Installs and configures MongoDB, supporting:
- Single MongoDB
- Replication
- Sharding
- Replication and Sharding
- 10gen repository package installation
The cookbook aims to be platform independant, but is best tested on debian squeeze systems.
The 10gen_repo
recipe configures the package manager to use 10gen's
official package reposotories on Debian, Ubuntu, Redhat, CentOS, Fedora, and
Amazon linux distributions.
This cookbook contains a definition mongodb_instance
which can be used to configure
a certain type of mongodb instance, like the default mongodb or various components
of a sharded setup.
For examples see the USAGE section below.
This cookbook defines the LWRP mongodb_user
to manage users and authentication
in mongodb.
For examples see the USAGE section below.
mongodb[:auth]
- Enable or disable authmongodb[:keyfile]
- keyFile contents for auth w/ replicasetsmongodb[:dbpath]
- Location for mongodb data directory, defaults to "/var/lib/mongodb"mongodb[:logpath]
- Path for the logfiles, default is "/var/log/mongodb"mongodb[:port]
- Port the mongod listens on, default is 27017mongodb[:client_role]
- Role identifing all external clients which should have access to a mongod instancemongodb[:cluster_name]
- Name of the cluster, all members of the cluster must reference to the same name, as this name is used internally to identify all members of a cluster.mongodb[:shard_name]
- Name of a shard, default is "default"mongodb[:sharded_collections]
- Define which collections are shardedmongodb[:replicaset_name]
- Define name of replicatsetmongodb[:package_version]
- Version of the MongoDB package to install, default is nil
Adds the stable 10gen repo for the corresponding platform. Currently only implemented for the Debian and Ubuntu repository.
Usage: just add recipe[mongodb::10gen_repo]
to the node run_list before any other
MongoDB recipe, and the mongodb-10gen stable packages will be installed instead of the distribution default.
Simply add
include_recipe "mongodb::default"
to your recipe. This will run the mongodb instance as configured by your distribution.
You can change the dbpath, logpath and port settings (see ATTRIBUTES) for this node by
using the mongodb_instance
definition:
mongodb_instance "mongodb" do
port node['application']['port']
end
This definition also allows you to run another mongod instance with a different name on the same node
mongodb_instance "my_instance" do
port node['mongodb']['port'] + 100
dbpath "/data/"
end
The result is a new system service with
/etc/init.d/my_instance <start|stop|restart|status>
Add mongodb::replicaset
to the node's run_list. Also choose a name for your
replicaset cluster and set the value of node[:mongodb][:cluster_name]
for each
member to this name.
You need a few more components, but the idea is the same: identification of the
members with their different internal roles (mongos, configserver, etc.) is done via
the node[:mongodb][:cluster_name]
and node[:mongodb][:shard_name]
attributes.
Let's have a look at a simple sharding setup, consisting of two shard servers, one config server and one mongos.
First we would like to configure the two shards. For doing so, just use
mongodb::shard
in the node's run_list and define a unique mongodb[:shard_name]
for each of these two nodes, say "shard1" and "shard2".
Then configure a node to act as a config server - by using the mongodb::configserver
recipe.
And finally you need to configure the mongos. This can be done by using the
mongodb::mongos
recipe. The mongos needs some special configuration, as these
mongos are actually doing the configuration of the whole sharded cluster.
Most importantly you need to define what collections should be sharded by setting the
attribute mongodb[:sharded_collections]
:
{
"mongodb": {
"sharded_collections": {
"test.addressbook": "name",
"mydatabase.calendar": "date"
}
}
}
Now mongos will automatically enable sharding for the "test" and the "mydatabase" database. Also the "addressbook" and the "calendar" collection will be sharded, with sharding key "name" resp. "date". In the context of a sharding cluster always keep in mind to use a single role which is added to all members of the cluster to identify all member nodes. Also shard names are important to distinguish the different shards. This is esp. important when you want to replicate shards.
The setup is not much different to the one described above. All you have to do is adding the
mongodb::replicaset
recipe to all shard nodes, and make sure that all shard
nodes which should be in the same replicaset have the same shard name.
For more details, you can find a tutorial for Sharding + Replication in the wiki.
To enable authentication on a database, you need to enable authentication by
setting the auth
attribute to true
{
"mongodb": {
"auth": true
}
}
If you are using replicasets, you only need to define the keyFile secret string.
{
"mongodb": {
"keyfile": "5UJ1aZngn9OM/YN0IsiivhQUmGY..."
}
}
The keyFile contents is arbitrary, but a long string is suggested. openssl
can be used to generate a good string:
$ openssl rand -base64 753
require 'openssl'
default['mongodb']['keyfile'] = ::OpenSSL::Random.random_bytes(753).gsub(/\W/, '')
Once authentication has been enabled, you can then define users simply by including
mongodb
in the run list and defining your user.
mongodb_user "someguy" do
password "n0ts3cur3"
database "gbs_for_dbs"
end
Actions available:
:add
- default:delete
:update
Author:: Markus Korn markus.korn@edelight.de
Copyright:: 2011, edelight GmbH
Licensed under the Apache License, Version 2.0 (the "License"); you may not use this file except in compliance with the License. You may obtain a copy of the License at
http://www.apache.org/licenses/LICENSE-2.0
Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the License for the specific language governing permissions and limitations under the License.