This plugin was incorported into the official Ansible 2.9 release. See:
The Ansible version of the source code for this plugin has moved to the Ansible Commnunity Collection and is available at:
A dynamic inventory plugin for Ansible. Tested with Ansible 2.7.10 and Python 3.6.7 on Ubuntu Linux 18.04 LTS with Digital Ocean Droplets created by Docker Machine.
This plugin is based on and is very similar to the Ansible 2.8 docker_swarm and aws_ec2 Dynamic Inventory plugins.
While there are other similar solutions out there, I did not find an existing Docker Machine dynamic inventory plugin, only scripts (e.g. this).
This plugin was created for and is being used by the NLnet Labs Gantry project.
This plugin has been submitted for inclusion in Ansible core. See PR #54946.
This plugin teaches Ansible about which Docker Machine machines exist and how to connect to them via SSH so that the Ansible controller can execute commands on the remote machine.
It also makes available the DOCKER_xxx environment variables output by docker-machine env <machine-name> as Ansible host variables, prefixed with dm_. These are intended to be used to set Ansible or environment variables such that Docker commands (e.g. docker ps, docker-compose up) will be executed against the Docker daemon running on the remote machine and not the local Docker daemon.
Like the Docker Swarm and AWS EC2 plugins it supports keyed_groups and other constructable ways of dynamically defining Ansible host groups.
If verbose_output is enabled it also exposes the entire Docker Machine inspect output as an ansible inventory host variable called docker_machine_node_attributes (matching the naming style used by the Docker Swarm inventory plugin).
Docker Machine must be installed on the Ansible controller.
- Tell Ansible where to find the plugin. See the Ansible documentation for all the possible ways that you can do this.
- Tell Ansible how to configure the plugin by pointing it to a
docker_machine.ymlconfiguration file that you create.
E.g.
$ export ANSIBLE_INVENTORY_ENABLED=docker_machine
$ export ANSIBLE_INVENTORY_PLUGINS=/tmp/ansible-docker-machine-inventory-plugin
$ git clone https://github.com/ximon18/ansible-docker-machine-inventory-plugin.git ${ANSIBLE_INVENTORY_PLUGINS}
$ vi ${ANSIBLE_INVENTORY_PLUGINS}/docker_machine.yml
$ ansible -i ${ANSIBLE_INVENTORY_PLUGINS}/docker_machine.yml -m ping all
To find out what the plugin has discovered do:
$ ansible-inventory -i ${ANSIBLE_INVENTORY_PLUGINS}/docker_machine.yml --graph --vars
To read the full plugin docs you can use the ansible-doc command:
$ ansible-doc -t inventory docker_machine
E.g. at the time of writing on my local Ansible 2.7.10 installation this produces:
> INVENTORY (/root/.ansible/plugins/inventory/docker_machine.py)
Get inventory hosts from Docker Machine. Uses a YAML configuration file that ends with docker_machine.(yml|yaml). The plugin sets standard host variables
`ansible_host', `ansible_port', `ansible_user' and `ansible_ssh_private_key'. The plugin stores the Docker Machine 'env' output variables in `dm_' prefixed host
variables.
OPTIONS (= is mandatory):
- cache
Toggle to enable/disable the caching of the inventory's source data, requires a cache plugin setup to work.
[Default: False]
set_via:
env:
- name: ANSIBLE_INVENTORY_CACHE
ini:
- key: cache
section: inventory
type: boolean
- cache_connection
Cache connection data or path, read cache plugin documentation for specifics.
[Default: (null)]
set_via:
env:
- name: ANSIBLE_INVENTORY_CACHE_CONNECTION
ini:
- key: cache_connection
section: inventory
- cache_plugin
Cache plugin to use for the inventory's source data.
[Default: (null)]
set_via:
env:
- name: ANSIBLE_INVENTORY_CACHE_PLUGIN
ini:
- key: cache_plugin
section: inventory
- cache_timeout
Cache duration in seconds
[Default: 3600]
set_via:
env:
- name: ANSIBLE_INVENTORY_CACHE_TIMEOUT
ini:
- key: cache_timeout
section: inventory
type: integer
- compose
create vars from jinja2 expressions
[Default: {}]
type: dictionary
- daemon_required
when true, hosts for which Docker Machine cannot output Docker daemon connection environment variables will be skipped.
[Default: True]
type: bool
- groups
add hosts to group based on Jinja2 conditionals
[Default: {}]
type: dictionary
- keyed_groups
add hosts to group based on the values of a variable
[Default: []]
type: list
= plugin
token that ensures this is a source file for the `docker_machine' plugin.
(Choices: docker_machine)
- running_required
when true, hosts which Docker Machine indicates are in a state other than `running' will be skipped.
[Default: True]
type: bool
- strict
If true make invalid entries a fatal error, otherwise skip and continue
Since it is possible to use facts in the expressions they might not always be available and we ignore those errors by default.
[Default: False]
type: boolean
- verbose_output
when true, include all available nodes metadata (e.g. Image, Region, Size) as a JSON object.
[Default: True]
type: bool
REQUIREMENTS: L(Docker Machine,https://docs.docker.com/machine/)
AUTHOR: Ximon Eighteen (@ximon18)
NAME: docker_machine
PLUGIN_TYPE: inventory
EXAMPLES:
# Minimal example
plugin: docker_machine
# Example using constructed features to create a group per Docker Machine driver
# (https://docs.docker.com/machine/drivers/), e.g.:
# $ docker-machine create --driver digitalocean ... mymachine
# $ ansible-inventory -i ./path/to/docker-machine.yml --host=mymachine
# {
# ...
# "digitalocean": {
# "hosts": [
# "mymachine"
# ]
# ...
# }
strict: no
keyed_groups:
- separator: ''
key: docker_machine_node_attributes.DriverName
# Example grouping hosts by Digital Machine tag
strict: no
keyed_groups:
- prefix: tag
key: 'dm_tags'
# Example using compose to override the default SSH behaviour of asking the user to accept the remote host key
compose:
ansible_ssh_common_args: '"-o StrictHostKeyChecking=accept-new"'
END