mongo-orchestration 0.6.6

Restful service for managing MongoDB servers

See the wiki for documentation.

Mongo Orchestration is an HTTP server that provides a REST API for creating and managing MongoDB configurations on a single host.

THIS PROJECT IS FOR TESTING OF MONGODB DRIVERS.

Features

• Start and stop mongod servers, replica sets, and sharded clusters on the host running mongo-orchestration.

• Add and remove replica set members.

• Add and remove shards and mongos routers.

• Reset replica sets and clusters to restart all members that were stopped.

• Freeze secondary members of replica sets.

• Retrieve information about MongoDB resources.

• Interaction all through REST interface.

Requires

• Python 2.6, 2.7, or >= 3.2

• bottle >= 0.12.7

• PyMongo >= 3.0.2

• CherryPy >= 3.5.0, < 7.1

• argparse >= 1.2.1 (Python 2.6 only)

Installation

The easiest way to install Mongo Orchestration is with pip:

pip install mongo-orchestration

You can also install the development version of Mongo Orchestration manually:

git clone https://github.com/10gen/mongo-orchestration.git
cd mongo-orchestration
python setup.py install

Cloning the repository this way will also give you access to the tests for Mongo Orchestration as well as the mo script. Note that you may have to run the above commands with sudo, depending on where you’re installing Mongo Orchestration and what privileges you have. Installation will place a mongo-orchestration script on your path.

Usage

mongo-orchestration [-h] [-f CONFIG] [-e ENV] [--no-fork] [-b BIND IP="localhost"] [-p PORT]
                    [-s {cherrypy,wsgiref}] [--socket-timeout-ms MILLIS]
                    [--pidfile PIDFILE] [--enable-majority-read-concern] {start,stop,restart}

Arguments:

• -h - show help

• -f, –config - path to config file

• -e, –env - default release to use, as specified in the config file

• –no-fork - run server in foreground

• -b, –bind - host on which Mongo Orchestration and subordinate mongo processes should listen for requests. Defaults to “localhost”.

• -s, –server - HTTP backend to use: one of cherrypy or wsgiref

• -p - port number (8889 by default)

• –socket-timeout-ms - socket timeout when connecting to MongoDB servers

• –pidfile - location where mongo-orchestration should place its pid file

• –enable-majority-read-concern - enable “majority” read concern on server versions that support it.

• start/stop/restart: start, stop, or restart the server, respectively

In addition, Mongo Orchestration can be influenced by the MONGO_ORCHESTRATION_HOME environment variable, which informs the server where to find the “configurations” directory for presets as well as where to put the log and pid files.

Examples

mongo-orchestration start

Starts Mongo Orchestration as service on port 8889.

mongo-orchestration stop

Stop the server.

mongo-orchestration -f mongo-orchestration.config -e 30-release -p 8888 --no-fork start

Starts Mongo Orchestration on port 8888 using 30-release defined in mongo-orchestration.config. Stops with Ctrl+C.

Configuration File

Mongo Orchestration may be given a JSON configuration file with the --config option specifying where to find MongoDB binaries. See mongo-orchestration.config for an example. When no configuration file is provided, Mongo Orchestration uses whatever binaries are on the user’s PATH.

Predefined Configurations

Mongo Orchestration has a set of predefined configurations that can be used to start, restart, or stop MongoDB processes. You can use a tool like curl to send these files directly to the Mongo Orchestration server, or use the mo script in the scripts directory (in the repository only). Some examples:

• Start a single node without SSL or auth:

  mo configurations/servers/clean.json start

• Get the status of a single node without SSL or auth:

  mo configurations/servers/clean.json status

• Stop a single node without SSL or auth:

  mo configurations/servers/clean.json stop

• Start a replica set with ssl and auth:

  mo configurations/replica_sets/ssl_auth.json start

• Use curl to create a basic sharded cluster with the id “myCluster”:

  curl -XPUT http://localhost:8889/v1/sharded_clusters/myCluster \
             -d@configurations/sharded_clusters/basic.json

Note that in order to run the mo script, you need to be in the same directory as “configurations”.

Helpful hint: You can prettify JSON responses from the server by piping the response into python -m json.tool, e.g.:

$ curl http://localhost:8889/v1/servers/myServer | python -m json.tool

{
    "id": "myServer",
    "mongodb_uri": "mongodb://localhost:1025",
    "orchestration": "servers",
    "procInfo": {
        "alive": true,
        "name": "mongod",
        "optfile": "/var/folders/v9/spc2j6cx3db71l/T/mongo-KHUACD",
        "params": {
            "dbpath": "/var/folders/v9/spc2j6cx3db71l/T/mongo-vAgYaQ",
            "ipv6": true,
            "journal": true,
            "logappend": true,
            "noprealloc": true,
            "oplogSize": 100,
            "port": 1025,
            "smallfiles": true
        },
        "pid": 51320
    },
    // etc.
}

Tests

In order to run the tests, you should first clone the repository. Running the tests has the following additional dependency:

• unittest2 >= 0.6 (Python 2.6 only)

Run all tests

python setup.py test

Run a test module

python -m unittest tests.test_servers

Run a single test case

python -m unittest tests.test_servers.ServerSSLTestCase

Run a single test method

python -m unittest tests.test_servers.ServerSSLTestCase.test_ssl_auth

Run a single test example for debugging with verbose and immediate stdout output

python -m unittest -v tests.test_servers.ServerSSLTestCase