droplets 1.3.0

Library for building Ansible dynamic inventories for Digital Ocean.

droplets

droplets aids in the management of Digital Ocean (DO) droplets. droplets also has the ability to act as an inventory script for an Ansible playbook.

Using droplets

The basic command line usage allows you to build your own script to connect to your DO environments. The script serve two purposes. First, the creation and destruction of droplets based on defined blueprint. Second, is to act as an Ansible invetory script for those droplets.

Creating the script

Using droplets you will define your own script for your different DO deployments. In your script you will define a blueprint.

Calling main

main requires a blueprint and also accepts some optional arguments.

• image (default: ubuntu-14-04-x64)
• size (default: 512mb)
• region (default: nyc3)
• prefix (default: '')
• api_token (default: None)

The defaults will be used for every instance created unless you explicitly override the arguments in your call to main or in the group definition of your blueprint.

prefix is useful when creating scripts for different DO envrionments. prefix will be used when generating the names for the instances of your groups. For example if you call main with prefix=staging- and you have an app group defined in your blueprint, your instances will have the name staging-app00, staging-app01, etc...

Defining a blueprint

A blueprint is a standard python dictionary. A blueprint should contain a key entry for every group. The value of each group should be another dictionary that will act as a key/value store for hostvars that will be passed to an ansible playbook.

A group can contain some special keys which override the default values of main.

• size (default: 512mb)
• image (default: ubuntu-14-04-x64)
• region (default: nyc3)
• n (number of instances, default: 1)

Typical group names are things like app, db, and loadbalancer. The droplet script will use your group name along with the prefix (see below) when creating instance names

API Token

You must also provide an API token. There are two ways to do this. First, you can pass the api_token keyword argument to the main call. Second, you can set the environment variable 'DIGITAL_OCEAN_TOKEN' to your token value. The later option is useful if you intend to store your droplet script in a public repository.

Example

Here is a very simple example of a single application server and single database server blueprint. They override none of the default values and the app group passes the hostvar git_branch to the playbook.

#!/usr/bin/env python

import sys
from droplets import main

blueprint = {
    'app': {
        'git_branch': 'master',
    },
    'db': {},
}

if __name__ == '__main__':
    main(blueprint, api_token='XXX', prefix='staging-')

Running the script

Now with this staging script defined, we can issue commands on it directly. For example

./staging --help

Options:
    -h --help    Show this screen
    --list       Outputs inventory in json.
    --human      Outputs inventory in format suitable for human beings.
    --reconcile  Creates/deletes droplets as necessary to match blueprint.
    --hostkeys   Gets ssh host keys for hosts in inventory and install them in
                ~/.ssh/known_hosts
    --destroy=<group>    Destroys the droplets in the specified group.

By default, the script will execute as if you have passed in the --list option. This allows the droplets script to at as an inventory script for the ansible-playbook command.

If we wanted to create these instances to prepare for a playbook run we would use the --reconcile option. It is worth mentioning that --reconcile will not retroactively reconcile changes to size, image, or region. You will need to login to DO to make adjustments to size OR use --destroy on the group and then --reconsile.

Running with ansible-playbook

Now that you have your DO instances deployed how you want (using --reconcile). You can use your droplets script as an inventory script for your playbook.

ansible-playbook -i staging deploy-staging.yaml