single
// item, as key is {}.
return res[0] ? res[0].count : 0;
}
Examples
The examples assume data like this:
{ domain: "www.mongodb.org"
, invoked_at: {d:"2009-11-03", t:"17:14:05"}
, response_time: 0.05
, http_action: "GET /display/DOCS/Aggregation"
}
Show me stats for each http_action in November 2009:
db.test.group(
{ cond: {"invoked_at.d": {$gt: "2009-11", $lt: "2009-12"}}
, key: {http_action: true}
, initial: {count: 0, total_time:0}
, reduce: function(doc, out){ out.count++; out.total_time+=doc.response_time }
, finalize: function(out){ out.avg_time = out.total_time / out.count }
} );
[
{
"http_action" : "GET /display/DOCS/Aggregation",
"count" : 1,
"total_time" : 0.05,
"avg_time" : 0.05
}
]
Show me stats for each domain for each day in November 2009:
db.test.group(
{ cond: {"invoked_at.d": {$gt: "2009-11", $lt: "2009-12"}}
, key: {domain: true, invoked_at.d: true}
, initial: {count: 0, total_time:0}
, reduce: function(doc, out){ out.count++; out.total_time+=doc.response_time }
, finalize: function(out){ out.avg_time = out.total_time / out.count }
} );
[
{
"http_action" : "GET /display/DOCS/Aggregation",
"count" : 1,
"total_time" : 0.05,
"avg_time" : 0.05
}
]
Using Group from Various Languages
Some language drivers provide a group helper function. For those that don't, one can manually issue the db command for group. Here's an
example using the Mongo shell syntax:
> db.foo.find()
{"_id" : ObjectId( "4a92af2db3d09cb83d985f6f") , "x" : 1}
{"_id" : ObjectId( "4a92af2fb3d09cb83d985f70") , "x" : 3}
{"_id" : ObjectId( "4a92afdab3d09cb83d985f71") , "x" : 3}
> db.$cmd.findOne({group : {
... ns : "foo",
... cond : {},
... key : {x : 1},
... initial : {count : 0},
... $reduce : function(obj,prev){prev.count++;}}})
{"retval" : [{"x" : 1 , "count" : 1},{"x" : 3 , "count" : 2}] , "count" : 3 , "keys" : 2 , "ok" : 1}
If you use the database command with keyf (instead of key) it must be prefixed with a $. For example:
db.$cmd.findOne({group : {
... ns : "foo",
... $keyf : function(doc) { return {"x" : doc.x}; },
... initial : {count : 0},
... $reduce : function(obj,prev) { prev.count++; }}})
Map/Reduce
MongoDB provides a MapReduce facility for more advanced aggregation needs. CouchDB users: please note that basic queries in MongoDB do
not use map/reduce.
See Also
jstests/eval2.js for an example of group() usage

Advanced Queries
Removing
Removing Objects from a Collection
To remove objects from a collection, use the remove() function in the mongo shell. (Other drivers offer a similar
function, but may call the function "delete". Please check your driver's documentation ).
remove() is like find() in that it takes a JSON-style query document as an argument to select which documents are removed. If you call
remove() without a document argument, or with an empty document {}, it will remove all documents in the collection. Some examples :
db.things.remove({}); // removes all
db.things.remove({n:1}); // removes all where n == 1
If you have a document in memory and wish to delete it, the most efficient method is to specify the item's document _id value as a criteria:
db.things.remove({_id: myobject._id});
You may be tempted to simply pass the document you wish to delete as the selector, and this will work, but it's inefficient.
References
If a document is deleted, any existing references to the document will still exist in the database. These references will return null
when evaluated.
Concurrency and Remove
v1.3+ supports concurrent operations while a remove runs. If a simultaneous update (on the same collection) grows an object which matched the
remove criteria, the updated object may not be removed (as the operations are happening at approximately the same time, this may not even be
surprising). In situations where this is undesireable, pass {$atomic : true} in your filter expression:
db.videos.remove( { rating : { $lt : 3.0 }, $atomic : true } )
The remove operation is then completely atomic – however, it will also block other operations while executing.
Updating
MongoDB supports atomic, in-place updates as well as more traditional updates for replacing an entire document.
update()
save() in the mongo shell
Modifier Operations
$inc
$set
$unset
$push
$pushAll
$addToSet
$pop
$pull
$pullAll
The $ positional operator
Upserts with Modifiers
Pushing a Unique Value
Checking the Outcome of an Update Request
Notes
Object Padding
Blocking
See Also
update()
update() replaces the document matching criteria entirely with objNew. If you only want to modify some fields, you should use the atomic
modifiers below.
Here's the MongoDB shell syntax for update():
db.collection.update( criteria, objNew, upsert, multi )
Arguments:
criteria - query which selects the record to update;

objNew - updated object or $ operators (e.g., $inc) which manipulate the object
upsert - if this should be an "upsert"; that is, if the record does not exist, insert it
multi - if all documents matching criteria should be updated
If you are coming from SQL, be aware that by default, update() only modifies the first matched object. If you want to modify all
matched objects you need to use the multi flag
save() in the mongo shell
The save() command in the mongo shell provides a shorthand syntax to perform a single object update with upsert:
// x is some JSON style object

db.mycollection.save(x); // updates if exists; inserts if new
save() does an upsert if x has an _id field and an insert if it does not. Thus, normally, you will not need to explicitly request upserts, just use
save().
Upsert means "update if present; insert if missing".
myColl.update( { name: "Joe" }, { name: "Joe", age: 20 }, true );
Modifier Operations
Modifier operations are highly-efficient and useful when updating existing values; for instance, they're great for incrementing a number.
So, while a conventional implementation does work:
var j=myColl.findOne( { name: "Joe" } );

j.n++;
myColl.save(j);
a modifier update has the advantages of avoiding the latency involved in querying and returning the object. The modifier update also features
operation atomicity and very little network data transfer.
To perform an atomic update, simply specify any of the special update operators (which always start with a '$' character) with a relevant update
document:
db.people.update( { name:"Joe" }, { $inc: { n : 1 } } );
The preceding example says, "Find the first document where 'name' is 'Joe' and then increment 'n' by one."
While not shown in the examples, most modifier operators will accept multiple field/value pairs when one wishes to modify
multiple fields. For example, the following operation would set x to 1 and y to 2:
{ $set : { x : 1 , y : 2 } }
Also, multiple operators are valid too:
{ $set : { x : 1 }, $inc : { y : 1 } }
$inc
{ $inc : { field : value } }
increments field by the number value if field is present in the object, otherwise sets field to the number value.
$set
{ $set : { field : value } }
sets field to value. All datatypes are supported with $set.
$unset
{ $unset : { field : 1} }
Deletes a given field. v1.3+
$push
{ $push : { field : value } }
appends value to field, if field is an existing array, otherwise sets field to the array [value] if field is not present. If field is present
but is not an array, an error condition is raised.
$pushAll
{ $pushAll : { field : value_array } }
appends each value in value_array to field, if field is an existing array, otherwise sets field to the array value_array if field is not
present. If field is present but is not an array, an error condition is raised.
$addToSet
{ $addToSet : { field : value } }
Adds value to the array only if its not in the array already.
To add many valuest.update
{ $addToSet : { a : { $each : [ 3 , 5 , 6 ] } } }
$pop
{ $pop : { field : 1 } }
removes the last element in an array (ADDED in 1.1)
{ $pop : { field : -1 } }
removes the first element in an array (ADDED in 1.1) |
$pull
{ $pull : { field : _value } }
removes all occurrences of value from field, if field is an array. If field is present but is not an array, an error condition is raised.
$pullAll
{ $pullAll : { field : value_array } }

removes all occurrences of each value in value_array from field, if field is an array. If field is present but is not an array, an error
condition is raised.
The $ positional operator
Version 1.3.4+ only.
The $ operator (by itself) means "position of the matched array item in the query". Use this to find an array member and then manipulate it. For
example:
> t.find()
{ "_id" : ObjectId("4b97e62bf1d8c7152c9ccb74"), "title" : "ABC",
"comments" : [ { "by" : "joe", "votes" : 3 }, { "by" : "jane", "votes" : 7 } ] }
> t.update( {'comments.by':'joe'}, {$inc:{'comments.$.votes':1}}, false, true )
> t.find()
{ "_id" : ObjectId("4b97e62bf1d8c7152c9ccb74"), "title" : "ABC",
"comments" : [ { "by" : "joe", "votes" : 4 }, { "by" : "jane", "votes" : 7 } ] }
Currently the $ operator only applies to the first matched item in the query. For example:
> t.find();
{ "_id" : ObjectId("4b9e4a1fc583fa1c76198319"), "x" : [ 1, 2, 3, 2 ] }
> t.update({x: 2}, {$inc: {"x.$": 1}}, false, true);
> t.find();
{ "_id" : ObjectId("4b9e4a1fc583fa1c76198319"), "x" : [ 1, 3, 3, 2 ] }
The positional operator cannot be combined with an upsert since it requires a matching array element. If your update results in an insert then
the "$" will literally be used as the field name.
Using "$unset" with an expression like this "array.$" will result in the array item becoming null, not being removed. You can
issue an update with "{$pull:{x:null}}" to remove all nulls.
> t.insert({x: [1,2,3,4,3,2,3,4]})

> t.find()
{ "_id" : ObjectId("4bde2ad3755d00000000710e"), "x" : [ 1, 2, 3, 4, 3, 2, 3, 4 ] }
> t.update({x:3}, {$unset:{"x.$":1}})
> t.find()
{ "_id" : ObjectId("4bde2ad3755d00000000710e"), "x" : [ 1, 2, null, 4, 3, 2, 3, 4 ] }
Upserts with Modifiers
You may use upsert with a modifier operation. In such a case, the modifiers will be applied to the update criteria member and the resulting
object will be inserted. The following upsert example may insert the object {name:"Joe",x:1,y:1}.
db.people.update( { name:"Joe" }, { $inc: { x:1, y:1 } }, true );
There are some restrictions. A modifier may not reference the _id field, and two modifiers within an update may not reference the same field, for
example the following is not allowed:
db.people.update( { name:"Joe" }, { $inc: { x: 1 }, $set: { x: 5 } } );
Pushing a Unique Value
To add a value to an array only if not already present:
Starting in 1.3.3, you can do

update( {_id:'joe'},{"$addToSet": { tags : "baseball" } } );
For older versions, add $ne : <value> to your query expression:
update( {_id:'joe', tags: {"$ne": "baseball"}},

{"$push": { tags : "baseball" } } );
Checking the Outcome of an Update Request
As described above, a non-upsert update may or may not modify an existing object. An upsert will either modify an existing object or insert a new
object. The client may determine if its most recent message on a connection updated an existing object by subsequently issuing a
getlasterror command ( db.runCommand( "getlasterror" ) ). If the result of the getlasterror command contains an
updatedExisting field, the last message on the connection was an update request. If the updatedExisting field's value is true, that update
request caused an existing object to be updated; if updatedExisting is false, no existing object was updated. An upserted field will contain
the new _id value if an insert is performed (new as of 1.5.4).
Notes
Object Padding
When you update an object in MongoDB, the update occurs in-place if the object has not grown in size. This is good for insert performance if the
collection has many indexes.
Mongo adaptively learns if objects in a collection tend to grow, and if they do, it adds some padding to prevent excessive movements. This
statistic is tracked separately for each collection.
Blocking
Staring in 1.5.2, multi updates yield occasionally so you can safely update large amounts of data. If you want a multi update to be truly atomic,
you can use the $atomic flag.
See Also
Atomic Operations
Atomic Operations
Modifier operations
"Update if Current"
The ABA Nuance
"Insert if Not Present"
Find and Modify (or Remove)
Applying to Multiple Objects Atomically
MongoDB supports atomic operations on single documents. MongoDB does not support traditional locking and complex transactions for a
number of reasons:
First, in sharded environments, distributed locks could be expensive and slow. Mongo DB's goal is to be lightweight and fast.
We dislike the concept of deadlocks. We want the system to be simple and predictable without these sort of surprises.
We want Mongo DB to work well for realtime problems. If an operation may execute which locks large amounts of data, it might stop
some small light queries for an extended period of time. (We don't claim Mongo DB is perfect yet in regards to being "real-time", but we
certainly think locking would make it even harder.)
MongoDB does support several methods of manipulating single documents atomically, which are detailed below.
Modifier operations
The Mongo DB update command supports several modifiers, all of which atomically update an element in a document. They include:
$set - set a particular value

$unset - set a particular value (since 1.3.0)
$inc - increment a particular value by a certain amount
$push - append a value to an array
$pushAll - append several values to an array
$pull - remove a value(s) from an existing array
$pullAll - remove several value(s) from an existing array
These modifiers are convenient ways to perform certain operations atomically.
"Update if Current"
Another strategy for atomic updates is "Update if Current". This is what an OS person would call Compare and Swap. For this we
1. Fetch the object.

2. Modify the object locally.
3. Send an update request that says "update the object to this new value if it still matches its old value".
Should the operation fail, we might then want to try again from step 1.
For example, suppose we wish to fetch one object from inventory. We want to see that an object is available, and if it is, deduct it from the
inventory. The following code demonstrates this using mongo shell syntax (similar functions may be done in any language):
> t=db.inventory
> s = t.findOne({sku:'abc'})
{"_id" : "49df4d3c9664d32c73ea865a" , "sku" : "abc" , "qty" : 30}
> qty_old = s.qty;
> --s.qty;
> t.update({_id:s._id, qty:qty_old}, s); db.$cmd.findOne({getlasterror:1});
{"err" : , "updatedExisting" : true , "n" : 1 , "ok" : 1} // it worked
For the above example, we likely don't care the exact sku quantity as long as it is as least as great as the number to deduct. Thus the following
code is better, although less general -- we can get away with this as we are using a predefined modifier operation ($inc). For more general
updates, the "update if current" approach shown above is recommended.
> t.update({sku:"abc",qty:{$gt:0}}, { $inc : { qty : -1 } } ) ; db.$cmd.findOne({getlasterror:1})

{"err" : , "updatedExisting" : true , "n" : 1 , "ok" : 1} // it worked
> t.update({sku:"abcz",qty:{$gt:0}}, { $inc : { qty : -1 } } ) ; db.$cmd.findOne({getlasterror:1})
{"err" : , "updatedExisting" : false , "n" : 0 , "ok" : 1} // did not work
The ABA Nuance
In the first of the examples above, we basically did "update object if qty is unchanged". However, what if since our read, sku had been modified?
We would then overwrite that change and lose it!
There are several ways to avoid this problem ; it's mainly just a matter of being aware of the nuance.
1. Use the entire object in the update's query expression, instead of just the _id and qty field.
2. Use $set to set the field we care about. If other fields have changed, they won't be effected then.
3. Put a version variable in the object, and increment it on each update.
4. When possible, use a $ operator instead of an update-if-current sequence of operations.
"Insert if Not Present"
Another optimistic concurrency scenario involves inserting a value when not already there. When we have a unique index constraint for the
criteria, we can do this. The following example shows how to insert monotonically increasing _id values into a collection using optimistic
concurrency:
function insertObject(o) {
x = db.myCollection;
while( 1 ) {
// determine next _id value to try
var c = x.find({},{_id:1}).sort({_id:-1}).limit(1);
var i = c.hasNext() ? c.next()._id + 1 : 1;
o._id = i;
x.insert(o);
var err = db.getLastErrorObj();
if( err && err.code ) {
if( err.code == 11000 /* dup key */ )
continue;
else
print( "unexpected error inserting data: " + tojson(err));
}
break;
}
}
See the findandmodify Command documentation for more information.
Applying to Multiple Objects Atomically
You can use multi-update to apply the same modifier to each object. This will be pseudo-atomic by default. Its one operation, but occasonaly
someone else can get it. To make it full atomic you can use $atomic
non atomic
db.foo.update( { x : 1 } , { $inc : { y : 1 } } , false , true );
atomic
db.foo.update( { x : 1 , $atomic : 1 } , { $inc : { y : 1 } } , false , true );
v1.3.0 and higher
MongoDB 1.3+ supports a "find, modify, and return" command. This command can be used to atomically modify a document (at most one) and
return it. Note that, by default, the document returned will not include the modifications made on the update.
If you don't need to return the document, you can use Update (which can affect multiple documents, as well).
The general form is
db.runCommand( { findandmodify : <collection>,

<options> } )
The MongoDB shell includes a helper method, findAndModify(), for executing the command. Some drivers provide helpers also.
At least one of the update or remove parameters is required; the other arguments are optional.
Argument Description Default

query a filter for the query {}
sort if multiple docs match, choose the first one in the specified sort order as the object to manipulate {}
remove set to a true to remove the object before returning N/A
update a modifier object N/A
new set to true if you want to return the modified object rather than the original. Ignored for remove. false
fields see Retrieving a Subset of Fields (1.5.0+) All fields
upsert create object if it doesn't exist. examples (1.5.4+) false
The sort option is useful when storing queue-like data. Let's take the example of fetching the highest priority job that hasn't been grabbed yet and
atomically marking it as grabbed:
job = db.jobs.findAndModify({
query: {inprogress:false},
sort:{priority:-1},
update: {$set: {inprogress: true, started: new Date()}}
});
You could also simply remove the object to be returned, but be careful. If the client crashes before processing the job, the document will be lost
forever.
job = db.jobs.findAndModify({sort:{priority:-1}, remove:true}}});
See the tests for more examples.
If your driver doesn't provide a helper function for this command, run the command directly with something like this:
job = db.runCommand({ findandmodify : "jobs",

sort : { priority : -1 },
remove : true
}).value;
Sharding limitations
findandmodify will behave the same when called through a mongos as long as the collection it is modifying is unsharded. If the collection is
sharded, then the query must contain the shard key. This is the same as regular sharded updates.
See Also
Atomic Operations
Updating Data in Mongo

Updating a Document in the mongo Shell with save()
Embedding Documents Directly in Documents
Database References
Updating a Document in the mongo Shell with save()
As shown in the previous section, the save() method may be used to save a new document to a collection. We can also use save() to update
an existing document in a collection.
Continuing with the example database from the last section, lets add new information to the document {name:"mongo"} that already is in the
collection.
> var mongo = db.things.findOne({name:"mongo"});
> print(tojson(mongo));
{"_id" : "497dab624ee47b3a675d2d9c" , "name" : "mongo"}
> mongo.type = "database";
database
> db.things.save(mongo);
> db.things.findOne({name:"mongo"});
{"_id" : "497dab624ee47b3a675d2d9c" , "name" : "mongo" , "type" : "database"}
>
This was a simple example, adding a string valued element to the existing document. When we called save(), the method saw that the
document already had an "_id" field, so it simply performed an update on the document.
In the next two sections, we'll show how to embed documents within documents (there are actually two different ways), as well as show how to
query for documents based on values of embedded documents.
Embedding Documents Directly in Documents
As another example of updating an existing document, lets embed a document within an existing document in the collection. We'll keep working
with the original {name:"mongo"} document for simplicity.
> var mongo = db.things.findOne({name:"mongo"});

{"_id" : "497da93d4ee47b3a675d2d9b" , "name" : "mongo", "type" : "database"}
> mongo.data = { a:1, b:2};
{"a" : 1 , "b" : 2}
> db.things.findOne({name:"mongo"});
{"_id" : "497da93d4ee47b3a675d2d9b" , "name" : "mongo" , "type" : "database", "data" : {"a" : 1 , "b"
: 2}}
>
As you can see, we added new data to the mongo document, adding {a:1, b:2} under the key "data".
Note that the value of "data" is a document itself - it is embedded in the parent mongo document. With BSON, you may nest and embed
documents to any level. You can also query on embedded document fields, as shown here:
> db.things.findOne({ "data.a" : 1});

{"_id" : "497da93d4ee47b3a675d2d9b" , "name" : "mongo" , "data" : {"a" : 1 , "b" : 2}}
> db.things.findOne({ "data.a" : 2});
>
Note that the second findOne() doesn't return anything, because there are no documents that match.
Database References
Alternatively, a document can reference other documents which are not embedded via a database reference, which is analogous to a foreign key
in a relational database. A database reference (or "DBRef" for short), is a reference implemented according to the Database References. Most
drivers support helpers for creating DBRefs. Some also support additional functionality, like dereference helpers and auto-referencing. See
specific driver documentation for examples / more information
In the previous section we saw that the statement
obj_a.x = obj_b;
Lets repeat the above example, but create a document and place in a different collection, say otherthings, and embed that as a reference in our
favorite "mongo" object under the key "otherdata":
// first, save a new doc in the 'otherthings' collection
> var other = { s : "other thing", n : 1};

> db.otherthings.save(other);
> db.otherthings.find();
{"_id" : "497dbcb36b27d59a708e89a4" , "s" : "other thing" , "n" : 1}
// now get our mongo object, and add the 'other' doc as 'otherthings'
> var mongo = db.things.findOne();

{"_id" : "497dab624ee47b3a675d2d9c" , "name" : "mongo" , "type" : "database" , "data" : {"a" : 1 , "b"
: 2}}
> mongo.otherthings = new DBRef( 'otherthings' , other._id );
{"s" : "other thing" , "n" : 1 , "_id" : "497dbcb36b27d59a708e89a4"}
> db.things.findOne().otherthings.fetch();
: 2} , "otherthings" : {"_id" : "497dbcb36b27d59a708e89a4" , "s" : "other thing" , "n" : 1}}
// now, lets modify our 'other' document, save it again, and see that when the dbshell
// gets our mongo object and prints it, if follows the dbref and we have the new value
> other.n = 2;
2
> db.otherthings.save(other);
> db.otherthings.find();
{"_id" : "497dbcb36b27d59a708e89a4" , "s" : "other thing" , "n" : 2}
> db.things.findOne().otherthings.fetch();
: 2} , "otherthings" : {"_id" : "497dbcb36b27d59a708e89a4" , "s" : "other thing" , "n" : 2}}
>
MapReduce
Map/reduce in MongoDB is useful for batch manipulation of data and aggregation operations. It is similar in spirit to using something like Hadoop
with all input coming from a collection and output going to a collection. Often, in a situation where you would have used GROUP BY in SQL,
map/reduce is the right tool in MongoDB.
Indexing and standard queries in MongoDB are separate from map/reduce. If you have used CouchDB in the past, note this is a
big difference: MongoDB is more like MySQL for basic querying and indexing. See the queries and indexing documentation for
those operations.
Overview
Map Function
Reduce Function
Finalize Function
Examples
Shell Example 1
Shell Example 2
More Examples
Note on Permanent Collections
Parallelism
Presentations
See Also
Overview
Version 1.1.1 and above

map/reduce is invoked via a database command. The database creates a temporary collection to hold output of the operation. The collection is
cleaned up when the client connection closes, or when explicitly dropped. Alternatively, one can specify a permanent output collection name. map
and reduce functions are written in JavaScript and execute on the server.
Command syntax:
db.runCommand(
{ mapreduce : <collection>,
map : <mapfunction>,
reduce : <reducefunction>
[, query : <query filter object>]
[, sort : <sort the query. useful for optimization>]
[, limit : <number of objects to return from collection>]
[, out : <output-collection name>]
[, keeptemp: <true|false>]
[, finalize : <finalizefunction>]
[, scope : <object where fields go into javascript global scope >]
[, verbose : true]
}
);
keeptemp - if true, the generated collection is not treated as temporary. Defaults to false. When out is specified, the collection is
automatically made permanent.
finalize - function to apply to all the results when finished
verbose - provide statistics on job execution time
scope - can pass in variables that can be access from map/reduce/finalize example mr5
Result:
{ result : <collection_name>,
counts : {
input : <number of objects scanned>,
emit : <number of times emit was called>,
output : <number of items in output collection>
} ,
timeMillis : <job_time>,
ok : <1_if_ok>,
[, err : <errmsg_if_error>]
}
A command helper is available in the MongoDB shell :
db.collection.mapReduce(mapfunction,reducefunction[,options]);
map, reduce, and finalize functions are written in JavaScript.
Map Function
The map function references the variable this to inspect the current object under consideration. A map function must call emit(key,value) at
least once, but may be invoked any number of times, as may be appropriate.
function map(void) -> void
Reduce Function
The reduce function receives a key and an array of values. To use, reduce the received values, and return a result.
function reduce(key, value_array) -> value
The MapReduce engine may invoke reduce functions iteratively; thus, these functions must be idempotent. That is, the following must hold for
your reduce function:
for all k,vals : reduce( k, [reduce(k,vals)] ) == reduce(k,vals)
If you need to perform an operation only once, use a finalize function.
The output of emit (the 2nd param) and reduce should be the same format to make iterative reduce possible. If not, there will be
weird bugs that are hard to debug.
Currently, the return value from a reduce function cannot be an array (it's typically an object or a number).
Finalize Function
A finalize function may be run after reduction. Such a function is optional and is not necessary for many map/reduce cases. The finalize
function takes a key and a value, and returns a finalized value.
function finalize(key, value) -> final_value
In sharded environments, data processing of map/reduce operations runs in parallel on all shards.
Examples
Shell Example 1
The following example assumes we have an events collection with objects of the form:
{ time : <time>, user_id : <userid>, type : <type>, ... }
We then use MapReduce to extract all users who have had at least one event of type "sale":
> m = function() { emit(this.user_id, 1); }

> r = function(k,vals) { return 1; }
> res = db.events.mapReduce(m, r, { query : {type:'sale'} });
> db[res.result].find().limit(2)
{ "_id" : 8321073716060 , "value" : 1 }
{ "_id" : 7921232311289 , "value" : 1 }
If we also wanted to output the number of times the user had experienced the event in question, we could modify the reduce function like so:
> r = function(k,vals) {
... var sum=0;
... for(var i in vals) sum += vals[i];
... return sum;
... }
Note, here, that we cannot simply return vals.length, as the reduce may be called multiple times.
Shell Example 2
$ ./mongo
> db.things.insert( { _id : 1, tags : ['dog', 'cat'] } );
> db.things.insert( { _id : 2, tags : ['cat'] } );
> db.things.insert( { _id : 3, tags : ['mouse', 'cat', 'dog'] } );
> db.things.insert( { _id : 4, tags : [] } );
> // map function

> m = function(){
... this.tags.forEach(
... function(z){
... emit( z , { count : 1 } );
... }
... );
...};
> // reduce function

> r = function( key , values ){
... var total = 0;
... for ( var i=0; i<values.length; i++ )
... total += values[i].count;
... return { count : total };
...};
> res = db.things.mapReduce(m,r);

> res
{"timeMillis.emit" : 9 , "result" : "mr.things.1254430454.3" ,
"numObjects" : 4 , "timeMillis" : 9 , "errmsg" : "" , "ok" : 0}
> db[res.result].find()
{"_id" : "cat" , "value" : {"count" : 3}}
{"_id" : "dog" , "value" : {"count" : 2}}
{"_id" : "mouse" , "value" : {"count" : 1}}
> db[res.result].drop()
More Examples
example mr1
Finalize example: example mr2
Note on Permanent Collections
Even when a permanent collection name is specified, a temporary collection name will be used during processing. At map/reduce completion, the
temporary collection will be renamed to the permanent name atomically. Thus, one can perform a map/reduce job periodically with the same
target collection name without worrying about a temporary state of incomplete data. This is very useful when generating statistical output
collections on a regular basis.
Parallelism
As of right now, MapReduce jobs on a single mongod process are single threaded. This is due to a design limitation in current JavaScript engines.
We are looking into alternatives to solve this issue, but for now if you want to parallelize your MapReduce jobs, you will need to either use
sharding or do the aggregation client-side in your code.
Presentations
Map/reduce, geospatial indexing, and other cool features - Kristina Chodorow at MongoSF (April 2010)
See Also
Aggregation
Kyle's Map/Reduce basics

DRAFT - TO BE COMPLETED.
This guide provides instructions for using MongoDB batch data processing oriented features including map/reduce.
By "data processing", we generally mean operations performed on large sets of data, rather than small interactive operations.
Import
One can always write a program to load data of course, but the mongoimport utility also works for some situations. mongoimport supports
importing from json, csv, and tsv formats.
A common usage pattern would be to use mongoimport to load data in a relatively raw format and then use a server-side script ( db.eval() or
map/reduce ) to reduce the data to a more clean format.
See Also
Import/Export Tools
Server-Side Code Execution
Map/Reduce
mongo - The Interactive Shell

Introduction
More Information
Some Notes on Datatypes in the Shell
Numbers
Dates
BinData
Presentations
Introduction
The MongoDB distribution includes bin/mongo, the MongoDB interactive shell. This utility is a JavaScript shell that allows you to issue
commands to MongoDB from the command line. (Basically, it is an extended SpiderMonkey shell.)
The shell is useful for:
inspecting a database's contents

testing queries
creating indices
other administrative functions.
When you see sample code in this wiki and it looks like JavaScript, assume it is a shell example. See the driver syntax table for a chart that can
be used to convert those examples to any language.
More Information
Shell Overview
Shell Reference
Shell APIDocs
Some Notes on Datatypes in the Shell
Numbers
The shell treats all numbers as floating-point values. If you have long/integer BSON data from the database you may see something like this:
"bytes" : {
"floatApprox" : 575175
}
or something like this for larger numbers:

"bytes" : {
"floatApprox" : 5284376243087482000,
"top" : 1230364721,
"bottom" : 4240317554
}
In addition, setting/incrementing any number will (most likely) change the data type to a floating point value
Dates
The Date() function returns a string and a "new Date()" will return an object (which is what you should use to store values).
> Date()
Sun May 02 2010 19:07:40 GMT-0700 (Pacific Daylight Time)
> new Date()
"Sun May 02 2010 19:07:43 GMT-0700 (Pacific Daylight Time)"
> typeof(new Date())
object
> typeof(Date())
string
BinData
The BSON BinData datatype is represented via class BinData in the shell. Run help misc for more information.
Presentations
CRUD and the JavaScript Shell - Presentation by Mike Dirolf at MongoSF (April 2010)
Overview - The MongoDB Interactive Shell
Starting the Shell
The interactive shell is included in the standard MongoDB distribution. To start the shell, go into the root directory of the distribution and type
./bin/mongo
It might be useful to add mongo_distribution_root/bin to your PATH so you can just type mongo from anywhere.
If you start with no parameters, it connects to a database named "test" running on your local machine on the default port (27017). You can see the
db to which you are connecting by typing db:
./mongo
type "help" for help
> db
test
You can pass mongo an optional argument specifying the address, port and even the database to initially connect to:
./mongo foo connects to the foo database on your local machine
./mongo 192.168.13.7/foo connects to the foo database on 192.168.13.7
./mongo dbserver.mydomain.com/foo connects to the foo database on dbserver.mydomain.com
./mongo 192.168.13.7:9999/foo connects to the foo database on 192.168.13.7 on port 9999
Connecting
If you have not connected via the command line, you can use the following commands:
conn = new Mongo(host);

db = conn.getDB(dbname);
db.auth(username,password);
where host is a string that contains either the name or address of the machine you want to connect to (e.g. "192.168.13.7") or the machine and
port (e.g. "192.168.13.7:9999"). Note that host in an optional argument, and can be omitted if you want to connect to the database instance
running on your local machine. (e.g. conn = new Mongo() )
Alternatively you can use the connect helper method:
> db = connect("localhost:27020/mytestdb"); // example with a nonstandard port #
Basics Commands
The following are three basic commands that provide information about the available databases, and collections in a given database.
show dbs displays all the databases on the server you are connected to
use db_name switches to db_name on the same server
show collections displays a list of all the collections in the current database
Querying
mongo uses a JavaScript API to interact with the database. Because mongo is also a complete JavaScript shell, db is the variable that is the
current database connection.
To query a collection, you simply specify the collection name as a property of the db object, and then call the find() method. For example:
db.foo.find();
This will display the first 10 objects from the foo collection. Typing it after a find() will display the next 10 subsequent objects.
By setting the shellBatchSize you can change this:
DBQuery.shellBatchSize = #
Inserting Data
In order to insert data into the database, you can simply create a JavaScript object, and call the save() method. For example, to save an object {
name: "sara"} in a collection called foo, type:
db.foo.save({ name : "sara"});
Note that MongoDB will implicitly create any collection that doesn't already exist.
Modifying Data
Let's say you want to change someone's address. You can do this using the following mongo commands:
person = db.people.findOne( { name : "sara" } );

person.city = "New York";
db.people.save( person );
Deleting Data
db.foo.drop() drop the entire foo collection
db.foo.remove() remove all objects from the collection
db.foo.remove( { name : "sara" } ) remove objects from the collection where name is sara
Indexes
db.foo.getIndexKeys() get all fields that have indexes on them
db.foo.ensureIndex({ _field_ : 1 }) create an index on field if it doesn't exist
Line Continuation
If a line contains open '(' or '{' characters, the shell will request more input before evaluating:
> function f() {

... x = 1;
... }
>
You can press Ctrl-C to escape from "..." mode and terminate line entry.
See Also
MongoDB Shell Reference
dbshell Reference
Command Line
Special Command Helpers
Basic Shell Javascript Operations
Queries
Error Checking
Administrative Command Helpers
Opening Additional Connections
Miscellaneous
Examples
Command Line
--help Show command line options
--nodb Start without a db, you can connect later with new Mongo() or connect()
--shell After running a .js file from the command line, stay in the shell rather than terminating
Special Command Helpers
Non-javascript convenience macros:
help Show help
db.help() Show help on db methods
db.myColl.help() Show help on collection methods
show dbs Print a list of all databases on this server
use dbname Set the db variable to represent usage of dbname on the server
show collections Print a list of all collections for current database
show users Print a list of users for current database

show profile Print most recent profiling operations that took >= 1ms
Basic Shell Javascript Operations
db The variable that references the current database object / connection. Already defined for you in your
instance.
db.auth(user,pass) Authenticate with the database (if running in secure mode).
coll = db.collection Access a specific collection within the database.
cursor = coll.find(); Find all objects in the collection. See queries.
coll.remove(objpattern Remove matching objects from the collection.

); objpattern is an object specifying fields to match. E.g.: coll.remove( { name: "Joe" } );
coll.save(object) Save an object in the collection, or update if already there.

If your object has a presave method, that method will be called before the object is saved to the db (before
both updates and inserts)
coll.insert(object) Insert object in collection. No check is made (i.e., no upsert) that the object is not already present in the
collection.
coll.update(...) Update an object in a collection. See the Updating documentation; update() has many options.
coll.ensureIndex( { Creates an index on tab.name. Does nothing if index already exists.

name : 1 } )
coll.update(...)
coll.drop() Drops the collection coll
db.getSisterDB(name) Return a reference to another database using this same connection. Usage example:
db.getSisterDB('production').getCollectionNames()
Queries
coll.find() Find all.
it Continue iterating the last cursor returned from find().
coll.find( Find objects matching criteria in the collection. E.g.: coll.find( { name: "Joe" } );
criteria );
coll.findOne( Find and return a single object. Returns null if not found. If you want only one object returned, this is more efficient
criteria ); than just find() as limit(1) is implied. You may use regular expressions if the element type is a string, number,
or date: coll.find( { name: /joe/i } );
coll.find( Get just specific fields from the object. E.g.: coll.find( {}, {name:true} );
criteria, fields
);
coll.find().sort( Return results in the specified order (field ASC). Use -1 for DESC.
{field:1[, field
:1] });
coll.find( Return the objects matching criteria, sorted by field.

criteria ).sort(
{ field : 1 } )
coll.find( ... Limit result to n rows. Highly recommended if you need only a certain number of rows for best performance.
).limit(n)
coll.find( ... Skip n results.

).skip(n )
coll.count() Returns total number of objects in the collection.
coll.find( ... Returns the total number of objects that match the query. Note that the number ignores limit and skip; for example if
).count() 100 records match but the limit is 10, count() will return 100. This will be faster than iterating yourself, but still take
time.
More information: see queries.
Error Checking
db.getLastError() Returns error from the last operation.
db.getPrevError() Returns error from previous operations.
db.resetError() Clear error memory.
Administrative Command Helpers
db.cloneDatabase(fromhost) Clone the current database from the other host specified. fromhost database must be in
noauth mode.
db.copyDatabase(fromdb, todb, Copy fromhost/fromdb to todb on this server. fromhost must be in noauth mode.
fromhost)
db.repairDatabase() Repair and compact the current database. This operation can be very slow on large
databases.
db.addUser(user,pwd) Add user to current database.
db.getCollectionNames() get list of all collections.
db.dropDatabase() Drops the current database.
Opening Additional Connections
db = Open a new database connection. One may have multiple connections within a single shell,
connect("<host>:<port>/<dbname>") however, automatic getLastError reporting by the shell is done for the 'db' variable only. See
here for an example of connect().
conn = new Mongo("hostname") Open a connection to a new server. Use getDB() to select a database thereafter.
db = conn.getDB("dbname") Select a specific database for a connection
Miscellaneous
Object.bsonsize(db.foo.findOne()) prints the bson size of a db object (mongo version 1.3 and greater)
db.foo.findOne().bsonsize() prints the bson size of a db object (mongo versions predating 1.3)
For a full list of functions, see the shell API.
Examples
The MongoDB source code includes a jstests/ directory with many mongo shell scripts.
Developer FAQ
What's a "namespace"?
How do I copy all objects from one database collection to another?
If you remove an object attribute is it deleted from the store?
Are null values allowed?
Does an update fsync to disk immediately?
How do I do transactions/locking?
How do I do equivalent of SELECT count * and GROUP BY?
What are so many "Connection Accepted" messages logged?
What RAID should I use?
Can I run on Amazon EBS? Any issues?
Why are my data files so large?
Do I Have to Worry About SQL Injection

How does concurrency work
What is the Compare Order for BSON Types
Also check out Markus Gattol's excellent FAQ on his website.
What's a "namespace"?
MongoDB stores BSON objects in collections. The concatenation of the database name and the collection name (with a period in between) is
called a namespace.
For example, acme.users is a namespace, where acme is the database name, and users is the collection name. Note that periods can occur in
collection names, so a name such as acme.blog.posts is legal too (in that case blog.posts is the collection name.
How do I copy all objects from one database collection to another?
See below. The code below may be ran server-side for high performance with the eval() method.
db.myoriginal.find().forEach( function(x){db.mycopy.save(x)} );
If you remove an object attribute is it deleted from the store?
Yes, you remove the attribute and then re-save() the object.
Are null values allowed?
For members of an object, yes. You cannot add null to a database collection though as null isn't an object. You can add {}, though.
Does an update fsync to disk immediately?
No, writes to disk are lazy by default. A write may hit disk a couple of seconds later. For example, if the database receives a thousand increments
to an object within one second, it will only be flushed to disk once. (Note fsync options are available though both at the command line and via
getLastError.)
How do I do transactions/locking?
MongoDB does not use traditional locking or complex transactions with rollback, as it is designed to be lightweight and fast and predictable in its
performance. It can be thought of as analogous to the MySQL MyISAM autocommit model. By keeping transaction support extremely simple,
performance is enhanced, especially in a system that may run across many servers.
The system provides alternative models for atomically making updates that are sufficient for many common use cases. See the wiki page
Atomics Operations for detailed information.
How do I do equivalent of SELECT count * and GROUP BY?
See aggregation.
What are so many "Connection Accepted" messages logged?
If you see a tremendous number of connection accepted messages in the mongod log, that means clients are repeatedly connecting and
disconnected. This works, but is inefficient.
With CGI this is normal. If you find the speed acceptable for your purposes, run mongod with --quiet to suppress these messages in the log. If
you need better performance, which to a solution where connections are pooled -- such as an Apache module.
What RAID should I use?
We recommend not using RAID-5, but rather, RAID-10 or the like. Both will work of course.
Can I run on Amazon EBS? Any issues?
Works fine in our experience; more information here.
Why are my data files so large?
MongoDB does aggressive preallocation of reserved space to avoid file system fragmentation. This is configurable. More info here.

Generally, with MongoDB we are not building queries from strings, so traditional SQL Injection attacks are not a problem. More details and some
nuances are covered below.
MongoDB queries are represented as BSON objects. Typically the programming language gives a convenient way to build these objects that is
injection free. For example in C++ one would write:
BSONObj my_query = BSON( "name" << a_name );

auto_ptr<DBClientCursor> cursor = c.query("tutorial.persons", my_query);
my_query then will have a value such as { name : "Joe" }. If my_query contained special characters such as ", :, {, etc., nothing bad happens, they
are just part of the string.
Javascript
Some care is appropriate when using server-side Javascript. For example when using the $where statement in a query, do not concatenate user
supplied data to build Javascript code; this would be analogous to a SQL injection vulnerability. Fortunately, most queries in MongoDB can be
expressed without Javascript. Also, we can mix the two modes. It's a good idea to make all the user-supplied fields go straight to a BSON field,
and have your Javascript code be static and passed in the $where field.
If you need to pass user-supplied values into a $where clause, a good approach is to escape them using the CodeWScope mechanism. By
setting the user values as variables in the scope document you will avoid the need to have them eval'ed on the server-side.
If you need to use db.eval() with user supplied values, you can either use a CodeWScope or you can supply extra arguments to your function.
Something like: db.eval(function(userVal){...}, user_value); This will ensure that user_value gets sent as data rather than code.
User-Generated Keys
Sometimes it is useful to build a BSON object where the key is user-provided. In these situations, keys will need to have substitutions for the
reserved $ and . characters. If you are unsure what characters to use, the Unicode full width equivalents aren't a bad choice: U+FF04 () and
U+FFOE ()
For example:
BSONObj my_object = BSON( a_key << a_name );
The user may have supplied a $ value within a_key. my_object could be { $where : "things" }. Here we can look at a few cases:
Inserting. Inserting into the the database will do no harm. We are not executing this object as a query, we are inserting the data in the
database.
Note: properly written MongoDB client drivers check for reserved characters in keys on inserts.
Update. update(query, obj) allows $ operators in the obj field. $where is not supported in update. Some operators are possible that
manipulate the single document only -- thus, the keys should be escaped as mentioned above if reserved characters are possible.
Querying. Generally this is not a problem as for { x : user_obj }, dollar signs are not top level and have no effect. In theory one might let
the user build a query completely themself and provide it to the database. In that case checking for $ characters in keynames is
important. That however would be a highly unusual case.
One way to handle user-generated keys is to always put them in sub-objects. Then they are never at top level (where $operators live) anyway.
See Also
http://groups.google.com/group/mongodb-user/browse_thread/thread/b4ef57912cbf09d7

mongos
mongod
v1.0-v1.2 Concurrency
Viewing Operations in Progress
Read/Write Lock
Operations
On Javascript
Multicore
mongos
For sharded environments, mongos can perform any number of operations concurrently. This results in downstream operations to mongod
instances. Execution of operations at each mongod is independent; that is, one mongod does not block another.
mongod
The original mongod architecture is concurrency friendly; however, some work with respect to granular locking and latching is not yet done. This
means that some operations can block others. This is particular true in versions < 1.3. Version 1.3+ has improvements to concurrency, although
future work will make things even better.
v1.0-v1.2 Concurrency
In these versions of mongod, most operations prevent concurrent execution of other operations. In many circumstances, this worked reasonably
as most operations can be executed very quickly.
The following operations do have concurrent support in v1.2 and below:
1. db.currentOp() and db.killOp() commands

2. map/reduce
3. queries returning large amounts of data do interleave with other operations (but does block when scanning data that is not returned)
The rest of this document focuses on concurrency for v1.3+.
Viewing Operations in Progress
Use db.currentOp() to view operations in progress, and db.killOp() to terminate an operation.
You can also see operations in progress from the adminstrative Http Interface.
Read/Write Lock
mongod uses a read/write lock for many operations. Any number of concurrent read operations are allowed, but typically only one write operation
(although some write operations yield and in the future more concurrency will be added). The write lock acquisition is greedy: a pending write lock
acquisition will prevent further read lock acquisitions until fulfilled.
Operations
Operation Lock type Notes
OP_QUERY (query ) Acquires read lock see also: SERVER-517
OP_GETMORE (get more from cursor ) Acquires read lock
OP_INSERT (insert ) Acquires write lock Inserts are normally fast and short-lived operations
OP_DELETE (remove ) Acquires write lock Yields while running to allow other operations to interleave.
OP_UPDATE (update ) Acquires write lock Will yield for interleave (1.5.2+)
map/reduce At times locked Allows substantial concurrent operation.
create index See notes Batch build acquires write lock. But a background build option is available.
db.eval() Acquires write lock
getLastError command Non-blocking
ismaster command Non-blocking
serverStatus command Non-blocking
On Javascript
Only one thread in the mongod process executes Javascript at a time (other database operations are often possible concurrent with this).
Multicore
With read operations, it is easy for mongod 1.3+ to saturate all cores. However, because of the read/write lock above, write operations will not yet
fully utilize all cores. This will be improved in the future.

This page not done. Please help us finish it!
MySQL Program Mongo Program
mysqld mongod
mysql mongo
MongoDB queries are expressed as JSON (BSON) objects. This quick reference chart shows examples as both SQL and in Mongo Query
Language syntax.
The query expression in MongoDB (and other things, such as index key patterns) is represented as JSON. However, the actual verb (e.g. "find")
is done in one's regular programming language. The exact forms of these verbs vary by language. The examples below are Javascript and can
be executed from the mongo shell.
SQL Statement Mongo Query Language Statement
implicit; can be done explicitly

CREATE TABLE USERS (a Number, b Number)
INSERT INTO USERS VALUES(1,1) db.users.insert({a:1,b:1})
SELECT a,b FROM users db.users.find({}, {a:1,b:1})
SELECT * FROM users db.users.find()
SELECT * FROM users WHERE age=33 db.users.find({age:33})
SELECT a,b FROM users WHERE age=33 db.users.find({age:33}, {a:1,b:1})
SELECT * FROM users WHERE age=33 ORDER BY db.users.find({age:33}).sort({name:1})

name
SELECT * FROM users WHERE age>33 db.users.find({'age':{$gt:33}})})
SELECT * FROM users WHERE age<33 db.users.find({'age':{$lt:33}})})
SELECT * FROM users ORDER BY name DESC db.users.find().sort({name:-1})
CREATE INDEX myindexname ON users(name) db.users.ensureIndex({name:1})
SELECT * FROM users WHERE a=1 and b='q' db.users.find({a:1,b:'q'})
SELECT * FROM users LIMIT 10 SKIP 20 db.users.find().limit(10).skip(20)
SELECT * FROM users LIMIT 1 db.users.findOne()
EXPLAIN SELECT * FROM users WHERE z=3 db.users.find({z:3}).explain()
SELECT DISTINCT last_name FROM users db.users.distinct('last_name')
SELECT COUNT(*y) db.users.count()

FROM users
SELECT COUNT(*y) db.users.find({age: {'$gt': 30}}).count()

FROM users where AGE > 30
SELECT COUNT(AGE) from users db.users.find({age: {'$exists': true

}}).count()
UPDATE users SET a=1 WHERE b='q' db.users.update({b:'q'}, {$set:{a:1}},
false, true)
What is the Compare Order for BSON Types

MongoDB allows objects in the same collection which have values which may differ in type. When comparing values from different types, a
convention is utilized as to which value is less than the other. This (somewhat arbitary but well defined) ordering is listed below.
Note that some types are treated as equivalent for comparison purposes -- specifically numeric types which undergo conversion before
comparison.
See also the BSON specification.
Null
Numbers (ints, longs, doubles)
Symbol, String
Object
Array
BinData
ObjectID
Boolean
Date, Timestamp
Regular Expression
Example (using the mongo shell ):
> t = db.mycoll;
> t.insert({x:3});
> t.insert( {x : 2.9} );
> t.insert( {x : new Date()} );
> t.insert( {x : true } )
> t.find().sort({x:1})
{ "_id" : ObjectId("4b03155dce8de6586fb002c7"), "x" : 2.9 }
{ "_id" : ObjectId("4b03154cce8de6586fb002c6"), "x" : 3 }
{ "_id" : ObjectId("4b031566ce8de6586fb002c9"), "x" : true }
{ "_id" : ObjectId("4b031563ce8de6586fb002c8"), "x" : "Tue Nov 17 2009 16:28:03 GMT-0500 (EST)" }
MinKey and MaxKey
In addition to the above types MongoDB internally uses a special type for MinKey and MaxKey which are less than, and greater than all other
possible BSON element values, respectively.
From the mongo Javascript Shell
For example we can continue our example from above adding two objects which have x key values of MinKey and MaxKey respectively:
> t.insert( { x : MaxKey } )

> t.insert( { x : MinKey } )
> t.find().sort({x:1})
{ "_id" : ObjectId("4b04094b7c65b846e2090112"), "x" : { $minKey : 1 } }
{ "_id" : ObjectId("4b03155dce8de6586fb002c7"), "x" : 2.9 }
{ "_id" : ObjectId("4b03154cce8de6586fb002c6"), "x" : 3 }
{ "_id" : ObjectId("4b031566ce8de6586fb002c9"), "x" : true }
{ "_id" : ObjectId("4b031563ce8de6586fb002c8"), "x" : "Tue Nov 17 2009 16:28:03 GMT-0500 (EST)" }
{ "_id" : ObjectId("4b0409487c65b846e2090111"), "x" : { $maxKey : 1 } }
From C++
See also the Tailable Cursors page for an example of using MinKey from C++. See also minKey and maxKey definitions in jsobj.h.
Admin Zone
Production Notes
Replication
Sharding
Hosting Center
Monitoring and Diagnostics
Backups
Durability and Repair
Admin UIs
Starting and Stopping Mongo
GridFS Tools
DBA Operations from the Shell
Architecture and Components
Troubleshooting
Community Admin-Related Articles

boxedice.com - notes from a production deployment
Survey of Admin UIs for MongoDB
MongoDB Nagios Check
MongoDB Cacti Graphs
See Also
Commands in Developer Zone
Production Notes
Architecture
Production Options
Backups
Recommended Unix System Settings
TCP Port Numbers
File Systems
Tips
See Also
Architecture
Production Options
Master Slave
1 master, N slaves - failover is handled manually
Version 1.6: Replica Sets
N servers, 1 is always primary, auto-failover, auto-recovery
Backups
Import Export Tools
Recommended Unix System Settings

Turn off atime
Set file descriptor limit to 4k+ (see etc/limits and ulimit)
Do not use large VM pages with Linux (more info)
TCP Port Numbers

Default TCP port numbers for MongoDB processes:
Standalone mongod – 27017

mongos – 27017
shard server (mongod --shardsvr) – 27018
config server (mongod --configsvr) – 27019
web stats page for mongod – add 1000 to port number (28017, by default)
File Systems
MongoDB uses large files for storing data, and preallocates these. Some filesystems are much better at this
ext4
xfs
Tips
Handling Halted Replication
See Also
Starting and Stopping the Database
Replication
MongoDB supports asynchronous replication of data between servers for failover and redundancy. Only one server (in the set/shard) is active for
writes (the primary, or master) at a given time. With a single active master at any point in time, strong consistency semantics are available. One
can optionally send read operations to the slaves/secondaries when eventual consistency semantics are acceptable.
Master-Slave Replication
Replica Sets
Which should I use?
if using <v1.6 : master/slave

if need automatic fail-over and recovery (easy administration): replica sets
if using --auth (security) or --slavedelay : for now, master/slave
if using sharding : either, but replica sets are best for clusters that are not small
if risk averse : master/slave (replica sets are new to v1.6.0)
Verifying propagation of writes with getlasterror
A client can block until a write operation has been replicated to N servers -- read more here .
Presentations
Replication Video
Replication Slides Only
Verifying Propagation of Writes with getLastError

v1.5+.
A client can block until a write operation has been replicated to N servers. Use the getlasterror command with a new parameter w:
db.runCommand( { getlasterror : 1 , w : 2 } )
If w is not set, or equals 1, the command returns immediately, implying the data is on 1 server (itself). If w is 2, then the data is on the current
server and 1 other server (a secondary).
The higher w is, the longer acknowledgement may take. A recommended way of using this feature in a web context is to do all the write
operations for a page, then call this once if needed. That way you're only paying the cost once.
There is an optional wtimeout parameter that allows you to timeout after a certain number of milliseconds and perhaps return an error or
warning to a user. For example, the following will wait for 3 seconds before giving up:
> db.runCommand({getlasterror : 1, w : 40, wtimeout : 3000})

{
"err" : null,
"n" : 0,
"wtimeout" : true,
"waited" : 3006,
"errmsg" : "timed out waiting for slaves",
"ok" : 0
}
Note: the current implementation returns when the data has been delivered to w servers. Future versions will provide more options for delivery
vs. say, physical fsync at the server.
See also replica set configuration for information on how to change the getlasterror default parameters.
Replica Sets
v1.6.0 and higher.
Replica Sets are MongoDB's new method for replication. They are an elaboration on the existing master/slave replication, adding automatic
failover and automatic recovery of member nodes.
Replica Sets are "Replica Pairs version 2" and are available in MongoDB version 1.6. Replica Pairs will be deprecated.
Features
Supports 1-7 servers in the cluster

Automatic failover and recovery
Data center aware (v1.6.1) (rack aware too, eventually)
Supports passive set members (slaves) that are never primary
Docs
To get started:
Try it out
Learn how to configure your set
If you would like to start using replica sets with an existing system:
Learn how to migrate your existing setup

Upgrade your client code to use replica set connections (see also your driver's documentation for details)
When running replica sets, it is important to know about:
The admin UI
Administrative commands
More Docs
Sample Replica Set Config Session.pdf

Limits
Design Concepts
HowTo
Resyncing a Very Stale Replica Set Member
Adding a New Set Member
Adding an Arbiter
About the local database
Reconfiguring when members are down
Data Center Awareness
See Also
Replication Video
Replica Sets Slides
About the local database

mongod reserves the database local for special functionality. It is special in that its contents are never replicated.
Using the database for end-user data
You may place end user application data in local, if you would like it to not replicate to other servers. Put your collections under local.usr.*.
Replica Sets
Replica sets use the following collections in local:
local.system.replset the replica set's configuration object is stored here. (View via the rs.conf() helper in the shell – or query it
directly.)
local.oplog.rs is a capped collection that is the oplog. You can use the --oplogSize command line parameter to set the size of
this collection.
local.replset.minvalid sometimes contains an object used internally by replica sets to track sync status
Master/Slave Replication
Master
local.oplog.$main the "oplog"
local.slaves
Slave
local.sources
Other
local.me
local.pair.* (replica pairs, which are deprecated)
Data Center Awareness

The 1.6.0 build of replica sets does not support much in terms of data center awareness. However additional functionality will be added in the
future. Below are some suggestions configurations which work today.
Primary plus DR site
Use one site, with one or more set members, as primary. Have a member at a remote site with priority=0. For example:
{ _id: 'myset',
members: [
{ _id:0, host:'sf1', priority:1 },
{ _id:2, host:'ny1', priority:0 }
]
}
Multi-site with local reads

Another configuration would be to have one member in each of three data centers. One node arbitrarily becomes primary, the others though are
secondaries and can process reads locally.
{ _id: 'myset',
members: [
{ _id:1, host:'ny1', priority:1 },
{ _id:2, host:'uk1', priority:1 }
]
}
Reconfiguring a replica set when members are down

One may modify a set when some members are down if a majority is established. In this case, simply send the reconfig command to the current
primary.
If there is no primary (and this condition is not transient), no majority is available. Reconfiguring a minority partition would be dangerous as two
sides of a network partition won't both be aware of the reconfiguration. Thus, this is not allowed.
However, in some administrative circumstances we will want to take action even though there is no majority. Suggestions on how to deal with this
are outlined below.
Example 1
A replica set has three members, which in the past were healthy. Two of the servers are permanently destroyed. We wish to bring the remaining
member online immediately.
One option is to make the last standing mongod a standalone server and not a set member:
1. stop the surviving mongod

2. consider doing a backup...
3. delete the local.* datafiles in the data directory. this will prevent potential future confusion if it is ever restarted with --replSet in the
future.
4. restart mongod without the --replSet parameter.
We are now back online with a single node that is not a replica set member. Clients can use it for both reads and writes.
Example 2
A replica set has three members, which in the past were healthy. Two of the servers are permanently destroyed. We wish to bring the remaining
member online and add a new member to its set.
We cannot reconfigure the existing set with only 1 of 3 members available. However, we can "break the mirror" and start a new set:
1. stop the surviging mongod

3. delete the local.* datafiles in the data directory.
4. restart the mongod with a new replica set name
5. initiate this new set
6. then, add the new second member
Example 3
A replica set has five members, which in the past were healthy. Three of the servers are permanently destroyed. We wish to bring the remaining
members online.
As in example 2 we will use the "break the mirror" technique. Unfortunately one of the two members must be re-synced.
1. stop the surviging mongod's

3. delete the local.* datafiles on server 1
4. delete (ideally just move to a backup location) all the datafiles from server 2
5. restart both mongod's with the new replica set name on the command line for each
6. initiate this new set on server 1
7. then, add the new second member (server 2)
Replica Set Design Concepts

1.
A write is only truly committed once it has replicated to a majority of members of the set. For important writes, the client should request
acknowledgement of this with a getLastError({w:...}) call.
2.
Writes which are committed at the primary of the set may be visible before the true cluster-wide commit has occurred. This property, which is
more relaxed than some traditional RDBMS products, makes theoretically achievable performance and availability higher.
3.
On a failover, if there is data which has not replicated from the primary, the data is dropped (thus the use of getlasterror in #1 above).
Data is now backed up on rollback, although the assumption is that in most cases this data is never recovered as that would
require operator intervention: http://jira.mongodb.org/browse/SERVER-1512.
Replica Sets Troubleshooting

can't get local.system.replset config from self or any seed (EMPTYCONFIG)
Set needs to be initiated. Run rs.initiate() from the shell.
If the set is already initiated and this is a new node, verify it is present in the replica set's configuration and there are no typos in the host names:
> // send to a working node in the set:

> rs.conf()

This tutorial will guide you through the basic configuration of a replica set on a single machine. If you're attempting to deploy replica sets in
production, be sure to read the comprehensive replica set documentation. Also, do keep in mind that replica sets are production-ready as of
MongoDB 1.6.
Introduction
Starting the nodes
Initializing the set
Replication
Failing Over
Changing the replica set configuration
Running with two nodes
Drivers
Introduction
A replica set is group of N mongod nodes that work together to provide automated failover.
Setting up a replica set is a two-step process that requires starting each node and then formally initiating the set. Here, we'll be configuring a set
of three nodes, which is standard.
Once the mongod nodes are started, we'll issue a command to properly initialize the set. After a few seconds, one node will be elected master,
and you can begin writing to and querying the set.
Starting the nodes
First, create a separate data directory for each of the nodes in the set:
mkdir -p /data/r0
mkdir -p /data/r1
mkdir -p /data/r2
Next, start each mongod process with the --replSet parameter. The parameter requires that you specify the name of the replica set. Let's call
our replica set "foo." We'll launch our first node like so:
mongod --replSet foo --port 27017 --dbpath /data/r0
The second node gets launched on port 27018:
Let's start the second node on port 27018:
And, finally, we'll start a third node on port 27019.
You should now have three nodes running. At this point, each node should be printing the following warning:
Mon Aug 2 11:30:19 [startReplSets] replSet can't get local.system.replset config from self or any
seed (EMPTYCONFIG)
We can't use the replica set until we've initialized it, which we'll do next.
Initializing the set
We can initiate the replica set by connecting to one of the members and running the replSetInitiate command. This command takes a
configuration object that specifies the name of the set and each of the members.
mongo localhost:27017
[kyle@arete ~$]$ mongo localhost:27017
MongoDB shell version: 1.5.7
connecting to: localhost:27017/test
> config = {_id: 'foo', members: [
{_id: 0, host: 'localhost:27017'},
{_id: 2, host: 'localhost:27019'}]
}
> rs.initiate(config);
{
"info" : "Config now saved locally. Should come online in about a minute.",
"ok" : 1
}
We specify the config object and then pass it to rs.initiate(). Then, if everything is in order, we get a response saying that the replica set will
be online in a minute. During this time, one of the nodes will be elected master.
To check the status of the set, run rs.status:

> rs.status()
{
"set" : "foo",
"date" : "Mon Aug 02 2010 11:39:08 GMT-0400 (EDT)",
"myState" : 1,
"members" : [
{
"name" : "arete.local:27017",
"self" : true,
},
{
"name" : "localhost:27019",
"health" : 1,
"uptime" : 101,
"lastHeartbeat" : "Mon Aug 02 2010 11:39:07 GMT-0400",
},
{
"name" : "localhost:27018",
"health" : 1,
"uptime" : 107,
"lastHeartbeat" : "Mon Aug 02 2010 11:39:07 GMT-0400",
}
],
"ok" : 1
}
You'll see that both of the other members of the set are up. You may also notice that the myState value is 1, indicating that we're connected to
the master node; a value of 2 indicates a slave.
You can also check the set's status in the HTTP Admin UI.
Replication
Go ahead and write something to the master node:
db.messages.save({name: "ReplSet Tutorial"});
If you pay attention to the logs on the slave nodes, you'll see the write being replicated. This initial replication is essential for failover; the system
won't fail over to a new master until an initial sync between nodes is complete.
Failing Over
Now, the purpose of a replica set is to provide automated failover. This means that, if the master node is killed, a slave node can take over. To
see how this works in practice, go ahead and kill the master node with ^C:
^CMon Aug 2 11:50:16 got kill or ctrl c or hup signal 2 (Interrupt), will terminate after current cmd
ends
Mon Aug 2 11:50:16 [interruptThread] now exiting
Mon Aug 2 11:50:16 dbexit:
If you look at the logs on the slave nodes, you'll see a series of messages indicating failover. On our first slave, we see this:
Mon Aug 2 11:50:16 [ReplSetHealthPollTask] replSet info localhost:27017 is now down (or slow to
respond)
Mon Aug 2 11:50:17 [conn1] replSet info voting yea for 2
Mon Aug 2 11:50:17 [rs Manager] replSet not trying to elect self as responded yea to someone else
recently
Mon Aug 2 11:50:27 [rs_sync] replSet SECONDARY
And on the second, this:
Mon Aug 2 11:50:17 [ReplSetHealthPollTask] replSet info localhost:27017 is now down (or slow to
respond)
Mon Aug 2 11:50:17 [rs Manager] replSet info electSelf 2
Mon Aug 2 11:50:17 [rs Manager] replSet PRIMARY
Mon Aug 2 11:50:27 [initandlisten] connection accepted from 127.0.0.1:61263 #5
Both nodes notice that the master has gone down and, as a result, a new primary node is elected. In this case, the node at port 27019 is
promoted. If we bring the failed node on 27017 back online, it will come up as a slave.
Changing the replica set configuration
There are times when you'll want to change the replica set configuration. Suppose, for instance, that you want to adjust the number of votes
available to each node. To do this, you need to pass a new configuration object to the database's replSetReconfig command. Here's how.
First, define the new configuation:
new_config = {_id: 'foo', members: [

{_id: 0, host: 'localhost:27017', votes: 1},
{_id: 1, host: 'localhost:27018', votes: 2},
{_id: 2, host: 'localhost:27019', votes: 3}]
}
Then, add the version to the config object. To do this, you'll need to increment the old config version.
use local
old_config = db.system.replset.findOne();
new_config.version = old_config.version + 1;
Finally, reconfigure:
use admin
db.runCommand({replSetReconfig: new_config});
Running with two nodes
Suppose you want to run replica sets with just two database servers. This is possible as long as you also use an arbiter on a separate node; most
likely, running the arbiter on one or more application servers will be ideal. With an arbiter in place, the replica set will behave appropriately,
recovering automatically during both network partitions and node failures (e.g., machine crashes).
You start up an arbiter just as you would a standard replica set node, with the --replSet option. However, when initiating, you need to include
the arbiterOnly option in the config document.
With an arbiter, the configuration presented above would look like this instead:
config = {_id: 'foo', members: [

{_id: 2, host: 'localhost:27019', arbiterOnly: true}]
}
The other requirement here is that the total number of votes for the database nodes needs to consist of a majority. This means that if you have
two database nodes and ten arbiters, there's a total of twelve votes. So the best bet in this case it to give each database node enough votes so
that even if all but a single arbiter goes down, the master still has enough votes to stay up. In that situation, each database node would need at
least three votes.
For more information on arbiters and other interesting config options, see the replica set configuration docs.
Drivers
All of the MongoDB drivers are designed to take any number of replica set seed hosts from a replica set and then cache the hosts of any other
known members.
With this complete set of potential master nodes, the driver can automatically find the new master if the current master fails. See your driver's
documentation for details. If you happen to be using the Ruby driver, check out Replica Sets in Ruby.
Replica Set Configuration

Command Line
Initial Setup
The Replica Set Config Object
Shell Example 1
Shell Example 2
See Also
Command Line
Each mongod participating in the set should have a --replSet parameter on its command line. The syntax is
mongod --replSet setname
setname is the logical name of the set.
The --rest command line parameter is also recommended when using replica sets, as the web admin interface of mongod
(normally at port 28017) shows status information on the set.
Initial Setup
We use the initiate command for initial configuration of a replica set. Send the initiate command to a single server with the set to christen the set.
The member being initiated may have initial data; the other servers in the set should be empty.
> db.runCommand( { replSetInitiate : <config_object> } )
A shorthand way to type the above is via a helper method in the shell:
> rs.initiate(<config_object>)
A quick way to initiate a set is to leave out the config object parameter. The initial set will then consist of the member to which the shell is
communicating, along with all the seeds that member knows of. However, see the configuration object details below for more options.
> rs.initiate()
The Replica Set Config Object
local.system.replset holds a singleton object which contains the replica set configuration. The config object automatically propagates among
members of the set. The object is not directly manipulated, but rather changed via commands (such as replSetInitiate).
The config object has the form:

{
_id : <setname>,
members: [
{
_id : <ordinal>,
host : <hostname[:port]>,
[, priority: <priority>]
[, arbiterOnly : true]
[, votes : <n>]
[, hidden : true]
}
, ...
],
settings: {
[getLastErrorDefaults: <lasterrdefaults>]
[, heartbeatSleep : <seconds>]
[, heartbeatTimeout : <seconds>]
[, heartbeatConnRetries : <n>]
}
}
_id - the set name. must match command line setting.

members
_id - each member has an _id ordinal, typically beginning with zero and numbered in increasing order. when a node is retired
(removed from the config), its _id should not be reused.
host - host name and optionally the port for the member
priority - priority a server has for potential election as primary. The highest priority member which is up will become primary.
Default 1.0. Priority zero means server can never be primary.
arbiterOnly - participates in consensus, but receives no data.
votes - number of votes this set member gets in elections. Default 1.
hidden - when true, do not advertise the member's existence to clients in isMaster command responses. (v1.7+)
settings. Heartbeat params are optional as reasonable settings are predefined.
getLastErrorDefaults specifies defaults for the getlasterror command. If the client calls getLastError with no parameters, the
default object specified here is used. (v1.6.2+)
heartbeatSleep how frequently nodes should send a heartbeat to each other (default: 2 seconds, must be greater than 10
milliseconds).
heartbeatTimeout indicates how long a node needs to fail to send data before we note a problem (default: 10 seconds, must be
greater than 10 milliseconds).
heartbeatConnRetries is how many times after heartbeatTimeout to try connecting again and getting a new heartbeat (default: 3
tries).
Shell Example 1
> // all at once method

> cfg = {
... _id : "acme_a",
... members : [
... { _id : 0, host : "sf1.acme.com" },
... { _id : 1, host : "sf2.acme.com" },
... { _id : 2, host : "sf3.acme.com" } ] }
> rs.initiate(cfg)
> rs.status()
Shell Example 2
$ # incremental configuration method

$ mongo sf1.acme.com/admin
> rs.initiate();
> rs.add("sf2.acme.com");
> rs.add("sf3.acme.com");
> rs.status();
See Also
Adding a new node to an existing replica set is easy. The new node should either have an empty data directory or a recent copy of the data from
another set member. When we start the new node, we only need to provide the replica set name:
$ ./mongod --replSet foo
After bringing up the new server (we'll call it broadway:27017) we need to add it to the set - we connect to our primary server using the shell:
$ ./mongo
MongoDB shell version: ...
connecting to: test
> rs.add("broadway:27017");
{ "ok" : 1 }
After adding the node it will perform a full resync and come online as a secondary. If the node is started with a recent copy of data from another
node in the set it won't need a full resync.
See also:
Adding an Arbiter
Adding an Arbiter
Arbiters are nodes in a replica set that only participate in elections: they don't have a copy of the data and will never become the primary node (or
even a readable secondary). They are mainly useful for breaking ties during elections (e.g. if a set only has two members).
To add an arbiter, bring up a new node and point it at the replica set using the --replSet flag - this part is identical to when Adding a New Set
Member. To start as an arbiter, we'll use rs.addArb() instead of rs.add(). While connected to the current primary:
> rs.addArb("broadway:27017");
{ "ok" : 1 }
See Also
Upgrading to Replica Sets

Upgrading From a Single Server
Upgrading From Replica Pairs or Master/Slave
Upgrading Drivers
Upgrading From a Single Server
If you're running MongoDB on a single server, upgrading to replica sets is trivial (and a good idea!). First, we'll initiate a new replica set with a
single node. We need a name for the replica set - in this case we're using foo. Start by shutting down the server and restarting with the
--replSet option, and our set name:
$ ./mongod --replSet foo
Add the --rest option too (just be sure that port is secured): the <host>:28017/_replSet diagnostics page is incredibly useful.
The server will allocate new local data files before starting back up. Consider pre-allocating those files if you need to minimize downtime.
Next we'll connect to the server from the shell and initiate the replica set:
$ ./mongo
MongoDB shell version: ...
connecting to: test
> rs.initiate();
{
"info2" : "no configuration explicitly specified -- making one",
"info" : "Config now saved locally. Should come online in about a minute.",
"ok" : 1
}
The server should now be operational again, this time as the primary in a replica set consisting of just a single node. The next step is to add some
additional nodes to the set.
Upgrading From Replica Pairs or Master/Slave
Version 1.6.1 will have more seamless support for upgrading, track this case for details: http://jira.mongodb.org/browse/SERVER-1553.
With 1.6.0 the best way to upgrade is to simply restart the current master as a single server replica set, and then add any slaves after wiping their
data directory. To find the master in a replica pair, use the ismaster command.
Once you know the master, the process will look like this:
m$ # shutdown mongod master and slave

m$ killall mongod
s$ killall mongod
m$ # backup your /data/db directory on the master

m$ cp /data/db/* /to_somewhere_backup/
s$ # we start empty on the slave. so let's save the old data (assuming drive large enough)
s$ mv /data/db /data/old_slave_data
s$ mkdir /data/db
s$ # /data/db is now empty
m$ mongod --rest --replSet mysetname

m$ mongo
m> rs.initiate()
m> // try these:
m> db.isMaster()
m> rs.help()
m> rs.status()
m> rs.conf()
m> // see also http://localhost:28017/_replSet
s$ # start replica set member on the old slave.

s$ # it has no data and will do a full sync initially
s$ mongod --rest --replSet mysetname
s$ mongo m/admin
m> // still in the mongo shell on the master

m> rs.add("s") // "s" is your slave host name
m> rs.status(); // see also http://localhost:28017/_replSet
arb$ # we should now add an arbiter so break ties on elections and

arb$ # know who is up in a network partition.
arb$ # arbiter is very lightweight and can run on about any server
arb$ # including 32 bit servers.
arb$ # we use different directories and ports here so that the server
arb$ # is still available as a "normal" mongod server if that is
arb$ # desired and also to avoid confusion. the /data/arb directory
arb$ # will be very light in content.
arb$ mkdir /data/arb
arb$ mongod --rest --replSet mysetname --dbpath /data/arb --port 30000 --oplogSize 8
m> rs.addArb("arb:30000"); // replace 'arb' with your arb host name

m> rs.status()
Upgrading Drivers
There are new versions of most MongoDB Drivers which support replica sets elegantly. See the documentation pages for the specific driver of
interest.
Replica Set Admin UI

The mongod process includes a simple administrative UI for checking the status of a replica set.
To use, first enable --rest from the mongod command line. The rest port is the db port plus 1000 (thus, the default is 28017). Be sure this port
is secure before enabling this.
Then you can navigate to http://<hostname>:28017/ in your web browser. Once there, click Replica Set Status (/_replSet) to move to the
Replica Set Status page.
See Also
Http Interface
Replica Set Commands

Shell Helpers
Commands
{ isMaster : 1 }
{ replSetGetStatus : 1 }
{ replSetInitiate : <config> }
{ replSetStepDown : true }
{ replSetFreeze : <bool> }
{ replSetReconfig: <config> }
Shell Helpers
rs.help() show help

rs.status() { replSetGetStatus : 1 }
rs.initiate() { replSetInitiate : null } initiate
with default settings
rs.initiate(cfg) { replSetInitiate : cfg }
rs.add(hostportstr) add a new member to the set
rs.add(membercfgobj) add a new member to the set
rs.addArb(hostportstr) add a new member which is arbiterOnly:true
rs.stepDown() { replSetStepDown : true }
rs.conf() return configuration from local.system.replset
db.isMaster() check who is primary
Commands
{ isMaster : 1 }
Checks if the node to which we are connecting is currently primary. Most drivers do this check automatically and then send requires to the current
primary.
Returns an object that looks like:
{
"ismaster" : false,
"secondary" : true,
"hosts" : [
"sf1.example.com",
"sf4.example.com",
"ny3.example.com"
],
"passives" : [
"sf3.example.com",
"sf2.example.com",
"ny2.example.com",
],
"arbiters" : [
"ny1.example.com",
]
"primary" : "sf4.example.com",
"ok" : 1
}
The hosts array lists primary and secondary servers, the passives array lists passive servers, and the arbiters array lists arbiters.
If the "ismaster" field is false, there will be a "primary" field that indicates which server is primary.
{ replSetGetStatus : 1 }
Status information on the replica set from this node's point of view.
The output looks like:
{
"set" : "florble",
"date" : "Wed Jul 28 2010 15:01:01 GMT-0400 (EST)",
"myState" : 1,
"members" : [
{
"name" : "dev1.example.com",
"self" : true,
"errmsg" : ""
},
{
"name" : "dev2.example.com",
"health" : 1,
"uptime" : 13777,
"lastHeartbeat" : "Wed Jul 28 2010 15:01:01 GMT-0400 (EST)",
"errmsg" : "initial sync done"
}
],
"ok" : 1
}
The myState field indicates the state of this server. Valid states are:
0 Starting up, phase 1

1 Primary
2 Secondary
3 Recovering
4 Fatal error
5 Starting up, phase 2
6 Unknown state
7 Arbiter
8 Down
The health field is 1 if the server is up and 0 if it is down.
The errmsg field can contain informational messages, as shown above.
{ replSetInitiate : <config> }
Initiate a replica set. Run this command at one node only, to initiate the set. Whatever data is on the initiating node becomes the initial data for the
set. This is a one time operation done at cluster creation. See also Configuration.
{ replSetStepDown : true }
Step down as primary. Node will become eligible to be primary again after 1 minute. (Presumably, another node will take over by then if it were
eligible.)
This command will be enhanced later to allow specification of a min duration of the step-down.
{ replSetFreeze : <bool> }
Freezing a replica set prevents failovers from occurring. This can be useful during maintenance.
Note: As of v1.6.0, replSetFreeze is not yet implemented.
{ replSetReconfig: <config> }
Adjust configuration of a replica set (just like initialize)
db._adminCommand({replSetReconfig: cfg })
Note: db._adminCommand is short-hand for db.getSisterDB("admin").runCommand();
Replica Set FAQ

How long does failover take?
Failover thresholds are configurable. With the defaults, it may take 20-30 seconds for the primary to be declared down by the other members and
a new primary elected. During this window of time, the cluster is down for "primary" operations – that is, writes and strong consistent reads.
However, you may execute eventually consistent queries to secondaries at any time, including during this window.
Should I use replica sets or replica pairs?
After 1.6, use Replica Sets.
Connecting to Replica Sets from Clients

Most drivers have been updated to provide ways to connect to a replica set. In general, this is very similar to how the drivers support connecting
to a replica pair.
Instead of taking a pair of hostnames, the drivers will typically take a comma separated list of host[:port] names. This is a seed host list; it need
not be every member of the set. The driver then looks for the primary from the seeds. The seed members will report back other members of the
set that the client is not aware of yet. Thus we can add members to a replica set without changing client code.
With Sharding
With sharding, the client connects to a mongos process. The mongos process will then automatically find the right member(s) of the set.
See Also
Driver authors should review Connecting Drivers to Replica Sets.
Replica Sets Limits
v1.6
Authentication mode not supported. JIRA

Limits on config changes to sets at first. Especially when a lot of set members are down.
Map/reduce writes new collections to the server. Because of this, for now it may only be used on the primary. This will be enhanced
later.
Error RS102
MongoDB writes operations to an oplog. For replica sets this data is stored in collection local.oplog.rs. This is a capped collection and wraps
when full "RRD"-style. Thus, it is important that the oplog collection is large enough to buffer a good amount of writes when some members of a
replica set are down. If too many writes occur, the down nodes, when they resume, cannot catch up. In that case, a full resync would be
required.
Sizing the oplog
The command line --oplogSize parameter sets the oplog size. A good rule of thumb is 5 to 10% of total disk space. On 64 bit builds, the
default is large and similar to this percentage. You can check your existing oplog sizes from the mongo shell :
> use local

> db.oplog.rs.stats()
What to do on a sync error
If one of your members has been offline and is now too far behind to catch up, you will need to resync. There are a number of ways to do this.
1. Delete all data. If you stop the failed mongod, delete all data, and restart it, it will automatically resynchronize itself. Of course this may be slow
if the database is huge or the network slow.
2. Copy data from another member. You can copy all the data files from another member of the set IF you have a snapshot of that member's data
file's. This can be done in a number of ways. The simplest is to stop mongod on the source member, copy all its files, and then restart mongod on
both nodes. The Mongo fsync and lock feature is another way to achieve this. On a slow network, snapshotting all the datafiles from another
(inactive) member to a gziped tarball is a good solution. Also similar strategies work well when using SANs and services such as Amazon Elastic
Block Service snapshots.
3. Find a member with older data. Each member of the replica set has an oplog. It is possible that a member has a larger oplog than the current
primary.
Replica Set Internals

Design Concepts
Configuration
Command Line
Node Types
local.system.replset
Set Initiation (Initial Setup)
Design
Server States
Applying Operations
OpOrdinal
Picking Primary
Heartbeat Monitoring
Assumption of Primary
Failover
Resync (Connecting to a New Primary)
Consensus
Increasing Durability
Reading from Secondaries and Staleness
Example
Administration
Future Versions
Design Concepts
Check out the Replica Set Design Concepts for some of the core concepts underlying MongoDB Replica Sets.
Configuration
Command Line
We specify --replSet set_name/seed_hostname_list on the command line. seed_hostname_list is a (partial) list of some members of the set. The
system then fetches full configuration information from the collection local.system.replset. set_name is specified to help the system catch
misconfigurations.
Node Types
Conceptually, we have some different types of nodes:
Standard - a standard node as described above. Can transition to and from being a primary or a secondary over time. There is only one
primary (master) server at any point in time.
Passive - a server can participate as if it were a member of the replica set, but be specified to never be Primary.
Arbiter - member of the cluster for consensus purposes, but receives no data. Arbiters cannot be seed hosts.
Each node in the set has a priority setting. On a resync (see below), the rule is: choose as master the node with highest priority that is healthy. If
multiple nodes have the same priority, pick the node with the freshest data. For example, we might use 1.0 priority for Normal members, 0.0 for
passive (0 indicates cannot be primary no matter what), and 0.5 for a server in a less desirable data center.
local.system.replset
This collection has one document storing the replica set's configuration. See the configuration page for details.
Set Initiation (Initial Setup)
For a new cluster, on negotiation the max OpOrdinal is zero everywhere. We then know we have a new replica set with no data yet. A special
command
{replSetInitiate:1}
is sent to a (single) server to begin things.
Design
Server States
Primary - Can be thought of as "master" although which server is primary can vary over time. Only 1 server is primary at a given point in
time.
Secondary - Can be thought of as a slave in the cluster; varies over time.
Recovering - getting back in sync before entering Secondary mode.
Applying Operations
Secondaries apply operations from the Primary. Each applied operation is also written to the secondary's local oplog. We need only apply from
the current primary (and be prepared to switch if that changes).
OpOrdinal
We use a monotonically increasing ordinal to represent each operation.
These values appear in the oplog (local.oplog.$main). maxLocalOpOrdinal() returns the largest value logged. This value represents how
up-to-date we are. The first operation is logged with ordinal 1.
Note two servers in the set could in theory generate different operations with the same ordinal under some race conditions. Thus for full
uniqueness we must look at the combination of server id and op ordinal.
Picking Primary
We use a consensus protocol to pick a primary. Exact details will be spared here but that basic process is:
1. get maxLocalOpOrdinal from each server.
2. if a majority of servers are not up (from this server's POV), remain in Secondary mode and stop.
3. if the last op time seems very old, stop and await human intervention.
4. else, using a consensus protocol, pick the server with the highest maxLocalOpOrdinal as the Primary.
Any server in the replica set, when it fails to reach master, attempts a new election process.
Heartbeat Monitoring
All nodes monitor all other nodes in the set via heartbeats. If the current primary cannot see half of the nodes in the set (including itself), it will fall
back to secondary mode. This monitoring is a way to check for network partitions. Otherwise in a network partition, a server might think it is still
primary when it is not.
Assumption of Primary
When a server becomes primary, we assume it has the latest data. Any data newer than the new primary's will be discarded. Any discarded data
is backed up to a flat file as raw BSON, to allow for the possibility of manual recovery (see this case for some details). In general, manual
recovery will not be needed - if data must be guaranteed to be committed it should be written to a majority of the nodes in the set.
Failover
We renegotiate when the primary is unavailable, see Picking Primary.
Resync (Connecting to a New Primary)
When a secondary connects to a new primary, it must resynchronize its position. It is possible the secondary has operations that were never
committed at the primary. In this case, we roll those operations back. Additionally we may have new operations from a previous primary that never
replicated elsewhere. The method is basically:
for each operation in our oplog that DNE at the primary, (1) remove from oplog and (2) resync the document in question by a query to the
primary for that object. update the object, deleting if it does not exist at the primary.
We can work our way back in time until we find a few operations that are consistent with the new primary, and then stop.
Any data that is removed during the rollback is stored offline (see Assumption of Primary, so one can manually recover it. It can't be done
automatically because there may be conflicts.
Reminder: you can use w= to ensure writes make it to a majority of slaves before returning to the user, to ensure no writes need to be rolled back.
Consensus
Fancier methods would converge faster but the current method is a good baseline. Typically only ~2 nodes will be jockeying for primary status at
any given time so there isn't be much contention:
query all others for their maxappliedoptime

try to elect self if we have the highest time and can see a majority of nodes
if a tie on highest time, delay a short random amount first
elect (selfid,maxoptime) msg -> others
if we get a msg and our time is higher, we send back NO
we must get back a majority of YES
if a YES is sent, we respond NO to all others for 1 minute. Electing ourself counts as a YES.
repeat as necessary after a random sleep
Increasing Durability
We can trade off durability versus availability in a replica set. When a primary fails, a secondary will assume primary status with whatever data it
has. Thus, we have some desire to see that things replicate quickly. Durability is guaranteed once a majority of servers in the replica set have an
operation.
To improve durability clients can call getlasterror and wait for acknowledgement until replication of a an operation has occurred. The client can
then selectively call for a blocking, somewhat more synchronous operation.
Reading from Secondaries and Staleness
Secondaries can report via a command how far behind the primary they are. Then, a read-only client can decide if the server's data is too stale or
close enough for usage.
Example
server-a: secondary oplog: ()

server-b: secondary oplog: ()
server-c: secondary oplog: ()
…
server-a: primary oplog: (a1,a2,a3,a4,a5)
server-b: secondary oplog: ()
server-c: secondary oplog: ()
…
server-a: primary oplog: (a1,a2,a3,a4,a5)
server-b: secondary oplog: (a1)
server-c: secondary oplog: (a1,a2,a3)
…
// server-a goes down
…
server-c: secondary oplog: (a1,a2,a3)
...
server-c: primary oplog: (a1,a2,a3) // c has highest ord and becomes primary
...
server-b: secondary oplog: (a1,a2,a3)
server-c: primary oplog: (a1,a2,a3,c4)
...
server-a resumes
...
server-a: recovering oplog: (a1,a2,a3,a4,a5)
server-b: secondary oplog: (a1,a2,a3)
…
server-a: recovering oplog: (a1,a2,a3,c4)
server-b: secondary oplog: (a1,a2,a3,c4)
…
server-a: secondary oplog: (a1,a2,a3,c4)
server-b: secondary oplog: (a1,a2,a3,c4)
server-c: primary oplog: (a1,a2,a3,c4,c5,c6,c7,c8)
…
server-a: secondary oplog: (a1,a2,a3,c4,c5,c6,c7,c8)
server-b: secondary oplog: (a1,a2,a3,c4,c5,c6,c7,c8)
server-c: primary oplog: (a1,a2,a3,c4,c5,c6,c7,c8)
In the above example, server-c becomes primary after server-a fails. Operations (a4,a5) are lost. c4 and c5 are new operations with the same
ordinals.
Administration
See the Replica Set Commands page for full info.
Commands:
{ replSetFreeze : <bool> } "freeze" or unfreeze a set. When frozen, new nodes cannot be elected master. Used when doing
administration. Details TBD.
{ replSetGetStatus : 1 } get status of the set, from this node's POV
{ replSetInitiate : 1 }
{ ismaster : 1 } check if this node is master
Future Versions
add support for replication trees / hierarchies

replicating to a slave that is not a member of the set (perhaps we do not need this given we have the Passive set member type)
Master Slave
Configuration and Setup
Command Line Options
Master
Slave
--slavedelay
Diagnostics
Security
Administrative Tasks
Failing over to a Slave (Promotion)
Inverting Master and Slave
Creating a slave from an existing master's disk image
Creating a slave from an existing slave's disk image
Resyncing a slave that is too stale to recover
See Also
Configuration and Setup
To configure an instance of Mongo to be a master database in a master-slave configuration, you'll need to start two instances of the database,
one in master mode, and the other in slave mode.
Data Storage
The following examples explicitly specify the location of the data files on the command line. This is unnecessary if you are
running the master and slave on separate machines, but in the interest of the readers who are going try this setup on a single
node, they are supplied in the interest of safety.
$ bin/mongod --master [--dbpath /data/masterdb/]
As a result, the master server process will create a local.oplog.$main collection. This is the "transaction log" which queues operations which
will be applied at the slave.
To configure an instance of Mongo to be a slave database in a master-slave configuration:
$ bin/mongod --slave --source <masterhostname>[:<port>] [--dbpath /data/slavedb/]
Details of the source server are then stored in the slave's local.sources collection. Instead of specifying the --source parameter, one can
add an object to local.sources which specifies information about the master server:
$ bin/mongo <slavehostname>/local
> db.sources.find(); // confirms the collection is empty. then:
> db.sources.insert( { host: <masterhostname> } );
host: masterhostname is the IP address or FQDN of the master database machine. Append :port to the server hostname if you
wish to run on a nonstandard port number.
only: databasename (optional) if specified, indicates that only the specified database should replicate. NOTE: A bug with only is
fixed in v1.2.4+.
A slave can pull from multiple upstream masters. In such a situation add multiple configuration objects to the local.sources collection. See the
One Slave Two Masters doc page.
A slave may become out of sync with a master if it falls far behind the data updates available from that master, or if the slave is terminated and
then restarted some time later when relevant updates are no longer available from the master. If a slave becomes out of sync, replication will
terminate and operator intervention is required by default if replication is to be restarted. An operator may restart replication using the {resync:1}
command. Alternatively, the command line option --autoresync causes a slave to restart replication automatically (after ten second pause) if it
becomes out of sync. If the --autoresync option is specified, the slave will not attempt an automatic resync more than once in a ten minute
periond.
The --oplogSize command line option may be specified (along with --master) to configure the amount of disk space in megabytes which will
be allocated for storing updates to be made available to slave nodes. If the --oplogSize option is not specified, the amount of disk space for
storing updates will be 5% of available disk space (with a minimum of 1GB) for 64bit machines, or 50MB for 32bit machines.
Master
--master master mode

--oplogSize arg size limit (in MB) for op log
Slave
--slave slave mode
--source arg arg specifies master as <server:port>
--only arg arg specifies a single database to replicate
--slavedelay arg arg specifies delay (in seconds) to be used
when applying master ops to slave
--autoresync automatically resync if slave data is stale
--slavedelay
Sometimes its beneficial to have a slave that is purposefully many hours behind to prevent human error. In MongoDB 1.3.3+, you can specify this
with the --slavedelay mongod command line option. Specify the delay in seconds to be used when applying master operations to the slave.
Specify this option at the slave. Example command line:
mongod --slave --source mymaster.foo.com --slavedelay 7200
Diagnostics
Check master status from the mongo shell with:
// inspects contents of local.oplog.$main on master and reports status:

db.printReplicationInfo()
Check slave status from the mongo shell with:
// inspects contents of local.sources on the slave and reports status:

db.printSlaveReplicationInfo()
(Note you can evaluate the above functions without the parenthesis above to see their javascript source and a bit on the internals.)
As of 1.3.2, you can do this on the slave
db._adminCommand( { serverStatus : 1 , repl : N } )
N is the level of diagnostic information and can have the following values:
0 - none
1 - local (doesn't have to connect to other server)
2 - remote (has to check with the master)
Security
When security is enabled, one must configure a user account for the local database that exists on both servers.
The slave-side of a replication connection first looks for a user repl in local.system.users. If present, that user is used to authenticate against the
local database on the source side of the connection. If repl user does not exist, the first user object in local.system.users is tried.
The local database works like the admin database: an account for local has access to the entire server.
Example security configuration when security is enabled:
$ mongo <slavehostname>/admin -u <existingadminusername> -p<adminpassword>

> use local
> db.addUser('repl', <replpassword>);
^c
$ mongo <masterhostname>/admin -u <existingadminusername> -p<adminpassword>
> use local
Administrative Tasks
Failing over to a Slave (Promotion)
To permanently fail over from a down master (A) to a slave (B):
shut down A
stop mongod on B
backup or delete local.* datafiles on B
restart mongod on B with the --master option
Note that is a one time cutover and the "mirror" is broken. A cannot be brought back in sync with B without a full resync.
Inverting Master and Slave
If you have a master (A) and a slave (B) and you would like to reverse their roles, this is the recommended sequence of steps. Note the following
assumes A is healthy and up.
1. Halt writes on A (using the fsync command)

2. Make sure B is caught up
3. Shut down B
4. Wipe local.* on B to remove old local.sources
5. Start up B with the --master option
6. Do a write on B (primes the oplog)
7. Shut down B
8. Shut down A and replace A's local.* files with a copy of B's local.* files
9. Start B with the --master option
10. Start A with --fastsync
If A is not healthy but the hardware is okay (power outage, server crash, etc.):
Skip the first two steps

Replace all of A's files with B's files in step 7.
If the hardware is not okay, replace A with a new machine and then follow the instructions in the previous paragraph.
Creating a slave from an existing master's disk image
--fastsync is a way to start a slave starting with an existing master disk image/backup. This option declares that the adminstrator guarantees
the image is correct and completely up to date with that of the master. If you have a full and complete copy of data from a master (and the master
is not accepting new writes concurrently!) you can use this option to avoid a full synchronization upon starting the slave.
Creating a slave from an existing slave's disk image
You can just copy the other slave's data file snapshot without any special options. Note data snapshots should only be taken when a mongod
process is down or in fsync-and-lock state.
Resyncing a slave that is too stale to recover
Slaves asynchronously apply write operations from the master. These operations are stored in the master's oplog. The oplog is finite in length. If a
slave is too far behind, a full resync will be necessary. See the Halted Replication page.
See Also
Replica Sets
One Slave Two Masters

This document shows an example session with one slave pulling data from two different masters.
Despite the example shown here, it is better, simpler, and generally recommended to have multiple mongod --slave
processes on a single box than to have one process pull from multiple masters.
A few notes:
Each master host has a different hostname (hostname:port).

Pulling the same database from two different databases can have unexpected results. This can be done in certainly limited ways, as the
data will tend to be merged, but there are some edge cases: for example the two masters should have exactly the same set of collections
or else some may not show up. Generally, this is not recommended.
Slaving a replica pair is unfortunately not currently supported – see SERVER-30.
$ mkdir /data/1
$ mkdir /data/2
$ mkdir /data/3
$ ./mongod --port 27020 --dbpath /data/1 --master &
$ ./mongod --port 27021 --dbpath /data/2 --master &
$ ./mongod --port 27022 --dbpath /data/3 --slave &
$ # add some data to masters

$ ./mongo localhost:27020
> use db1
> db.foo.insert({x:1})
^C
$ # master 2
> use db2
> db.foo.insert({x:999, note:"in db2"})
^C
$ # configure slave
> use local
> db.sources.insert({host:"localhost:27020"})
> db.sources.insert({host:"localhost:27021"})
> db.sources.find()
{ "_id" : ObjectId("4b8ecfac0cb095ca52b62949"), "host" : "localhost:27020" }
{ "_id" : ObjectId("4b8ecfc30cb095ca52b6294a"), "host" : "localhost:27021" }
> // wait a little, still connected to slave
> use db1

> db.foo.count()
2
> use db2
> db.foo.find()
{ "_id" : ObjectId("4b8ed00a1d42d47b3afa3c47"), "x" : 999, "note" : "in db2" }
> db.printSlaveReplicationInfo()
source: localhost:27020
syncedTo: Wed Mar 03 2010 16:04:35 GMT-0500 (EST)
= 2717secs ago (0.75hrs)
source: localhost:27021
syncedTo: Wed Mar 03 2010 16:09:31 GMT-0500 (EST)
= 2421secs ago (0.67hrs)
>
Replica Pairs
Setup of Replica Pairs
Consistency
Security
Replacing a Replica Pair Server
Querying the slave
What is and when should you use an arbiter?
Working with an existing (non-paired) database
See Also
Setup of Replica Pairs
Replica Sets will soon replace replica pairs. If you are just now setting up an instance, you may want to wait for that and use
master/slave replication in the meantime.
Mongo supports a concept of replica pairs. These databases automatically coordinate which is the master and which is the slave at a given point
in time.
At startup, the databases will negotiate which is master and which is slave. Upon an outage of one database server, the other will automatically
take over and become master from that point on. In the event of another failure in the future, master status would transfer back to the other
server. The databases manage this themselves internally.
Note: Generally, start with empty /data/db directories for each pair member when creating and running the pair for the first time. See section on
Existing Databases below for more information.
To start a pair of databases in this mode, run each as follows:
$ ./mongod --pairwith <remoteserver> --arbiter <arbiterserver>
where
remoteserver is the hostname of the other server in the pair. Append :port to the server hostname if you wish to run on a
nonstandard port number.
arbiterserver is the hostname (and optional port number) of an arbiter. An arbiter is a Mongo database server that helps negotiate
which member of the pair is master at a given point in time. Run the arbiter on a third machine; it is a "tie-breaker" effectively in
determining which server is master when the members of the pair cannot contact each other. You may also run with no arbiter by not
including the --arbiter option. In that case, both servers will assume master status if the network partitions.
One can manually check which database is currently the master:
$ ./mongo
> db.$cmd.findOne({ismaster:1});
{ "ismaster" : 0.0 , "remote" : "192.168.58.1:30001" , "ok" : 1.0 }
(Note: When security is on, remote is only returned if the connection is authenticated for the admin database.)
However, Mongo drivers with replica pair support normally manage this process for you.
Consistency
Members of a pair are only eventually consistent on a failover. If machine L of the pair was master and fails, its last couple seconds of operations
may not have made it to R - R will not have those operations applied to its dataset until L recovers later.
Security
Example security configuration when security is enabled:
$ ./mongo <lefthost>/admin -u <adminusername> -p<adminpassword>

> use local
^c
$ ./mongo <righthost>/admin -u <adminusername> -p<adminpassword>
> use local
Replacing a Replica Pair Server
When one of the servers in a Mongo replica pair set fails, should it come back online, the system recovers automatically. However, should a
machine completely fail, it will need to be replaced, and its replacement will begin with no data. The following procedure explains how to replace
one of the machines in a pair.
Let's assume nodes (n1, n2) is the old pair and that n2 dies. We want to switch to (n1,n3).
1. If possible, assure the dead n2 is offline and will not come back online: otherwise it may try communicating with its old pair partner.
2. We need to tell n1 to pair with n3 instead of n2. We do this with a replacepeer command. Be sure to check for a successful return
value from this operation.
n1> ./mongo n1/admin

> db.$cmd.findOne({replacepeer:1});
{
"info" : "adjust local.sources hostname; db restart now required" ,
"ok" : 1.0
}
At this point, n1 is still running but is reset to not be confused when it begins talking to n3 in the future. The server is still up although
replication is now disabled.
3. Restart n1 with the right command line to talk to n3
n1> ./mongod --pairwith n3 --arbiter <arbiterserver>
4. Start n3 paired with n1.
n3> ./mongod --pairwith n1 --arbiter <arbiterserver>
Note that n3 will not accept any operations as "master" until fully synced with n1, and that this may take some time if there is a
substantial amount of data on n1.
Querying the slave
You can query the slave if you set the slave ok flag. In the shell:
db.getMongo().setSlaveOk()
What is and when should you use an arbiter?
The arbiter is used in some situations to determine which side of a pair is master. In the event of a network partition (left and right are both up, but
can't communicate) whoever can talk to the arbiter becomes master.
If your left and right server are on the same switch, an arbiter isn't necessary. If you're running on the same ec2 availability zone, probably not
needed as well. But if you've got left and right on different ec2 availability zones, then an arbiter should be used.
Working with an existing (non-paired) database
Care must be taken when enabling a pair for the first time if you have existing datafiles you wish to use that were created from a singleton
database. Follow the following procedure to start the pair. Below, we call the two servers "left" and "right".
assure no mongod processes are running on both servers

we assume the data files to be kept are on server left. Check that there is no local.* datafiles in left's /data/db (--dbpath) directory. If there
are, remove them.
check that there are no datafiles at all on right's /data/db directory
start the left process with the appropriate command line including --pairwith argument
start the right process with the appropriate paired command line
If both left and right servers have datafiles in their dbpath directories at pair initiation, errors will occur. Further, you do not want a local database
(which contains replication metadata) during initiation of a new pair.
See Also
Replica Pairs in Ruby
Master Master Replication

Mongo does not support full master-master replication. However, for certain restricted use cases master-master can be used. Generally
speaking, we don't recommend the using a master-master configuration with MongoDB.
Master-master usages is eventually consistent.
To configure master-master, simply run both databases with both the --master and --slave parameters. For example, to set up this configuration
on a single machine as a test one might run:
$ nohup mongod --dbpath /data1/db --port 27017 --master --slave --source localhost:27018 > /tmp/dblog1
&
$ nohup mongod --dbpath /data2/db --port 27018 --master --slave --source localhost:27017 > /tmp/dblog2
&
This mode is safe for:

insert operations
delete operations by _id;
any query
Master-master should not be used if:
concurrent updates of single object may occur (including $inc and other updates)
A sample test session on a single computer follows:
$ # terminal 1, we run a mongod on default db port (27017)

$ ./mongod --slave --master --source localhost:10000
$ # terminal 2, we run a mongod on port 10000

$ ./mongod --slave --master --dbpath /data/slave --port 10000 --source localhost
$ # terminal 3, we run the shell here

$ ./mongo
> // 'db' is now connected to localhost:27017/test
> z = connect("localhost:10000/test")
> // 'z' is now connected to localhost:10000/test db
> db.foo.insert({x:7});
> z.foo.find()
{"_id" : ObjectId( "4ab917d7c50e4c10591ce3b6") , "x" : 7}
> db.foo.find()
> db.foo.find()
{"_id" : ObjectId( "4ab9182a938798896fd8a906") , "x" : 8}
> z.foo.find()
> z.foo.save({x:9})
> z.foo.find()
{"_id" : ObjectId( "4ab9188ac50e4c10591ce3b7") , "x" : 9}
> db.foo.find()
> z.foo.remove({x:8})
> db.foo.find()
> z.foo.find()
> db.foo.drop()
{"nIndexesWas" : 1 , "msg" : "all indexes deleted for collection" , "ns" : "test.foo" , "ok" : 1}
> db.foo.find()
> z.foo.find()
>
Replication Oplog Length

Replication uses an operation log ("oplog") to store write operations. These operations replay asynchronously on other nodes.
The length of the oplog is important if a secondary is down. The larger the log, the longer the secondary can be down and still recover. Once the
oplog has exceeded the downtime of the secondary, there is no way for the secondary to apply the operations; it will then have to do a full
synchronization of the data from the primary.
By default, on 64 bit builds, oplogs are quite large - perhaps 5% of disk space. Generally this is a reasonable setting.
The mongod --oplogSize command line parameter sets the size of the oplog.
This collection is named:
local.oplog.$main for master/slave replication;

local.oplog.rs for replica sets
See also
The Halted Replication page

Halted Replication
These instructions are for master/slave replication. For replica sets, see Resyncing a Very Stale Replica Set Member instead.
If you're running mongod with master-slave replication, there are certain scenarios where the slave will halt replication because it hasn't kept up
with the master's oplog.
The first is when a slave is prevented from replicating for an extended period of time, due perhaps to a network partition or the killing of the slave
process itself. The best solution in this case is to resyc the slave. To do this, open the mongo shell and point it at the slave:
$ mongo <slave_host_and_port>
Then run the resync command:
> use admin

> db.runCommand({resync: 1})
This will force a full resync of all data (which will be very slow on a large database). The same effect can be achieve dby stopping mongod on the
slave, delete all slave datafiles, and restarting it.
Increasing the OpLog Size
Since the oplog is a capped collection, it's allocated to a fixed size; this means that as more data is entered, the collection will loop around and
overwrite itself instead of growing beyond its pre-allocated size. If the slave can't keep up with this process, then replication will be halted. The
solution is to increase the size of the master's oplog. There are a couple of ways to do this, depending on how big your oplog will be and how
much downtime you can stand. But first you need to figure out how big an oplog you need.
If the current oplog size is wrong, how do you figure out what's right? The goal is not to let the oplog age out in the time it takes to clone the
database. The first step is to print the replication info. On the master node, run this command:
> db.printReplicationInfo();
You'll see output like this:
configured oplog size: 1048.576MB

log length start to end: 7200secs (2hrs)
oplog first event time: Wed Mar 03 2010 16:20:39 GMT-0500 (EST)
oplog last event time: Wed Mar 03 2010 18:20:39 GMT-0500 (EST)
now: Wed Mar 03 2010 18:40:34 GMT-0500 (EST)
This indicates that you're adding data to the database at a rate of 524MB/hr. If an initial clone takes 10 hours, then the oplog should be at least
5240MB, so something closer to 8GB would make for a safe bet.
The standard way of changing the oplog size involves stopping the mongod master, deleting the local.* oplog datafiles, and then restarting with
the oplog size you need, measured in MB:
$ # Stop mongod - killall mongod or kill -2 or ctrl-c) - then:
$ rm /data/db/local.*
$ mongod --oplog=8038 --master
Once you've changed the oplog size, restart with slave with --autoresync:
mongod --slave --autoresync
This method of oplog creation might pose a problem if you need a large oplog (say, > 10GB), since the time it takes mongod to pre-allocate the
oplog files may mean too much downtime. If this is the case, read on.
Manually Allocating OpLog Files
An alternative approach is to create the oplog files manually before shutting down mongod. Suppose you need an 20GB oplog; here's how you'd
go about creating the files:
1. Create a temporary directory, /tmp/local.

2. You'll be creating ten 2GB datafiles. Here's a shell script for doing just that:
cd /tmp/local
for i in {0..9}
do
echo $i
head -c 2146435072 /dev/zero > local.$i
done
Note that the datafiles aren't exactly 2GB due MongoDB's max int size.
3. Shut down the mongod master (kill -2) and then replace the oplog files:
$ mv /data/db/local.* /safe/place
$ mv /tmp/local/* /data/db/
4. Restart the master with the new oplog size:
$ mongod --master --oplogSize=20000
5. Finally, resync the slave. This can be done by shutting down the slave, deleting all its datafiles, and restarting it.
Sharding
MongoDB scales horizontally via an auto-sharding architecture.
Sharding offers:
Scaling out to thousands of nodes

Easy addition of new machines
Automatic balancing for changes in load and data distribution
Zero single points of failure
Automatic failover
*Sharding will be production-ready in MongoDB v1.6, estimated to be released in Aug 6th, 2010. * Please see the limitations page for progress
updates and current restrictions.
Documentation
1. What Is Sharding?
Here we provide an introduction to MongoDB's auto-sharding, highlighting its philosophy, use cases, and its core components.
2. How To Set Up and Manage a Cluster

How to set up a sharding cluster and manage it.
Configuration
Administration
Failover
3. Sharding Internals
Auto-Sharding implementation details.
4. Restrictions and Limitations

Sharding in the 1.5.x development branch is not yet production-ready. Here you can find out the current limitations and keep track of progress
towards the 1.6 production release.
5. FAQ
Common questions.
Presentations and Further Materials
Sharding Presentation from MongoSF April 2010
Sharding Introduction
MongoDB supports an automated sharding architecture, enabling horizontal scaling across multiple nodes. For applications that outgrow the
resources of a single database server, MongoDB can convert to a sharded cluster, automatically managing failover and balancing of nodes, with
few or no changes to the original application code.
This document explains MongoDB's auto-sharding approach to scalability in detail and provides an architectural overview of the various
components that enable it.
Since auto-sharding as of the 1.5.x branch is still alpha, be sure to acquaint yourself with the current limitations.
MongoDB's Auto-Sharding
Sharding in a Nutshell
Balancing and Failover
Scaling Model
Architectural Overview
Shards
Shard Keys
Chunks
Config Servers
Routing Processes
Operation Types
Server Layout
Configuration
MongoDB's Auto-Sharding
Sharding in a Nutshell
Sharding is the partitioning of data among multiple machines in an order-preserving manner. To take an example, let's imagine sharding a
collection of users by their state of residence. If we designate three machines as our shard servers, the first of those machines might contain
users from Alaska to Kansas, the second from Kentucky to New York, and the third from North Carolina to Wyoming.
Our application connects to the sharded cluster through a mongos process, which routes operations to the appropriate shard(s). In this way, the
sharded MongoDB cluster continues to look like a single-node database system to our application. But the system's capacity is greatly enhanced.
If our users collection receives heavy writes, those writes are now distributed across three shard servers. Queries continue to be efficient, as
well, because they too are distributed. And since the documents are organized in an order-preserving manner, any operations specifying the state
of residence will be routed only to those nodes containing that state.
Sharding occurs on a per-collection basis, not on the database as a whole. This makes sense since, as our application grows, certain collections
will grow much larger than others. For instance, if we were building a service like Twitter, our collection of tweets would likely be several orders of
magnitude larger than the next biggest collection. The size and throughput demands of such a collection would be prime for sharding, whereas
smaller collections would still live on a single server. In the context on MongoDB's sharded architecture, non-sharded collections will reside on just
one of the sharded nodes.
Balancing and Failover
A sharded architecture needs to handle balancing and failover. Balancing is necessary when the load on any one shard node grows out of
proportion with the remaining nodes. In this situation, the data must be redistributed to equalize load across shards. A good portion of the work
being applied to the 1.5.x branch is devoted to auto-balancing.
Automated failover is also quite important since proper system functioning requires that each shard node be always online. In practice, this means
that each shard consists of more than one machine in a configuration known as a replica set. A replica set is a set of n servers, frequently three or
more, each of which contains a replica of the entire data set for the given shard. One of the n servers in a replica set will always be master. If the
master replica fails, the remaining replicas are capable of electing a new master. Thus is automated failover provided for the individual shard.
Replica sets are another focus of development in 1.5.x. See the documentation on replica sets for more details.
Scaling Model
MongoDB's auto-sharding scaling model shares many similarities with Yahoo's PNUTS and Google's BigTable. Readers interested in detailed
discussions of distributed databases using order-preserving partitioning are encouraged to look at the PNUTS and BigTable white papers.
Architectural Overview
A MongoDB shard cluster consists of two or more shards, one or more config servers, and any number of routing processes to which the
application servers connect. Each of these components is described below in detail.
Shards
Each shard consists of one or more servers and stores data using mongod processes (mongod being the core MongoDB database process). In a
production situation, each shard will consist of multiple replicated servers per shard to ensure availability and automated failover. The set of
servers/mongod process within the shard comprise a replica set.
Replica sets, as discussed earlier, represent an improved version of MongoDB's replication (SERVER-557).
For testing, you can use sharding with a single mongod instance per shard. If you need redundancy, use one or more slaves for each shard's
mongod master. This configuration will require manual failover until replica sets become available.
Shard Keys
To partition a collection, we specify a shard key pattern. This pattern is similar to the key pattern used to define an index; it names one or more
fields to define the key upon which we distribute data. Some example shard key patterns include the following:
{ state : 1 }
{ name : 1 }
{ _id : 1 }
{ lastname : 1, firstname : 1 }
{ tag : 1, timestamp : -1 }
MongoDB's sharding is order-preserving; adjacent data by shard key tends to be on the same server. The config database stores all the metadata
indicating the location of data by range:
Chunks
A chunk is a contiguous range of data from a particular collection. Chunks are described as a triple of collection, minKey, and maxKey.
Thus, the shard key K of a given document assigns that document to the chunk where minKey <= K < maxKey.
Chunks grow to a maximum size, usually 200MB. Once a chunk has reached that approximate size, the chunk splits into two new chunks. When
a particular shard has excess data, chunks will then migrate to other shards in the system. The addition of a new shard will also influence the
migration of chunks.
When choosing a shard key, keep in mind that these values should be granular enough to ensure an even distribution of data. For instance, in the
above example, where we're sharding on name, we have to be careful that we don't have a disproportionate number of users with the same
name. In that case, the individual chunk can become too large and find itself unable to split, e.g., where the entire range comprises just a single
key.
Thus, if it's possible that a single value within the shard key range might grow exceptionally large, it's best to use a compound shard key instead
so that further discrimination of the values will be possible.
Config Servers
The config servers store the cluster's metadata, which includes basic information on each shard server and the chunks contained therein.
Chunk information is the main data stored by the config servers. Each config server has a complete copy of all chunk information. A two-phase
commit is used to ensure the consistency of the configuration data among the config servers.
If any of the config servers is down, the cluster's meta-data goes read only. However, even in such a failure state, the MongoDB cluster can still
be read from and written to.
Routing Processes
The mongos process can be thought of as a routing and coordination process that makes the various components of the cluster look like a single
system. When receiving client requests, the mongos process routes the request to the appropriate server(s) and merges any results to be sent
back to the client.
mongos processes have no persistent state; rather, they pull their state from the config server on startup. Any changes that occur on the the
config servers are propagated to each mongos process.
mongos processes can run on any server desired. They may be run on the shard servers themselves, but are lightweight enough to exist on each
application server. There are no limits on the number of mongos processes that can be run simultaneously since these processes do not
coordinate between one another.
Operation Types
Operations on a sharded system fall into one of two categories: global and targeted.
For targeted operations, mongos communicates with a very small number of shards -- often a single shard. Such targeted operations are quite
efficient.
Global operations involve the mongos process reaching out to all (or most) shards in the system.
The following table shows various operations and their type. For the examples below, assume a shard key of { x : 1 }.
Operation Type Comments
db.foo.find( { x : 300 } ) Targeted Queries a single shard.
db.foo.find( { x : 300, age : 40 } ) Targeted Queries a single shard.
db.foo.find( { age : 40 } ) Global Queries all shards.
db.foo.find() Global sequential

db.foo.find(...).count() Variable Same as the corresponding find() operation
db.foo.find(...).sort( { age : 1 } ) Global parallel
db.foo.find(...).sort( { x : 1 } ) Global sequential
db.foo.count() Global parallel
db.foo.insert( <object> ) Targeted
db.foo.update( { x : 100 }, <object> ) Targeted

db.foo.remove( { x : 100 } )
db.foo.update( { age : 40 }, <object> ) Global
db.foo.remove( { age : 40 } )
db.getLastError()
db.foo.ensureIndex(...) Global
Server Layout
Machines may be organized in a variety of fashions. For instance, it's possible to have separate machines for each config server process,
mongos process, and mongod process. However, this can be overkill since the load is almost certainly low on the config servers. Here, then, is
an example where some sharing of physical machines is used to lay out a cluster.
Yet more configurations are imaginable, especially when it comes to mongos. For example, it's possible to run mongos processes on all of
servers 1-6. Alternatively, as suggested earlier, the mongos processes can exists on each application server (server 7). There is some potential
benefit to this configuration, as the communications between app server and mongos then can occur over the localhost interface.
Configuration
Sharding becomes a bit easier to understand one you've set it up. It's even possible to set up a sharded cluster on a single machine. To try it for
yourself, see the configuration docs.
Configuring Sharding
Introduction
This document describes the steps involved in setting up a basic sharding cluster. A sharding cluster requires, at minimum, three components:
1. Two or more shards.

2. At least one config server.
3. A mongos routing process.
For testing purposes, it's possible to start all the required processes on a single server, whereas in a production situation, a number of server
configurations are possible.
Once the shards, config servers, and mongos processes are running, configuration is simply a matter of issuing a series of commands to
establish the various shards as being part of the cluster. Once the cluster has been established, you can begin sharding individual collections.
This document is fairly detailed; if you're the kind of person who prefers a terse, code-only explanation, see the sample shard configuration. If
you'd like a quick script to set up a test cluster on a single machine, we have a python sharding script that can do the trick.
Introduction
1. Sharding Components
Shard Servers
Config Servers
mongos Router
2. Configuring the Shard Cluster
Adding shards
Optional Parameters
Listing shards
Removing a shard
Enabling Sharding on a Database
3. Sharding a Collection
Relevant Examples and Docs
1. Sharding Components
First, start the individual shards, config servers, and mongos processes.
Shard Servers
Run mongod on the shard servers. Use the --shardsvr command line parameter to indicate this mongod is a shard. For auto failover support,
replica sets will be required. See replica sets for more info.
Note that replica pairs will never be supported as shards.
To get started with a simple test, we recommend running a single mongod process per shard, as a test configuration doesn't demand automated
failover.
Config Servers
Run mongod on the config server(s) with the --configsvr command line parameter. If the config servers are running on a shared machine, be sure
to provide a separate dbpath for the config data (--dbpath command line parameter).
mongos Router
Run mongos on the servers of your choice. Specify the --configdb parameter to indicate location of the config database(s).
2. Configuring the Shard Cluster
Once the shard components are running, issue the sharding commands. You may want to automate or record your steps below in a .js file for
replay in the shell when needed.
Start by connecting to one of the mongos processes, and then switch to the admin database before issuing any commands.
Keep in mind that once these commands are run, the configuration data will be persisted to the config servers. So, regardless of the number of
mongos processes you've launched, you'll only need run these commands on one of those processes.
You can connect to the admin database via mongos like so:
./mongo <mongos-hostname>:<mongos-port>/admin
> db
admin
Adding shards
Each shard can consist of more than one server (see replica sets); however, for testing, only a single server with one mongod instance need be
used.
You must explicitly add each shard to the cluster's configuration using the addshard command:
> db.runCommand( { addshard : "<serverhostname>[:<port>]" } );

{"ok" : 1 , "added" : ...}
Run this command once for each shard in the cluster.
If the individual shards consist of replica sets, they can be added by specifying replicaSetName
/<serverhostname>[:port][,serverhostname2[:port],...], where at least one server in the replica set is given.
> db.runCommand( { addshard : "foo/<serverhostname>[:<port>]" } );

{"ok" : 1 , "added" : "foo"}
Optional Parameters
name
Each shard has a name, which can be specified using the name option. If no name is given, one will be assigned automatically.
maxSize
The addshard command accepts an optional maxSize parameter. This parameter lets you tell the system a maximum amount of disk space in
megabytes to use on the specified shard. If unspecified, the system will use the entire disk. maxSize is useful when you have machines with
different disk capacities or when you want to prevent storage of too much data on a particular shard.
As an example:
> db.runCommand( { addshard : "sf103", maxSize:100000 } );
Listing shards
To see current set of configured shards, run the listshards command:
> db.runCommand( { listshards : 1 } );
This way, you can verify that all the shard have been committed to the system.
Removing a shard
Before a shard can be removed, we have to make sure that all the chunks and databases that once lived there were relocated to other shards.
The 'removeshard' command takes care of "draining" the chunks out of a shard for us. To start the shard removal, you can issue the command
> db.runCommand( { removeshard : "localhost:10000" } );

{ msg : "draining started succesfully" , state: "started" , shard :"localhost:10000" , ok : 1
That will put the shard in "draining mode". Its chunks are going to be moved away slowly over time, so not to cause any disturbance to a running
system. The command will return right away but the draining task will continue on the background. If you issue the command again during it, you'll
get a progress report instead:

{ msg: "draining ongoing" , state: "ongoing" , remaining : { chunks :23 , dbs : 1 }, ok : 1 }
Whereas the chunks will be removed automatically from that shard, the databases hosted there will need to be moved manually. (This has to do
with a current limitation that will go away eventually):
> db.runCommand( { moveprimary : "test", to : "localhost:10001" } );

{ "primary" : "localhost:10001", "ok" : 1 }
When the shard is empty, you could issue the 'removeshard' command again and that will clean up all metadata information:

{ msg: "remove shard completed succesfully" , stage: "completed", host: "localhost:10000", ok : 1 }
After the 'removeshard' command reported being done with that shard, you can take that process down.
Enabling Sharding on a Database
Once you've added the various shards, you can enable sharding on a database. This is an important step; without it, all collection in the database
will be stored on the same shard.
> db.runCommand( { enablesharding : "<dbname>" } );
Once enabled, mongos will place different collections for the database on different shards, with the caveat that each collection will still exists on
one shard only. To enable real partitioning of data, we have to shard an individual collection.
3. Sharding a Collection
Use the shardcollection command to shard a collection. When you shard a collection, you must specify the shard key. If there is data in the
collection, mongo will require an index to be create upfront (it speeds up the chunking process) otherwise an index will be automatically created
for you.
> db.runCommand( { shardcollection : "<namespace>",

key : <shardkeypatternobject> });
For example, let's assume we want to shard a GridFS chunks collection stored in the test database. We'd want to shard on the files_id key,
so we'd invoke the shardcollection command like so:
> db.runCommand( { shardcollection : "test.fs.chunks", key : { files_id : 1 } } )

{"ok" : 1}
One note: a sharded collection can have only one unique index, which must exist on the shard key. No other unique indexes can exist on the
collection.
Of course, a unique shard key wouldn't make sense on the GridFS chunks collection. But it'd be practically a necessity for a users collection
sharded on email address:
db.runCommand( { shardcollection : "test.users" , key : { email : 1 } , unique : true } );
Relevant Examples and Docs
Examples
Sample configuration session

The following example shows how to run a simple shard setup on a single machine for testing purposes: Sharding JS Test.
Docs
Sharding Administration
Notes on TCP Port Numbers
A Sample Configuration Session

The following example uses two shards (one server each), one config db, and one mongos process, all running on a single test server. In addition
to the script below, a python script for starting and configuring shard components on a single machine is available.
Creating the Shards
First, start up a couple _mongod_s to be your shards.
$ mkdir /data/db/a /data/db/b

$ ./mongod --shardsvr --dbpath /data/db/a --port 10000 > /tmp/sharda.log &
$ cat /tmp/sharda.log
$ ./mongod --shardsvr --dbpath /data/db/b --port 10001 > /tmp/shardb.log &
$ cat /tmp/shardb.log
Now you need a configuration server and mongos:
$ mkdir /data/db/config
$ ./mongod --configsvr --dbpath /data/db/config --port 20000 > /tmp/configdb.log &
$ cat /tmp/configdb.log
$ ./mongos --configdb localhost:20000 > /tmp/mongos.log &
$ cat /tmp/mongos.log
mongos does not require a data directory, it gets its information from the config server.
In a real production setup, mongod's, mongos's and configs would live on different machines. The use of hostnames or IP
addresses is mandatory in that case. 'localhost' appearance here is merely illustrative – but fully functional – and should be
confined to single-machine, testing scenarios only.
You can toy with sharding by using a small --chunkSize, e.g. 1MB. This is more satisfying when you're playing around, as you won't have to
insert 200MB of documents before you start seeing them moving around. It should not be used in production.
$ ./mongos --configdb localhost:20000 --chunkSize 1 > /tmp/mongos.log &
Setting up the Cluster
We need to run a few commands on the shell to hook everything up. Start the shell, connecting to the mongos process (at localhost:27017 if you
followed the steps above).
To set up our cluster, we'll add the two shards (a and b).
$ ./mongo
connecting to: test
> use admin
> db.runCommand( { addshard : "localhost:10000" } )
{ "shardadded" : "shard0000", "ok" : 1 }
> db.runCommand( { addshard : "localhost:10001" } )
{ "shardadded" : "shard0001", "ok" : 1 }
Now you need to tell the database that you want to spread out your data at a database and collection level. You have to give the collection a key
(or keys) to partition by.
This is similar to creating an index on a collection.
> db.runCommand( { enablesharding : "test" } )

{"ok" : 1}
> db.runCommand( { shardcollection : "test.people", key : {name : 1} } )
{"ok" : 1}
Administration
To see what's going on in the cluster, use the config database.
> use config

switched to db config
> show collections
chunks
databases
lockpings
locks
mongos
settings
shards
system.indexes
version
These collections contain all of the sharding configuration information.
Upgrading from a Non-Sharded System

A mongod process can become part of a sharded cluster without any change to that process or downtime. If you haven't done so yet, feel free to
have a look at the Sharding Introduction to familiarize yourself with the components of a sharded cluster and at the Sample Configuration
Sesssion to get to know the basic commands involved.
Sharding is a new feature introduced at the 1.6.0 release. This page assumes your non-sharded mongod is on that release.
Adding the mongod process to a cluster
If you haven't changed the mongod default port, it would be using port 27017. You care about this now because a mongo shell will always try to
connect to it by default. But in a sharded environment, you want your shell to connect to a mongos instead.
If the port 27017 is taken by a mongod process, you'd need to bring up the mongos in a different port. Assuming that port is 30000 you can
connect your shell to it by issuing:
$ mongo <mongos-host-address>:30000/admin
We're switching directly to the admin database on the mongos process. That's where we will be able to issue the following command

connecting to: <mongos-address>:30000/admin
> db.runCommand( { addshard : "192.168.25.203:27017" } )
> { "shardAdded" : "shard0000", "ok" : 1 }
The host address and port you see on the command are the original mongod's. All the databases of that process were added to the cluster and
are accessible now through mongos.
> db.runCommand( { listdatabases : 1 } )

{
"databases" : [
{
"name" : "mydb"
...
"shards" : {
"shard0000" : <size-in-shard00000>
}
},
...
Note that that doesn't mean that the database or any of its collections is sharded. They haven't moved (see next). All we did so far is to make
them visible within the cluster environment.
You should stop accessing the former stand-alone mongod directly and should have all the clients connect to a mongos process, just as we've
been doing here.
Sharding a collection
All the databases of your mongod-process-turned-shard can be chunked and balanced among the cluster's shards. The commands and examples
to do so are listed at the
Configuring Sharding page. Note that a chunk size defaults to 200MB in version 1.6.0, so if you want to change that – for testing purposes, say –
you would do so by starting the mongos process with the additional --chunkSize parameter.
Difference between upgrading and starting anew
You should pay attention to the host addresses and ports when upgrading, is all.
Again, if you haven't changed the default ports of your mongod process, it would be listening on 27017, which is the port that mongos would try to
bind by default, too.
Sharding Administration
Here we present a list of useful commands for obtaining information about a sharding cluster.
To set up a sharding cluster, see the docs on sharding configuration.
Identifying a Shard Cluster

List Existing Shards
List Which Databases are Sharded
View Sharding Details
Chunk Operations
Identifying a Shard Cluster
// Test if we're speaking to a mongos process or

// straight to a mongod process
> db.runCommand({ isdbgrid : 1});
// if connected to mongos, this command returns { ismaster: 0.0, msg: "isdbgrid" }

> db.runCommand({ismaster:1});
List Existing Shards
> db.runCommand({ listshards : 1});

{"servers" :
[{"_id" : ObjectId( "4a9d40c981ba1487ccfaa634") ,
"host" : "localhost:10000"},
{"_id" : ObjectId( "4a9d40df81ba1487ccfaa635") ,
"host" : "localhost:10001"}
],
"ok" : 1
}
List Which Databases are Sharded
Here we query the config database, albeit through mongos. The getSisterDB command is used to return the config database.
> config = db.getSisterDB("config")

> config.system.namespaces.find()
View Sharding Details

> use admin
> db.printShardingStatus();
// A very basic sharding configuration on localhost

sharding version: { "_id" : 1, "version" : 2 }
shards:
{ "_id" : ObjectId("4bd9ae3e0a2e26420e556876"), "host" : "localhost:30001" }
{ "_id" : ObjectId("4bd9ae420a2e26420e556877"), "host" : "localhost:30002" }
{ "_id" : ObjectId("4bd9ae460a2e26420e556878"), "host" : "localhost:30003" }
databases:
{ "name" : "admin", "partitioned" : false,
"primary" : "localhost:20001",
"_id" : ObjectId("4bd9add2c0302e394c6844b6") }
my chunks
{ "name" : "foo", "partitioned" : true,

"primary" : "localhost:30002",
"sharded" : { "foo.foo" : { "key" : { "_id" : 1 }, "unique" : false } },
"_id" : ObjectId("4bd9ae60c0302e394c6844b7") }
my chunks
foo.foo { "_id" : { $minKey : 1 } } -->> { "_id" : { $maxKey : 1 } }
on : localhost:30002 { "t" : 1272557259000, "i" : 1 }
Notice the output to the printShardingStatus command. First, we see the locations the the three shards comprising the cluster. Next, the
various databases living on the cluster are displayed.
The first database shown is the admin database, which has not bee partitioned. The primary field indicates the location of the database, which, in
the case of the admin database, is on the config server running on port 20001.
The second database is partitioned, and it's easy to see the shard key and the location and ranges of chunks comprising the partition. Since
there's no data in the foo database, only a single chunk exists. That single chunk includes the entire range of possible shard keys.
Chunk Operations
MongoDB v1.6 will managing the arrangement chunks automatically. However, it may be desirable to move a chunk manually; here's the
command to do that:
db.runCommand( { movechunk : "test.blog.posts" ,

find : { author : "eliot" } ,
to : "shard1" } )
Parameters:
movechunk: a full collection namespace, including the database name

find: a query expression that falls within the chunk to be moved
to: shard id where the chunk will be moved
Sharding and Failover

A properly-configured MongoDB shard cluster will have no single point of failure.
This document describes the various potential failure scenarios of components within a shard cluster, and how failure is handled in each situation.
1. Failure of a mongos routing process.
One mongos routing process will be run on each application server, and that server will communicate to the cluster exclusively through the
mongos process. mongos process aren't persistent; rather, they gather all necessary config data on startup from the config server.
This means that the failure of any one application server will have no effect on the shard cluster as a whole, and all other application servers will
continue to function normally. Recovery is simply a matter starting up a new app server and mongos process.
2. Failure of a single mongod server within a shard.

Each shard will consist of a group of n servers in a configuration known as a replica set. If any one server in the replica set fails, read and write
operations on the shard are still permitted. What's more, no data need be lost on the failure of a server because the replica allows an option on
write that forces replication of the write before returning. This is similar to setting W to 2 on Amazon's Dynamo.
Replica sets will be available as of MongoDB v1.6. Read more about replica set internals or follow the jira issue.
3. Failure of all mongod servers comprising a shard.
If all replicas within a shard are down, the data within that shard will be unavailable. However, operations that can be resolved at other shards will
continue to work properly. See the documentation on global and targeted operations to see why this is so.
If the shard is configured as a replica set, with at least one member of the set in another data center, then an outage of an ensure shard is
extremely unlikely. This will be the recommended configuration for maximum redundancy.
4. Failure of a config server.
A production shard cluster will have three config server processes, each existing on a separate machine. Writes to config servers use a
two-phase commit to ensure an atomic and replicated transaction of the shard cluster's metadata.
On the failure of any one config server, the system's metadata becomes read-only. The system will continue to function, but chunks will be unable
to split within a single shard or migrate across shards. For most use cases, this will present few problems, since changes to the chunk metadata
will be infrequent.
That said, it will be important that the down config server be restored in a reasonable time period (say, a day) so that shards do not become
unbalanced due to lack of migrates (again, for many production situations, this may not be an urgent matter).
Sharding Limits
Sharding Release 1 (MongoDB v1.6.0)
Differences from Unsharded Configurations
Sharding must be ran in trusted security mode, without explicit security.

Shard keys are immutable in the current version.
All (non-multi)updates, upserts, and inserts must include the current shard key. This may cause issues for anyone using a mapping
library since you don't have full control of updates.
$where
$where works with sharding. However do not reference the db object from the $where function (one normally does not do this anyway).
db.eval
db.eval() may not be used with sharded collections. However, you may use db.eval() if the evaluation function accesses unsharded collections
within your database. Use map/reduce in sharded environments.
getPrevError
getPrevError is unsupported for sharded databases, and may remain so in future releases (TBD). Let us know if this causes a problem for you.
Unique Indexes
For a sharded collection, you may only (optionally) specify a unique constraint on the shard key. Other secondary indexes work (via a global
operation) as long as no unique constraint is specified.
Counts
Count is supported with sharding; however, a "count all in collection" will not be instantaneous for a sharded collection as it is for an unsharded
collection.
Scale Limits
Goal is support of systems of up to 1,000 shards. Testing so far has been limited to clusters with a modest number of shards (e.g., 20). More
information will be reported here later on any scaling limitations which are encountered.
MongoDB sharding supports two styles of operations -- targeted and global. On giant systems, global operations will be of less applicability.
Sharding Internals
This section includes internal implementation details for MongoDB auto sharding. See also the main sharding documentation.
Note: some internals docs could be out of date -- if you see that let us know so we can fix.
Internals
Moving Chunks
Sharding Config Schema
Sharding Design
Sharding Use Cases
Shard Ownership
Splitting Chunks
Unit Tests
./mongo --nodb jstests/sharding/*js
Moving Chunks
inc version
try to set on from
if set is successful, have it "locked"
start transfer
finish transfer
commit result
max version for a shard is MAX( chunks on shard )

this poses slight problem when moving last chunk off of a shard, so add a special marker
Sharding Config Schema

Sharding configuration schema. This lives in the config servers.
Collections
version
This is a singleton that contains the current meta-data version number.
> db.getCollection("version").findOne()
{ "_id" : 1, "version" : 2 }
settings
Key/Value table for configurable options (chunkSize)
> db.settings.find()
{ "_id" : "chunksize", "value" : 200 }
{ "_id" : "balancer", "who" : "ubuntu:27017", "x" : ObjectId("4bd0cb39503139ae28630ee9") }
shards
Stores information about the shards.
> db.shards.findOne()
{ "_id" : "shard0", "host" : "localhost:30001" }
databases
{
"_id" : "admin",
"partitioned" : false,
"primary" : "localhost:20001"
}
chunks
{
"_id" : "test.foo-x_MinKey",
"lastmod" : {
"t" : 1271946858000,
"i" : 1
},
"ns" : "test.foo",
"min" : {
"x" : { $minKey : 1 }
},
"max" : {
"x" : { $maxKey : 1 }
},
"shard" : "localhost:30002"
}
mongos
Record of all mongos affiliated with this cluster. mongos will ping every 30 seconds so we know who is alive.
> db.mongos.findOne()
{
"_id" : "erh-wd1:27017",
"ping" : "Fri Apr 23 2010 11:08:39 GMT-0400 (EST)",
"up" : 30
}
changelog
Human readable log of all meta-data changes. Capped collection that defaults to 10mb.
> db.changelog.findOne()
{
"_id" : "erh-wd1-2010-3-21-17-24-0",
"server" : "erh-wd1",
"time" : "Wed Apr 21 2010 13:24:24 GMT-0400 (EST)",
"what" : "split",
"ns" : "test.foo",
"details" : {
"before" : {
"min" : {
"x" : { $minKey : 1 }
},
"max" : {
"x" : { $maxKey : 1 }
}
},
"left" : {
"min" : {
"x" : { $minKey : 1 }
},
"max" : {
"x" : 5
}
},
"right" : {
"min" : {
"x" : 5
},
"max" : {
"x" : { $maxKey : 1 }
}
}
}
}
Changes
2 (<= 1.5.0) -> 3 (1.5.1)
shards : _id is now the name

databases : _id is now the db name
general : all references to a shard can be via name or host
Sharding Design
concepts
config database - the top level database that stores information about servers and where things live.
shard. this can be either a single server or a replica pair.
database - one top level namespace. a database can be partitioned or not
chunk - a region of data from a particular collection. A chunk can be though of as ( collectionname,fieldname,lowvalue,highvalue). The
range is inclusive on the low end and exclusive on the high end, i.e., [lowvalue,highvalue).
components and database collections
config database
config.servers - this contains all of the servers that the system has. These are logical servers. So for a replica pair, the entry would be
192.168.0.10,192.168.0.11
config.databases - all of the databases known to the system. This contains the primary server for a database, and information about
whether its partitioned or not.
config.shards - a list of all database shards. Each shard is a db pair, each of which runs a db process.
config.homes - specifies which shard is home for a given client db.
shard databases
client.system.chunklocations - the home shard for a given client db contains a client.system.chunklocations collection. this
collection lists where to find particular chunks; that is, it maps chunk->shard.
mongos process
"routes" request to proper db's, and performs merges. can have a couple per system, or can have 1 per client server.
gets chunk locations from the client db's home shard. load lazily to avoid using too much mem.
chunk information is cached by mongos. This information can be stale at a mongos (it is always up to date at the owning
shard; you cannot migrate an item if the owning shard is down). If so, the shard contacted will tell us so and we can then
retry to the proper location.
db operations
moveprimary - move a database's primary server

migrate - migrate a chunk from one machine to another.
lock and migrate
shard db's coordinate with home shard to atomically pass over ownership of the chunk (two phase commit)
split - split a chunk that is growing too large into pieces. as the two new chunks are on the same machine after the split, this is really just
a metadata update and very fast.
reconfiguration operations
add shard - dbgrid processes should lazy load information on a new (unknown) shard when encountered.
retire shard - in background gradually migrate all chunks off
minimizing lock time
If a chunk is migrating and is 50MB, that might take 5-10 seconds which is too long for the chunk to be locked.
We could perform the migrate much like Cloner works,where we copy the objects and then apply all operations that happened during copying.
This way lock time is minimal.
Sharding Use Cases

What specific use cases do we want to address with db partioning (and other techniques) that are challenging to scale? List here for discussion.
video site (e.g., youtube) (also, GridFS scale-up)

seems straightforward: partition by video
for related videos feature, see search below
social networking (e.g., facebook)
this can be quite hard to partition, because it is difficult to cluster people.
very high RPS sites with small datasets
N replicas, instead of partioning, might help here
replicas only work if the dataset is really small as we are using/wasting the same RAM on each replica. thus, partioning
might help us with ram cache efficiency even if entire data set fits on one or two drives.
twitter
search & tagging
Log Processing
Use cases related to map-reduce like things.
massive sort
top N queries per day
compare data from two nonadjacent time periods
Shard Ownership
By shard ownership we mean which server owns a particular key range.
Early draft/thoughts will change:
Contract
the master copy of the ownership information is in the config database

mongos instances have cached info on which server owns a shard. this information may be stale.
mongod instances have definitive information on who owns a shard (atomic with the config db) when they know about a shards
ownership
mongod
The mongod processes maintain a cache of shards the mongod instance owns:
map<ShardKey,state> ownership
State values are as follows:
missing - no element in the map means no information available. In such a situation we should query the config database to get the state.
1 - this instance owns the shard
0 - this instance does not own the shard (indicates we queried the config database and found another owner, and remembered that fact)
Initial Assignment of a region to a node.
This is trivial: add the configuration to the config db. As the ShardKey is new, no nodes have any cached information.
Splitting a Key Range
The mongod instance A which owns the range R breaks it into R1,R2 which are still owned by it. It updates the config db. We take care to handle
the config db crashing or being unreachable on the split:
lock(R) on A
update the config db -- ideally atomically perhaps with eval(). await return code.
ownership[R].erase
unlock(R) on A
After the above the cache has no information on the R,R1,R2 ownerships, and will requery configdb on the next request. If the config db crashed
and failed to apply the operation, we are still consistent.
Migrate ownership of keyrange R from server A->B. We assume here that B is the coordinator of the job:
B copies range from A

lock(R) on A and B
B copies any additional operations from A (fast)
clear ownership maps for R on A and B. B waits for a response from A on this operation.
B then updates the ownership data in the config db. (Perhaps even fsyncing.) await return code.
unlock(R) on B
delete R on A (cleanup)
unlock (R) on A
We clear the ownership maps first. That way, if the config db update fails, nothing bad happens, IF mongos filters data upon receipt for being in
the correct ranges (or in its query parameters).
R stays locked on A for the cleanup work, but as that shard no longer owns the range, this is not an issue even if slow. It stays locked for that
operation in case the shard were to quickly migrate back.
Migrating Empty Shards
Typically we migrate a shard after a split. After certain split scenarios, a shard may be empty but we want to migrate it.
Splitting Chunks
Normally, splitting chunks is done automatically for you. Currently, the splits happen as a side effect of inserting (and are transparent). In the
future, there may be other cases where a chunk is automatically split.
A recently split chunk may be moved immediately to a new shard if the system finds that future insertions will benefit from that move. (Chunk
moves are transparent, too.)
Moreover, MongoDB has a sub-system called Balancer, which constantly monitors shards loads and, as you guessed, moves chunks around if it
finds an imbalance. Balancing chunks automatically helps incremental scalability. If you add a new shard to the system, some chunks will
eventually be moved to that shard to spread out the load.
That all being said, in certain circumstances one may need to force a split manually.
The Balancer will treat all chunks the same way, regardless if they were generated by a manual or an automatic split.
The following command splits the chunk where the _id 99 would reside using that key as the split point. Note that a key need not exist for a chunk
to use it in its range.
> use admin

> db.runCommand( { split : "test.foo" , middle : { _id : 99 } } )
...
The following command splits the chunk where the _id 99 would reside in two. The key used as the middle key is computed internally to roughly
divide the chunk in equally sized parts.
> use admin

> db.runCommand( { split : "test.foo" , find : { _id : 99 } } )
...
Sharding FAQ
How does sharding work with replication?
Where do unsharded collections go if sharding is enabled for a database?
When will data be on more than one shard?
What happens if I try to update a document on a chunk that is being migrated?
What if a shard is down or slow and I do a query?
How do queries distribute across shards?
Now that I sharded my collection, how do I <...> (e.g. drop it)?
If I don't shard on _id how is it kept unique?
Why is all my data on one server?
How does sharding work with replication?
Each shard is a logical collection of partitioned data. The shard could consist of a single server or a cluster of replicas. Typically in production one
would use a replica set for each shard.
Where do unsharded collections go if sharding is enabled for a database?
In alpha 2 unsharded data goes to the "primary" for the database specified (query config.databases to see details). Future versions will parcel out
unsharded collections to different shards (that is, a collection could be on any shard, but will be on only a single shard if unsharded).
When will data be on more than one shard?
MongoDB sharding is range based. So all the objects in a collection get put into a chunk. Only when there is more than 1 chunk is there an option
for multiple shards to get data. Right now, the chunk size is 50mb, so you need at least 50mb for a migration to occur.
What happens if I try to update a document on a chunk that is being migrated?
The update will go through immediately on the old shard, and then the change will be replicated to the new shard before ownership transfers.
What if a shard is down or slow and I do a query?
If a shard is down, the query will return an error. If a shard is responding slowly, mongos will wait for it. You won't get partial results.
How do queries distribute across shards?
There are a few different cases to consider, depending on the query keys and the sort keys. Suppose 3 distinct attributes, X, Y, and Z, where X is
the shard key. A query that keys on X and sorts on X will translate straightforwardly to a series of queries against successive shards in X-order. A
query that keys on X and sorts on Y will execute in parallel on the appropriate shards, and perform a merge sort keyed on Y of the documents
found. A query that keys on Y must run on all shards: if the query sorts by X, the query will serialize over shards in X-order; if the query sorts by Z,
the query will parallelize over shards and perform a merge sort keyed on Z of the documents found.
Now that I sharded my collection, how do I <...> (e.g. drop it)?
Even if chunked, your data is still part of a collection and so all the collection commands apply.
If I don't shard on _id how is it kept unique?
If you don't use _id as the shard key then it is your responsibility to keep the _id unique. If you have duplicate _id values in your collection bad
things will happen (as mstearn says).
Why is all my data on one server?
MongoDB sharding breaks data into chunks. By default, these chunks are 200mb. Sharding will keep chunks balanced across shards. This
means that you need many chunks to trigger balancing, typically 2gb of data or so. db.printShardingStatus() will tell you how many chunks you
have, typically need 10 to start balancing.
Hosting Center
Cloud-Style
MongoHQ provides cloud-style hosted MongoDB instances
Mongo Machine is currently in private beta
Dedicated Servers
ServerBeach offers preconfigured, dedicated MongoDB servers Blog
EngineYard supports MongoDB on its private cloud.
VPS
(mt) Media Temple's (ve) server platform is an excellent choice for easy MongoDB deployment.
Dreamhost offers instant configuration and deployment of MongoDB
LOCUM Hosting House is a project-oriented shared hosting and VDS. MongoDB is available for all customers as a part of their
subscription plan.
Setup Instructions for Others

Amazon EC2
Joyent
Linode
Webfaction
Amazon EC2
Instance Types
Linux
EC2 TCP Port Management
EBS Snapshotting
EBS vs. Local Drives
MongoDB runs well on Amazon EC2 . This page includes some notes in this regard.
Instance Types
MongoDB works on most EC2 types including Linux and Windows. We recommend you use a 64 bit instance as this is required for all MongoDB
databases of significant size. Additionally, we find that the larger instances tend to be on the freshest ec2 hardware.
Linux
One can download a binary or build from source. Generally it is easier to download a binary. We can download and run the binary without being
root. For example on 64 bit Linux:
[~]$ curl -O http://downloads.mongodb.org/linux/mongodb-linux-x86_64-1.0.1.tgz

[~]$ tar -xzf mongodb-linux-x86_64-1.0.1.tgz
[~]$ cd mongodb-linux-x86_64-1.0.1/bin
[bin]$ ./mongod --version
Before running the database one should decide where to put datafiles. Run df -h to see volumes. On some images /mnt will be the many locally
attached storage volume. Alternatively you may want to use Elastic Block Store which will have a different mount point. Regardless, create a
directory in the desired location and then run the database:
mkdir /mnt/db
./mongod --fork --logpath ~/mongod.log --dbpath /mnt/db/
EC2 TCP Port Management
By default the database will now be listening on port 27017. The web administrative UI will be on port 28017.
EBS Snapshotting
v1.3.1+
If your datafiles are on an EBS volume, you can snapshot them for backups. Use the fsync lock command to lock the database to prevent writes.
Then, snapshot the volume. Then use the unlock command to allow writes to the database again. See the fsync documentation for more
information.
This method may also be used with slave databases.
EBS vs. Local Drives
Local drives may be faster than EBS; however, they are impermanent. One strategy is to have a hot server which uses local drives and a slave
which uses EBS for storage.
We have seen sequential read rates by MongoDB from ebs (unstriped) of 400Mbps on an extra large instance box. (YMMV)
Joyent
The prebuilt MongoDB Solaris 64 binaries work with Joyent accelerators.
Some newer gcc libraries are required to run -- see sample setup session below.
$ # assuming a 64 bit accelerator

$ /usr/bin/isainfo -kv
64-bit amd64 kernel modules
$ # get mongodb
$ # note this is 'latest' you may want a different version
$ curl -O http://downloads.mongodb.org/sunos5/mongodb-sunos5-x86_64-latest.tgz
$ gzip -d mongodb-sunos5-x86_64-latest.tgz
$ tar -xf mongodb-sunos5-x86_64-latest.tar
$ mv "mongodb-sunos5-x86_64-2009-10-26" mongo
$ cd mongo
$ # get extra libraries we need (else you will get a libstdc++.so.6 dependency issue)
$ curl -O http://downloads.mongodb.org.s3.amazonaws.com/sunos5/mongo-extra-64.tgz
$ gzip -d mongo-extra-64.tgz
$ tar -xf mongo-extra-64.tar
$ # just as an example - you will really probably want to put these somewhere better:
$ export LD_LIBRARY_PATH=mongo-extra-64
$ bin/mongod --help
Query Profiler
Http Console
db.serverStatus() from mongo shell
Trending/Monitoring Adaptors
Database Record/Replay
Checking Server Memory Usage

Database Profiler
Http Interface
mongostat
mongosniff
Admin UIs
Query Profiler
Use the Database Profiler to analyze slow queries.
Http Console
The mongod process includes a simple diagnostic screen at http://localhost:28017/. See the Http Interface docs for more information.
db.serverStatus() from mongo shell
> db.stats()
> db.serverStatus()
> db.foo.find().explain()
> help
> db.help()
> db.foo.help()
Server Status Fields
globalLock - totalTime & lockTime are total microsecs since startup that there has been a write lock
mem - current usage in megabytes
indexCounters - counters since startup, may rollover
opcounters - operation counters since startup
asserts - assert counters since startup
Trending/Monitoring Adaptors
munin, and another munin

ganglia
scout app slow queries
cacti
Chris Lea from (mt) Media Temple has made an easy to install Ubuntu package for the munin plugin.
Database Record/Replay
Recording database operations, and replaying them later, is sometimes a good way to reproduce certain problems in a controlled environment.
To enable logging:
db._adminCommand( { diagLogging : 1 } )
To disable:
db._adminCommand( { diagLogging : 0 } )
Values for diagLogging:
0 off. Also flushes any pending data to the file.

1 log writes
2 log reads
3 log both
Note: if you log reads, it will record the findOnes above and if you replay them, that will have an effect!
Output is written to diaglog.bin_ in the /data/db/ directory (unless --dbpath is specified).
To replay the logged events:

nc ''database_server_ip'' 27017 < ''somelog.bin'' | hexdump -c
Checking Server Memory Usage
Checking using DB Commands
The serverStatus() command provides memory usage information.
> db.serverStatus()
add: how to interpret. what would indicate a memory leak
Checking via Unix Commands
mongod uses memory-mapped files; thus the memory stats in top are not that useful. On a large database, virtual bytes/VSIZE will tend to be the
size of the entire database, and if the server doesn't have other processes running, resident bytes/RSIZE will be the total memory of the machine
(as this counts file system cache contents).
vmstat can be useful. Try running vmstat 2. On OS X, use vm_stat.
Checking in Windows
Help, any Windows admin experts our there? What should be here?
Historial Memory Leak Bugs (that are fixed)
(2 issues)
Key FixVersion Summary
SERVER-768 1.3.4 Memory leak and high memory usage from snapshots thread
SERVER-774 MessagingPorts are leaking
Database Profiler
Mongo includes a profiling tool to analyze the performance of database operations.
See also the currentOp command.
Enabling Profiling
To enable profiling, from the mongo shell invoke:
> db.setProfilingLevel(2);
{"was" : 0 , "ok" : 1}
> db.getProfilingLevel()
2
Profiling levels are:
0 - off
1 - log slow operations (>100ms)
2 - log all operations
Starting in 1.3.0, you can also enable on the command line, --profile=1
When profiling is enabling, there is continual writing to the system.profile table. This is very fast but does use a write lock which
has certain implications for concurrency.
Viewing
Profiling data is recorded in the database's system.profile collection. Query that collection to see the results.
> db.system.profile.find()
{"ts" : "Thu Jan 29 2009 15:19:32 GMT-0500 (EST)" , "info" : "query test.$cmd ntoreturn:1 reslen:66
nscanned:0 <br>query: { profile: 2 } nreturned:1 bytes:50" , "millis" : 0}
...
To see output without $cmd (command) operations, invoke:
db.system.profile.find( function() { return this.info.indexOf('$cmd')<0; } )
To view operations for a particular collection:
> db.system.profile.find( { info: /test.foo/ } )

{"ts" : "Thu Jan 29 2009 15:19:40 GMT-0500 (EST)" , "info" : "insert test.foo" , "millis" : 0}
{"ts" : "Thu Jan 29 2009 15:19:42 GMT-0500 (EST)" , "info" : "insert test.foo" , "millis" : 0}
{"ts" : "Thu Jan 29 2009 15:19:45 GMT-0500 (EST)" , "info" : "query test.foo ntoreturn:0 reslen:102
nscanned:2 <br>query: {} nreturned:2 bytes:86" , "millis" : 0}
{"ts" : "Thu Jan 29 2009 15:21:17 GMT-0500 (EST)" , "info" : "query test.foo ntoreturn:0 reslen:36
nscanned:2 <br>query: { $not: { x: 2 } } nreturned:0 bytes:20" , "millis" : 0}
{"ts" : "Thu Jan 29 2009 15:21:27 GMT-0500 (EST)" , "info" : "query test.foo ntoreturn:0 exception
bytes:53" , "millis" : 88}
To view operations slower than a certain number of milliseconds:
> db.system.profile.find( { millis : { $gt : 5 } } )

{"ts" : "Thu Jan 29 2009 15:21:27 GMT-0500 (EST)" , "info" : "query test.foo ntoreturn:0 exception
bytes:53" , "millis" : 88}
To see newest information first:
db.system.profile.find().sort({$natural:-1})
The mongo shell includes a helper to see the most recent 5 profiled events that took at least 1ms to execute. Type show profile at the command
prompt to use this feature.
Understanding the Output
The output reports the following values:
ts Timestamp of the profiled operation.

millis Time, in milliseconds, to perform the operation.
info Details on the operation.
query A database query operation. The query info field includes several additional terms:
ntoreturn Number of objects the client requested for return from a query. For example, <code>findOne()</code> sets
ntoreturn to 1. <code>limit()</code> sets the appropriate limit. Zero indicates no limit.
query Details of the query spec.
nscanned Number of objects scanned in executing the operation.
reslen Query result length in bytes.
nreturned Number of objects returned from query.
update A database update operation. <code>save()</code> calls generate either an update or insert operation.
fastmod Indicates a fast modify operation. See Updates. These operations are normally quite fast.
fastmodinsert - indicates a fast modify operation that performed an upsert.
upsert Indicates on upsert performed.
moved Indicates the update moved the object on disk (not updated in place). This is slower than an in place update, and
normally occurs when an object grows.
insert A database insert.
getmore For large queries, the database initially returns partial information. getmore indicates a call to retrieve further
information.
Optimizing Query Performance

If nscanned is much higher than nreturned, the database is scanning many objects to find the target objects. Consider creating an
index to improve this.
reslen A large number of bytes returned (hundreds of kilobytes or more) causes slow performance. Consider passing
<code>find()</code> a second parameter of the member names you require.
Note: There is a cost for each index you create. The index causes disk writes on each insert and some updates to the collection. If a rare query,
it may be better to let the query be "slow" and not create an index. When a query is common relative to the number of saves to the collection, you
will want to create the index.
Optimizing Update Performance
Examine the nscanned info field. If it is a very large value, the database is scanning a large number of objects to find the object to
update. Consider creating an index if updates are a high-frequency operation.
Use fast modify operations when possible (and usually with these, an index). See Updates.
Profiler Performance
When enabled, profiling affects performance, although not severely.
Profile data is stored in the database's system.profile collection, which is a Capped Collection. By default it is set to a very small size and
thus only includes recent operations.
Configuring "Slow"
Since 1.3.0 there are 2 ways to configure "slow"
--slowms on the command line when starting mongod (or file config)
db.setProfilingLevel( level , slowms )
db.setProfilingLevel( 1 , 10 );
will log all queries over 10ms to system.profile
See Also
Optimization
explain()
Http Interface
REST Interfaces
Sleepy Mongoose (Python)
MongoDB Rest (Node.js)
HTTP Console
HTTP Console Security
Simple REST Interface
JSON in the simple REST interface
See Also
REST Interfaces
Sleepy Mongoose (Python)
Sleepy Mongoose is a full featured REST interface for MongoDB which is available as a separate project.
MongoDB Rest (Node.js)
MongoDB Rest is an alpha REST interface to MongoDB, which uses the MongoDB Node Native driver.
HTTP Console
MongoDB provides a simple http interface listing information of interest to administrators. This interface may be accessed at the port with numeric
value 1000 more than the configured mongod port; the default port for the http interface is 28017. To access the http interface an administrator
may, for example, point a browser to http://localhost:28017 if mongod is running with the default port on the local machine.
Here is a description of the informational elements of the http interface:
element description
db version database version information
git hash database version developer tag
sys info mongod compilation environment
dblocked indicates whether the primary mongod mutex is held
uptime time since this mongod instance was started
assertions any software assertions that have been raised by this mongod instance
replInfo information about replication configuration
currentOp most recent client request
# databases number of databases that have been accessed by this mongod instance
curclient last database accessed by this mongod instance
Cursors describes outstanding client cursors
master whether this mongod instance has been designated a master
slave whether this mongod instance has been designated a slave
initialSyncCompleted whether this slave or repl pair node has completed an initial clone of the mongod instance it is replicating
DBTOP Displays the total time the mongod instance has devoted to each listed collection, as well as the percentage of available
time devoted to each listed collection recently and the number of reads, writes, and total calls made recently
dt Timing information about the primary mongod mutex
HTTP Console Security
If security is configured for a mongod instance, authentication is required for a client to access the http interface from another machine.
Simple REST Interface
The mongod process includes a simple read-only REST interface for convenience. For full REST capabilities we recommend using an external
tool such as Sleepy.Mongoose.
Note: in v1.3.4+ of MongoDB, this interface is disabled by default. Use --rest on the command line to enable.
To get the contents of a collection (note the trailing slash):
http://127.0.0.1:28017/databaseName/collectionName/
To add a limit:
http://127.0.0.1:28017/databaseName/collectionName/?limit=-10
To skip:
http://127.0.0.1:28017/databaseName/collectionName/?skip=5
To query for {a : 1}:
http://127.0.0.1:28017/databaseName/collectionName/?filter_a=1
Separate conditions with an &:
http://127.0.0.1:28017/databaseName/collectionName/?filter_a=1&limit=-10
JSON in the simple REST interface
The simple ReST interface uses strict JSON (as opposed to the shell, which uses Dates, regular expressions, etc.). To display non-JSON types,
the web interface wraps them in objects and uses the key for the type. For example:
# ObjectIds just become strings

"_id" : "4a8acf6e7fbadc242de5b4f3"
# dates
"date" : { "$date" : 1250609897802 }
# regular expressions
"match" : { "$regex" : "foo", "$options" : "ig" }
The code type has not been implemented yet and causes the DB to crash if you try to display it in the browser.
See Mongo Extended JSON for details.
See Also
Replica Set Admin UI

Diagnostic Tools
mongostat
Use the mongostat utility to quickly view statistics on a running mongod instance.
Starting in 1.3.3; this output is from 1.5.5

insert/s query/s update/s delete/s getmore/s command/s flushes/s mapped vsize res faults/s locked
% idx miss % q t|r|w conn time
0 0 0 0 0 1 0 0 90 4 0
0 0 0|0|0 1 07:59:30
0 0 0 0 0 1 0 0 90 4 0
0 0 0|0|0 1 07:59:31
0 0 0 0 0 1 0 0 90 4 0
0 0 0|0|0 1 07:59:32
0 0 0 0 0 1 0 0 90 4 0
0 0 0|0|0 1 07:59:33
0 0 0 0 0 1 0 0 90 4 0
0 0 0|0|0 1 07:59:34
0 0 0 0 0 1 0 0 90 4 0
0 0 0|0|0 1 07:59:35
Fields:
inserts/s : # of inserts per second
query/s - # of queries per second
update/s - # of updates per second
delete/s - # of deletes per second
getmore/s - # of get mores (cursor batch) per second
command/s - # of commands per second
flushes/s - # of fsync flushes per second
mapped - amount of data mmaped (total data size) megabytes
visze - virtual size of process in megabytes
res - resident size of process in megabytes
faults/s - # of pages faults/sec (linux only)
locked - percent of time in global write lock
idx miss - percent of btree page misses (sampled)
q t|r|w - lock queue lengths (total|read|write)
conn - number of open connections
mongosniff
Unix releases of MongoDB include a utility called mongosniff. This utility is to MongoDB what tcpdump is to TCP/IP; that is, fairly low level and for
complex situations. The tool is quite useful for authors of driver tools.
$ ./mongosniff --help
Usage: mongosniff [--help] [--forward host:port] [--source (NET <interface> | FILE <filename>)]
[<port0> <port1> ...]
--forward Forward all parsed request messages to mongod instance at
specified host:port
--source Source of traffic to sniff, either a network interface or a
file containing perviously captured packets, in pcap format.
If no source is specified, mongosniff will attempt to sniff
from one of the machine's network interfaces.
<port0>... These parameters are used to filter sniffing. By default,
only port 27017 is sniffed.
--help Print this help message.
Building
mongosniff is including in the binaries for Unix distributions. The MongoDB SConstruct only builds mongosniff if libpcap is installed.
sudo yum install libpcap-devel

scons mongosniff
Example
To monitor localhost:27017, run ifconfig to find loopback's name (usually something like lo or lo0). Then run:
mongosniff --source NET lo
If you get the error message "error opening device: socket: Operation not permitted" or "error finding device: no suitable device found", try running
it as root.
Backups
Fsync, Write Lock and Backup
Shutdown and Backup
Exports
Slave Backup
Community Stuff
Several strategies exist for backing up MongoDB databases. A word of warning: it's not safe to back up the mongod data files (by default in
/data/db/) while the database is running and writes are occurring; such a backup may turn out to be corrupt. See the fsync option below for a way
around that.
Fsync, Write Lock and Backup
MongoDB v1.3.1 and higher supports an fsync and lock command with which we can flush writes, lock the database to prevent writing, and then
backup the datafiles.
While in this locked mode, all writes will block. If this is a problem consider one of the other methods below.
For example, you could use LVM2 to create a snapshot after the fsync+lock, and then use that snapshot to do an offsite backup in the
background. This means that the server will only be locked while the snapshot is taken. Don't forget to unlock after the backup/snapshot is taken.
Shutdown and Backup
A simple approach is just to stop the database, back up the data files, and resume. This is safe but of course requires downtime.
Exports
The mongodump utility may be used to dump an entire database, even when the database is running and active. The dump can then be restored
later if needed.
Slave Backup
Another good technique for backups is replication to a slave database. The slave polls master continuously and thus always has a
nearly-up-to-date copy of master.
We then have several options for backing up the slave:
1. Fsync, write lock, and backup the slave.

2. Shut it down, backup, and restart.
3. Export from the slave.
For methods 1 and 2, after the backup the slave will resume replication, applying any changes made to master in the meantime.
Using a slave is advantageous because we then always have backup database machine ready in case master fails (failover). But a slave also
gives us the chance to back up the full data set without affecting the performance of the master database.
Community Stuff
http://github.com/micahwedemeyer/automongobackup
How to do Snapshotted Queries in the Mongo Database
This document refers to query snapshots. For backup snapshots of the database's datafiles, see the fsync lock page.
MongoDB does not support full point-in-time snapshotting. However, some functionality is available which is detailed below.
Cursors
A MongoDB query returns data as well as a cursor ID for additional lookups, should more data exist. Drivers lazily perform a "getMore" operation
as needed on the cursor to get more data. Cursors may have latent getMore accesses that occurs after an intervening write operation on the
database collection (i.e., an insert, update, or delete).
Conceptually, a cursor has a current position. If you delete the item at the current position, the cursor automatically skips its current position
forward to the next item.
Mongo DB cursors do not provide a snapshot: if other write operations occur during the life of your cursor, it is unspecified if your application will
see the results of those operations. In fact, it is even possible (although unlikely) to see the same object returned twice if the object were updated
and grew in size (and thus moved in the datafile). To assure no update duplications, use snapshot() mode (see below).
Snapshot Mode
snapshot() mode assures that objects which update during the lifetime of a query are returned once and only once. This is most important when
doing a find-and-update loop that changes the size of documents that are returned ($inc does not change size).
> // mongo shell example

> var cursor = db.myCollection.find({country:'uk'}).snapshot();
Even with snapshot mode, items inserted or deleted during the query may or may not be returned; that is, this mode is not a true point-in-time
snapshot.
Because snapshot mode traverses the _id index, it may not be used with sorting or explicit hints. It also cannot use any other index for the query.
You can get the same effect as snapshot by using any unique index on a field(s) that will not be modified (probably best to use explicit hint() too).
If you want to use a non-unique index (such as creation time), you can make it unique by appending _id to the index at creation time.
Import Export Tools

mongoimport
mongoexport
mongodump
Example: Dumping Everything
Example: Dumping a Single Collection
Example: Dumping a Single Collection to Stdout
mongorestore
bsondump
If you just want to do Clone Database from one server to another you don't need these tools.
These tool just work with the raw data (the documents in the collection); they do not save, or load, the metadata like the defined
indexes or (capped) collection properties. You will need to (re)create those yourself in a separate step, before loading that data.
Vote here to change this.
mongoimport
This utility takes a single file that contains 1 JSON/CSV/TSV string per line and inserts it. You have to specify a database and a collection.
options:
--help produce help message
-v [ --verbose ] be more verbose (include multiple times for more
verbosity e.g. -vvvvv)
-h [ --host ] arg mongo host to connect to ( "left,right" for pairs)
-d [ --db ] arg database to use
-c [ --collection ] arg collection to use (some commands)
-u [ --username ] arg username
-p [ --password ] arg password
--dbpath arg directly access mongod data files in the given path,
instead of connecting to a mongod instance - needs to
lock the data directory, so cannot be used if a
mongod is currently accessing the same path
--directoryperdb if dbpath specified, each db is in a separate
directory
-f [ --fields ] arg comma seperated list of field names e.g. -f name,age
--fieldFile arg file with fields names - 1 per line
--ignoreBlanks if given, empty fields in csv and tsv will be ignored
--type arg type of file to import. default: json (json,csv,tsv)
--file arg file to import from; if not specified stdin is used
--drop drop collection first
--headerline CSV,TSV only - use first line as headers
mongoexport
This utility takes a collection and exports to either JSON or CSV. You can specify a filter for the query, or a list of fields to output.
Neither JSON nor TSV/CSV can represent all data types. Please be careful not to lose or change data (types) when using this.
For full fidelity please use mongodump.
If you want to output CSV, you have to specify the fields in the order you want them.
Example
options:
directory
-q [ --query ] arg query filter, as a JSON string
-f [ --fields ] arg comma seperated list of field names e.g. -f name,age
--csv export to csv instead of json
-o [ --out ] arg output file; if not specified, stdout is used
mongodump
This takes a database and outputs it in a binary representation. This is mostly used for doing hot backups of a database.
If you're using sharding and try to migrate data this way, this will dump shard configuration information and overwrite
configurations upon restore.
options:
instead of connecting to a mongod instance - needs
to lock the data directory, so cannot be used if a
directory
-o [ --out ] arg (=dump) output directory
Example: Dumping Everything
To dump all of the collections in all of the databases, run mongodump with just the --host:
$ ./mongodump --host prod.example.com

connected to: prod.example.com
all dbs
DATABASE: log to dump/log
log.errors to dump/log/errors.bson
713 objects
log.analytics to dump/log/analytics.bson
234810 objects
DATABASE: blog to dump/blog
blog.posts to dump/log/blog.posts.bson
59 objects
DATABASE: admin to dump/admin
You'll then have a folder called "dump" in your current directory.
If you're running mongod locally on the default port, you can just do:
$ ./mongodump
Example: Dumping a Single Collection
If we just want to dump a single collection, we can specify it and get a single .bson file.
$ ./mongodump --db blog --collection posts

connected to: 127.0.0.1
DATABASE: blog to dump/blog
blog.posts to dump/blog/posts.bson
59 objects
Example: Dumping a Single Collection to Stdout
In version 1.7.0+, you can use stdout instead of a file by specifying --out stdout:
$ ./mongodump --db blog --collection posts --out - > blogposts.bson

mongodump creates a file for each database collection, so we can only dump one collection at a time to stdout.
mongorestore
This takes the output from mongodump and restores it.
usage: ./mongorestore [options] [directory or filename to restore from]

options:
directory
--drop drop each collection before import
--objcheck validate object before inserting
bsondump
Added in 1.6
This takes a bson file and converts it to json/debug output.
usage: ./bsondump [options] [filename]

options:
--type arg (=json) type of output: json,debug

Single Server Durability
Repair Command
Validate Command
--syncdelay Command Line Option
See Also
Single Server Durability
The v1.8 release of MongoDB will have single server durability. You can follow the Jira here : http://jira.mongodb.org/browse/SERVER-980.
We recommend using replication to keep copies of data for now – and likely forever – as a single server could fail catastrophically regardless.
Repair Command
There is a bug with repair and replica sets in MongoDB v1.6.0. Please see this Jira for information:
http://jira.mongodb.org/browse/SERVER-1614. Do NOT run repair without reading this first. This bug applies to 1.6.0 only. Will
be fixed in 1.6.1. In the meantime there are workarounds.
After a machine crash or or kill -9 termination, consider running the repairDatabase command. This command will check all data for
corruption, remove any corruption found, and compact data files a bit.
In the event of a hard crash, we recommend running a repair – analogous to running fsck. If a slave crashes, another option is just to restart the
slave from scratch.
From the command line:
mongod --repair
From the shell (you have to do for all dbs including local if you go this route):
> db.repairDatabase();
During a repair operation, mongod must store temporary files to disk. By default, mongod creates temporary directories under the dbpath for this
purpose. Alternatively, the --repairpath command line option can be used to specify a base directory for temporary repair files.
Note that repair is a slow operation which inspects the entire database.
If the databases exited uncleanly and you attempt to restart the database, mongod will print:
**************
old lock file: /data/db/mongod.lock. probably means unclean shutdown
recommend removing file and running --repair
see: http://dochub.mongodb.org/core/repair for more information
*************
Then it will exit. After running with --repair, mongod will start up normally.
Validate Command
Alternatively one could restart and run the validate command on select tables. The validate command checks if the contents of a collection are
valid.
For example, here we validate the users collection:
> db.users.validate();
{
"ns" : "test.users",
"result" : " validate
details: 0x1243dbbdc ofs:740bdc
firstExtent:0:178b00 ns:test.users
lastExtent:0:178b00 ns:test.users
# extents:1
padding:1
first extent:
loc:0:178b00 xnext:null xprev:null
nsdiag:test.users
size:8192 firstRecord:0:178bb0 lastRecord:0:178bb0
deletedList: 0000000010000000000
nIndexes:2
test.users.$_id_ keys:1
test.users.$username_1 keys:1 ",
"ok" : 1,
"valid" : true,
"lastExtentSize" : 8192
}
--syncdelay Command Line Option
Since 1.1.4, the --syncdelay option controls how often changes are flushed to disk (the default is 60 seconds). If replication is not being used, it
may be desirable to reduce this default.
See Also
What About Durability? (MongoDB Blog)

fsync Command
MongoDB (Single-Server) Data Durability Guide

Running Without Security (Trusted Environment)
This is the default option and is recommended.
One valid way to run the Mongo database is in a trusted environment, with no security and authentication (much like how one would use, say,
memcached). Of course, in such a configuration, one must be sure only trusted machines can access database TCP ports.
The current versions of sharding and replica sets requires trusted (nonsecure) mode.
Mongo Security
The current version of Mongo supports only very basic security. One authenticates a username and password in the context of a particular
database. Once authenticated, a normal user has full read and write access to the database in question while a read only user only has read
access.
The admin database is special. In addition to several commands that are administrative being possible only on admin, authentication on admin
gives one read and write access to all databases on the server. Effectively, admin access means root access to the server process.
Run the database (mongod process) with the --auth option to enable security. You must either have added a user to the admin db before
starting the server with --auth, or add the first user from the localhost interface.
Configuring Authentication and Security

Authentication is stored in each database's system.users collection. For example, on a database projectx, projectx.system.users will
contain user information.
We should first configure an administrator user for the entire db server process. This user is stored under the special admin database.
If no users are configured in admin.system.users, one may access the database from the localhost interface without authenticating. Thus,
from the server running the database (and thus on localhost), run the database shell and configure an administrative user:
$ ./mongo
> use admin
> db.addUser("theadmin", "anadminpassword")
We now have a user created for database admin. Note that if we have not previously authenticated, we now must if we wish to perform further
operations, as there is a user in admin.system.users.
> db.auth("theadmin", "anadminpassword")
We can view existing users for the database with the command:
> db.system.users.find()
Now, let's configure a "regular" user for another database.
> use projectx

> db.addUser("joe", "passwordForJoe")
Finally, let's add a readonly user. (only supported in 1.3.2+)

> use projectx
> db.addUser("guest", "passwordForGuest", true)
Changing Passwords
The shell addUser command may also be used to update a password: if the user already exists, the password simply updates.
Many Mongo drivers provide a helper function equivalent to the db shell's addUser method.
Deleting Users
To delete a user:
db.system.users.remove( { user: username } )
Admin UIs
Several administrative user interfaces, or GUIs, are available for MongoDB. Tim Gourley's blog has a good summary of the tools.
Fang of Mongo
Futon4Mongo
Mongo3
MongoHub
MongoVUE
Mongui
Myngo
Opricot
PHPMoAdmin
RockMongo
The built-in replica set admin UI
Details
Fang of Mongo
A web-based user interface for MongoDB build with django and jquery.
It will allow you to explore content of mongodb with simple but (hopefully) pleasant user interface.
Features:
field name autocompletion in query builder
data loading indicator
human friendly collection stats
disabling collection windows when there is no collection selected
twitter stream plugin
many more minor usability fixes
works well on recent chrome and firefox
See it in action at: http://blueone.pl:8001/fangofmongo/

Get it from github: http://github.com/Fiedzia/Fang-of-Mongo
Or track progress on twitter: @fangofmongo
MongoHub
MongoHub is a native OS X GUI.
MongoVUE
MongoVUE is a .NET GUI for MongoDB.

Opricot
Opricot is a hybrid GUI/CLI/Scripting web frontend implemented in PHP to manage your MongoDB servers and databases. Use as a
point-and-click adventure for basic tasks, utilize scripting for automated processing or repetitive things.
Opricot combines the following components to create a fully featured administration tool:
An interactive console that allows you to either work with the database through the UI, or by using custom Javascript.
A set of simple commands that wrap the Javascript driver, and provide an easy way to complete the most common tasks.
Javascript driver for Mongo that works on the browser and talks with the AJAX interface.
Simple server-side AJAX interface for communicating with the MongoDB server (currently available for PHP).
PHPMoAdmin
PHPMoAdmin is a MongoDB administration tool for PHP built on a stripped-down version of the Vork high-performance framework.
Nothing to configure - place the moadmin.php file anywhere on your web site and it just works!
Fast AJAX-based XHTML 1.1 interface operates consistently in every browser!
Self-contained in a single 95kb file!
Works on any version of PHP5 with the MongoDB NoSQL database installed & running.
Super flexible - search for exact-text, text with * wildcards, regex or JSON (with Mongo-operators enabled)
Option to enable password-protection for one or more users; to activate protection, just add the username-password(s) to the array at the
top of the file.
E_STRICT PHP code is formatted to the Zend Framework coding standards + fully-documented in the phpDocumentor DocBlock
standard.
Textareas can be resized by dragging/stretching the lower-right corner.
Free & open-source! Release under the GPLv3 FOSS license!
Option to query MongoDB using JSON or PHP-array syntax
Multiple design themes to choose from
Instructional error messages - phpMoAdmin can be used as a PHP-Mongo connection debugging tool
PHPMoAdmin can help you discover the source of connection issues between PHP and Mongo. Download phpMoAdmin, place the moadmin.php
file in your web site document directory and navigate to it in a browser. One of two things will happen:
You will see an error message explaining why PHP and Mongo cannot connect and what you need to do to fix it
You will see a bunch of Mongo-related options, including a selection of databases (by default, the "admin" and "local" databases always
exist) - if this is the case your installation was successful and your problem is within the PHP code that you are using to access
MongoDB, troubleshoot that from the Mongo docs on php.net
RockMongo
RockMongo is a MongoDB management tool, written in PHP 5.
Main features:
easy to install, and open source

multiple hosts, and multiple administrators for one host
password protection
query dbs
advanced collection query tool
read, insert, update, duplicate and remove single row
query, create and drop indexes
clear collection
remove and change (only work in higher php_mongo version) criteria matched rows
view collection statistics

Starting Mongo
Default Data Directory, Default Port
Alternate Data Directory, Default Port
Alternate Port
Running as a Daemon
Stopping Mongo
Control-C
Sending shutdownServer() message from the mongo shell
Sending a Unix INT or TERM signal
Memory Usage
MongoDB is run as a standard program from the command line. Please see Command Line Parameters for more information on those options.
The following examples assume that you are in the directory where the Mongo executable is, and the Mongo executable is called mongod.
Starting Mongo
Default Data Directory, Default Port
To start Mongo in default mode, where data will be stored in the /data/db directory (or c:\data\db on Windows), and listening on port 27017,
just type
$ ./mongod
Alternate Data Directory, Default Port
To specify a directory for Mongo to store files, use the --dbpath option:
$ ./mongod --dbpath /var/lib/mongodb/
Note that you must create the directory and set its permissions appropriately ahead of time -- Mongo will not create the directory if it doesn't exist.
Alternate Port
You can specify a different port for Mongo to listen on for connections from clients using the --port option
$ ./mongod --port 12345
This is useful if you want to run more than one instance of Mongo on a machine (e.g., for running a master-slave pair).
Running as a Daemon
Note: these options are only available in MongoDB version 1.1 and later.
This will fork the Mongo server and redirect its output to a logfile. As with --dbpath, you must create the log path yourself, Mongo will not create
parent directories for you.
$ ./mongod --fork --logpath /var/log/mongodb.log --logappend
Stopping Mongo
Control-C
If you have Mongo running in the foreground in a terminal, you can simply "Ctrl-C" the process. This will cause Mongo to do a clean exit, flushing
and closing it's data files. Note that it will wait until all ongoing operations are complete.
Sending shutdownServer() message from the mongo shell
The shell can request that the server terminate.
$ ./mongo
> db.shutdownServer()
This command only works from localhost, or, if one is authenticated.
From a driver (where the helper function may not exist), one can run the command
{ "shutdown" : 1 }
Sending a Unix INT or TERM signal
You can cleanly stop mongod using a SIGINT or SIGTERM signal on Unix-like systems. Either ^C, "kill -2 PID," or kill -15 PID will work.
Sending a KILL signal kill -9 will probably cause damage as mongod will not be able to cleanly exit. (In such a scenario, run
the repairDatabase command.)
After a unclean shutdown, MongoDB will say it was not shutdown cleanly, and ask you to do a repair. This is absolutely not the
same as corruption, this is MongoDB saying it can't 100% verify what's going on, and to be paranoid, run a repair.
Memory Usage
Mongo uses memory mapped files to access data, which results in large numbers being displayed in tools like top for the mongod process. This is
not a concern, and is normal when using memory-mapped files. Basically, the size of mapped data is shown in the virtual size parameter, and
resident bytes shows how much data is being cached in RAM.
You can get a feel for the "inherent" memory footprint of Mongo by starting it fresh, with no connections, with an empty /data/db directory and
looking at the resident bytes.
Logging
MongoDB outputs some important information to stdout while its running. There are a number of things you can do to control this
--quiet - less verbose output

-v - more verbose output. use more v's (such as -vvvvvv) for higher levels of verbosity
--logpath <file> output to file instead of stdout
if you use logpath, you can rotate the logs by either running the logRotate command (1.3.4+) or sending SIGUSR1
> db.runCommand("logRotate");
Command Line Parameters

MongoDB can be configured via command line parameters in addition to File Based Configuration. You can see the currently supported set of
command line options by running the database with -h [ --help ] as a single parameter:
$ ./mongod --help
Information on usage of these parameters can be found in Starting and Stopping Mongo.
The following list of options is not complete; for the complete list see the usage information as described above.
Basic Options
-h | --help Shows all options
-f | --config <file> Specify a configuration file to use
--port <portno> Specifies the port number on which Mongo will listen for client connections. Default is 27017
--dbpath <path> Specifies the directory for datafiles. Default is /data/db or c:\data\db
--fork Fork the server process

--bind_ip <ip> Specifies a single IP that the database server will listen for
--directoryperdb Specify use of an alternative directory structure, in which files for each database are kept in a unique
directory. (since 1.3.2)
--quiet Reduces amount of log output
--nohttpinterface Disable the HTTP interface (localhost:27018)
--rest Allow extended operations at the Http Interface
--logpath <file> File to write logs to (instead of stdout). You can rotate the logs by sending SIGUSR1 to the server.
--logappend Append to existing log file, instead of overwritting
--repairpath <path> Root path for temporary files created during database repair. Default is dbpath value.
--cpu Enables periodic logging of CPU utilization and I/O wait
--noauth Turns off security. This is currently the default
--auth Turn on security
-v[v[v[v[v]]]] | Verbose logging output (-vvvvv is most verbose, -v == --verbose)

--verbose
--objcheck Inspect all client data for validity on receipt (useful for developing drivers)
--quota Enable db quota management
--diaglog <n> Set oplogging level where n is 0=off (default) 1=W 2=R 3=both 7=W+some reads
--nocursors Diagnostic/debugging option
--nohints Ignore query hints
--noscripting Turns off server-side scripting. This will result in greatly limited functionality
--notablescan Turns off table scans. Any query that would do a table scan fails
--noprealloc Disable data file preallocation
--smallfiles Use a smaller default file size
--nssize <MB> Specifies .ns file size for new databases
--sysinfo Print system info as detected by Mongo and exit
--upgrade Upgrade database files to new format if necessary

(required when upgrading from <= 1.0 to 1.1+)
Master/Slave Replication Options
--master Designate this server as a master in a master-slave setup
--slave Designate this server as a slave in a master-slave setup
--source <server:port> Specify the source (master) for a slave instance
--only <db> Slave only: specify a single database to replicate
--arbiter <server:port> Address of arbiter server
--autoresync Automatically resync if slave data is stale
--oplogSize <MB> Custom size for replication operation log
Replica Set Options
--replSet <setname> Use replica sets with the specified logical set name
--oplogSize <MB> Custom size for replication operation log

File Based Configuration
In addition to accepting Command Line Parameters, MongoDB can also be configured using a configuration file. A configuration file to use can be
specified using the -f or --config command line options. The following example configuration file demonstrates the syntax to use:
# This is an example config file for MongoDB.
dbpath = /var/lib/mongodb
bind_ip = 127.0.0.1
noauth = true # use 'true' for options that don't take an argument
verbose = true # to disable, comment out.
Notes
Lines starting with octothorpes (#) are comments

Options are case sensitive
The syntax is assignment of a value to an option name
All command line options are accepted
GridFS Tools
File Tools
mongofiles is a tool for manipulating GridFS from the command line.
Example:
$ ./mongofiles list
$ ./mongofiles put libmongoclient.a

done!
$ ./mongofiles list
libmongoclient.a 12000964
$ cd /tmp/
$ ~/work/mon/mongofiles get libmongoclient.a
$ ~/work/mongo/mongofiles get libmongoclient.a

done write to: libmongoclient.a
$ md5 libmongoclient.a
MD5 (libmongoclient.a) = 23a52d361cfa7bad98099c5bad50dc41
$ md5 ~/work/mongo/libmongoclient.a
MD5 (/Users/erh/work/mongo/libmongoclient.a) = 23a52d361cfa7bad98099c5bad50dc41

This page lists common DBA-class operations that one might perform from the MongoDB shell.
Note one may also create .js scripts to run in the shell for administrative purposes.
help show help
show dbs show database names
show collections show collections in current database
show users show users in current database
show profile show most recent system.profile entries with time >= 1ms
use <db name> set curent database to <db name>
db.addUser (username, password)

db.removeUser(username)
db.cloneDatabase(fromhost)
db.copyDatabase(fromdb, todb, fromhost)
db.createCollection(name, { size : ..., capped : ..., max : ... } )
db.getName()
db.dropDatabase()
db.printCollectionStats()
db.currentOp() displays the current operation in the db

db.killOp() kills the current operation in the db
db.getProfilingLevel()
db.setProfilingLevel(level) 0=off 1=slow 2=all
db.getReplicationInfo()
db.printReplicationInfo()
db.printSlaveReplicationInfo()
db.repairDatabase()
db.version() current version of the server
db.shutdownServer()
Commands for manipulating and inspecting a collection:

db.foo.drop() drop the collection
db.foo.dropIndex(name)
db.foo.dropIndexes()
db.foo.getIndexes()
db.foo.ensureIndex(keypattern,options) - options object has these possible
fields: name, unique, dropDups
db.foo.find( [query] , [fields]) - first parameter is an optional

query filter. second parameter
is optional
set of fields to return.
e.g. db.foo.find(
{ x : 77 } ,
{ name : 1 , x : 1 } )
db.foo.find(...).count()
db.foo.find(...).limit(n)
db.foo.find(...).skip(n)
db.foo.find(...).sort(...)
db.foo.findOne([query])
db.foo.getDB() get DB object associated with collection
db.foo.count()
db.foo.group( { key : ..., initial: ..., reduce : ...[, cond: ...] } )
db.foo.renameCollection( newName ) renames the collection
db.foo.stats()
db.foo.dataSize()
db.foo.storageSize() - includes free space allocated to this collection
db.foo.totalIndexSize() - size in bytes of all the indexes
db.foo.totalSize() - storage allocated for all data and indexes
db.foo.validate() (slow)
db.foo.insert(obj)
db.foo.update(query, object[, upsert_bool])
db.foo.save(obj)
db.foo.remove(query) - remove objects matching query
remove({}) will remove all

MongoDB has two primary components to the database server. The first is the mongod process which is the core database server. In many
cases, mongod may be used as a self-contained system similar to how one would use mysqld on a server. Separate mongod instances on
different machines (and data centers) can replicate from one to another.
Another MongoDB process, mongos, facilitates auto-sharding. mongos can be thought of as a "database router" to make a cluster of mongod
processes appear as a single database. See the sharding documentation for more information.
Database Caching
With relational databases, object caching is usually a separate facility (such as memcached), which makes sense as even a RAM page cache hit
is a fairly expensive operation with a relational database (joins may be required, and the data must be transformed into an object representation).
Further, memcached type solutions are more scaleable than a relational database.
Mongo eliminates the need (in some cases) for a separate object caching layer. Queries that result in file system RAM cache hits are very fast as
the object's representation in the database is very close to its representation in application memory. Also, the MongoDB can scale to any level
and provides an object cache and database integrated together, which is very helpful as there is no risk of retrieving stale data from the cache. In
addition, the complex queries a full DBMS provides are also possible.
Troubleshooting
Excessive Disk Space
Too Many Open Files
mongod process "disappeared"
See Also
mongod process "disappeared"

Scenario here is the log ending suddenly with no error or shutdown messages logged.
On Unix, check /var/log/messages:
$ grep mongod /var/log/messages

$ grep score /var/log/messages
See Also
Diagnostic Tools
Excessive Disk Space

You may notice that for a given set of data the MongoDB datafiles in /data/db are larger than the data set inserted into the database. There are
several reasons for this.
Preallocation
Each datafile is preallocated to a given size. (This is done to prevent file system fragmentation, among other reasons.) The first file for a
database is <dbname>.0, then <dbname>.1, etc. <dbname>.0 will be 64MB, <dbname>.1 128MB, etc., up to 2GB. Once the files reach 2GB in
size, each successive file is also 2GB.
Thus if the last datafile present is say, 1GB, that file might be 90% empty if it was recently reached.
Additionally, on Unix, mongod will preallocate an additional datafile in the background and do background initialization of this file. These files are
prefilled with zero bytes. This inititialization can take up to a minute (less on a fast disk subsystem) for larger datafiles; without prefilling in the
background this could result in significant delays when a new file must be prepopulated.
You can disable preallocation with the --noprealloc option to the server. This flag is nice for tests with small datasets where you drop the db after
each test. It shouldn't be used on production servers.
For large databases (hundreds of GB or more) this is of no signficant consequence as the unallocated space is small.
Deleted Space
MongoDB maintains deleted lists of space within the datafiles when objects or collections are deleted. This space is reused but never freed to the
operating system.
To compact this space, run db.repairDatabase() from the mongo shell (note this operation will block and is slow).
When testing and investigating the size of datafiles, if your data is just test data, use db.dropDatabase() to clear all datafiles and start fresh.
Checking Size of a Collection
Use the validate command to check the size of a collection -- that is from the shell run:
> db.<collectionname>.validate();
> // these are faster:

> db.<collectionname>.dataSize(); // just data size for collection
> db.<collectionname>.storageSize(); // allocation size including unused space
> db.<collectionname>.totalSize(); // data + index
> db.<collectionname>.totalIndexSize(); // index data size
This command returns info on the collection data but note there is also data allocated for associated indexes. These can be checked with validate
too, if one looks up the index's namespace name in the system.namespaces collection. For example:
> db.system.namespaces.find()
{"name" : "test.foo"}
{"name" : "test.system.indexes"}
{"name" : "test.foo.$_id_"}
> > db.foo.$_id_.validate()
{"ns" : "test.foo.$_id_" , "result" : "
validate
details: 0xb3590b68 ofs:83fb68
firstExtent:0:8100 ns:test.foo.$_id_
lastExtent:0:8100 ns:test.foo.$_id_
# extents:1
padding:1
first extent:
loc:0:8100 xnext:null xprev:null
ns:test.foo.$_id_
size:131072 firstRecord:0:81b0 lastRecord:0:81b0
deletedList: 0000000000001000000
nIndexes:0
" , "ok" : 1 , "valid" : true , "lastExtentSize" : 131072}
Too Many Open Files

If you receive the error "too many open files" or "too many open connections" in the mongod log, there are a couple of possible reasons for this.
First, to check what file descriptors are in use, run lsof (some variations shown below):
lsof | grep mongod

lsof | grep mongod | grep TCP
lsof | grep mongod | grep data | wc
If most lines include "TCP", there are many open connections from client sockets. If most lines include the name of your data directory, the open
files are mostly datafiles.
ulimit
If the numbers from lsof look reasonable, check your ulimit settings. The default for file handles (often 1024) might be too low for production
usage. Run ulimit -a (or limit -a depending on shell) to check.
Use ulimit -n X to change the max number of file handles to X. If your OS is configured to not allow modifications to that setting you might need to
reconfigure first. On ubuntu you'll need to edit /etc/security/limits.conf and add a line something like the following (where user is the username and
X is the desired limit):
user hard nofile X
Upstart uses a different mechanism for setting file descriptor limits - add something like this to your job file:
limit nofile X
High TCP Connection Count
If lsof shows a large number of open TCP sockets, it could be that one or more clients is opening too many connections to the database. Check
that your client apps are using connection pooling.
Contributors
MongoDB kernel code development rules
Project Ideas
UI
Source Code
Building
Database Internals
Contributing to the Documentation
10gen Contributor Agreement

Git Commit Rules
Kernel class rules
Kernel code style
Kernel concurrency rules
Kernel exception architecture
Kernel string manipulation
Writing Tests
Git Commit Rules

commit messages should have the case in the message SERVER-XXX
commit messages should be descriptive enough that a glance can tell the basics
commits should only include 1 thought.
Kernel class rules

new classes
By default, use explicit constructors

Inherit from boost::noncopyable unless you have implemented copy constructor and assignment.
inheritance
No multiple inheritance.
Be very careful about adding the FIRST virtual function to a class as you then have a vtable entry for every object.
If anything is virtual, make your destructor virtual.
destructor guards? what are those for?
Kernel code style

case
strings
brackets
class members
templates
namespaces
start of file
assertions
case
Use camelCase for
most varNames
--commandLineOptions
{ commandNames : 1 }
strings
See
utils/mongoutils/str.h
bson/stringdata.h
Use startsWith(), endsWith(), not strstr().
Use << 'c' not << "c".
Use str[0] == '\0' not strlen(str) == 0.
brackets
if ( 0 ) {
}
class members
class Foo {
int _bar;
};
templates
set<int> s;
namespaces
namespace foo {
int foo;
namespace bar {
int bar;
}
}
start of file
// @file <filename>
license
assertions
There are several different types of assertions used in the MongoDB code. In brief:
assert should be used for internal assertions.

massert is an internal assertion with a message.
uassert is used for a user error
wassert warn and continue
Both massert and uassert take error codes, so that all errors have codes associated with them. These error codes are assigned randomly, so
there aren't segments that have meaning. scons checks for duplicates, but if you want the next available code you can run:
python buildscripts/errorcodes.py
Kernel concurrency rules

All concurrency code must be placed under utils/concurrency. You will find several helper libraries there.
Several rules are listed below. Don't break them. If you think there is a real need let's have the group weigh in and get a concensus on the
exception.
Do not use/add recursive locks.

Do not use rwlocks.
Always acquire locks in a consistent order. In fact, the MutexDebugger can assist with verification of this.
Kernel exception architecture

For string manipulation, use the util/mongoutils/str.h library.
mongoutils
MongoUtils has its own namespace. Its code has these basic properties:
1. are not database specific, rather, true utilities

2. are cross platform
3. may require boost headers, but not libs (header-only works with mongoutils)
4. are clean and easy to use in any c++ project without pulling in lots of other stuff
5. apache license
str.h
mongoutils/str.h provides string helper fucntions for each manipulation. Add new functions here rather than lines and lines of code to your
app that are not generic.
Typically these fucntions return a string and take two as paramters : string f(string,string). Thus we wrap them all in a namespace called str.
Writing Tests
We have three general flavors of tests currently.
Lightweight startup test.
You can inherit from class UnitTest and make a test that runs at program startup. These tests run EVERY TIME the program starts. Thus, they
should be minimal: the test should ideally take 1ms or less to run. Why run the tests in the general program? This gives some validation at
program run time that the build is reasonable. For example, we test that pcre supports UTF8 regex in one of these tests at startup. If someone
had built the server with other settings, this would be flagged upon execution, even if the test suite has not been invoked.
dbtests
jstests
See Also
Smoke Tests
Project Ideas
If you're interested in getting involved in the MongoDB community (or the open source community in general) a great way to do so is by starting or
contributing to a MongoDB related project. Here we've listed some project ideas for you to get started on. For some of these ideas projects are
already underway, and for others nothing (that we know of) has been started yet.
A GUI
One feature that is often requested for MongoDB is a GUI, much like CouchDB's futon or phpMyAdmin. There are a couple of projects working on
this sort of thing that are worth checking out:
http://github.com/sbellity/futon4mongo
http://www.mongodb.org/display/DOCS/Http+Interface
http://www.mongohq.com
We've also started to spec out the features that a tool like this should provide.
Try Mongo!
It would be neat to have a web version of the MongoDB Shell that allowed users to interact with a real MongoDB instance (for doing the tutorial,
etc). A project that does something similar (using a basic MongoDB emulator) is here:
http://github.com/banker/mongulator
Real-time Full Text Search Integration
It would be interesting to try to nicely integrate a search backend like Xapian, Lucene or Sphinx with MongoDB. One idea would be to use
MongoDB's oplog (which is used for master-slave replication) to keep the search engine up to date.
GridFS FUSE
There is a project working towards creating a FUSE filesystem on top of GridFS - something like this would create a bunch of interesting potential
uses for MongoDB and GridFS:
http://github.com/mikejs/gridfs-fuse
GridFS Web Server Modules
There are a couple of modules for different web servers designed to allow serving content directly from GridFS:
Nginx: http://github.com/mdirolf/nginx-gridfs
Lighttpd: http://bitbucket.org/bwmcadams/lighttpd-gridfs
Framework Adaptors
Working towards adding MongoDB support to major web frameworks is a great project, and work has been started on this for a variety of different
frameworks (please use google to find out if work has already been started for your favorite framework).
Logging and Session Adaptors
MongoDB works great for storing logs and session information. There are a couple of projects working on supporting this use case directly.
Logging:
Zend: http://raphaelstolt.blogspot.com/2009/09/logging-to-mongodb-and-accessing-log.html
Python: http://github.com/andreisavu/mongodb-log
Rails: http://github.com/peburrows/mongo_db_logger
Sessions:
web.py: http://github.com/whilefalse/webpy-mongodb-sessions
Beaker: http://pypi.python.org/pypi/mongodb_beaker
Package Managers
Add support for installing MongoDB with your favorite package manager and let us know!
Locale-aware collation / sorting
MongoDB doesn't yet know how to sort query results in a locale-sensitive way. If you can think up a good way to do it and implement it, we'd like
to know!
Drivers
If you use an esoteric/new/awesome programming language write a driver to support MongoDB! Again, check google to see what people have
started for various languages.
Some that might be nice:
Scheme (probably starting with PLT)

GNU R
Visual Basic
Lisp (e.g, Common Lisp)
Delphi
Falcon
Write a killer app that uses MongoDB as the persistance layer!
UI
Spec/requirements for a future MongoDB admin UI.
list databases
repair, drop, clone?
collections
validate(), datasize, indexsize, clone/copy
indexes
queries - explain() output
security: view users, adjust
see replication status of slave and master
sharding
system.profile viewer ; enable disable profiling
curop / killop support
Source Code
All source for MongoDB, it's drivers, and tools is open source and hosted at Github .
Mongo Database (includes C++ driver)

Python Driver
PHP Driver
Ruby Driver
Java Driver
Perl Driver
(Additionally, community drivers and tools also exist and will be found in other places.)
See Also
Building
License
Building
This section provides instructions on setting up your environment to write Mongo drivers or other infrastructure code. For specific instructions, go
to the document that corresponds to your setup.
Note: see the Downloads page for prebuilt binaries!
Sub-sections of this section:
Building Boost
Building for FreeBSD
Building for Linux
Building for OS X
Building for Solaris
Building for Windows
Building Spider Monkey
scons
See Also

Building with V8
Building Boost
Windows
MongoDB uses the [www.boost.org] C++ libraries.
Windows
See also the prebuilt libraries page.
By default c:\boost\ is checked for the boost files. Include files should be under \boost\boost, and libraries in \boost\lib.
First download the boost source. Then use the 7 Zip utility to extra the files. Place the extracted files in C:\boost.
Then we will compile the required libraries.

See buildscripts/buildboost.bat and buildscripts/buildboost64.bat for some helpers.
> rem set PATH for compiler:

> "C:\Program Files (x86)\Microsoft Visual Studio 10.0\VC\vcvarsall.bat"
>
> rem build the bjam make tool:
> cd \boost\tools\jam\src\
> build.bat
>
> cd \boost
> tools\jam\src\bin.ntx86\bjam --help
> rem see also mongo/buildscripts/buildboost*.bat
> rem build DEBUG libraries:
> tools\jam\src\bin.ntx86\bjam variant=debug threading=multi --with-program_options --with-filesystem
--with-date_time --with-thread
> mkdir lib
> move stage\lib\* lib\
Building for FreeBSD

On FreeBSD 8.0 and later, there is a mongodb port you can use.
For FreeBSD <= 7.2:
1. Get the database source: http://www.github.com/mongodb/mongo.

2. Update your ports tree:
$ sudo portsnap fetch && portsnap extract
The packages that come by default on 7.2 and older are too old, you'll get weird errors when you try to run the database)
3. Install SpiderMonkey:
$ cd /usr/ports/lang/spidermonkey && make && make install
4. Install scons:
$ cd /usr/ports/devel/scons && make && make install
5. Install boost: (it will pop up an X "GUI", select PYTHON)
$ cd /usr/ports/devel/boost-all && make && make install
6. Install libexecinfo:
$ cd /usr/ports/devel/libexecinfo && make && make install
7. Change to the database source directory

8. scons .
See Also
Building for Linux - many of the details there including how to clone from git apply here too.
Building for Linux

General Instructions
Special Notes about Spider Monkey
Package Requirements
Fedora
Fedora 8 or 10
Ubuntu
Ubuntu 8.04
Ubuntu 9.04 and 9.10
Ubuntu 10.04
See Also
General Instructions
1. Install Dependencies - see platform specific below

2. get source
git clone git://github.com/mongodb/mongo.git

# pick a stable version unless doing true dev
git tag -l
# Switch to a stable branch (unless doing development) --
# an even second number indicates "stable". (Although with
# stharding you will want the latest if the latest is less
# than 1.6.0.) For example:
git checkout r1.4.1
3. build
scons all
4. install
scons --prefix=/opt/mongo install
Special Notes about Spider Monkey
Most pre-built spider monkey binaries don't have UTF8 compiled in. Additionally, ubuntu has a weird version of spider monkey that doesn't
support everything we use. If you get any warnings during compile time or runtime, we highly recommend building spider monkey from source.
See Building Spider Monkey for more information.
We currently support spider monkey 1.6 and 1.7, although there is some degredation with 1.6, so we recommend using 1.7. We have not yet
tested 1.8, but will once it is officially released.
Package Requirements
Fedora
Fedora 8 or 10
sudo yum -y install git tcsh scons gcc-c++ glibc-devel

sudo yum -y install boost-devel pcre-devel js-devel readline-devel
#for release builds:
sudo yum -y install boost-devel-static readline-static ncurses-static
Ubuntu
See spider monkey note above.
Ubuntu 8.04
apt-get -y install tcsh git-core scons g++

apt-get -y install libpcre++-dev libboost-dev libreadline-dev xulrunner-1.9-dev
apt-get -y install libboost-program-options-dev libboost-thread-dev libboost-filesystem-dev
libboost-date-time-dev
Ubuntu 9.04 and 9.10

apt-get -y install libpcre++-dev libboost-dev libreadline-dev xulrunner-1.9.1-dev
Ubuntu 10.04

apt-get -y install libpcre++-dev libboost-dev libreadline-dev xulrunner-dev
See Also
The Building page for setup information for other operating systems
Building for OS X
Upgrading to Snow Leopard
Setup
Package Manager Setup (32bit)
Manual Setup
Install Apple developer tools
Install libraries (32bit option)
Compiling
XCode
Troubleshooting
To set up your OS X computer for MongoDB development:
Upgrading to Snow Leopard
If you have installed Snow Leopard, the builds will be 64 bit -- so if moving from a previous OS release, a bit more setup may be required than
one might first expect.
1. Install XCode tools for Snow Leopard.

2. Install MacPorts (snow leopard version). If you have MacPorts installed previously, we've had the most success by running rm -rf
/opt/local first.
3. Update/install packages: sudo port install boost pcre.
4. Update/install SpiderMonkey with sudo port install spidermonkey. (If this fails, see the note on #2 above.)
Setup
1. Install git. If not already installed, download the source and run ./configure; make; sudo make install
Then: git clone git://github.com/mongodb/mongo.git (more info)
Then: git tag -l to see tagged version numbers
Switch to a stable branch (unless doing development) -- an even second number indicates "stable". (Although with stharding you
will want the latest if the latest is less than 1.6.0.) For example:
git checkout r1.4.1
If you do not wish to install git you can instead get the source code from the Downloads page.
1. Install gcc.
gcc version 4.0.1 (from XCode Tools install) works, but you will receive compiler warnings. The easiest way to upgrade gcc is to install
the iPhone SDK.
Package Manager Setup (32bit)
1. Install libraries (using macports)
port install boost pcre++ spidermonkey
Manual Setup
Install Apple developer tools
1. Download boost {{boost 1.37.0 http://downloads.sourceforge.net/boost/boost_1_37_0.tar.gz}}Apply the following patch:
diff -u -r a/configure b/configure

--- a/configure 2009-01-26 14:10:42.000000000 -0500
+++ b/configure 2009-01-26 10:21:29.000000000 -0500
@@ -9,9 +9,9 @@
BJAM=""
TOOLSET=""
-BJAM_CONFIG=""
+BJAM_CONFIG="--layout=system"
BUILD=""
PREFIX=/usr/local
EPREFIX=
diff -u -r a/tools/build/v2/tools/darwin.jam b/tools/build/v2/tools/darwin.jam
--- a/tools/build/v2/tools/darwin.jam 2009-01-26 14:22:08.000000000 -0500
+++ b/tools/build/v2/tools/darwin.jam 2009-01-26 10:22:08.000000000 -0500
@@ -367,5 +367,5 @@
actions link.dll bind LIBRARIES

{
- "$(CONFIG_COMMAND)" -dynamiclib -Wl,-single_module -install_name "$(<:B)$(<:S)" -L
"$(LINKPATH)" -o "$(<)" "$(>)" "$(LIBRARIES)" -l$(FINDLIBS-SA) -l$(FINDLIBS-ST)
$(FRAMEWORK_PATH) -framework$(_)$(FRAMEWORK:D=:S=) $(OPTIONS) $(USER_OPTIONS)
+ "$(CONFIG_COMMAND)" -dynamiclib -Wl,-single_module -install_name
"/usr/local/lib/$(<:B)$(<:S)" -L"$(LINKPATH)" -o "$(<)" "$(>)" "$(LIBRARIES)" -l$(FINDLIBS-SA)
-l$(FINDLIBS-ST) $(FRAMEWORK_PATH) -framework$(_)$(FRAMEWORK:D=:S=) $(OPTIONS) $(USER_OPTIONS)
}
then,
./configure; make; sudo make install
2. Install pcre http://www.pcre.org/ (must enable UTF8)
./configure --enable-utf8 --enable-unicode-properties --with-match-limit=200000

--with-match-limit-recursion=4000; make; sudo make install
3. Install c++ unit test framework http://unittest.red-bean.com/ (optional)
(The 64bit libraries will be installed in /usr/64/{include,lib}.)
1. Download SpiderMonkey: ftp://ftp.mozilla.org/pub/mozilla.org/js/js-1.7.0.tar.gz
Apply the following patch:

diff -u -r js/src/config/Darwin.mk js-1.7.0/src/config/Darwin.mk
--- js/src/config/Darwin.mk 2007-02-05 11:24:49.000000000 -0500
+++ js-1.7.0/src/config/Darwin.mk 2009-05-11 10:18:37.000000000 -0400
@@ -43,7 +43,7 @@
# Just ripped from Linux config
#
-CC = cc
+CC = cc -m64
CCC = g++
CFLAGS += -Wall -Wno-format
OS_CFLAGS = -DXP_UNIX -DSVR4 -DSYSV -D_BSD_SOURCE -DPOSIX_SOURCE -DDARWIN
@@ -56,9 +56,9 @@
#.c.o:
# $(CC) -c -MD $*.d $(CFLAGS) $<
-CPU_ARCH = $(shell uname -m)

+CPU_ARCH = "X86_64"
ifeq (86,$(findstring 86,$(CPU_ARCH)))
-CPU_ARCH = x86
+CPU_ARCH = x86_64
OS_CFLAGS+= -DX86_LINUX
endif
GFX_ARCH = x
@@ -81,3 +81,14 @@
# Don't allow Makefile.ref to use libmath
NO_LIBM = 1
+ifeq ($(CPU_ARCH),x86_64)
+# Use VA_COPY() standard macro on x86-64
+# FIXME: better use it everywhere
+OS_CFLAGS += -DHAVE_VA_COPY -DVA_COPY=va_copy
+endif
+
+ifeq ($(CPU_ARCH),x86_64)
+# We need PIC code for shared libraries
+# FIXME: better patch rules.mk & fdlibm/Makefile*
+OS_CFLAGS += -DPIC -fPIC
+endif
compile and install
cd src
make -f Makefile.ref
sudo JS_DIST=/usr/64 make -f Makefile.ref export
remove the dynamic library
sudo rm /usr/64/lib64/libjs.dylib
# Download boost {{boost 1.37.0 http://downloads.sourceforge.net/boost/boost_1_37_0.tar.gz}}Apply the following patch:

diff -u -r a/configure b/configure
--- a/configure 2009-01-26 14:10:42.000000000 -0500
+++ b/configure 2009-01-26 10:21:29.000000000 -0500
@@ -9,9 +9,9 @@
BJAM=""
TOOLSET=""
-BJAM_CONFIG=""
+BJAM_CONFIG="architecture=x86 address-model=64 --layout=system"
BUILD=""
-PREFIX=/usr/local
+PREFIX=/usr/64
EPREFIX=
LIBDIR=
INCLUDEDIR=
diff -u -r a/tools/build/v2/tools/darwin.jam b/tools/build/v2/tools/darwin.jam
--- a/tools/build/v2/tools/darwin.jam 2009-01-26 14:22:08.000000000 -0500
+++ b/tools/build/v2/tools/darwin.jam 2009-01-26 10:22:08.000000000 -0500
@@ -367,5 +367,5 @@
actions link.dll bind LIBRARIES

{
- "$(CONFIG_COMMAND)" -dynamiclib -Wl,-single_module -install_name "$(<:B)$(<:S)" -L"$(LINKPATH)"
-o "$(<)" "$(>)" "$(LIBRARIES)" -l$(FINDLIBS-SA) -l$(FINDLIBS-ST) $(FRAMEWORK_PATH)
-framework$(_)$(FRAMEWORK:D=:S=) $(OPTIONS) $(USER_OPTIONS)
+ "$(CONFIG_COMMAND)" -dynamiclib -Wl,-single_module -install_name "/usr/64/lib/$(<:B)$(<:S)" -L
"$(LINKPATH)" -o "$(<)" "$(>)" "$(LIBRARIES)" -l$(FINDLIBS-SA) -l$(FINDLIBS-ST) $(FRAMEWORK_PATH)
-framework$(_)$(FRAMEWORK:D=:S=) $(OPTIONS) $(USER_OPTIONS)
}
then,
# Install pcre http://www.pcre.org/ (must enable UTF8)
CFLAGS="-m64" CXXFLAGS="-m64" LDFLAGS="-m64" ./configure --enable-utf8 --with-match-limit=200000

--with-match-limit-recursion=4000 --enable-unicode-properties --prefix /usr/64; make; sudo make
install
# Install unit test framework http://unittest.red-bean.com/ (optional)
CFLAGS="-m64" CXXFLAGS="-m64" LDFLAGS="-m64" ./configure --prefix /usr/64; make; sudo make install
Compiling
To compile 32bit, just run
scons
To compile 64bit on 10.5 (64 is default on 10.6), run
scons --64
XCode
You can open the project with:

$ open mongo.xcodeproj/
You need to add an executable target.:
1. In the mongo project window, go to the Executables, right click and choose Add->NewCustomExecutable.
2. Name it db. Path is ./db/db.
It will appear under Executables".
3. Double-click on it.
4. Under general, set the working directory to the project directory.
5. Under arguments, add run.
6. Go to general prefs (cmd ,), go to debugging and turn off lazy load.
(Seems to be an issue that prevents breakpoints from working in debugger?)
Troubleshooting
Undefined symbols: "_PR_NewLock", referenced from: _JS_Init in libjs.a.

Try not using the scons --release option (if you are using it). That option attempts to use static libraries.

MongoDB server currently supports little endian Solaris operation. (Although most drivers – not the database server – work on both.)
Community: Help us make this rough page better please! (And help us add support for big endian please...)
Prerequisites:
g++ 4.x (SUNWgcc)

scons (need to install from source)
spider monkey Building Spider Monkey
pcre (SUNWpcre)
boost (need to install from source)
See Also
Joyent
Building for Linux - many of the details there including how to clone from git apply here too

MongoDB can be compiled for Windows (32 and 64 bit) using Visual C++. SCons is the make mechanism, although a .vcproj/.sln is also included
in the project for convenience when using the Visual Studio IDE.
There are several dependencies exist which are listed below; you may find it easier to simply download a pre-built binary.
Building with Visual Studio 2008

Building the Shell
See Also
Prebuilt Boost Libraries

Prebuilt SpiderMonkey for VS2010
Building Boost
Building SpiderMonkey
Windows Quick Links
scons
Boost 1.41.0 Visual Studio 2010 Binary
This is OLD and was for the VS2010 BETA. See the new Boost and Windows page instead.
The following is a prebuilt boost binary (libraries) for Visual Studio 2010 beta 2.
The MongoDB vcxproj files assume this package is unzipped under c:\Program Files\boost\boost_1_41_0\.
http://downloads.mongodb.org/misc/boost_1_41_0_binary_vs10beta2.zipx
Note: we're not boost build gurus please let us know if there are things wrong with the build.
See also the prebuilt boost binaries at http://www.boostpro.com/download.
Boost and Windows

Visual Studio 2010
Prebuilt from mongodb.org
Building Yourself
Visual Studio 2008
Prebuilt from boostpro.com
Building Yourself
Additional Notes
Visual Studio 2010
Click here for a prebuilt boost library for Visual Studio 2010. 7zip format.
Building Yourself
Download the boost source from boost.org. Move it to C:\boost\.

Run C:\Program Files (x86)\Microsoft Visual Studio 10.0\vc\vcvarsall.bat.
From the MongoDB source project, run buildscripts\buildboost.bat. Or, buildboost64.bat for the 64 bit version.
We have successfully compiled version 1.42 – you might want to try that version or higher. See additional notes section at end of this
page too.
Visual Studio 2008
Click here for a prebuilt boost library for Visual Studio 2008. 7zip format. This file has what you need to build MongoDB, but not some other boost
libs, so it's partial.
Prebuilt from boostpro.com
Or, you can download a complete prebuilt boost library for 32 bit VS2008 at http://www.boostpro.com/products/free. Install the prebuilt libraries for
Boost version 1.35.0 (or higher - generally newer is better). During installation, for release builds choose static multithread libraries
for installation. The Debug version of the project uses the DLL libraries; choose all multithread libraries if you plan to do development.
From the BoostPro installer, be sure to select all relevant libraries that mongodb uses -- for example, you need Filesystem, Regex, Threads, and
ProgramOptions (and perhaps others).
Building Yourself
Download the boost source from boost.org. Move it to C:\boost\.

From the Visual Studio 2008 IDE, choose Tools.Visual Studio Command Prompt to get a command prompt with all PATH variables set
nicely for the C++ compiler.
From the MongoDB source project, run buildscripts\buildboost.bat. Or, buildboost64.bat for the 64 bit version.
Additional Notes
When using bjam, MongoDB expects
variant=debug for debug builds, and variant=release for release builds

threading=multi
link=static runtime-link=static for release builds
address-model=64 for 64 bit
Building the Mongo Shell on Windows

You can build the mongo shell with either scons or a Visual Studio 2010 project file.
Scons
scons mongo
Visual Studio 2010 Project File
A VS2010 vcxproj file is availabe for building the shell. From the mongo directory open shell/msvc/mongo.vcxproj.
The project file assumes that GNU readline is installed in ../readline/ relative to the mongo project. If you would prefer to build without having to
install readline, remove the definition of USE_READLINE in the preprocessor definitions section of the project file, and exclude readline.lib from
the project.
The project file currently only supports 32 bit builds of the shell (scons can do 32 and 64 bit). However this seems sufficient given there is no real
need for a 64 bit version of the shell.
Readline Library
The shell uses the GNU readline library to facilitate command line editing and history. You can build the shell without readline but would then lose
that functionality. USE_READLINE is defined when building with readline. SCons will look for readline and if not found build without it.
See Also
Prebuilt readline for Windows 32 bit at SourceForge (DLL version)

Get the MongoDB Source Code
Get Boost Libraries
Get SpiderMonkey
Building MongoDB from the IDE
Install SCons
Building MongoDB with SCons
Troubleshooting
MongoDB can be compiled for Windows (32 and 64 bit) using Visual C++. SCons is the make mechanism, although a solution file is also included
Download the source code from Downloads.
Or install Git. Then:
git clone git://github.com/mongodb/mongo.git (more info)

git tag -l to see tagged version numbers
Switch to a stable branch (unless doing development) -- an even second number indicates "stable". (Although with sharding you will want
the latest if the latest is less than 1.6.0.) For example:
git checkout r1.4.1
Get Boost Libraries
Click here for a prebuilt boost library for Visual Studio. 7zip format. This file has what you need to build MongoDB, but not some other
boost libs, so it's partial.
See the Boost and Windows page for other options.
The Visual Studio project files are setup to look for boost in the following locations:
c:\program files\boost\latest
c:\boost
\boost
You can unzip boost to c:\boost, or use an NTFS junction point to create a junction point to one of the above locations. Some versions of windows
come with linkd.exe, but others require you to download Sysinternal's junction.exe to accomplish this task. For example, if you installed boost 1.42
via the installer to the default location of c:\Program Files\boost\boost_1_42, You can create a junction point with the following command:
junction "c:\Program Files\boost\latest" "c:\Program Files\boost\boost_1_42"
This should return the following output:
Junction v1.05 - Windows junction creator and reparse point viewer

Copyright (C) 2000-2007 Mark Russinovich
Systems Internals - http://www.sysinternals.com
Created: c:\Program Files\boost\latest

Targetted at: c:\Program Files\boost\boost_1_42
Get SpiderMonkey
Build a SpiderMonkey js engine library (js.lib) – details here.
Open the db/db.sln solution file.
Note: currently the mongo shell and C++ client libraries must be built from scons. Also, for the VS2008 project files, 64 bit must be built from
scons (although you can do 64 bit with db_10.sln in vs2010).
Install SCons
If building with scons, install SCons:
First install Python: http://www.python.org/download/releases/2.6.4/.

Then SCons itself: http://sourceforge.net/projects/scons/files/scons/1.2.0/scons-1.2.0.win32.exe/download.
Add the python scripts directory (e.g., C:\Python26\Scripts) to your PATH.
The SConstruct file from the MongoDB project is the preferred way to perform production builds. Run scons in the mongo project directory to
build.
If scons does not automatically find Visual Studio, preset your path for it by running the VS2010 vcvars*.bat file.
To build:
scons // build mongod

scons mongoclient.lib // build C++ client driver library
scons all // build all end user components
scons . // build all including unit test
Troubleshooting
If you are using scons, check the file config.log which is generated.
Can't find jstypes.h when compiling. This file is generated when building SpiderMonkey. See the Building SpiderMonkey page for
more info.
Can't find / run cl.exe when building with scons. See troubleshooting note on the Building SpiderMonkey page.
Error building program database. (VS2008.) Try installing the Visual Studio 2008 Service Pack 1.

Get Boost Libraries
Get SpiderMonkey
Install SCons
Troubleshooting
MongoDB can be compiled for Windows (32 and 64 bit) using Visual C++. SCons is the make mechanism, although a solution file is also included
Download the source code from Downloads.
Or install Git. Then:
git clone git://github.com/mongodb/mongo.git (more info)

git tag -l to see tagged version numbers
Switch to a stable branch (unless doing development) -- an even second number indicates "stable". (Although with sharding you will want
the latest if the latest is less than 1.6.0.) For example:
git checkout r1.4.1
Get Boost Libraries
Click here for a prebuilt boost library for Visual Studio. 7zip format. This file has what you need to build MongoDB, but not some other
boost libs, so it's partial.
See the Boost and Windows page for other options. Use v1.42 or higher with VS2010.
Get SpiderMonkey
Download prebuilt libraries and headers here for VS2010. Place these files in ../js/ relative to your mongo project directory.
Or (more work) build SpiderMonkey js.lib yourself – details here.
Open the db/db_10.sln solution file.
Note: a separate project file exists for the mongo shell. Currently the C++ client libraries must be built from scons (this obviously needs to be
fixed...)
Install SCons
If building with scons, install SCons:
First install Python: http://www.python.org/download/releases/2.6.4/.

Then SCons itself: http://sourceforge.net/projects/scons/files/scons/1.2.0/scons-1.2.0.win32.exe/download.
Add the python scripts directory (e.g., C:\Python26\Scripts) to your PATH.
The SConstruct file from the MongoDB project is the preferred way to perform production builds. Run scons in the mongo project directory to
build.
If scons does not automatically find Visual Studio, preset your path for it by running the VS2010 vcvars*.bat file.
To build:
scons // build mongod

scons mongoclient.lib // build C++ client driver library
scons all // build all end user components
scons . // build all including unit test
Troubleshooting
If you are using scons, check the file config.log which is generated.
Can't find jstypes.h when compiling.

This file is generated when building SpiderMonkey. See the Building SpiderMonkey page for more info.
Can't find / run cl.exe when building with scons.
See troubleshooting note on the Building SpiderMonkey page.
LINK : fatal error LNK1104: cannot open file js64d.lib js64r.lib js32d.lib js32r.lib
Get the prebuilt spidermonkey libraries -- or copy your self-built js.lib to the above name.

Building js.lib - Unix
Download
Build
Install
Building js.lib - Windows
Prebuilt
Download
Build
Troubleshooting scons
See Also
MongoDB uses SpiderMonkey for server-side Javascript execution. The mongod project requires a file js.lib when linking. This page details how
to build js.lib.
Note: V8 Javascript support is under development.
Building js.lib - Unix
Download
curl -O ftp://ftp.mozilla.org/pub/mozilla.org/js/js-1.7.0.tar.gz
tar zxvf js-1.7.0.tar.gz
Build
cd js/src
export CFLAGS="-DJS_C_STRINGS_ARE_UTF8"
SpiderMonkey does not use UTF-8 by default, so we enable before building.
An experimental SConstruct build file is available here.
Install
JS_DIST=/usr make -f Makefile.ref export
By default, the mongo scons project expects spidermonkey to be located at ../js/.
Building js.lib - Windows
Prebuilt
VS2008: a prebuilt SpiderMonkey library and headers for Win32 is attached to this document (this file may or may not work depending on
your compile settings and compiler version).
VS2010 prebuilt libraries
Alternatively, follow the steps below to build yourself.
Download
From an msysgit or cygwin shell, run:
curl -O ftp://ftp.mozilla.org/pub/mozilla.org/js/js-1.7.0.tar.gz
tar zxvf js-1.7.0.tar.gz
Build
cd js/src
export CFLAGS="-DJS_C_STRINGS_ARE_UTF8"
If cl.exe is not found, launch Tools...Visual Studio Command Prompt from inside Visual Studio -- your path should then be correct for make.
If you do not have a suitable make utility installed, you may prefer to build using scons. An experimental SConstruct file to build the js.lib is
available in the mongodb/snippets project. For example:
cd
git clone git://github.com/mongodb/mongo-snippets.git
cp mongo-snippets/jslib-sconstruct js/src/SConstruct
cd js/src
scons
Troubleshooting scons
Note that scons does not use your PATH to find Visual Studio. If you get an error running cl.exe, try changing the following line in the msvc.py
scons source file from:
MVSdir = os.getenv('ProgramFiles') + r'\Microsoft Visual Studio 8'
to
MVSdir = os.getenv('ProgramFiles') + r'\Microsoft Visual Studio ' + version
See Also
Building MongoDB
scons
Use scons to build MongoDB, and related utilities and libraries. See the SConstruct file for details.
Run scons --help to see all options.
Targets
Run scons <target>.
scons .
scons all
scons mongod build mongod
scons mongo build the shell
scons shell build just the generated shell .cpp files
scons mongoclient build just the client lib
Options
--d debug build

--dd debug build with _DEBUG defined (extra asserts etc.)
--release
--32 force 32 bit
--64 force 64 bit
--clean
Troubleshooting
scons generates a config.log file. See this file when there are problems building.
See Also
Smoke Tests
Database Internals
This section provides information for developers who want to write drivers or tools for MongoDB, \ contribute code to the MongoDB codebase
itself, and for those who are just curious how it works internally.
Sub-sections of this section:
Caching
Cursors
Error Codes
Internal Commands
Replication Internals
Smoke Tests
Pairing Internals
Caching
Memory Mapped Storage Engine
This is the current storage engine for MongoDB, and it uses memory-mapped files for all disk I/O. Using this strategy, the operating system's
virtual memory manager is in charge of caching. This has several implications:
There is no redundancy between file system cache and database cache: they are one and the same.
MongoDB can use all free memory on the server for cache space automatically without any configuration of a cache size.
Virtual memory size and resident size will appear to be very large for the mongod process. This is benign: virtual memory space will be
just larger than the size of the datafiles open and mapped; resident size will vary depending on the amount of memory not used by other
processes on the machine.
Caching behavior (such as LRU'ing out of pages, and laziness of page writes) is controlled by the operating system: quality of the VMM
implementation will vary by OS.
Cursors
Redirection Notice
This page should redirect to Internals.
Error Codes
Error Code Description Comments
10003 objects in a capped ns cannot grow
11000 duplicate key error _id values must be unique in a collection
11001 duplicate key on update
12000 idxNo fails an internal error
12001 can't sort with $snapshot the $snapshot feature does not support sorting yet
12010, 12011, 12012 can't $inc/$set an indexed field
13440 bad offset accessing a datafile Run a database --repair
Internal Commands
Most commands have helper functions and do not require the $cmd.findOne() syntax. These are primarily internal and administrative.
> db.$cmd.findOne({assertinfo:1})
{
"dbasserted" : false , // boolean: db asserted
"asserted" : false , // boolean: db asserted or a user assert have happend
"assert" : "" , // regular assert
"assertw" : "" , // "warning" assert
"assertmsg" : "" , // assert with a message in the db log
"assertuser" : "" , // user assert - benign, generally a request that was not meaningful
"ok" : 1.0
}
> db.$cmd.findOne({serverStatus:1})
{
"uptime" : 6 ,
"globalLock" : {
"totalTime" : 6765166 ,
"lockTime" : 2131 ,
"ratio" : 0.00031499596610046226
} ,
"mem" : {
"resident" : 3 ,
"virtual" : 111 ,
"mapped" : 32
} ,
"ok" : 1
}
> admindb.$cmd.findOne({replacepeer:1})
{
"info" : "adjust local.sources hostname; db restart now required" ,
"ok" : 1.0
}
// close all databases. a subsequent request will reopen a db.

> admindb.$cmd.findOne({closeAllDatabases:1});
On the master mongod instance, the local database will contain a collection, oplog.$main, which stores a high-level transaction log. The
transaction log essentially describes all actions performed by the user, such as "insert this object into this collection." Note that the oplog is not a
low-level redo log, so it does not record operations on the byte/disk level.
The slave mongod instance polls the oplog.$main collection from master. The actual query looks like this:
local.oplog.$main.find({ ts: { $gte: ''last_op_processed_time'' } }).sort({$natural:1});
where 'local' is the master instance's local database. oplog.$main collection is a capped collection, allowing the oldest data to be aged out
automatically.
See the Replication section of the Mongo Developers' Guide for more information.
OpTime
An OpTime is a 64-bit timestamp that we use to timestamp operations. These are stored as Javascript Date datatypes but are not JavaScript
Date objects. Implementation details can be found in the OpTime class in repl.h.
Applying OpTime Operations
Operations from the oplog are applied on the slave by reexecuting the operation. Naturally, the log includes write operations only.
Note that inserts are transformed into upserts to ensure consistency on repeated operations. For example, if the slave crashes, we won't know
exactly which operations have been applied. So if we're left with operations 1, 2, 3, 4, and 5, and if we then apply 1, 2, 3, 2, 3, 4, 5, we should
achieve the same results. This repeatability property is also used for the initial cloning of the replica.
Tailing
After applying operations, we want to wait a moment and then poll again for new data with our $gteoperation. We want this operation to be fast,
quickly skipping past old data we have already processed. However, we do not want to build an index on ts, as indexing can be somewhat
expensive, and the oplog is write-heavy. Instead, we use a table scan in [natural] order, but use a tailable cursor to "remember" our position.
Thus, we only scan once, and then when we poll again, we know where to begin.
Initiation
To create a new replica, we do the following:
t = now();
cloneDatabase();
end = now();
applyOperations(t..end);
cloneDatabaseeffectively exports/imports all the data in the database. Note the actual "image" we will get may or may not include data
modifications in the time range (t..end). Thus, we apply all logged operations from that range when the cloning is complete. Because of our
repeatability property, this is safe.
See class Cloner for more information.
Smoke Tests
smoke.py lets you run a subsets of the tests in jstests. When it is running tests, it starts up an instance of mongod, runs the tests, and then shuts
it down again. For the moment, smoke.py must be run from the top-level directory of a mongo source repository.
To see the possible options, run:
$ python buildscripts/smoke.py --help

Usage: smoke.py [OPTIONS] ARGS*
Options:
-h, --help show this help message and exit
--mode=MODE If "files", ARGS are filenames; if "suite", ARGS are
sets of tests. ( default "suite")
--mongod=MONGODEXECUTABLE
Path to mongod to run ( default "./mongod")
--port=MONGODPORT Port the mongod will bind to (default 32000)
--mongo=SHELLEXECUTABLE
Path to mongo, for .js test files (default "./mongo")
--continue-on-failure
If supplied, continue testing even after a test fails
--one-mongod-per-test
If supplied, run each test in a fresh mongod
--from-file=FILE Run tests/suites named in FILE, one test per line, '-'
means stdin
--smoke-db-prefix=SMOKEDBPREFIX
Prefix to use for the mongods' dbpaths.
--small-oplog Run tests with master/slave replication & use a small
oplog
To run specific tests, use the --mode=files option:
python buildscripts/smoke.py --mode=files jstests/find1.js
You can specify as many files as you want.
You can also run a suite of tests. Suites are predefined and include:
smoke
smokeAll
smokePerf
smokeJs
smokeQuota
smokeJsPerf
smokeDisk
smokeJsSlowNightly
smokeJsSlowWeekly
smokeParallel
smokeClone
smokeRepl
smokeAuth
smokeSharding
smokeTool
smokeClient
mongosTest
To run a suite, specify the suite's name:
python buildscripts/smoke.py smokeJs
The old way to run tests is using scons, then the suite name:
// run basic c++ unit tests:

scons smoke
// run all unit tests (slow)

// NOTE This scons target starts a mongod instance needed by some tests.
scons smokeAll
// to run basic jstests

// first start mongod, then:
scons smokeJs
Full list of scons targets for running smoke tests: smoke, smokePerf, smokeClient, mongosTest, smokeJs, smokeClone, smokeRepl, smokeDisk,
smokeSharding, smokeJsPerf, smokeQuota, smokeTool.
Pairing Internals
Policy for reconciling divergent oplogs
In a paired environment, a situation may arise in which each member of a pair has logged operations as master that have not been applied to the
other server. In such a situation, the following procedure will be used to ensure consistency between the two servers:
1. The new master will scan through its own oplog from the point at which it last applied an operation from it's peer's oplog to the end. It will
create a set C of object ids for which changes were made. It will create a set M of object ids for which only modifier changes were made.
The values of C and M will be updated as client operations are applied to the new master.
2. The new master will iterate through its peer's oplog, applying only operations that will not affect an object having an id in C.
3. For any operation in the peer's oplog that may not be applied due to the constraint in the previous step, if the id of the of the object in
question is in M, the value of the whole object on the new master is logged to the new master's oplog.
4. The new slave applies all operations from the new master's oplog.

Qualified volunteers are welcome to assist in editing the wiki documentation. Contact us for more information.
Emacs tips for MongoDB work

You can edit confluence directly from emacs:
First, follow the basic instructions on http://code.google.com/p/confluence-el/
Change the confluence-url in their sample setup to http://mongodb.onconfluence.com/rpc/xmlrpc
Might also want to change the default space to DOCS or DOCS-ES or whatever space you edit the most.
etags setup (suggested by mstearn)
First, install "exuberant ctags", which has nicer features than GNU etags.
http://ctags.sourceforge.net/
Then, run something like this in the top-level mongo directory to make an emacs-style TAGS file:
ctags -e --extra=+qf --fields=+iasnfSKtm --c++-kinds=+p --recurse .
Then you can use M-x visit-tags-table, M-., M-* as normal.
Mongo Documentation Style Guide

This page provides information for everyone adding to the Mongo documentation on Confluence. It covers:
General Notes on Writing Style

Guide to Confluence markup for specific situations
Some general notes about doc production
General Notes on Writing Style
Voice
Active voice is almost always preferred to passive voice.
To make this work, however, you may find yourself anthropromorphizing components of the system - that is, treating the driver or the database as
an agent that actually does something. ("The dbms writes the new record to the collection" is better than "the new record is written to the
database", but some purists may argue that the dbms doesn't do anything - it's just code that directs the actions of the processor - but then
someone else says "yes, but does the processor really do anything?" and so on and on.) It is simpler and more economical to write as if these
components are actually doing things, although you as the infrastructure developers might have to stop and think about which component is
actually performing the action you are describing.
Tense
Technical writers in general prefer to keep descriptions of processes in the present tense: "The dbms writes the new collection to disk" rather than
"the dbms will write the new collection to disk." You save a few words that way.
MongoDB Terminology
It would be good to spell out precise definitions of technical words and phrases you are likely to use often, like the following:
Mongo
database (do you want "a Mongo database"? Or a Mongo database instance?)
dbms (I have't seen this term often - is it correct to talk about "the Mongo DBMS"?)
Document
Record
Transaction (I stopped myself from using this term because my understanding is the Mongo doesn't support "transactions" in the sense of
operations that are logged and can be rolled back - is this right?)
These are just a few I noted while I was editing. More should be added. It would be good to define these terms clearly among yourselves, and
then post the definitions for outsiders.
Markup for terms
It's important to be consistent in the way you treat words that refer to certain types of objects. The following table lists the types you will deal with
most often, describes how they should look, and (to cut to the chase) gives you the Confluence markup that will achieve that appearance.
Type Appearance Markup
Object name (the type of "object" that "object-oriented programming" deals with) monospace {{ term }}
short code fragment inline monospace {{ term }}
file path/name, extension italic _term_
programming command, statement or expression monospace {{ term }}

variable or "replaceable item" monospace italic _term_
Placeholders in paths, directories, or other text that would be italic anyway angle brackets around <item> <item>
GUI element (menus menu items, buttons bold *term*
First instance of a technical term italic _term_
tag (in HTML or XML, for example) monospace {{ term }}
Extended code sample code block {code}

program code
{code}
In specifying these, I have relied on the O'Reilly Style Guide, which is at:
http://oreilly.com/oreilly/author/stylesheet.html
This guide is a good reference for situations not covered here.
I should mention that for the names of GUI objects I followed the specification in the Microsoft Guide to Technical Publications.
Other Confluence markup
If you are editing a page using Confluence's RTF editor, you don't have to worry about markup. Even if you are editing markup directly,
Confluence displays a guide on the right that shows you most of the markup you will need.
References and Links
Confluence also provides you with a nice little utility that allows you to insert a link to another Confluence page by searching for the page by title
or by text and choosing it from a list. Confluence handles the linking markup. You can even use it for external URLs.
The one thing this mechanism does NOT handle is links to specific locations within a wiki page. Here is what you have to know if you want to
insert these kinds of links:
Every heading you put in a Confluence page ("h2.Title", "h3.OtherTitle", etc.) becomes an accessible "anchor" for linking.
You can also insert an anchor anywhere else in the page by inserting "{anchor\:anchorname}" where _anchorname is the unique name
you will use in the link.
To insert a link to one of these anchors, you must go into wiki markup and add the anchor name preceded by a "#". Example: if the page
MyPage contains a heading or an ad-hoc anchor named GoHere, the link to that anchor from within the same page would look like
[#GoHere], and a link to that anchor from a different page would look like [MyPage#GoHere]. (See the sidebar for information about
adding other text to the body of the link.)
Special Characters
You will often need to insert code samples that contain curly braces. As Dwight has pointed out, Confluence gets confused by this unless
you "escape" them by preceding them with a backslash, thusly:
\{ \}
You must do the same for "[", "]", "_" and some others.
Within a {code} block you don't have to worry about this. If you are inserting code fragments inline using {{ and }}, however, you still need
to escape these characters. Further notes about this:
If you are enclosing a complex code expression with {{ and }}, do NOT leave a space between the last character of the
expression and the }}. This confuses Confluence.
Confluence also gets confused (at least sometimes) if you use {{ and }}, to enclose a code sample that includes escaped curly
brackets.
About MongoDB's Confluence wiki
Confluence has this idea of "spaces". Each person has a private space, and there are also group spaces as well.
The MongoDB Confluence wiki has three group spaces defined currently:
Mongo Documentation - The publicly accessible area for most Mongo documentation
Contributor - Looks like, the publicly accessible space for information for "Contributors"
Private - a space open to MongoDB developers, but not to the public at large.
As I said in my email on Friday, all of the (relevant) info from the old wiki now lives in the "Mongo Documentation"
Standard elements of Wiki pages
You shouldn't have to spend a lot of time worrying about this kind of thing, but I do have just a few suggestions:
Since these wiki pages are (or can be) arranged hierarchically, you may have "landing pages" that do little more than list their child
pages. I think Confluence actually adds a list of children automatically, but it only goes down to the next hierarchical level. To insert a
hierarchical list of a page's children, all you have to do is insert the following Confluence "macro":
{children:all=true}
See the Confluence documentation for more options and switches for this macro.
For pages with actual text, I tried to follow these guidelines:

For top-level headings, I used "h2" not "h1"
I never began a page with a heading. I figured the title of the page served as one.
I always tried to include a "See Also" section that listed links to other Mongo docs.
I usually tried to include a link to the "Talk to us about Mongo" page.
Community
General Community Resources
User Mailing List
The user list is for general questions about using, configuring, and running MongoDB and the associated tools and drivers. The list is
open to everyone
IRC chat
irc://irc.freenode.net/#mongodb
Blog
http://blog.mongodb.org/
Bugtracker
File, track, and vote on bugs and feature requests. There is issue tracking for MongoDB and all supported drivers
Announcement Mailing List
http://groups.google.com/group/mongodb-announce - for release announcement and important bug fixes.
Store
Visit our Cafepress store for Mongo-related swag.
Resources for Driver and Database Developers
Developer List
This mongodb-dev mailing list is for people developing drivers and tools, or who are contributing to the MongoDB codebase itself.
Source
The source code for the database and drivers is available at the http://github.com/mongodb.
Job Board
Click Here to access the Job Board. The Board is a community resource for all employers to post MongoDB-related jobs. Please feel free
to post/investigate positions!
MongoDB Commercial Services Providers
Note: if you provide consultative or support services for MongoDB and wish to be listed here, just let us know.
Support
10gen
Training
Hosting
Consulting
10gen
Hashrocket
LightCube Solutions
Squeejee
Mijix
Support
10gen
10gen began the MongoDB project, and offers commercial MongoDB support services.
Training
10gen offers MongoDB training.
Hosting
See the MongoDB Hosting Center.
Consulting
10gen
10gen offers consulting services for MongoDB application design, development, and production operation. These services are typically advisory in
nature with the goal of building higher in-house expertise on MongoDB for the client.
Hashrocket
Hashrocket is a full-service design and development firm that builds successful web businesses. Hashrocket continually creates and follows best
practices and surround themselves with passionate and talented craftsmen to ensure the best results for you and your business.
LightCube Solutions
LightCube Solutions provides PHP development and consulting services, as well as a lightweight PHP framework designed for MongoDB called
'photon'
Squeejee
{partnerlink:Squeejee|http://squeejee.com/|squeejee] builds web applications on top of MongoDB with multiple sites already in production.
Mijix
MijiX, a software development studio based on Indonesia, provides consulting for MongoDB in Asia-Pacific area.
User Feedback
"I just have to get my head around that mongodb is really _this_ good"
-muckster, #mongodb
"Guys at Redmond should get a long course from you about what is the software development and support "
[email protected], mongodb-user list
"#mongoDB keep me up all night. I think I have found the 'perfect' storage for my app "
-elpargo, Twitter
"Maybe we can relax with couchdb but with mongodb we are completely in dreams"
-namlook, #mongodb
"Dude, you guys are legends!"

-Stii, mongodb-user list
"Times I've been wowed using MongoDB this week: 7."

-tpitale, Twitter
Community Blog Posts

B is for Billion
-Wordnik (July 9, 2010)
[Reflections on MongoDB]
-Brandon Keepers, Collective Idea (June 15, 2010)
Building a Better Submission Form

-New York Times Open Blog (May 25, 2010)
Notes from a Production MongoDB Deployment

-Boxed Ice (February 28, 2010)
NoSQL in the Real World

-CNET (February 10, 2010)
Why I Think Mongo is to Databases what Rails was to Frameworks

-John Nunemaker, Ordered List (December 18, 2009)
MongoDB a Light in the Darkness...

-EngineYard (September 24, 2009)
Introducing MongoDB
-Linux Magazine (September 21, 2009)
Choosing a non-relational database; why we migrated from MySQL to MongoDB

-Boxed Ice (July 7, 2010)
The Other Blog - The Holy Grail of the Funky Data Model
-Tom Smith (June 6, 2009)
GIS Solved - Populating a MongoDb with POIs

-Samuel
Community Presentations
Scalable Event Analytics with MongoDb and Ruby on Rails
Jared Rosoff at RubyConfChina (June 2010)
How Python, TurboGears, and MongoDB are Transforming SourceForge.net

Rick Copeland at PyCon 2010
MongoDB
Adrian Madrid at Mountain West Ruby Conference 2009, video
MongoDB - Ruby friendly document storage that doesn't rhyme with ouch
Wynn Netherland at Dallas.rb Ruby Group, slides
MongoDB
jnunemaker at Grand Rapids RUG, slides
Developing Joomla! 1.5 Extensions, Explained (slide 37)

Mitch Pirtle at Joomla!Day New England 2009, slides
Drop Acid (slide 31) (video)

Bob Ippolito at Pycon 2009
Python and Non-SQL Databases (in French, slide 21)

Benoit Chesneau at Pycon France 2009, slides
Massimiliano Dessì at the Spring Framework Italian User Group
MongoDB (in Italian)

MongoDB and Scala (in Italian)
Presentations and Screencasts at Learnivore
Frequently-updated set of presentations and screencasts on MongoDB.
Benchmarking
We keep track of user benchmarks on the Benchmarks page.
Job Board
Redirecting...
Redirection Notice
This page should redirect to http://jobs.mongodb.org/ in about 2 seconds.
About
Philosophy
Use Cases
Production Deployments
Mongo-Based Applications
Events
Slide Gallery
Articles
Benchmarks
FAQ
Product Comparisons
Licensing
Philosophy
Design Philosophy
Databases are specializing - the "one size fits all" approach no

longer applies.
By reducing transactional semantics the db provides, one can
still solve an interesting set of problems where performance is
very important, and horizontal scaling then becomes easier.
The (JSON) document data model is easy to code to, easy to
manage (schemaless), and yields excellent performance by
grouping relevant data together internally.
A non-relational approach is the best path to database
solutions which scale horizontally to many machines.
While there is an opportunity to relax certain capabilities for
better performance, there is also a need for deeper
functionality than that provided by pure key/value stores.
Database technology should run anywhere, being available
both for running on your own servers or VMs, and also as a
cloud pay-for-what-you-use service.
Use Cases
See also the Production Deployments page for a discussion of how companies like Shutterfly, foursquare, bit.ly, Etsy, SourceForge, etc. use
MongoDB.
Use Case Articles
Using MongoDB for Real-time Analytics

Using MongoDB for Logging
MongoDB and E-Commerce
Well Suited
Operational data store of a web site. MongoDB is very good at real-time inserts, updates, and queries. Scalability and replication are
provided which are necessary functions for large web sites' real-time data stores. Specific web use case examples:
content management
comment storage, management, voting
real time page view counters
user registration, profile, session data
Caching. With its potential for high performance, MongoDB works well as a caching tier in an information infrastructure. The persistent
backing of Mongo's cache assures that on a system restart the downstream data tier is not overwhelmed with cache population activity.
High volume problems. Problems where a traditional DBMS might be too expensive for the data in question. In many cases developers
would traditionally write custom code to a filesystem instead using flat files or other methodologies.
Storage of program objects and JSON data (and equivalent). Mongo's BSON data format makes it very easy to store and retrieve data in
a document-style / "schemaless" format. Addition of new properties to existing objects is easy and does not require blocking "ALTER
TABLE" style operations.
Document and Content Management Systems - as a document-oriented (JSON) database, MongoDB's flexible schemas are a good fit
for this.
Electronic record keeping - similar to document management.
Less Well Suited
Systems with a heavy emphasis on complex transations such as banking systems and accounting. These systems typically require
multi-object transactions, which MongoDB doesn't support. It's worth noting that, unlike many "NoSQL" solutions, MongoDB does support
atomic operations on single documents. As documents can be fairly rich entities, for many use cases this is sufficient.
Traditional Business Intelligence. Data warehouses are more suited to new, problem-specific BI databases. However note that
MongoDB can work very well for several reporting and analytics problems where data is predistilled or aggregated in runtime -- but
classic, nightly batch load business intelligence, while possible, is not necessarily a sweet spot.
Problems requiring SQL.
Use Case - Session Objects

MongoDB is a good tool for storing HTTP session objects.
One implementation model is to have a sessions collection, and store the session object's _id value in a browser cookie.
With its update-in-place design and general optimization to make updates fast, the database is efficient at receiving an update to the session
object on every single app server page view.
Aging Out Old Sessions
The best way to age out old sessions is to use the auto-LRU facility of capped collections. The one complication is that objects in capped
collections may not grow beyond their initial allocation size. To handle this, we can "pre-pad" the objects to some maximum size on initial
addition, and then on further updates we are fine if we do not go above the limit. The following mongo shell javascript example demonstrates
padding.
(Note: a clean padding mechanism should be added to the db so the steps below are not necessary.)
> db.createCollection('sessions', { capped: true, size : 1000000 } )
{"ok" : 1}
> p = "";
> for( x = 0; x < 100; x++ ) p += 'x';
> s1 = { info: 'example', _padding : p };
{"info" : "example" , "_padding" :
"xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx"
}
> db.sessions.save(s1)
> s1
{"info" : "example" , "_padding" :
, "_id" : ObjectId( "4aafb74a5761d147677233b0") }
> // when updating later

> s1 = db.sessions.find( { _id : ObjectId( "4aafb74a5761d147677233b0") } )
{"_id" : ObjectId( "4aafb74a5761d147677233b0") , "info" : "example" , "_padding" :
}
> delete s._padding;
true
> s.x = 3; // add a new field
3
> db.sessions.save(s);
> s
{"_id" : ObjectId( "4aafb5a25761d147677233af") , "info" : "example" , "x" : 3}
Production Deployments
If you're using MongoDB in production, we'd love to list you here! Email [email protected].
Company Use Case
Shutterfly is an Internet-based social expression and personal publishing service.

MongoDB is used for various persistent data storage requirements within Shutterfly.
MongoDB helps Shutterfly build an unrivaled service that enables deeper, more
personal relationships between customers and those who matter most in their lives. For
more information, view Kenny Gorman's presentation Implementing MongoDB at
Shutterfly from MongoSF (April 2010): Slides and Video.
Foursquare is a location based social network that incorporates gaming elements.
MongoDB at foursquare presentation: Slides and Video (May 2010)
bit.ly allows users to shorten, share, and track links. bit.ly uses MongoDB to store user
history. For more information:
bit.ly user history, auto-sharded presentation at MongoNYC (May 2010)
MongoDB is used for back-end storage on the SourceForge front pages, project pages,
and download pages for all projects.
Scaling SourceForge with MongoDB - OSCON Presentation (July 2010)

MongoDB at SourceForge - QCon London Presentation (March 2010)
How Python, TurboGears, and MongoDB are Transforming SourceForge.net -
PyCon (February 2010)
SourceForge.net releases Ming - SourceForge blog (December 2009)
TurboGears on Sourceforge - Compound Thinking (July 2009)
Etsy is a website that allows users to buy and sell handmade items. Read the
MongoDB at Etsy blog series:
Part 1 - May 19, 2010

Part 2 - July 3, 2010
The New York Times is using MongoDB in a form-building application for photo
submissions. Mongo's lack of schema gives producers the ability to define any
combination of custom form fields. For more information:
Building a Better Submission Form - NYTimes Open Blog (May 25, 2010)
A Behind the Scenes Look at the New York Times Moment in Time Project -
Hacks/Hackers Blog (July 28, 2010)
Examiner.com is the fastest-growing local content network in the U.S., powered by the
largest pool of knowledgeable and passionate contributors in the world. Launched in
April 2008 with 60 cities, Examiner.com now serves hundreds of markets across the
U.S. and Canada.
Examiner.com migrated their site from Cold Fusion and SQL Server to Drupal 7 and
MongoDB. Details of the deployment are outlined in an Acquia case study
BoxedIce's server monitoring solution - Server Density - stores 600 million+ documents
in MongoDB.
BoxedIce blog posts:

Automating partitioning, sharding and failover with MongoDB
Why we migrated from mysql to mongodb
Notes from a production deployment
Humongous Data at Server Density: Approaching 1 Billion Documents
in MongoDB
Presentations:
Humongous Data at Server Density - MongoUK Presentation (June
2010)
MongoDB in Production at Boxed Ice - Webinar (May 2010)
Wordnik stores its entire text corpus in MongoDB - 1.2TB of data in over 5 billion
records. The speed to query the corpus was cut to 1/4 the time it took prior to migrating
to MongoDB. More about MongoDB at Wordnik:
B is for Billion - Wordnik Blog (July 2010)

MongoDB: Migration from Mysql at Wordnik - Scalable Web Architectures
(May 2010)
Tony Tam's Presentation at MongoSF (April 2010)
What has technology done for words lately? - Wordnik blog (February 2010)
ShareThis makes it easy to share ideas and get to the good stuff online. ShareThis is
the world’s largest sharing network reaching over 400 million users across 150,000
sites and 785,000 domains across the web
Business Insider has been using MongoDB since the beginning of 2008. All of the site's
data, including posts, comments, and even the images, are stored on MongoDB. For
more information:
How This Web Site Uses MongoDB (November 2009 Article)

How Business Insider Uses MongoDB (May 2010 Presentation)
GitHub, the social coding site, is using MongoDB for an internal reporting application.
Gilt Groupe is an invitation only luxury shopping site. Gilt uses MongoDB for real time
ecommerce analytics.
Gilt CTO Mike Bryzek's presentation at MongoSF in April 2010.

Hummingbird - a real-time web traffic visualization tool developed by Gilt and
powered by MongoDB
IGN Entertainment, a unit of News Corporation, is a leading Internet media and

services provider focused on the videogame and entertainment enthusiast markets.
IGN’s properties reached more than 37.3 million unique users worldwide February
2010, according to Internet audience measurement firm comScore Media Metrix.
MongoDB powers IGN’s real-time traffic analytics and RESTful Content APIs.
CollegeHumor is a comedy website. MongoDB is used in CollegeHumor for internal
analytics and link exchange application.
Evite uses MongoDB for analytics and quick reporting.
Tracking and visualizing mail logs with MongoDB and gviz_api - Grig
Gheorghiu's blog (July 2010)
Disqus is an innovative blog-commenting system.
MongoHQ provides a hosting platform for MongoDB and also uses MongoDB as the
back-end for its service. Our hosting centers page provides more information about
MongoHQ and other MongoDB hosting options.
Justin.tv is the easy, fun, and fast way to share live video online. MongoDB powers
Justin.tv's internal analytics tools for virality, user retention, and general usage stats
that out-of-the-box solutions can't provide. Read more about Justin.tv's broadcasting
architecture.
Chartbeat is a revolutionary real-time analytics service that enables people to

understand emergent behaviour in real-time and exploit or mitigate it. Chartbeat stores
all historical analytics data in MongoDB.
The Secret Weapons Behind Chartbeat - Kushal's coding blog (April 2010)
Kushal Dave's Presentation at MongoNYC (May 2010)
Hot Potato is a social tool that organizes conversations around events. For more
information:
Hot Potato's presentation about using Scala and MongoDB at the New York
Tech Talks Meetup (March 2010)
Hot Potato presentation at MongoSF (April 2010)
Hot Potato Infrastructure from Hot Potato blog (May 2010)
Hot Potato presentation at MongoNYC (May 2010)
Eventbrite gives you all the online tools you need to bring people together for an event
and sell tickets. EventBrite uses MongoDB to track page views.
Why you should track page views with MongoDB - EventBrite Blog (June
2010)
Flowdock is a modern web-based team messenger, that helps your team to become
more organized simply by chatting. Flowdock backend uses MongoDB to store all
messages.
Why Flowdock migrated from Cassandra to MongoDB - Flowdock Blog (July

2010)
The Chicago Tribune uses MongoDB in its Illinois School Report Cards application,
which is generated from a nearly 9,000 column denormalized database dump produced
annually by the State Board of Education. The application allows readers to search by
school name, city, county, or district and to view demographic, economic, and
performance data for both schools and districts.
Sugar CRM uses MongoDB to power the backend of its preview feedback mechanism.
It captures users' comments and whether they like or dislike portions of the application
all from within beta versions of Sugar.
WHERE® is a local search and recommendation service that helps people discover
places, events and mobile coupons in their area. Using WHERE, people can find
everything from the weather, news, and restaurant reviews, to the closest coffee shop,
cheapest gas, traffic updates, movie showtimes and offers from local merchants.
WHERE is available as a mobile application and as a web service at Where.com. here,
Inc. uses MongoDB to store geographic content for the WHERE application and for
WHERE Ads™ - a hyper-local ad network.
PhoneTag is a service that automatically transcribes voicemail to text and delivers it in

real-time via e-mail and SMS. PhoneTag stores the metadata and transcriptions for
every voicemail it process in MongoDB.
Harmony is a powerful web-based platform for creating and managing websites. It

helps developers with content editors work together with unprecedented flexibility and
simplicity. From stylesheets, images and templates, to pages, blogs, and comments,
every piece of Harmony data is stored in MongoDB. Switching to MongoDB from
MySQL drastically simplified Harmony's data model and increased the speed at which
we can deliver features.
Steve Smith's presentation about Harmony at MongoSF (April 2010)
Hashrocket is an expert web design and development group. Hashrocket built

PharmMD, a fully-featured Medication Management application in Ruby on Rails. The
system contains functionality for identifying and resolving drug-related problems for
millions of patients.
Yottaa offers Performance Analytics, a cloud service that monitors, ranks and analyzes
the performance of millions of web sites, providing an open database to answer
questions such as “why performance matters” and “how fast is my site”. Yottaa is using
Ruby on Rails and MongoDB to build their scalable analytics engine.
Scalable Event Analytics with MongoDB and Ruby - Jared Rosoff's

presentation at RubyConfChina (June 2010)
BuzzFeed is a trends aggregator that uses a web crawler and human editors to find
and link to popular stories around the web. BuzzFeed moved an analytics system
tracking over 400 million monthly events from MySQL to MongoDB.
The Mozilla open-source Ubiquity Herd project uses MongoDB for back-end storage.
Source code is available on bitbucket.
Codaset is an open system where you can browse and search through open source
projects, and check out what your friends are coding.
The awesomeness that is MongoDB and NoSQL, is taking over Codaset -

Codaset Blog (May 2010)
Handling Dates in MongoDB - Codaset Blog (July 2010)
Shopwiki uses Mongo as a data store for its shopping search engine, where they
commit all the data generated, such as custom analytics. Mongo's performance is such
that ShopWiki uses it in cases where MySQL would just not be practical. ShopWiki is
also using it as a storage engine for all R&D and data-mining efforts where MongoDB's
document oriented architecture offers maximum flexibility.
Avery's Talk at MongoNYC - ShopWiki Dev Blog (June 2010)
MyPunchbowl.com is a start to finish party planning site that uses MongoDB for
tracking user behavior and datamining.
Ryan Angilly on Replacing MySQL with MongoDB (Zero to Mongo) on The

Bitsource
MongoDB for Dummies: How MyPunchbowl went from 0 to production in under
3 days - Presentation at MongoNYC (May 2010)
Sunlight Labs is a community of open source developers and designers dedicated to
opening up our government to make it more transparent, accountable and responsible.
MongoDB powers the National Data Catalog, and the Drumbone API, which is an
aggregator of data about members of Congress.
Civic Hacking - Video from Luigi Montanez's presentation at MongoNYC (May

2010)
How We Use MongoDB at Sunlight blog post (May 2010)
photostre.am streams image data from flickr and uses MongoDB as it's only database.
MongoDB in Production at photostre.am - photostre.am blog (June 2010)
Fotopedia uses MongoDB as storage backend for its copy of wikipedia data, storage for
users and albums timelines, a feature that is currently under heavy refactoring, and as
the "metacache", an index of every tiny html fragment in its varnish cache for proactive
invalidation of stale content.
MongoDB: Our Swiss Army Datastore - Presentation at MongoFR in June

2010: Slides and Video
Grooveshark currently uses Mongo to manage over one million unique user sessions
per day.
Stickybits is a fun and social way to attach digital content to real world objects.
MongoDB is being used for the game feeds component. It caches game data from
different sources which gets served to ea.com, rupture.com and the EA download
manager.
Struq develops technology that personalises the contents and design of online display
advertising in real time.
Pitchfork is using MongoDB for their year-end readers survey and internal analytics.
Floxee, a web toolkit for creating Twitter directories, leverages MongoDB for back-end
storage. The award-winning TweetCongress is powered by Floxee.
Sailthru is an email service provider that uses MongoDB for click-stream analysis and
reporting.
Silentale keeps track of your contacts and conversations from multiple platforms and
allows you to search and access them from anywhere. Silentale is using MongoDB as
the back-end for indexing and searching on millions of stored messages of different
types. More details on Silentale can be found in this TechCrunch article.
One Year with MongoDB presentation from MongoUK (June 2010): Slides and
Video
TeachStreet helps people find local and online classes by empowering teachers with
robust tools to manage their teaching businesses. MongoDB powers our real-time
analytics system which provide teachers with insight into the performance and
effectiveness of their listings on TeachStreet.
Slides from Mongo Seattle - TeachStreet blog (July 2010)

Defensio is a comment-spam blocker that uses MongoDB for back-end storage.
TweetSaver is a web service for backing up, searching, and tagging your tweets.
TweetSaver uses MongoDB for back-end storage.
Bloom Digital's AdGear platform is a next-generation ad platform. MongoDB is used for

back-end reporting storage for AdGear.
KLATU Networks designs, develops and markets asset monitoring solutions which
helps companies manage risk, reduce operating costs and streamline operations
through proactive management of the status, condition, and location of cold storage
assets and other mission critical equipment. KLATU uses MongoDB to store
temperature, location, and other measurement data for large wireless sensor networks.
KLATU chose MongoDB over competitors for scalability and query capabilities.
songkick lets you track your favorite artists so you never miss a gig again.
Speeding up your Rails app with MongoDB - Presentation at MongoUK (June

2010)
Detexify is a cool application to find LaTeX symbols easily. It uses MongoDB for
back-end storage. Check out the blog post for more on why Detexfy is using MongoDB.
http://sluggy.com/ is built on MongoDB, mongodb_beaker, and MongoKit.
From MySQL to MongoDB at Sluggy.com - Brendan McAdams' presentation at

MongoNYC (May 2010)
Stylesignal is using MongoDB to store opinions from social media, blogs, forums and
other sources to use in their sentiment analysis system, Zeitgeist.
@trackmeet helps you take notes with twitter, and is built on MongoDB
eFlyoverleverages the Google Earth Browser Plugin and MongoDB to provide

interactive flyover tours of over two thousand golf courses worldwide.
Shapado is a multi-topic question and answer site in the style of Stack Overflow.
Shapado is written in Rails and uses MongoDB for back-end storage.
Sifino enables students to help each other with their studies. Students can share
notes, course summaries, and old exams, and can also ask and respond to questions
about particular courses.
GameChanger provides mobile apps that replace pencil-and-paper scorekeeping and

online tools that distribute real-time game updates for amateur sports.
Tornado, MongoDB, and the Realtime Web - Kiril Savino's presentation at

MongoNYC (May 2010)
GameChanger and MongoDB: a case study in MySQL conversion - Kiril
Savino's blog (September 2010)
soliMAP is a map-based ad listings site that uses MongoDB for storage.
MyBankTracker iPhone App uses MongoDB for the iPhone app's back-end server.
BillMonitor uses MongoDB to store all user data, including large amounts of billing
information. This is used by the live site and also by BillMonitor's internal data analysis
tools.
Tubricator allows you to create easy to remember links to YouTube videos. It's built on
MongoDB and Django.
Mu.ly uses MongoDB for user registration and as a backend server for its iPhone Push
notification service. MongoDB is mu.ly's Main backend database and absolute mission
critical for mu.ly.
Avinu is a Content Management System (CMS) built on the Vork enterprise framework
and powered by MongoDB.
edelight is a social shopping portal for product recommendations.
Topsy is a search engine powered by Tweets that uses Mongo for realtime log
processing and analysis.
Codepeek is using MongoDB and GridFS for storing pastes.
Similaria.pl is an online platform, created to connect users with people and products
that match them.
ToTuTam uses Mongo to store information about events in its portal and also to store
and organise information about users preferences.
themoviedb.org is a free, user driven movie database that uses MongoDB as its
primary database.
OCW Search is a search engine for OpenCourseWare. It stores all the course materials
in MongoDB and uses Sphinx to index these courses.
Full Text Search with Sphinx - Presentation from MongoUK (June 2010)
Mixero is the new generation Twitter client for people who value their time and are tired
of information noise. Mixero uses Mongo to store users' preferences and data.
Biggo is an advanced site constructor with e-commerce modules. Biggo uses

MongoDB for stats collection.
Kabisa is a web development firm specializing in Ruby on Rails and Java / J2EE.
Kabisa uses MongoDB for many of its client projects, including a mobile news
application for iPhone and Android.
DokDok makes it easy and automatic for users to find, work on and share the latest
version of any document - right from their inbox. DokDok migrated to a Mongo backend
in August 2009. See Bruno Morency's presentation Migrating to MongoDB for more
information.
Enbil is a swedish website for finding, and comparing, rental cars. MongoDB is used for
storing and querying data about geographical locations and car rental stations.
Websko is a content management system designed for individual Web developers and
cooperative teams. MongoDB's lack of schema gives unlimited possibilities for defining
manageable document oriented architecture and is used for back-end storage for all
manageable structure and data content. Websko is written in Rails, uses MongoMapper
gem and in-house crafted libraries for dealing with Mongo internals.
markitfor.me is a bookmarking service that makes your bookmarks available via full-text
search so you don't have to remember tags or folders. You can just search for what
you're looking for and the complete text of all of your bookmarked pages will be
searched. MongoDB is used as the datastore for the marked pages.
Backpage Pics is a website that displays backpage.com adult classified listings as an

image gallery. MongoDB is used to store listing data. Please note that this website is
NSFW.
Joomla Ads uses MongoDB for its back-end reporting services.
musweet keeps track of what artists and bands publish on the social web.
Eiwa System Management, Inc. is a software development firm that has been using
MongoDB for various projects since January 2010.
Morango is an internet strategy consultancy based in London, which uses MongoDB in

production on several client projects.
Building a Content Management System with MongoDB - Presentation from

MongoUK (June 2010)
PeerPong discovers everyone's expertise and connects you to the best person to
answer any question. We index users across the entire web, looking at public profiles,
real-time streams, and other publicly available information to discover expertise and to
find the best person to answer any question.
ibibo ("I build, I bond") is a social network using MongoDB for its dashboard feeds.
Each feed is represented as a single document containing an average of 1000 entries;
the site currently stores over two million of these documents in MongoDB.
MediaMath is the leader in the new and rapidly growing world of digital media trading.
Zoofs is a new way to discover YouTube videos that people are talking about on
Twitter. Zoofs camps in Twitter searching for tweets with YouTube video links, and then
ranks them based on popularity.
Oodle is an online classifieds marketplace that serves up more than 15 million visits a
month and is the company behind the popular Facebook Marketplace. Oodle is using
Mongo for storing user profile data for our millions of users and has also open sourced
its Mongo ORM layer.
Funadvice relaunched using the MongoDB and MongoMapper. Read the Funadvice
CTO's post to MongoDB User Forum from May 2010 for more details.
Ya Sabe is using MongoDB for the backend storage of business listings. Yasabe.com
is the first local search engine built for Hispanics in the US with advanced search
functionality. You can find and discover more than 14 million businesses via the web or
your mobile phone. All the information is in both Spanish and in English.
LoteriaFutbol.com is a Fantasy Soccer Portal recently launched for the World Soccer
Cup: South Africa 2010. Mongo has been used entirely to store data about users,
groups, news, tournaments and picks. It uses the PHP driver with a Mongo module for
Kohana v3 (Mango).
Kehalim switched over to MongoDB 1 year ago after exhausting other cloud and
relational options. As a contextual affiliate network, Kehalim stores all of its advertisers,
ads and even impressions on top of MongoDB. MongoDB has outed both MySQL and
memcached completely and also provides great hadoop-like alternative with its own
map-reduce.
Squarespace is an innovative web publishing platform that consists of a fully hosted

and managed GUI environment for creating and maintaining websites. Squarespace's
new social modules utilize Mongo to store large amounts of social data that is pulled in
from around the Internet and displayed in native widgets that are fully integrated with
the platform.
Givemebeats.net is an e-commerce music site that allows people to buy beats (music
instrumentals) produced by some of the best producers in the world. Now we entirely
use MongoDB to store users profile, beats information, and transaction statistics.
Cheméo, a search engine for chemical properties, is built on top of MongoDB. For a
fairly extensive explanation of the tools and software used and some MongoDB tips,
please go to chemeo.com/doc/technology.
Planetaki is place were you can read all your favourite websites in one place. MongoDB
has replaced MySQL for the storage backend that does all the heavy lifting and caching
of each website's news feed.
Planetaki Powered by MongoDB - SamLown.com (June 2010)
[ChinaVisual.com] is the leading and largest vertical portal and community for creative
people in China. ChinaVisual.com moved from mysql to mongoDB in early 2009.
Currently MongoDB powers its most major production and service, like file storage,
session server, and user tracking.
RowFeeder is an easy social media monitoring solution that allows people to track
tweets and Facebook posts in a spreadsheet. RowFeeder uses MongoDB to keep up
with the high volume of status updates across multiple social networks as well as
generate basic stats.
MongoDB for Real Time Data Collection and Stats Generation - Presentation
at Mongo Seattle (July 2010)
Mini Medical Record is designed to facilitate medical care for all members of the public.
While useful for everyone, it is especially useful for travelers, professional road
warriors, homeless, substance dependent, and other members of the public who
receive care through multiple medical systems.
Open Dining Network is a restaurant data and food ordering platform that provides a
RESTful API to take web and mobile orders. MongoDB is used to manage all
restaurant, customer, and order information.
URLi.st is a small web application to create and share list of links. The web application
is coded in Python (using the pylons framework) and uses MongoDB (with pymongo
1.6) in production to power its data layer.
Pinterest is a community to share collections of things you love. Pinterest is built in

Python and uses MongoDB for its internal analytics tools and huge data sets like
contacts imported from gmail and yahoo.
LearnBoost is a free and amazing gradebook web app that leverages MongoDB for its
data storage needs. LearnBoost is the creator of Mongoose, a JavaScript async ORM
for MongoDB that is flexible, extensible and simple to use.
Mongoose - LearnBoost blog (May 2010)
Kidiso is a safe online playground for children up to 13, with advanced parental
controls. In the current setup, we are using MongoDB for logging, analysis tasks, and
background jobs that aggregate data for performance (ie search results and allowed
content).
Carbon Calculated provides an open platform that aggregates carbon and green house
gas emissions for everything in the world, from passenger transport, raw materials,
through to consumer goods. Built on top of this platform, Carbon Calculated offers a
suite of products that make carbon calculation accessible and intuitive.
Vowch is a simple platform for telling the world about all the people, places and things
that matter most to you. It is a platform for making positive, public endorsements for
anyone or anything from a Twitter account.
View a vowch for MongoDB: http://vow.ch/2ij
HolaDoctor.com is the most comprehensive health and wellness portal available in

Spanish for the global online Hispanic community. MongoDB is being used to store all
the content for the site, including GridFS to store article images. Session data is also
being persisted on our MongoDB cluster using a custom PHP save handler.
Ros Spending is the first Russian public spending monitoring project. It includes
information about 1,400,000 federal government and 210,000 regional government
contracts, as well as information about more than 260,000 suppliers and 26,000
customers. MongoDB stores all reports, customer and supplier information, stats and
pre-cached queries. The project was initiated by the Institute of Contemporary
Development and launched publicly in July 2010 during the Tver economic forum.
BlueSpark designs and develops iPhone and iPad applications and specializes in
Adobe Flash development, we have a passion for creating great user experiences and
products that feel simple to use.
[Aghora] is a time attendance application specially designed for the requirements of the
Brasilian governmental requirements. Our whole application is based on PHP and
MongoDB. Click here for more information.
Man of the House is the real man's magazine, a guide for the jack of all trades trying to
be better – at work and at home, as a father and as a husband. The entire backend of
the site depends on MongoDB.
PeerIndex is an algorithmic authority ranking web service that uses MongoDB to scale
processing of the firehose of social media, as a distributed data store and middle cache
for fast site performance.
sahibinden.com is an online classifieds marketplace that serves more than 14.5 million
unique visitors and over 1.5 billion pageviews a month. sahibinden.com is using
MongoDB for storing classifieds data and caching.
Remembersaurus is a flashcard website targeted at language learners which helps the

learners focus on the cards that they are having the hardest time remembering.We're
using MongoDB for all of our backend storage, but it's been particularly useful for
helping log how well each user knows each of the cards they are studying.
Shadelight is a unique fantasy roleplaying game where you play one of the legendary
Guardians of Elumir. Set out on magical quests, battle mysterious creatures and
explore a truly unique fantasy world.
Ylastic is using MongoDB extensively in production. For example, MongoDB powers

Ylastic's monitors capability.
BRAINREPUBLIC is a social network for anyone to wants to talk face-to-face - or just

audio or chat - with like-minded people from anywhere at anytime.
Friendmaps is a tool that allows users to view all of their social networks on a single
map.
The affiliate marketing platform Jounce has gone live using MongoDB as the main
storage solution for its search data. As of August 2010, ~10 million offers are stored in
the database.
Virb Looking for a place to park your portfolio, your band, your website? Build an
elegantly simple website with Virb. You provide the content, we’ll help with the rest —
for only $10/month.
Deal Machine is a streamlined CRM that makes sales fun and effective. We use
MongoDB as our main storage. It has helped us a lot to make the web app better and
more scalable.
arrivalguides.com is the largest network of free online (and pdf) travel guides.
arrivalguides.com recently launched a new site where they rewrote the whole
application switching from SQL server to MongoDB using the NoRM Driver for C#. The
website is purely driven by MongoDB as the database backend.
The Hype Machine keeps track of emerging music on the web. We use MongoDB to
accelerate storage and retrieval of user preferences, and other core site data.
MongoDB's web-native design and high performance in our workloads was what got
our attention. It's from the future!
Scrabbly is a massively multiplayer online scrabble crossword. Uses MongoDB

geospatial indexing.
Building a Scrabble MMO in 48 Hours - Startup Monkeys Blog (September

2010)
See also
MongoDB Apps
Use Cases
User Feedback
Please list applications that leverage MongoDB here. If you're using MongoDB for your application, we'd love to list you here! Email
[email protected].
See Also
Production Deployments - Companies and Sites using MongoDB
Hosting Center
Applications Using MongoDB
CMS
HarmonyApp
Harmony is a powerful web-based platform for creating and managing websites. It helps connect developers with content editors, for
unprecedented flexibility and simplicity. For more information, view Steve Smith's presentation on Harmony at MongoSF (April 2010).
c5t
Content-management using TurboGears and Mongo
Websko
Websko is a content management system designed for individual Web developers and cooperative teams.
Graylog2
Graylog2 is an open source syslog server implementation that stores logs in MongoDB and provides a Rails frontend.
Analytics
Hummingbird
Hummingbird is a real-time web traffic visualization tool developed by Gilt Groupe
Events
Follow us on Facebook and Twitter to get all of the latest updates!
Mongo Conferences 10gen Weekly "Office
Hours"
Submit a proposal to present at an upcoming MongoDB conference! 10gen holds weekly open "office hours" with whiteboarding, hack
sessions, etc., in NYC. Come over to 10gen headquarters to
meet the MongoDB team.
Mongo Boston September 20
One day conference in Cambridge, MA. New York

Click here for conference agenda and registration.
Wednesdays, 4pm - 6:30pm ET
17 West 18th Street - 8th Floor
Mongo Berlin October 4
Between 5th & 6th Ave
New York, NY
One day conference in Berlin, Germany. * Please note that the doorbell says "ShopWiki"
Click here for conference agenda and registration.
San Francisco
Mongo Hamburg October 5
On the west coast? Stop by the Epicenter Cafe in San
Free, evening meetup in Hamburg, Germany, hosted by Jimdo Francisco on Mondays to meet 10gen Software Engineer
Click here for conference agenda and registration. Aaron Staple. Ask questions, hack, have some coffee.
Look for a laptop with a "Powered by MongoDB" sticker.
Mongo Chicago October 20
Mondays 5pm - 7pm PT
One day conference in Chicago, IL. Epicenter Cafe
Click here for conference agenda and registration. 764 Harrison St
Between 4th St & Lapu St
San Francisco, CA
MongoDB Webinars
Deployment Best Practices
September 14 at 12:30pm ET / 9:30am PT

Register
Scaling with MongoDB
September 17 at 1pm ET / 10am PT

Register
Data Migration: Importing your Data into MongoDB
October 12 at 12:30pm ET / 9:30am PT

Register
Check out the MongoDB Meetup groups:
New York MongoDB User Group San Francisco MongoDB User Group
Conferences and
Meetups
United States Europe Asia
Dinner in the sky with MongoDB

Northeast West Coast
Marc Boeker, ONchestra
NoSQL Franfurt
Schema Design with MongoDB Inside MongoDB: The Internals
Kyle Banker, Software Engineer, of an Open Source Database Franfurt, Germany
10gen Alvin Richards & Aaron Staple, September 28
New York MongoDB User Group 10gen
September 21 San Francisco MongoDB User Mongo - one year later
Group Marcin Szajek (Programa.pl) and
September 21 Mciej Dziardziel (Adwertajzing)
PyCon Ukraine
Midwest
ZendCon 2010 October 23-24
Converting your MySQL app to
Mongo Chicago October 20 NoSQL with MongoDB Building Web Applications with
One day conference in Chicago, MongoDB for Mobile MongoDB
IL. Applications Roger Bodamer, SVP Products
Click here for conference agenda Alvin Richards, 10gen and Engineering, 10gen
and registration. Santa Clara, CA Devoxx
November 2 Metropolis Antwerp, Belgium
Inside MongoDB: The Internals November 15
of an Open Source Database QCon 2010
Eliot Horowitz, CTO, 10gen Consistency Models in New
ChicagoDB Generation Databases
October 18 Dwight Merriman, 10gen
San Francisco, CA
November 4
Southwest
South
Introduction to MongoDB
Mohammad Azam, Sogeti
Houston Tech Fest
October 9
Media from Recent Events and Conferences
Slide Gallery | More Presentations and Video
If you're interested in having someone present MongoDB at your conference or meetup, or if you would like to list your MongoDB event on this
page, contact meghan at 10gen dot com. Want some MongoDB stickers to give out at your talk? Complete the Swag Request Form.
Video & Slides from Recent Events and Presentations

Table of Contents:
[ MongoDB Conferences ] [ Ruby/Rails ] [ Python ] [ Alt.NET ] [ User Experiences ] [ More about MongoDB ]
MongoDB Conferences
One-day conferences hosted by 10gen. 10gen develops and supports MongoDB.
MongoUK Video (June 2010)

MongoFR Video (June 2010)
MongoNYC (May 2010) and MongoSF (April 2010) Video
MongoSF (April 2010) Slides & Video
Ruby/Rails
Practical Ruby Projects with MongoDB

Alex Sharp, OptimisCorp
Ruby Midwest - June 2010
Scalable Event Analytics with MongoDB and Ruby

Jared Rosoff, Yottaa
RubyConfChina - June 26, 2010
The MongoDB Metamorphosis (Kyle Banker, 10gen)

Million Dollar Mongo (Obie Fernandez & Durran Jordan, Hashrocket)
Analyze This! (Blythe Dunham)
RailsConf
Baltimore, MD
June 7-10
MongoDB
Seth Edwards
London Ruby Users Group
London, UK
Wednesday April 14
Video & Slides
MongoDB: The Way and its Power

Kyle Banker, Software Engineer, 10gen
RubyNation
Friday April 9 & Saturday April 10
Reston, VA
Slides | Video
MongoDB Rules
Mountain West Ruby Conference
Salt Lake City, UT
Thursday March 11 & Friday March 12
Slides
MongoDB & Mongoid

Durran Jordan, Hashrocket
RubyJax
February 23, 2010
Video
MongoDB Isn't Water

Chicago Ruby
February 2, 2010
Video | Slides | Photos
Introduction to Mongo DB
Joon Yu, RubyHead
teachmetocode.com
Nov-Dec, 2009
Screencasts
Python
How Python, TurboGears, and MongoDB are Transforming SourceForge.net

Rick Copeland, SourceForge.net
PyCon - Atlanta, GA
February 21, 2010
Slides
Alt.NET
.NET and MongoDB - Building Applications with NoRM and MongoDB

Alex Hung
July 28, 2010
User Experiences
The Future of Content Technologies

Scaling Web Applications with NonSQL Databases: Business Insider Case Study
Ian White, Lead Developer, Business Insider
Gilbane Conference
San Francisco, CA
Thursday, May 20
Slides
Chartbeat and MongoDb - a perfect marriage

Kushal Dave, CTO, Chartbeat & Mike Dirolf, Software Engineer, 10gen
New York City Cloud Computing Meetup
New York, NY
May 18
Slides
Why MongoDB is Awesome

John Nunemaker, CTO, Ordered List
DevNation Chicago
May 15
Slides
Humongous Data at Server Density: Approaching 1 Billion Documents in MongoDB

David Mytton, Founder, Boxed Ice
Webinar
Wednesday May 5
Recording & Slides
Humongous Drupal
DrupalCon San Francisco
Karoly Negyesi, Examiner.com
Saturday April 17
Slides | Video
MongoDB: huMONGOus Data at SourceForge

Mark Ramm, Web Developer, SourceForge
QCon London
Thursday March 11
Slides
Migrating to MongoDB
Bruno Morency, DokDok
Confoo.ca
March 10 - 12
Slides
More about MongoDB
Recording of Michael Dirolf on MongoDB @ E-VAN 07 June 2010
NoSQL-Channeling the Data Explosion

Dwight Merriman, CEO, 10gen
Inside MongoDB: the Internals of an Open-Source
Mike Dirolf, Software Engineer, 10gen
Gluecon
Denver, CO
Wednesday May 26 & Thursday May 27
Schema Design with MongoDB

Webinar
Tuesday April 27
Recording and Slides
Dropping ACID with MongoDB

Kristina Chodorow, Software Engineer, 10gen
San Francisco MySQL Meetup
San Francisco, CA
Monday, April 12
Video
Emerging Technologies for the Enterprise Conference
Philadelphia, PA
Friday, April 9
Slides
Indexing with MongoDB

Aaron Staple, Software Engineer, 10gen
Webinar
Tuesday April 6, 2010
Video | Slides
TechZing Interview with Mike Dirolf, Software Engineer, 10gen

Monday, April 5
Podcast
Hot Potato and MongoDB
New York Tech Talks Meetup
Justin Shaffer and Lincoln Hochberg
New York, NY
Tuesday March 30
Video
MongoDB Day
Geek Austin Data Series
Austin, TX
Saturday March 27
Photo
Mongo Scale!
Webcast
Friday March 26
Webcast
NoSQL Live Boston

Boston, MA
Thursday March 11
Recap with slides and MP3
MongoDB: How it Works

Monday March 8, 12:30 PM Eastern Time
Slides
Intro to MongoDB
Alex Sharp, Founder / Lead Software Architect, FrothLogic
LA WebDev Meetup
February 23, 2010
Slides
FOSDEM - Brussels, Belgium
February 7, 2010
Video | Slides | Photos
If you're interested in having someone present MongoDB at your conference or meetup, or if you would like to list your MongoDB event on this
page, contact meghan at 10gen dot com.
Slide Gallery
Click here to visit our full listing of videos & slides from recent events and presentations.
Introduction to MongoDB User Experience
| Get your SlideShare Playlist | Get your SlideShare Playlist
Ruby/Rails Python
Java PHP
More About MongoDB

| Get your SlideShare Playlist
|
Articles
See also the User Feedback page for community presentations, blog posts, and more.
Best of the MongoDB Blog
What is the Right Data Model? - (for non-relational databases)

Why Schemaless is Good
The Importance of Predictability of Performance
Capped Collections - one of MongoDB's coolest features
Using MongoDB for Real-time Analytics
Using MongoDB for Logging
http://blog.mongodb.org/tagged/best+of
Articles / Key Doc Pages
On Atomic Operations
Reaching into Objects - how to do sophisticated query operations on nested JSON-style objects
Schema Design
MongoDB Production Deployments
Videos
MongoDB Blip.tv Channel

MongoDB for Rubyists (February 2010 Chicago Ruby Meetup)
Introduction to MongoDB (FOSDEM February 2010)
NY MySql Meetup - NoSQL, Scaling, MongoDB
Teach Me To Code - Introduction to MongoDB
DCVIE
Benchmarks
If you've done a benchmark, we'd love to hear about it! Let us know at kristina at 10gen dot com.
March 9, 2010 - Speed test between django_mongokit and postgresql_psycopg2 benchmarks creating, editing, and deleting.
February 15, 2010 - Benchmarking Tornado's Sessions flatfile, Memcached, MySQL, Redis, and MongoDB compared.
January 23, 2010 - Inserts and queries against MySQL, CouchDB, and Memcached.
May 10, 2009 - MongoDB vs. CouchDB vs. Tokyo Cabinet
July 2, 2009 - MongoDB vs. MySQL
September 25, 2009 - MongoDB inserts using Java.
August 11, 2009 - MySQL vs. MongoDB vs. Tokyo Tyrant vs. CouchDB inserts and queries using PHP.
August 23, 2009 - MySQL vs. MongoDB in PHP: Part 1 (inserts), Part 2 (queries), aginst InnoDB with and without the query log and MyISAM.
November 9, 2009 - MySQL vs. MongoDB in PHP and Ruby inserts (original Russian, English translation)
Disclaimer: these benchmarks were created by third parties not affiliated with MongoDB. MongoDB does not guarantee in any way the
correctness, thoroughness, or repeatability of these benchmarks.
See Also
http://blog.mongodb.org/post/472834501/mongodb-1-4-performance
FAQ
This FAQ answers basic questions for new evaluators of MongoDB. See also:
Developer FAQ
Sharding FAQ
MongoDB Intro FAQ
MongoDB Intro FAQ

What kind of database is the Mongo database?
What languages can I use to work with the Mongo database?
Does it support SQL?
Is caching handled by the database?
What language is MongoDB written in?
What are the 32-bit limitations?
What kind of database is the Mongo database?
MongoDB is an document-oriented DBMS. Think of it as MySQL but JSON (actually, BSON ) as the data model, not relational. There are no
joins. If you have used object-relational mapping layers before in your programs, you will find the Mongo interface similar to use, but faster, more
powerful, and less work to set up.
What languages can I use to work with the Mongo database?
Lots! See the drivers page.
Does it support SQL?
No, but MongoDB does support ad hoc queries via a JSON-style query language. See the Tour and Advanced Queries pages for more
information on how one performs operations.
Is caching handled by the database?
For simple queries (with an index) Mongo should be fast enough that you can query the database directly without needing the equivalent of
memcached. The goal is for Mongo to be an alternative to an ORM/memcached/mysql stack. Some MongoDB users do like to mix it with
memcached though.
What language is MongoDB written in?
The database is written in C++. Drivers are usually written in their respective languages, although some use C extensions for speed.
What are the 32-bit limitations?
MongoDB uses memory-mapped files. When running on a 32-bit operating system, the total storage size for the server (data, indexes,
everything) is 2gb. If you are running on a 64-bit os, there is virtually no limit to storage size. See the blog post for more information.
Product Comparisons
Interop Demo (Product Comparisons)
Interop 2009 MongoDB Demo
Code: http://github.com/mdirolf/simple-messaging-service/tree/master
MongoDB, CouchDB, MySQL Compare Grid

pending...
CouchDB MongoDB MySQL
Data Model Document-Oriented (JSON) Document-Oriented (BSON) Relational
Data Types string,number,boolean,array,object string, int, double, boolean, date, link

bytearray, object, array, others
Large Objects Yes (attachments) Yes (GridFS) blobs?
(Files)
Horizontal CouchDB Lounge Auto-sharding (v1.6) ?

partitioning
scheme
Replication Master-master (with developer supplied Master-slave (and "replica sets") Master-slave
conflict resolution)
Object(row) One large repository Collection based Table based

Storage
Query Method Map/reduce of javascript functions to lazily Dynamic; object-based query Dynamic; SQL
build an index per query language
Secondary Yes Yes Yes

Indexes
Atomicity Single document Single document Yes - advanced
Interface REST Native drivers ; REST add-on Native drivers
Server-side ? Map/Reduce, server-side javascript Yes (SQL)

batch data
manipulation
Written in Erlang C++ C++
Concurrency MVCC Update in Place

Control
Geospatial GeoCouch Yes. (As of June 2010, coordinate ?

Indexes system is cartesian. Spherical coming
soon.)
Distributed Eventually consistent (master-master Strong consistency. Eventually Strong consistency. Eventually
Consistency replication with versioning and version consistent reads from secondaries are consistent reads from secondaries are
Model reconciliation) available. available.
See Also
Comparing Mongo DB and Couch DB

We are getting a lot of questions "how are mongo db and couch different?" It's a good question: both are document-oriented databases with
schemaless JSON-style object data storage. Both products have their place -- we are big believers that databases are specializing and "one size
fits all" no longer applies.
We are not CouchDB gurus so please let us know in the forums if we have something wrong.
MVCC
One big difference is that CouchDB is MVCC based, and MongoDB is more of a traditional update-in-place store. MVCC is very good for certain
classes of problems: problems which need intense versioning; problems with offline databases that resync later; problems where you want a large
amount of master-master replication happening. Along with MVCC comes some work too: first, the database must be compacted periodically, if
there are many updates. Second, when conflicts occur on transactions, they must be handled by the programmer manually (unless the db also
does conventional locking -- although then master-master replication is likely lost).
MongoDB updates an object in-place when possible. Problems require high update rates of objects are a great fit; compaction is not necessary.
Mongo's replication works great but, without the MVCC model, it is more oriented towards master/slave and auto failover configurations than to
complex master-master setups. With MongoDB you should see high write performance, especially for updates.
Horizontal Scalability
One fundamental difference is that a number of Couch users use replication as a way to scale. With Mongo, we tend to think of replication as a
way to gain reliability/failover rather than scalability. Mongo uses (auto) sharding as our path to scalabity (sharding is GA as of 1.6). In this sense
MongoDB is more like Google BigTable. (We hear that Couch might one day add partitioning too.)
Query Expression
Couch uses a clever index building scheme to generate indexes which support particular queries. There is an elegance to the approach, although
one must predeclare these structures for each query one wants to execute. One can think of them as materialized views.
Mongo uses traditional dynamic queries. As with, say, MySQL, we can do queries where an index does not exist, or where an index is helpful but
only partially so. Mongo includes a query optimizer which makes these determinations. We find this is very nice for inspecting the data
administratively, and this method is also good when we don't want an index: such as insert-intensive collections. When an index corresponds
perfectly to the query, the Couch and Mongo approaches are then conceptually similar. We find expressing queries as JSON-style objects in
MongoDB to be quick and painless though
Atomicity
Both MongoDB and CouchDB support concurrent modifications of single documents. Both forego complex transactions involving large numbers
of objects.
Durability
The products take different approaches to durability. CouchDB is a "crash-only" design where the db can terminate at any time and remain
consistent. MongoDB take a different approach to durability. On a machine crash, one then would run a repairDatabase() operation when
starting up again (similar to MyISAM). MongoDB recommends using replication -- either LAN or WAN -- for true durability as a given server could
permanently be dead. To summarize: CouchDB is better at durability when using a single server with no replication.
Map Reduce
Both CouchDB and MongoDB support map/reduce operations. For CouchDB map/reduce is inherent to the building of all views. With MongoDB,
map/reduce is only for data processing jobs but not for traditional queries.
Javascript
Both CouchDB and MongoDB make use of Javascript. CouchDB uses Javascript extensively including in the building of views .
MongoDB supports the use of Javascript but more as an adjunct. In MongoDB, query expressions are typically expressed as JSON-style query
objects; however one may also specify a javascript expression as part of the query. MongoDB also supports running arbitrary javascript functions
server-side and uses javascript for map/reduce operations.
REST
Couch uses REST as its interface to the database. With its focus on performance, MongoDB relies on language-specific database drivers for
access to the database over a proprietary binary protocol. Of course, one could add a REST interface atop an existing MongoDB driver at any
time -- that would be a very nice community project. Some early stage REST implementations exist for MongoDB.
Performance
Philosophically, Mongo is very oriented toward performance, at the expense of features that would impede performance. We see Mongo DB
being useful for many problems where databases have not been used in the past because databases are too "heavy". Features that give
MongoDB good performance are:
client driver per language: native socket protocol for client/server interface (not REST)
use of memory mapped files for data storage
collection-oriented storage (objects from the same collection are stored contiguously)
update-in-place (not MVCC)
written in C++
Use Cases
It may be helpful to look at some particular problems and consider how we could solve them.
if we were building Lotus Notes, we would use Couch as its programmer versioning reconciliation/MVCC model fits perfectly. Any
problem where data is offline for hours then back online would fit this. In general, if we need several eventually consistent master-master
replica databases, geographically distributed, often offline, we would use Couch.
if we had very high performance requirements we would use Mongo. For example, web site user profile object storage and caching of
data from other sources.
if we were building a system with very critical transactions, such as bond trading, we would not use MongoDB for those transactions --
although we might in hybrid for other data elements of the system. For something like this we would likely choose a traditional RDBMS.
for a problem with very high update rates, we would use Mongo as it is good at that. For example, updating real time analytics counters
for a web sites (pages views, visits, etc.)
Generally, we find MongoDB to be a very good fit for building web infrastructure.
Licensing
If you are using a vanilla MongoDB server from either source or binary packages you have NO obligations. You can ignore the
rest of this page.
MongoDB Database
Free Software Foundation's GNU AGPL v3.0.
Commercial licenses are also available from 10gen.
Drivers:
mongodb.org "Supported Drivers": Apache License v2.0.
Third parties have created drivers too; licenses will vary there.
Documentation: Creative Commons.
From our blog post on the AGPL:
Our goal with using AGPL is to preserve the concept of copyleft with MongoDB. With traditional GPL, copyleft was associated
with the concept of distribution of software. The problem is that nowadays, distribution of software is rare: things tend to run in
the cloud. AGPL fixes this “loophole” in GPL by saying that if you use the software over a network, you are bound by the
copyleft. Other than that, the license is virtually the same as GPL v3.
Note however that it is never required that applications using mongo be published. The copyleft applies only to the mongod and
mongos database programs. This is why Mongo DB drivers are all licensed under an Apache license. Your application, even
though it talks to the database, is a separate program and “work”.
If you intend to modify the server and distribute or provide access to your modified version you are required to release the full source code for the
modified MongoDB server. To reiterate, you only need to provide the source for the MongoDB server and not your application (assuming you use
the provided interfaces rather than linking directly against the server).
A few example cases of when you'd be required to provide your changes to MongoDB to external users:
Case Required
Hosting company providing access MongoDB servers yes
Public-facing website using MongoDB for content yes
Internal use website using MongoDB no
Internal analysis of log files from a web site no
Regardless of whether you are required to release your changes we request that you do. The preferred way to do this is via a github fork. Then
we are likely to include your changes so everyone can benefit.
Windows
Windows Quick Links and Reference Center
Running MongoDB on Windows

See the Quickstart page for info on how to install and run the database for the first time.
Running as a Service
See the Windows Service page.
The MongoDB Server

Get pre-built binaries on the Downloads page. Binaries are available for both 32 bit and 64 bit Windows. MongoDB uses memory-mapped files for
data storage, so for servers managing more than 2GB of data you will definitely need the 64 bit version (and a 64 bit version of Windows).
Writing Apps
You can write apps in almost any programming language – see the Drivers page. In particular C#, .NET, PHP, C and C++ work just fine.
C Sharp Language Center

Building
We recommend using the pre-built binaries, but Mongo builds fine with Visual Studio 2008 and 2010. See the Building for Windows page.
Versions of Windows
We have successfully ran MongoDB (mongod etc.) on:
Windows Server 2008 R2 64 bit

Windows 7 (32 bit and 64 bit)
Windows XP
Vista
International Docs
Most documentation for MongoDB is currently written in English. We are looking for volunteers to contribute documentation in
other languages. If you're interested in contributing to documentation in another language please email roger at 10gen dot
com.
Language Homepages
Deutsch
Español
Français
[!hu.png! Magyar]
Italiano
Português
Books
Available for Pre-Order
The Definitive Guide to MongoDB: The NoSQL Database for Cloud and Desktop Computing
By Peter Membrey
MongoDB for Web Development

By Mitch Pirtle
MongoDB: The Definitive Guide

By Kristina Chodorow and Mike Dirolf
Coming Soon
MongoDB in Action
by Kyle Banker
First two chapters available through Manning Early Access Program
Doc Index
Space Index
0-9 ... 5 A ... 13 B ... 19 C ... 28 D ... 27

F ... 7 G ... 9 H ... 7 I ... 16 J ... 9
L ... 8 M ... 30 N ... 2 O ... 12 P ... 14
R ... 26 S ... 28 T ... 7 U ... 13 V ... 6
X ... 0 Y ... 0 Z ... 0 !@#$ ... 0
0-9 A
1.0 Changelist A Sample Configuration Session

Wrote MongoDB. See documentation following example uses two shards (one server
mongos process, all running on a single test se
1.1 Development Cycle
below, a python script for starting and configuri
1.2.x Release Notes machine ...
New Features More indexes per collection Faster index creation Map/Reduce Stored
About
JavaScript functions Configurable fsync time Several small features and fixes DB Upgrade
Required There are some changes that will require doing an upgrade ... About the local database
mongod}} reserves the database {{local}} for sp
1.4 Release Notes
that its contents are never replicated. Using the
We're pleased to announce the 1.4 release of MongoDB. 1.4 is a drop in replacement for 1.2.
may place end user application data in local, if
To upgrade you just need to shutdown mongod, then restart with the new binaries. (Users
to other servers. Put ...
upgrading from release 1.0 should review the 1.2 release notes 1.2.x ...
1.6 Release Notes
Adding a new node to an existing replica set is
MongoDB 1.6 is a dropin replacement for 1.4. To upgrade, simply shutdown {{mongod}} then
either have an empty data directory or a recent
restart with the new binaries. Sharding Sharding is now productionready, making MongoDB
set member. When we start the new node, we
horizontally scalable, with no single point of failure. A single instance of mongod can now be
upgraded to a distributed ... Adding an Arbiter
Arbiters are nodes in a replica set that only par
have a copy of the data and will never become
readable secondary). They are mainly useful fo
(e.g. if a set only has two members ...
Admin UIs
Several administrative user interfaces, or GUIs
Gourley's blog
http://blog.timgourley.com/post/453680012/tue
has a good summary of the tools. #Fang of Mo
http://github.com/sbellity/futon4mongo Mongo3
http://github.com/derailed/mongo3 #MongoHub
Admin Zone
Community AdminRelated Articles boxedice.co
deployment
http://blog.boxedice.com/2010/02/28/notesfrom
Survey of Admin UIs for MongoDB
http://blog.timgourley.com/post/453680012/tue
MongoDB Nagios Check http://tag1consulting.c
MongoDB Cacti Graphs http://tag1consulting.c
Advanced Queries
Introduction MongoDB offers a rich query envir
page lists some of those features. Queries in M
JSONstyle objects, very much like the docume
database. For example: // i.e., select ...
Aggregation
Mongo includes utility functions which provide
{{group by}} operations. More advanced
crafted using MapReduce http://www.mongodb
Count {{count()}} returns the number of objects
Amazon EC2
MongoDB runs well on Amazon EC2 http://aws
page includes some notes in this regard. Instan
most EC2 types including Linux and Windows.
64 ...
MongoDB has two primary components to the
is the mongod process which is the core datab
mongod may be used as a selfcontained syste
mysqld on a server ...
Articles
See also the User Feedback DOCS:User Feed
presentations, blog posts, and more. Best of th
Right Data Model?
http://blog.mongodb.org/post/142940558/whati
nonrelational databases) Why Schemaless is G
Atomic Operations
MongoDB supports atomic operations on single
does not support traditional locking and comple
reasons: First, in sharded environments, distrib
and slow. Mongo DB's goal is to be ligh
B C
Backups C Language Center

Several strategies exist for backing up MongoDB databases. A word of warning: it's not C Driver {}The MongoDB C Driver is the 10gen
safe to back up the mongod data files (by default in /data/db/) while the database is running It's written in pure C. The goal is to be super st
and writes are occurring; such a backup may turn out to be corrupt.&nbsp ... dependencies, and very embeddable anywhere
README http://github.com/mongodb/mongocd
Benchmarks
you've done a benchmark, we'd love to hear about it\! Let us know at kristina at 10gen C Sharp Language Center
dot com. March 9, 2010 Speed test between djangomongokit and postgresqlpsycopg2 C# Drivers mongodbcsharp driver
http://www.peterbe.com/plog/speedtestbetweendjangomongokitandpostgresqlpsycopg2 http://github.com/samus/mongodbcsharp/tree/m
benchmarks creating, editing, and deleting ... http://code.google.com/p/simplemongodb/ NoR
http://github.com/atheken/NoRM F# F# Examp
Books
Community Articles A List ...
Available for PreOrder By Peter Membrey By Mitch Pirtle By Kristina Chodorow and Mike Dirolf
Coming Soon !bankercover.jpg! MongoDB in Action http://manning.com/banker/ by Kyle C Tutorial
Banker First two chapters available through Manning Early Access Program document is an introduction to usage of the Mo
program. First, install Mongo \\ see the Quickst
Boost 1.41.0 Visual Studio 2010 Binary
wish to take a look at the Developer's Tour DO
OLD and was for the VS2010 BETA. See the new Boost and Windows page instead. The
guide for a language independent ...
following is a prebuilt boost http://www.boost.org/ binary (libraries) for Visual Studio 2010 beta
2. The MongoDB vcxproj files assume this package is unzipped under c:\Program ... C++ BSON Library
Overview The MongoDB C driver library includ
Boost and Windows
implements the BSON specification (see http://
Visual Studio 2010 Prebuilt from mongodb.org Click here
can be used standalone for object serialization
http://www.mongodb.org/pages/viewpageattachments.action?pageId=12157032 for a prebuilt
one is not using MongoDB at all. Include {{bson
boost library for Visual Studio 2010. 7zip http://www.7zip.org/ format. Building Yourself
Download the boost source ... C++ Language Center
C\ driver is available for communicating with th
BSON
database is written in C, the driver actually use
bsonspec.org http://www.bsonspec.org/ BSON is a bin-aryen-coded seri-al-iz-a-tion of JSONlike
this is the same driver that the database uses i
doc-u-ments. BSON is designed to be lightweight, traversable, and efficient. BSON, like JSON,
has been compiled successfully on Linux, OS X
supports the embedding of objects and arrays within other objects ...
C++ Tutorial
bsonspec.org
document is an introduction to usage of the Mo
Building program. First, install Mongo \\ see the Quickst
section provides instructions on setting up your environment to write Mongo drivers or other wish to take a look at the Developer's Tour DO
infrastructure code. For specific instructions, go to the document that corresponds to your guide for a language independent ...
setup. Note: see the Downloads DOCS:Downloads page for prebuilt binaries\! Subsections of
Caching
this section ...
Memory Mapped Storage Engine This is the cu
Building Boost MongoDB, and it uses memorymapped files fo
MongoDB uses the www.boost.org Boost C\ libraries. Windows See also the prebuilt libraries strategy, the operating system's virtual memory
http://www.mongodb.org/pages/viewpageattachments.action?pageId=12157032 page. By caching. This has several implications: T
default c:\boost\ is checked for the boost files. Include files should be under \boost\boost ...
Capped Collections
Building for FreeBSD Capped collections are fixed sized collections t
FreeBSD 8.0 and later, there is a mongodb port you can use. For FreeBSD <= 7.2: # Get the autoLRU ageout feature (age out is based on i
database source: http://www.github.com/mongodb/mongo. # Update your ports tree: $ sudo capped collections automatically, with high per
portsnap fetch && portsnap extract The packages that come by default on 7.2 ... order for the objects in the collection; this is ve
Building for Linux CentOS and Fedora Packages
General Instructions # Install Dependencies see platform specific below # get source git clone 10gen now publishes yuminstallable RPM pack
git://github.com/mongodb/mongo.git # pick a stable version unless doing true dev git tag l # x8664) and Fedora 11, 12, and 13 (x6464 only
Switch to a stable branch (unless ... revision in stable, unstable, and snapshot, ther
mongostable, mongostableserver, mongostabl
Building for OS X
for each ...
set up your OS X computer for MongoDB development: Upgrading to Snow Leopard If you
have installed Snow Leopard, the builds will be 64 bit \\ so if moving from a previous OS Checking Server Memory Usage
release, a bit more setup may be required ... Checking using DB Commands The serverStat
usage information. > db.serverStatus() {}add: h
would indicate a memory leak{} Checking via U
MongoDB server currently supports little endian Solaris operation. (Although most
memorymapped files; thus the memory ...
drivers not the database server work on both.) Community: Help us make this rough page
better please\! (And help us add support for big ... Clone Database
MongoDB includes commands for copying a da
another. // copy an entire database from one n
MongoDB can be compiled for Windows (32 and 64 bit) using Visual C. SCons
name on another server. omit <fromhostname>
http://www.scons.org/ is the make mechanism, although a .vcproj/.sln is also included in the
another on the same ...
project for convenience when using the Visual Studio IDE. There are several dependencies ...
Collections
MongoDB collections are essentially named gr
MongoDB uses SpiderMonkey http://www.mozilla.org/js/spidermonkey/ for serverside
think of them as roughly equivalent to relationa
Javascript execution. The mongod project requires a file js.lib when linking. This page details
MongoDB collection is a collection of BSON BS
how to build js.lib. Note: V8 http://code.google.com/p/v8/ Javascript support is under ...
documents are usually have the same structure
Building SpiderMonkey since MongoDB ...
Building the Mongo Shell on Windows Command Line Parameters
You can build the mongo shell with either scons or a Visual Studio 2010 project file. Scons MongoDB can be configured via command line
scons mongo Visual Studio 2010 Project File A VS2010 vcxproj file is availabe for building the based configuration. You can see the currently
shell. From the mongo directory open ... options by running the database with {{\h \ \hel
./mongod help Information on usage ...
MongoDB can be compiled for Windows (32 and 64 bit) using Visual C. SCons Commands
http://www.scons.org/ is the make mechanism, although a solution file is also included in the Introduction The Mongo database has a conce
project for convenience when using the Visual Studio IDE. There are several dependencies Database commands are ways to ask the data
exist ... operations, or to request information about its c
Database Commands A command is sent to th
special ...
MongoDB can be compiled for Windows (32 and 64 bit) using Visual C. SCons
http://www.scons.org/ is the make mechanism, although a solution file is also included in the Community
project for convenience when using the Visual Studio IDE. There are several dependencies General Community Resources User Mailing L
exist ... http://groups.google.com/group/mongodbuser
using, configuring, and running MongoDB and
The list is open to everyone IRC chat irc://irc.fr
Community Info
We are getting a lot of questions "how are mon
different?" It's a good question: both are
with schemaless JSONstyle object data storag
place \\ we are big believers that databases are
Configuring Sharding
Introduction This document describes the steps
sharding cluster. A sharding cluster requires, a
Two or more shards. 2. At least one config serv
process. For testing ...
Connecting
C\ driver includes several classes for managing
class DBClientInterface. In general, you will wa
DBClientConnection object, or a DBClientPaire
DBClientConnection is our normal connection c
MongoDB database ...
Connecting Drivers to Replica Sets
Ideally a MongoDB driver can connect to a clus
replica set DOCS:Replica Sets, and automatica
which to communicate. Failover should
general steps are: # The user ...
Connecting to Replica Sets from Clients
Most drivers have been updated to provide wa
general, this is very similar to how the drivers s
pair. Instead of taking a pair of hostnames, the
comma separated list of host ...
Connections
MongoDB is a database server: it runs in the fo
waits for connections from the user. Thus, whe
see something like: /$ ./mongod # # some logg
waiting ...
Qualified volunteers are welcome to assist in e
Contact us for more information
Contributing to the Perl Driver
easiest way to contribute is to file bugs and fea
http://jira.mongodb.org/browse/PERL. If you wo
read on... Finding Something to Help With Fixin
on Jira http ...
Contributors
10gen Contributor Agreement http://www.10ge
Conventions for Mongo Drivers
Interface Conventions It is desirable to keep dr
possible. Of course, idioms vary by language, a
appropriate. However, when the idiom is the sa
consistent across drivers is desirable. Termino
Creating and Deleting Indexes
Cursors
D E
Data Center Awareness Emacs tips for MongoDB work

1.6.0 build of replica sets does not support much in terms of data center awareness. However You can edit confluence directly from emacs: F
additional functionality will be added in the future. Below are some suggestions configurations on http://code.google.com/p/confluenceel/ Cha
which work today. Primary plus DR site Use one site ... sample setup to http://mongodb.onconfluence.
change the default space to DOCS ...
DRAFT TO BE COMPLETED. This guide provides instructions for using MongoDB batch data Error Codes
processing oriented features including map/reduce DOCS:MapReduce. By "data processing", Error Code \\ Description \\ Comments \\ 10003
we generally mean operations performed on large sets of data, rather than small ... grow 11000 duplicate key error \\ \id values mu
11001 duplicate key on update \\ 12000 idxNo
MongoDB (BSON) Data Types Mongo uses special data types in addition to the basic JSON Error Handling in Mongo Drivers
types of string, integer, boolean, double, null, array, and object. These types include date, an error occurs on a query (or getMore operati
object id Object IDs, binary data, regular ... object instead of user data. The error object ha
the reserved key {{$err}}. For example: The {{$
Database Internals
section provides information for developers who want to write drivers or tools for MongoDB, \ Events
contribute code to the MongoDB codebase itself, and for those who are just curious how it Follow us on Facebook
works internally. Subsections of this section http://www.facebook.com/#!/pages/MongoDB/3
http://www.twitter.com/mongodb to get all of th
Database Profiler
Conferences 10gen Weekly "Office Hours" Sub
Mongo includes a profiling tool to analyze the performance of database operations. See also
http://www.10gen.com/talkproposal to present
the currentOp DOCS:Viewing and Terminating Current Operation command. Enabling Profiling
To enable profiling, from the {{mongo}} shell invoke: > db.setProfilingLevel(2); > Excessive Disk Space
db.getProfilingLevel() 2 Profiling ... You may notice that for a given set of data the
are larger than the data set inserted into the da
Database References
reasons for this. Preallocation Each datafile is
MongoDB is nonrelational (no joins), references ("foreign keys") between documents are
size. (This is done to prevent file system
generally resolved clientside by additional queries to the server. Two conventions are common
for references in MongoDB: first simple manual references, and second, the DBRef standard,
which many drivers support ...
Databases
Each MongoDB server can support multiple databases. Each database is independent, and
the data for each database is stored separately, for security and ease of management. A
database consists of one or more collections, the documents (objects) in those collections, and
an optional set ...
page lists common DBAclass operations that one might perform from the MongoDB shell
DOCS:mongo The Interactive Shell. Note one may also create .js scripts to run in the shell for
administrative purposes. help show help show ...
dbshell Reference
Command Line {{\help}} Show command line options {{\nodb}} Start without a db, you can
connect later with {{new Mongo()}} or {{connect()}} {{\shell}} After running a .js file from the
command line, stay in the shell rather than ...
Design Overview
Developer FAQ
Also check out Markus Gattol's excellent FAQ on his website
http://sunoano.name/ws/publicxhtml/mongodb.html. What's a "namespace"? MongoDB stores
BSON objects in collections. The concatenation of the database name and the collection name
(with a period in between) is called a namespace ...
Developer Zone
Tutorial Shell mongo The Interactive Shell Manual Databases Collections Indexes Data Types
and Conventions GridFS Inserting Updating Querying Removing Optimization Developer FAQ
If you have a comment or question about anything, please contact us through ...
Diagnostic Tools
Django and MongoDB
Generally, with MongoDB we are not building queries from strings, so traditional SQL Injection
http://en.wikipedia.org/wiki/SQLinjection attacks are not a problem. More details and some
nuances are covered below. MongoDB queries are represented as BSON objects. Typically
the programming ...
Doc Index
Document-Oriented Datastore
Documentation
Dot Notation
Dot Notation
MongoDB is designed for store JSONstyle objects. The database understands the
structure of these objects and can reach into them to evaluate query expressions. Let's
suppose we have some objects of the form: > db.persons.findOne() { name: "Joe", address: ,
likes: 'scuba', 'math ...
Downloads
See also #Packages. Version OS X 32 bit OS X 64 bit Linux 32 bit Linux 64 bit Windows 32 bit
Windows 64bit Solaris i86pc Solaris 64 Source Date Change ...
Driver and Integration Center
Driver Syntax Table
wiki generally gives examples in JavaScript, so this chart can be used to convert those
examples to any language. JavaScript Python PHP Ruby Java C\ Perl \ \ \\ {{array()}} \\
{{BasicDBList}} BSONObj \ \ {} {} {{new stdClass}} {} {{BasicDBObject}} BSONObj {} \ \
{{array('x ...
Driver Testing Tools
Object IDs driverOIDTest for testing toString > db.runCommand
Drivers
MongoDB currently has client support for the following programming languages: mongodb.org
Supported C C Language Center C\ C Language Center Java Java Language Center
Javascript Javascript Language Center Perl Perl Language Center ...
Single Server Durability The v1.8 release of MongoDB will have single server durability. You
can follow the Jira here : http://jira.mongodb.org/browse/SERVER980. We recommend using
replication to keep copies of data for now and likely forever as a single server ...
F G
FAQ Geospatial Indexing

FAQ answers basic questions for new evaluators of MongoDB. See also: Developer FAQ v1.3.3\ MongoDB supports twodimensional geo
DOCS:Sharding FAQ MongoDB Intro FAQ What kind of database is the Mongo database? with locationbased queries in mind, such as "fin
MongoDB is an documentoriented DBMS. Think of it as MySQL but JSON (actually ... location." It can also efficiently filter on addition
closest N museums to my location ...
Feature Checklist for Mongo Drivers
Functionality Checklist This section lists tasks the driver author might handle. Essential BSON getLastError
serialization/deserialization Basic operations: {{query}}, {{save}}, {{update}}, {{remove}}, Most drivers, and the db shell, support a getlas
{{ensureIndex}}, {{findOne}}, {{limit}}, {{sort}} Fetch more data from a cursor when necessary ... one check the error code on the last operation.
DOCS:Commands, as well as queries, have a
File Based Configuration
getlasterror is primarily useful for write ...
addition to accepting command line parameters, MongoDB can also be configured using a
configuration file. A configuration file to use can be specified using the {{\f}} or {{\\config}} Getting Started
command line options. The following example configuration file demonstrates the syntax to use
Getting the Software
...
Placeholder $$$ TODO
Git Commit Rules
Find and Modify (or Remove) v1.3.0 and higher MongoDB 1.3\ supports a "find, modify, and
commit messages should have the case in the
return" command. This command can be used to atomically modify a document (at most
messages should be descriptive enough that a
one) and return it. Note that, by default, the document returned will not include the
commits should only include 1 thought
modifications made on the update ...
GridFS
Frequently Asked Questions - Ruby
GridFS is a specification for storing large files i
list of frequently asked questions about using Ruby with MongoDB. If you have a question
supported driver implement the GridFS spec D
you'd like to have answered here, please add it in the comments. Can I run \insert command
Rationale The database supports native storag
name here\ from the Ruby driver? Yes ...
DOCS:BSON objects. However ...
fsync Command
GridFS in Ruby
Version 1.3.1 and higher The fsync command allows us to flush all pending writes to
GridFS, which stands for "Grid File Store," is a
datafiles. More importantly, it also provides a lock option that makes backups easier.
in MongoDB. It works by dividing a file into man
fsync Command The fsync command forces the database to flush all ...
each of those chunks as a separate document.
Full Text Search in Mongo to achieve this: one collection stores ...
Introduction Mongo provides some functionality that is useful for text search and tagging.
GridFS Specification
Multikeys (Indexing Values in an Array) The Mongo multikey feature can automatically index
arrays of values. Tagging is a good example of where this feature is useful. Suppose you ... Introduction GridFS is a storage specification fo
works by splitting large object into small chunk
chunk is stored as a separate document in a {{
about the file, including the filename, content ty
GridFS Tools
File Tools {{mongofiles}} is a tool for manipulat
from the command line. Example: $ ./mongofile
./mongofiles put libmongoclient.a connected to
list connected to: 127.0.0.1 libmongoclient.a 12
H I
Halted Replication Implementing Authentication in a Driver

instructions are for master/slave replication. For replica sets, see DOCS:Resyncing a Very current version of Mongo supports only very ba
Stale Replica Set Member instead. If you're running mongod with masterslave replication authenticates a username and password in the
DOCS:Master Slave, there are certain scenarios where the slave will halt replication because database. Once authenticated, the user
... the database in question. The {{admin}} databa
Home Import Export Tools
Events Follow us on Facebook http://www.facebook.com/#!/pages/MongoDB/397955415556 you just want to do DOCS:Clone Database from
and Twitter http://www.twitter.com/mongodb to get all of the latest updates\! Coming Soon: need these tools. These tool just work with the
Mongo Boston http://bit.ly/mongoboston \ September 20 Mongo Berlin http://bit.ly/mongoberlin collection); they do not save, or load, the metad
... (capped ...
Hosting Center Index-Related Commands
CloudStyle provides cloudstyle hosted MongoDB instances is currently in private beta Create Index {{ensureIndex()}} is the helper fun
Dedicated Servers offers preconfigured, dedicated MongoDB servers supports MongoDB on creates an index by adding its info to the {{syst
its private cloud. VPS 's is an excellent choice for . offers instant configuration and deployment db.myCollection.ensureIndex(<keypattern>); >
of MongoDB ... db.system.indexes.insert( ); Note: Once you've
subsequent document inserts for the given coll
mongos For sharded DOCS:Sharding environments, mongos can perform any number of Indexes
operations concurrently. This results in downstream operations to mongod instances. Indexes enhance query performance, often dra
Execution of operations at each mongod is independent; that is, one mongod does not block about the kinds of queries your application will
another. mongod The original mongod ... relevant indexes. Once that's done, actually cre
relatively easy. Indexes in MongoDB ...
How to do Snapshotted Queries in the Mongo Database
document refers to query snapshots. For backup snapshots of the database's datafiles, see the Indexes in Mongo
fsync lock page fsync Command. MongoDB does not support full pointintime snapshotting.
However, some functionality is available which is detailed below. Cursors A MongoDB query ...
We get a lot of questions about indexing. Here
HowTo of these. There are a couple of points to keep i
MongoDB work quite similarly to indexes in My
Http Interface
techniques for building efficient indexes ...
REST Interfaces Sleepy Mongoose (Python) Sleepy Mongoose
http://www.snailinaturtleneck.com/blog/2010/02/22/sleepymongooseamongodbrestinterface/ is Indexing as a Background Operation
a full featured REST interface for MongoDB which is available as a separate project. MongoDB Slaves and replica secondaries build all indexe
Rest (Node.js) MongoDB Rest http://github.com/tdegrunt/mongodbrest is an alpha ... releases (including the latest). Thus even when
primary, the slave/secondary will be unavailabl
index builds there. By default the {{ensureIndex
Inserting
When we insert data into MongoDB, that data w
Documents are data structure analogous to JS
Ruby hashes, to take just a few examples. Her
documentorientation and describe how to inser
Internal Commands
Most commands Commands have helper funct
{{$cmd.findOne()}} syntax. These are primarily
db.$cmd.findOne( ) > db.$cmd.findOne( ) { "up
, "ok" : 1 } > admindb.$cmd.findOne( ) // close a
request ...
Internals
Cursors Tailable Cursors See p/db/dbclient.h fo
side, to support tailable cursors. Set OptionCur
{{queryOptions int}} field to indicate you want a
results when ...
International Docs
Most documentation for MongoDB is currently
for volunteers to contribute documentation in o
interested in contributing to documentation in a
roger at 10gen dot com. Language Homepage
International Documentation
MongoDB supports UTF8 for strings in stored o
(Specifically, BSON DOCS:BSON strings are U
programming language convert from the langu
UTF8 when serializing and deserializing BSON
Interop Demo (Product Comparisons)
Interop 2009 MongoDB Demo Code:
http://github.com/mdirolf/simplemessagingserv
Introduction - How Mongo Works
J K
Java - Saving Objects Using DBObject Kernel class rules

Java driver provides a DBObject interface to save custom objects to the database. For new classes By default, use {{explicit}} constru
example, suppose one had a class called Tweet that they wanted to save: public class Tweet boost::noncopyable unless you have implemen
implements DBObject Then you can say: Tweet myTweet = new Tweet ... assignment. inheritance No multiple inheritance
the FIRST virtual function to a class as you the
Java Driver Concurrency
Java MongoDB driver is thread safe. If you are using in a web serving environment, for Kernel code style
example, you should create a single Mongo instance, and you can use it in every case Use camelCase for most {{varNames}} {{\
request. The Mongo object maintains an internal pool of connections ... See utils/mongoutils/str.h bson/stringdata.h Us
strstr(). Use {{<< 'c'}} not {{<< "c"}}. Use {{str\0\
Java Language Center
brackets if ( 0 ) class ...
Basics Tutorial Java Tutorial API Documentation http://api.mongodb.org/java/index.html
Downloads http://github.com/mongodb/mongojavadriver/downloads Specific Topics Kernel concurrency rules
Concurrency Java Driver Concurrency Saving Objects Java Saving Objects Using DBObject All concurrency code must be placed under {{u
Data ... several helper libraries there. Several rules are
you think there is a real need let's have the gro
Java Tutorial
...
Introduction This page is a brief overview of working with the MongoDB Java Driver. For more
information about the Java API, please refer to the online API Documentation for Java Driver Kernel exception architecture
http://api.mongodb.org/java/index.html A Quick Tour Using the Java driver is very ...
Java Types string manipulation, use the {{util/mongoutils/st
Object Ids {{com.mongodb.ObjectId}} MongoUtils has its own namespace. Its code h
http://api.mongodb.org/java/0.11/com/mongodb/ObjectId.html is used to autogenerate unique not database specific, rather, true utilities # are
ids. ObjectId id = new ObjectId(); ObjectId copy = new ObjectId(id); Regular Expressions The boost headers, but not libs ...
Java driver uses {{java.util.regex.Pattern}} http://java.sun.com ...
Javascript Language Center
MongoDB can be Used by clients written in Javascript; Uses Javascript internally serverside
for certain options such as map/reduce; Has a shell DOCS:mongo The Interactive Shell that is
based on Javascript for administrative purposes. node.JS and V8 See the node.JS page.
node.JS ...
Job Board
Redirecting
Joyent
prebuilt DOCS:Downloads MongoDB Solaris 64 binaries work with Joyent accelerators.
Some newer gcc libraries are required to run \\ see sample setup session below. $ # assuming
a 64 bit accelerator $ /usr/bin/isainfo kv ...
JVM Languages
moved to Java Language Center
L M
Language Support Manual

MongoDB manual. Except where otherw
Last Error Commands
JavaScript for use with the mongo shell mongo
Since MongoDB doesn't wait for a response by default when writing to the database, a couple
There is a table Driver Syntax Table available g
commands exist for ensuring that these operations have succeeded. These commands can be
each of the drivers
invoked automatically with many of the drivers when saving and updating in "safe" mode. But
what's really happening ... MapReduce
Map/reduce in MongoDB is useful for batch ma
Legal Key Names
aggregation operations. It is similar in spirit to u
Key names in inserted documents are limited as follows: The '$' character must not be the first
all input coming from a collection and output go
character in the key name. The '.' character must not appear anywhere in the key name
situation where you would have used ...
Licensing
Master Master Replication
you are using a vanilla MongoDB server from either source or binary packages you have NO
Mongo does not support full mastermaster repl
obligations. You can ignore the rest of this page. MongoDB Database Free Software
certain restricted use cases mastermaster can
Foundation's GNU AGPL v3.0 http://www.fsf.org/licensing/licenses/agpl3.0 ...
speaking, we don't recommend the using a ma
List of Database Commands MongoDB. Mastermaster usages is eventually
iframe src ="http://api.mongodb.org/internal/current/commands.html" width="100%"
Master Slave
height="1000px" frameborder="0"> <a
Configuration and Setup To configure an instan
href="http://api.mongodb.org/internal/current/commands.html">List of MongoDB
database in a masterslave configuration, you'll
Commands</a> </iframe> See the Commands page for details on how to invoke a command
database, one in master mode, and the other in
...
examples explicitly specify the location of the d
Locking min and max Query Specifiers
min()}} and {{max()}} functions may be used in
Locking in Mongo
constrain query matches to those having index
Logging keys specified. The {{min()}} and {{max()}} func
MongoDB outputs some important information to stdout while its running. There are a number in conjunction. The index to be used may ...
of things you can do to control this Command Line Options \quiet less verbose output \v more
mongo - The Interactive Shell
verbose output. use more v's (such as vvvvvv ...
Introduction The MongoDB distribution Downlo
MongoDB interactive shell. This utility is a Java
issue commands to MongoDB from the comma
extended SpiderMonkey
https://developer.mozilla.org/En/SpiderMonkey
...
Mongo Administration Guide
Mongo Concepts and Terminology
Mongo Database Administration
Mongo Developers' Guide
Mongo Documentation Style Guide
page provides information for everyone adding
Confluence. It covers: #General Notes on Writi
markup for specific situations Some general no
Notes on Writing Style Voice Active ...
Mongo Driver Requirements
highlevel list of features that a driver for Mongo
group those features by priority. This list should
probably used more for inspiration than as law
way to learn about ...
Mongo Extended JSON
Mongo's REST interface supports storage and
Special representations are used for BSON typ
JSON mappings, and multiple representations
The REST interface supports three different mo
Mongo Metadata
dbname>.system. namespaces in MongoDB ar
system information. System collections
lists all namespaces. {{system.indexes}} lists a
/ index metadata exists in the database.ns files
...
Queries in MongoDB are expressed as JSON (
object as the equivalent of a SQL "WHERE" cla
select from users where x=3 and y='abc' order
server ...
Mongo Usage Basics
Mongo Wire Protocol
Introduction The Mongo Wire Protocol is a sim
requestresponse style protocol. Clients commu
through a regular TCP/IP socket. The default p
configurable and will vary. Clients should conn
TCP/IP ...
Please list applications that leverage MongoDB
for your application, we'd love to list you here\!&
[email protected]. See Also Production De
using MongoDB Hosting Center Applications ..
MongoDB - A Developer's Tour
MongoDB Commercial Services Providers
Note: if you provide consultative or support ser
be listed here, just let us know. Support 10gen
offers commercial MongoDB support services.
See the MongoDB ...
tutorial discusses the development of a web ap
MongoDB. MongoMapper will serve as
goal is to provide some insight into the design
MongoDB. To that end, we'll be constru
...
MongoDB Language Support
MongoDB, CouchDB, MySQL Compare Grid
pending... CouchDB \\ MongoDB \\ MySQL \\ D
(JSON http://www.json.org/) \\ DocumentOrien
http://blog.mongodb.org/post/114440717/bson)
string,number,boolean,array,object string, int, d
object ...
mongosniff
Unix releases of MongoDB include a utility call
MongoDB what tcpdump is to TCP/IP; that is, f
situations. The tool is quite useful for authors o
Usage: mongosniff help forward ...
mongostat
Use the mongostat utility to quickly view statist
instance. Starting in 1.3.3; this output is from 1
delete/s getmore/s command/s flushes/s mapp
miss % q t r ...
Monitoring
Admin UIs Query Profiler Use the Database Pr
Http Console The mongod process includes a
http://localhost:28017/. See the Http Interface d
db.serverStatus() from mongo ...
Moving Chunks
inc version try to set on from if set is successfu
finish transfer commit result max version for a s
this poses slight problem when moving last chu
Multikeys
MongoDB provides an interesting "multikey" fe
arrays of an object's values. A good example is
article tagged with some category names: $ db
db.articles.find() We can ...
N O
node.JS Object IDs

Node.js is used to write eventdriven, scalable network programs in serverside JavaScript. It is Documents in MongoDB are required to have a
similar in purpose to Twisted, EventMachine, etc. It runs on Google's V8. Web Frameworks identifies them. Document IDs: \id Every Mong
ExpressJS http://expressjs.com Mature web framework with MongoDB session support. 3rd its first attribute. This value usually a BS
Party ...
Object Mappers for Ruby and MongoDB
Notes on Pooling for Mongo Drivers Although it's possible to use the Ruby driver by
Note that with the db write operations can be sent asynchronously or synchronously (the latter validations, associations, and many of the othe
indicating a getlasterror request after the write). When asynchronous, one must be careful to ActiveRecord. Here, then, is a list of the most p
continue using the same connection (socket). This ensures that the next operation will not for working with Ruby and MongoDB. Recomm
begin until after ...
Old Pages
Older Downloads
One Slave Two Masters
mms.png align=center! This document shows a
slave pulling data from two different masters. D
it is better, simpler, and generally recommende
slave}} processes on a single ...
Online API Documentation
MongoDB API and driver documentation is ava
Java Driver API Documentation http://api.mong
Documentation http://api.mongodb.org/cplusplu
Documentation http://api.mongodb.org/python
Optimization
Optimizing A Simple Example This section des
optimizing database performance. Let's consid
is to display the front page of a blog we wish to
recent posts. Let's ...
Optimizing Mongo Performance
MongoDB records have a certain amount of ov
DOCS:BSON document) in a collection.
insignificant, but if your objects are tiny (just a f
fields) it would not be. Below ...
Query objects in Mongo by default AND expres
1.5.3 MongoDB did not include an "$or" operat
there are ways to express such queries. $in Th
value in ..." expression. For expressions
Overview - The MongoDB Interactive Shell
Starting the Shell The interactive shell is includ
distribution. To start the shell, go into the root d
type ./bin/mongo It might be useful to add {{{}m
{{PATH}} so you can just type {{mongo ...
Overview - Writing Drivers and Tools
section contains information for developers tha
protocols of Mongo people who are writing driv
Documents of particular interest : BSON http://
BSON binary document format. Fundamental t
software ...
P Q
Pairing Internals Queries and Cursors

Policy for reconciling divergent oplogs In a paired environment, a situation may arise in which Queries to MongoDB return a cursor, which ca
each member of a pair has logged operations as master that have not been applied to the The exact way to query will vary with language
other server. In such a situation, the following procedure will be used to ensure consistency queries from the MongoDB shell DOCS:mongo
between the two ... {{mongo}} process). The shell ...
Perl Language Center Query Optimizer
Installing Start a MongoDB server instance ({{mongod}}) before installing so that the tests will MongoDB query optimizer generates query pla
pass. The {{mongod}} cannot be running as a slave for the tests to pass. Some tests may be client. These plans are executed to retu
skipped, depending on the version of the database you are running. CPAN $ sudo cpan supports ad hoc queries much like say, MySQL
MongoDB ... interesting approach ...
Perl Tutorial Querying
One of MongoDB's best capabilities is its supp
Philosophy
Systems that support dynamic queries don't re
Design Philosophy !featuresPerformace.png align=right! Databases are specializing the "one
data; users can find data using any criteria. Fo
size fits all" approach no longer applies. By reducing transactional semantics the db provides,
one can still solve an interesting set of problems where performance is very ... Quickstart
Quickstart OS X Quickstart Unix Quickstart Win
PHP - Storing Files and Big Data
go to http://try.mongodb.org/. See Also DOCS:
PHP Language Center
Quickstart OS X
Using MongoDB in PHP To access MongoDB from PHP you will need: The MongoDB server
Install MongoDB The easiest way to install Mon
running the server is the "mongo{}d" file, not the "mongo" client (note the "d" at the end) The
manager or the prebuilt binaries: Package man
MongoDB PHP driver installed Installing the PHP Driver \NIX Run ...
http://mxcl.github.com/homebrew/ package ma
PHP Libraries, Frameworks, and Tools mongodb If you use MacPorts http://www.macp
PHP community has created a huge number of libraries to make working with MongoDB easier
Quickstart Unix
and integrate it with existing frameworks. CakePHP MongoDB datasource
Install MongoDB Note: If you are running an ol
http://github.com/ichikaway/mongoDBDatasource/downloads for CakePHP. There's also an
database doesn't start, or gives a floating point
introductory blog post http://markstory.com ...
version on the Downloads page instead of the
Product Comparisons managers Ubuntu and Debian ...
Production Deployments Quickstart Windows
you're using MongoDB in production, we'd love to list you here\! Email Download The easiest (and recommended) wa
[email protected]. <DIV mcestyle="textalign:center;margin:0" style="margin: 0pt; textalign: prebuilt binaries. 32bit binaries Download Dow
center;">Company</DIV> Use Case !shutterfly.jpg align=center ... The "Production" build is recommended. 64bit
and extract the 64bit .zip. Note: 64bit is recomm
Production Notes
Architecture Production Options Master Slave 1 master, N slaves failover is handled manually
Version 1.6: Replica Sets N servers, 1 is always primary, autofailover, autorecovery Backups
Import Export Tools Recommended Unix System ...
Project Ideas
you're interested in getting involved in the MongoDB community (or the open source
community in general) a great way to do so is by starting or contributing to a MongoDB related
project. Here we've listed some project ideas for you to get started on. For some of these ideas
...
PyMongo and mod_wsgi
Python Language Center
Python Tutorial
R S
Rails - Getting Started Schema Design

Using Rails 3? See Rails 3 Getting Started This tutorial describes how to set up a simple Rails Introduction With Mongo, you do less "normaliz
application with MongoDB, using MongoMapper as an object mapper. We assume you're using designing a relational schema because there a
Rails versions prior to 3.0 ... Generally, you will want one database collectio
objects. You do not want ...
Rails 3 - Getting Started
difficult to use MongoDB with Rails 3. Most of it comes down to making sure that you're not scons
loading ActiveRecord and understanding how to use Bundler Use scons to build MongoDB, and related utilit
http://github.com/carlhuda/bundler/blob/master/README.markdown, the new Ruby SConstruct file for details. Run {{scons \help}} t
dependency manager. Install the Rails ... {{scons <target>}}. {{scons .}} {{scons all}} {{sco
{{scons ...
Recommended Production Architectures
Searching and Retrieving
Reconfiguring a replica set when members are down
One may modify a set when some members are down if a majority is established. In this case, Security and Authentication
simply send the reconfig command to the current primary. If there is no primary (and this Running Without Security (Trusted Environmen
condition is not transient), no majority is available. Reconfiguring a minority partition would be recommended. One valid way to run the Mong
dangerous as two sides of a network ... environment, with no security and authenticatio
say, memcached). Of course, in such a
Removing
Removing Objects from a Collection To remove objects from a collection, use the {{remove()}} Server-side Code Execution
function in the mongo shell mongo The Interactive Shell. (Other drivers offer a similar function, Mongo supports the execution of code inside th
but may call the function "delete". Please check your driver's documentation ... Clauses and Functions in Queries In addition to
specification for {{find()}} operations, you can a
Replica Pairs
string containing a SQLstyle WHERE predicate
Setup of Replica Pairs Replica Sets will soon replace replica pairs. If you are just now
setting up an instance, you may want to wait for that and use master/slave replication in the Server-Side Processing
meantime. Mongo supports a concept of replica ...
Shard Ownership
Replica Pairs in Ruby shard ownership we mean which server owns a
Replica Sets will replace replica pairs in MongoDB 1.6. If you are just now setting up an draft/thoughts will change: Contract the master
instance, you may want to wait for that and use master/slave replication in the meantime. Here information is in the config database mongos in
follow a few considerations for those using ... which server owns a shard ...
Replica Set Admin UI Sharding
mongod}} process includes a simple administrative UI for checking the status of a replica set. MongoDB scales horizontally via an autoshard
To use, first enable {{\rest}} from the {{mongod}} command line. The rest port is the db port Scaling out to thousands of nodes Easy additio
plus 1000 (thus, the default is 28017). Be sure this port is secure ... balancing for changes in load and data distribu
Automatic failover Sharding ...
Replica Set Commands
Shell Helpers rs.help() show help rs.status() rs.initiate() initiate with default settings Sharding Administration
rs.initiate(cfg) rs.add(hostportstr) add a new member to the set rs.add(membercfgobj) add a Here we present a list of useful commands for
new member to the set rs.addArb(hostportstr) add a new member which ... sharding cluster. To set up a sharding cluster,
configuration DOCS:Configuring Sharding. Ide
Replica Set Configuration
we're speaking to a mongos process ...
Command Line Each {{mongod}} participating in the set should have a {{\replSet}} parameter
on its command line. The syntax is mongod replSet setname {}setname is the logical name of Sharding and Failover
the set. The rest command line parameter is also recommended when using replica ... properlyconfigured MongoDB shard cluster will
This document describes the various potential
Replica Set Design Concepts
within a shard cluster, and how failure is handle
1. A write is only truly committed once it has replicated to a majority of members of the set. For
a {{mongos}} routing process. One {{mongos ..
important writes, the client should request acknowledgement of this with a {{getLastError(\ )}}
DOCS:Verifying Propagation of Writes with getLastError call. 2. Writes which are committed at Sharding Config Schema
the primary of the set ... Sharding configuration schema. This lives in th
version This is a singleton that contains the cur
Replica Set FAQ
db.getCollection("version").findOne() settings K
How long does failover take? Failover thresholds are configurable. With the defaults, it may
options (chunkSize) > db.settings.find() shards
take 2030 seconds for the primary to be declared down by the other members and a new
shards ...
primary elected. During this window of time, the cluster is down for "primary" operations that is,
writes and strong ... Sharding Design
concepts config database \ the top level databa
Replica Set Internals
servers and where things live. shard. this can b
Design Concepts Check out the Replica Set Design Concepts for some of the core concepts
replica pair. database \ one top level namespac
underlying MongoDB Replica Sets. Configuration Command Line We specify \replSet
or not chunk \ a region ...
setname/seedhostnamelist on the command line. seedhostnamelist is a (partial) list of some
members ... Sharding FAQ
How does sharding work with replication? Each
partitioned data. The shard could consist of a s
tutorial will guide you through the basic configuration of a replica set on a single machine. If
replicas. Typically in production one would use
you're attempting to deploy replica sets in production, be sure to read the comprehensive
shard. Where ...
replica set documentation Replica Sets. Also, do keep in mind that replica sets ...
Sharding Internals
Replica Sets
section includes internal implementation details
v1.6.0 and higher. Replica Sets are MongoDB's new method for replication. They are an
sharding. See also the main sharding do
elaboration on the existing master/slave replication DOCS:Replication, adding automatic
Note: some internals docs could be out of date
failover and automatic recovery of member nodes. Replica Sets are "Replica Pairs version 2"
we ...
and are available ...
Sharding Introduction
Replica Sets in Ruby
MongoDB supports an automated sharding arc
Here follow a few considerations for those using the Ruby driver Ruby Tutorial with MongoDB
scaling across multiple nodes. For applications
and replica sets DOCS:Replica Sets. Setup First, make sure that you've configured and
single database server, MongoDB can convert
initialized a replica set. Connecting to a replica set from the Ruby ...
automatically managing failover and balancing
Replica Sets Limits ...
v1.6 Authentication mode not supported. JIRA http://jira.mongodb.org/browse/SERVER1567
Sharding Limits
Limits on config changes to sets at first. Especially when a lot of set members are down. Sharding Release 1 (MongoDB v1.6.0) Differen
Map/reduce writes new collections to the server. Because of this, for now it may only ... Configurations Sharding must be ran in trusted
security DOCS:Security and Authentication. Sh
Replica Sets Troubleshooting
current version. All (nonmulti)updates, upserts
can't get local.system.replset config from self or any seed (EMPTYCONFIG) Set needs to be
initiated. Run {{rs.initiate()}} from the shell. If the set is already initiated and this is a new node, Sharding Use Cases
verify it is present in the replica set's configuration and there are no typos in the host names: > What specific use cases do we want to addres
// send ... techniques) that are challenging to scale? List
(e.g., youtube) (also, GridFS scaleup) seems s
Replication
for related videos ...
MongoDB supports asynchronous replication of data between servers for failover and
redundancy. Only one server (in the set/shard) is active for writes (the primary, or master) at a Slide Gallery
given time. With a single active master at any point in time, strong consistency semantics are Click here http://www.mongodb.org/pages/view
available ... to visit our full listing of videos & slides from rec
Introduction to MongoDB User Experience <div
style="width:422px;margin:auto;"><object style
master mongod instance, the {{local}} database will contain a collection, {{oplog.$main}}, which
stores a highlevel transaction log. The transaction log essentially describes all actions Smoke Tests
performed by the user, such as "insert this object into this collection." Note that the oplog is not smoke.py lets you run a subsets of the tests in
a lowlevel redo log ... starts up an instance of mongod, runs the tests
again. For the moment, smoke.py must
Replication Oplog Length
of a mongo source ...
Replication uses an operation log ("oplog") to store write operations. These operations replay
asynchronously on other nodes. The length of the oplog is important if a secondary is down. Sorting and Natural Order
The larger the log, the longer the secondary can be down and still recover. Once the oplog has Natural order" is defined as the database's nat
... collection. When executing a {{find()}} with no p
objects in forward natural order. For standard t
particularly useful because, although the order
Error RS102 MongoDB writes operations to an oplog. For replica sets this data is stored
in collection local.oplog.rs. This is a capped collection and wraps when full Source Code
"RRD"style. Thus, it is important that the oplog collection is large enough to buffer ... All source for MongoDB, it's drivers, and tools i
Github http://github.com/mongodb. Mongo Dat
http://github.com/mongodb/mongo/tree/master
default on a find operation, the entire object is returned. However we may also request that
http://github.com/mongodb/mongopythondriver
only certain fields be returned. This is somewhat analogous to the list of column specifiers in a
SQL SELECT statement (projection). Regardless of what field specifiers are included ... Spec, Notes and Suggestions for Mongo Drivers
Assume that the BSON DOCS:BSON objects r
Ruby External Resources
be up to 4MB. This size may change over time
number of good resources appearing all over the web for learning about MongoDB and Ruby.
object. We recommend you test your driver wit
A useful selection is listed below. If you know of others, do let us know. Screencasts
Introduction to MongoDB Part I http://www.teachmetocode.com/screencasts ... Splitting Chunks
Normally, splitting chunks is done automatically
Ruby Language Center
happen as a side effect of inserting (and are tra
an overview of the available tools and suggested practices for using Ruby with MongoDB.
may be other cases where a chunk is automati
Those wishing to skip to more detailed discussion should check out the Ruby Driver Tutorial
may be moved immediately to a new shard if th
Ruby Tutorial, Getting started with Rails Rails Getting Started Rails ...
Ruby Tutorial
page not done. Please help us finish it\! MySQ
tutorial gives many common examples of using MongoDB with the Ruby driver. If you're
{{mysqld}} \\ {{mongod}} \\ {{mysql}} \\ {{mongo}
looking for information on data modeling, see MongoDB Data Modeling and Rails. Links to the
expressed as JSON (BSON DOCS:BSON) obj
various object mappers are listed on our object mappers page http://www.mongodb.org ...
chart shows examples as both SQL ...
MongoDB is run as a standard program from th
Command Line Parameters for more informatio
examples assume that you are in the directory
and the Mongo executable is called {{mongod}
Storing Data
Storing Files
Structuring Data for Mongo
T U
Tailable Cursors Ubuntu and Debian packages

MongoDB has a feature known as tailable cursors which are similar to the Unix "tail \f" Please read the notes on the Downloads page
command. Tailable means the cursor is not closed once all data is retrieved. Rather, the cursor packages are updated daily, and so if you find
marks the last known object's position and you can resume ... try updating your apt package lists, e.g., with 'a
10gen ...
Too Many Open Files
you receive the error "too many open files" or "too many open connections" in the mongod log, UI
there are a couple of possible reasons for this. First, to check what file descriptors are in use, Spec/requirements for a future MongoDB admi
run lsof (some variations shown below): lsof grep ... clone? collections validate(), datasize, indexsiz
explain() output security: view users, adjust see
TreeNavigation
master ...
Trees in MongoDB
Updates
best way to store a tree usually depends on the operations you want to perform; see below for
some different options. In practice, most developers find that one of the "Full Tree in Updating
Single Document", "Parent Links", and "Array of Ancestors" patterns ... MongoDB supports atomic, inplace updates as
for replacing an entire document. update() {{up
matching criteria entirely with objNew. If you on
Troubleshooting
you should use ...
mongod process "disappeared" Scenario here is the log ending suddenly with no error or
shutdown messages logged. On Unix, check /var/log/messages: $ grep mongod Updating Data in Mongo
/var/log/messages $ grep score /var/log/messages See Also Diagnostic ... Updating a Document in the mongo Shell with
previous section, the {{save()}} method may be
Troubleshooting the PHP Driver
a collection. We can also use {{save()}} to upda
Tutorial collection. Continuing with the example databa
Getting the Database First, run through the Quickstart guide for your platform to get up and
Upgrading from a Non-Sharded System
running. Getting A Database Connection Let's now try manipulating the database with the
mongod}} process can become part of a shard
database shell DOCS:mongo The Interactive Shell . (We could perform similar ...
that process or downtime. If you haven't done s
the Sharding Introduction
http://www.mongodb.org/display/DOCS/Shardi
yourself ...
Upgrading to Replica Sets
Upgrading From a Single Server If you're runni
upgrading to replica sets is trivial (and a good i
replica set with a single node. We need a name
we're ...
Use Case - Session Objects
MongoDB is a good tool for storing HTTP sess
model is to have a sessions collection, and sto
a browser cookie. With its updateinplace desig
make updates fast, the database is efficient ...
Use Cases
See also the Production Deployments DOCS:P
a discussion of how companies like Shutterfly,
SourceForge, etc. use MongoDB. Use Case Ar
Realtime Analytics
http://blog.mongodb.org/post/171353301/using
User Feedback
I just have to get my head around that mongod
\muckster, #mongodb "Guys at Redmond shou
about what is the software development and su
\[email protected], mongodbuser list ...
technique one can use with MongoDB in certai
collections to store information instead of a sin
this, certain repeating data no longer needs to
index on that key may be eliminated ...
Using Mongoid
Mongoid is a mature ODM for MongoDB. Much
and it sports an active user community and exc
we've seen a few of Mongoid's design decision
production. This page is an attempt ...
Using Multikeys to Simulate a Large Number of Ind
One way to work with data that has a high deg
use the multikey DOCS:Multikeys indexing fea
objects. For example: > x = { > ... id : "a
... , > ... , > ... , > ... > ... }; > db.foo.insert(x); >
db.foo.find ...
V W
v0.8 Details What is the Compare Order for BSON Types

Existing Core Functionality Basic Mongo database functionality: inserts, deletes, queries, MongoDB allows objects in the same collection
indexing. Master / Slave Replication Replica Pairs Serverside javascript code execution New to differ in type. When comparing values fr
v0.8 Drivers for Java, C, Python, Ruby. db shell utility ... is utilized as to which value is less than the oth
arbitary but well ...
Validate Command
Use this command to check that a collection is valid (not corrupt) and to get various statistics. Why are my datafiles so large?
This command scans the entire collection and its indexes and will be very slow on large
Why so many "Connection Accepted" messages lo
datasets. From the {{mongo}} shell: > db.foo.validate() From a driver one might invoke the
driver's equivalent ... Windows
Windows Quick Links and Reference Center R
Verifying Propagation of Writes with getLastError
See the Quickstart page DOCS:Quickstart Win
v1.5. A client can block until a write operation has been replicated to N servers. Use the
and run the database for the first time. Running
getlasterror command with a new parameter {{w}}: db.runCommand( ) If {{w}} is not set, or
Service page. The MongoDB Server Get ...
equals 1, the command returns immediately, implying the data is on 1 server ...
Windows Service
Version Numbers
windows mongod.exe has native support for in
MongoDB uses the oddnumbered versions for development releases
service. Service Related Commands The servi
http://en.wikipedia.org/wiki/Softwareversioning#Oddnumberedversionsfordevelopmentreleases.
mongod install mongod service mongod remov
There are 3 numbers in a MongoDB version: A.B.C A is the major version. This will rarely option pass the following to \install ...
change and signify very large changes B is the release number. This will include many
Working with Mongo Objects and Classes in Ruby
changes ...
Video & Slides from Recent Events and Presentations
See Also DOCS:Mongo Query Language mong
Table of Contents: MongoDB Conferences Oneday conferences hosted by 10gen
parameter DOCS:Command Line Parameters
http://www.10gen.com/. 10gen develops and supports MongoDB. MongoUK Video (June 2010)
http://skillsmatter.com/event/cloudgrid/mongouk MongoFR Video (June 2010) Writing Tests
http://lacantine.ubicast.eu/categories ... We have three general flavors of tests currently
can inherit from class {{UnitTest}} and make a
These tests run EVERY TIME the program sta
View Current Operation(s) in Progress > db.currentOp(); > // same as:
...
db.$cmd.sys.inprog.findOne() { inprog: { "opid" : 18 , "op" : "query" , "ns" : "mydb.votes" ,
"query" : " " , "inLock" : 1 } } Fields: opid an incrementing operation number. Use with
killOp(). op the operation type ...
X Y
Z !@#$

Mongodb Docs 2010 09 03

Uploaded by

Copyright:

Available Formats

Mongodb Docs 2010 09 03

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

Mongodb Docs 2010 09 03

Uploaded by

Copyright:

Available Formats

.bookmarks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

The 15 most recent bookmarks in MongoDB

1.1 Development Cycle

Creating and Deleting Indexes

Django and MongoDB

PyMongo and mod_wsgi

Recommended Production Architectures

Wanted, but may not make it

Getting the Software

Mongo Administration Guide

Working with Mongo Objects and Classes in Ruby

MongoDB Language Support

in the queryOptions int field to indicate you want a tailable cursor.

Searching and Retrieving

Mongo Developers' Guide

Mongo Database Administration

Mongo Concepts and Terminology

MongoDB - A Developer's Tour

Structuring Data for Mongo

Why so many "Connection Accepted" messages logged?

Why are my datafiles so large?

Introduction - How Mongo Works

Optimizing Mongo Performance

Mongo Usage Basics

Mongo Boston - September 20

For an even quicker start go to http://try.mongodb.org/.

SQL to Mongo Mapping Chart

If you use the Homebrew package manager, run:

$ brew install mongodb

If you use MacPorts you can install with:

$ sudo port install mongodb

This will take a while to install.

Note: 64-bit is recommended.

$ curl http://downloads.mongodb.org/osx/mongodb-osx-i386-1.4.4.tgz > mongo.tgz

$ curl http://downloads.mongodb.org/osx/mongodb-osx-x86_64-1.4.4.tgz > mongo.tgz

Create a data directory

Run and connect to the server

32-bit Linux binaries

Note: 64 bit is recommended.

$ curl http://downloads.mongodb.org/linux/mongodb-linux-i686-1.4.4.tgz > mongo.tgz

64-bit Linux binaries

$ curl http://downloads.mongodb.org/linux/mongodb-linux-x86_64-1.4.4.tgz > mongo.tgz

Create a data directory

$ sudo mkdir -p /data/db/

Run and connect to the server

Download and extract the 64-bit .zip.

Create a data directory

C:\> mkdir \data

Or you can do this from the Windows Explorer, of course.

Run and connect to the server

mongod.exe - the database server

Mongo Shell Info

Version OS X 32 OS X 64 Linux 32 Linux 64 Windows Windows Solaris Solaris Source Date

See Version Numbers

The MongoDB database server

Language Packages Source API Reference

Python bundles github api

PHP pecl github api

Ruby gemcutter github api