:dfn:`Ansible Galaxy` refers to the `Galaxy <https://galaxy.ansible.com>`_ website, a free site for finding, downloading, and sharing community developed roles.
Use Galaxy to jump-start your automation project with great content from the Ansible community. Galaxy provides pre-packaged units of work such as :ref:`roles <playbooks_reuse_roles>`, and new in Galaxy 3.2, :ref:`collections <collections>`
You can find roles for provisioning infrastructure, deploying applications, and all of the tasks you do everyday. The collection format provides a comprehensive package of automation that may include multiple playbooks, roles, modules, and plugins.
..contents::
:local:
:depth:2
.._finding_galaxy_collections:
Finding collections on Galaxy
=============================
To find collections on Galaxy:
#. Click the :guilabel:`Search` icon in the left-hand navigation.
#. Set the filter to *collection*.
#. Set other filters and press :guilabel:`enter`.
Galaxy presents a list of collections that match your search criteria.
You can download collections from Automation Hub at the command line. Automation Hub content is available to subscribers only, so you must download an API token and configure your local environment to provide it before you can you download collections. To download a collection from Automation Hub with the ``ansible-galaxy`` command:
1. Get your Automation Hub API token. Go to https://cloud.redhat.com/ansible/automation-hub/token/ and click :guilabel:`Get API token` from the version dropdown to copy your API token.
2. Configure Red Hat Automation Hub server in the ``server_list`` option under the ``[galaxy]`` section in your :file:`ansible.cfg` file.
The command line tool by default communicates with the Galaxy website API using the server address *https://galaxy.ansible.com*. If you run your own internal Galaxy server
and want to use it instead of the default one, pass the ``--server`` option following the address of this galaxy server. You can set permanently this option by setting
the Galaxy server value in your ``ansible.cfg`` file to use it . For information on setting the value in *ansible.cfg* see :ref:`galaxy_server`.
Use the ``ansible-galaxy`` command to download roles from the `Galaxy website <https://galaxy.ansible.com>`_
..code-block:: bash
$ ansible-galaxy install namespace.role_name
Setting where to install roles
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
By default, Ansible downloads roles to the first writable directory in the default list of paths ``~/.ansible/roles:/usr/share/ansible/roles:/etc/ansible/roles``. This installs roles in the home directory of the user running ``ansible-galaxy``.
You can override this with one of the following options:
* Set the environment variable :envvar:`ANSIBLE_ROLES_PATH` in your session.
* Use the ``--roles-path`` option for the ``ansible-galaxy`` command.
#. Click on the name to view more details, including the available versions.
You can also navigate directly to the role using the /<namespace>/<role name>. For example, to view the role geerlingguy.apache, go to `<https://galaxy.ansible.com/geerlingguy/apache>`_.
To install a specific version of a role from Galaxy, append a comma and the value of a GitHub release tag. For example:
You can install multiple roles by including the roles in a :file:`requirements.yml` file. The format of the file is YAML, and the
file extension must be either *.yml* or *.yaml*.
Use the following command to install roles included in :file:`requirements.yml:`
..code-block:: bash
$ ansible-galaxy install -r requirements.yml
Again, the extension is important. If the *.yml* extension is left off, the ``ansible-galaxy`` CLI assumes the file is in an older, now deprecated,
"basic" format.
Each role in the file will have one or more of the following attributes:
src
The source of the role. Use the format *namespace.role_name*, if downloading from Galaxy; otherwise, provide a URL pointing
to a repository within a git based SCM. See the examples below. This is a required attribute.
scm
Specify the SCM. As of this writing only *git* or *hg* are allowed. See the examples below. Defaults to *git*.
version:
The version of the role to download. Provide a release tag value, commit hash, or branch name. Defaults to the branch set as a default in the repository, otherwise defaults to the *master*.
name:
Download the role to a specific name. Defaults to the Galaxy name when downloading from Galaxy, otherwise it defaults
to the name of the repository.
Use the following example as a guide for specifying roles in *requirements.yml*:
Embedding credentials into a SCM URL is not secure. Make sure to use safe auth options for security reasons. For example, use `SSH <https://help.github.com/en/github/authenticating-to-github/connecting-to-github-with-ssh>`_, `netrc <https://linux.die.net/man/5/netrc>`_ or `http.extraHeader <https://git-scm.com/docs/git-config#Documentation/git-config.txt-httpextraHeader>`_/`url.<base>.pushInsteadOf <https://git-scm.com/docs/git-config#Documentation/git-config.txt-urlltbasegtpushInsteadOf>`_ in Git config to prevent your creds from being exposed in logs.
For large projects, the ``include`` directive in a :file:`requirements.yml` file provides the ability to split a large file into multiple smaller files.
For example, a project may have a :file:`requirements.yml` file, and a :file:`webserver.yml` file.
Below are the contents of the :file:`webserver.yml` file:
Roles can also be dependent on other roles, and when you install a role that has dependencies, those dependencies will automatically be installed to the ``roles_path``.
You can create the file ``meta/requirements.yml`` and define dependencies in the same format used for :file:`requirements.yml` described in the `Installing multiple roles from a file`_ section.
From there, you can import or include the specified roles in your tasks.
Using ``meta/main.yml``
^^^^^^^^^^^^^^^^^^^^^^^
Alternatively, you can specify role dependencies in the ``meta/main.yml`` file by providing a list of roles under the ``dependencies`` section. If the source of a role is Galaxy, you can simply specify the role in
the format ``namespace.role_name``. You can also use the more complex format in :file:`requirements.yml`, allowing you to provide ``src``, ``scm``, ``version``, and ``name``.
Dependencies installed that way, depending on other factors described below, will also be executed **before** this role is executed during play execution.
To better understand how dependencies are handled during play execution, see :ref:`playbooks_reuse_roles`.
The following shows an example ``meta/main.yml`` file with dependent roles:
..code-block:: yaml
---
dependencies:
- geerlingguy.java
galaxy_info:
author: geerlingguy
description: Elasticsearch for Linux.
company: "Midwestern Mac, LLC"
license: "license (BSD, MIT)"
min_ansible_version: 2.4
platforms:
- name: EL
versions:
- all
- name: Debian
versions:
- all
- name: Ubuntu
versions:
- all
galaxy_tags:
- web
- system
- monitoring
- logging
- lucene
- elk
- elasticsearch
Tags are inherited *down* the dependency chain. In order for tags to be applied to a role and all its dependencies, the tag should be applied to the role, not to all the tasks within a role.
Roles listed as dependencies are subject to conditionals and tag filtering, and may not execute fully depending on
what tags and conditionals are applied.
If the source of a role is Galaxy, specify the role in the format *namespace.role_name*:
..code-block:: yaml
dependencies:
- geerlingguy.apache
- geerlingguy.ansible
Alternately, you can specify the role dependencies in the complex form used in :file:`requirements.yml` as follows:
