Roles

Roles let you automatically load related vars, files, tasks, handlers, and other Ansible artifacts based on a known file structure. After you group your content into roles, you can easily reuse them and share them with other users.

Role directory structure 

An Ansible role has a defined directory structure with seven main standard directories. You must include at least one of these directories in each role. You can omit any directories the role does not use. For example:

# playbooks
site.yml
webservers.yml
fooservers.yml

roles/
    common/               # this hierarchy represents a "role"
        tasks/            #
            main.yml      #  <-- tasks file can include smaller files if warranted
        handlers/         #
            main.yml      #  <-- handlers file
        templates/        #  <-- files for use with the template resource
            ntp.conf.j2   #  <------- templates end in .j2
        files/            #
            bar.txt       #  <-- files for use with the copy resource
            foo.sh        #  <-- script files for use with the script resource
        vars/             #
            main.yml      #  <-- variables associated with this role
        defaults/         #
            main.yml      #  <-- default lower priority variables for this role
        meta/             #
            main.yml      #  <-- role dependencies
        library/          # roles can also include custom modules
        module_utils/     # roles can also include custom module_utils
        lookup_plugins/   # or other types of plugins, like lookup in this case

    webtier/              # same kind of structure as "common" was above, done for the webtier role
    monitoring/           # ""
    fooapp/               # ""

By default, Ansible will look in most role directories for a main.yml file for relevant content (also main.yaml and main):

tasks/main.yml - A list of tasks that the role provides to the play for execution.
handlers/main.yml - handlers that are imported into the parent play for use by the role or other roles and tasks in the play.
defaults/main.yml - very low precedence values for variables provided by the role (see Using Variables for more information). A role’s own defaults will take priority over other role’s defaults, but any/all other variable sources will override this.
vars/main.yml - high precedence variables provided by the role to the play (see Using Variables for more information).
files/stuff.txt - one or more files that are available for the role and it’s children.
templates/something.j2 - templates to use in the role or child roles.
meta/main.yml - metadata for the role, including role dependencies and optional Galaxy metadata such as platforms supported. This is required for uploading into galaxy as a standalone role, but not for using the role in your play.

Note

None of the files above are required for a role. For example, you can just provide files/something.txt or vars/for_import.yml and it will still be a valid role.
On stand alone roles you can also include custom modules and/or plugins, for example library/my_module.py, which may be used within this role (see Embedding modules and plugins in roles for more information).
A ‘stand alone’ role refers to role that is not part of a collection but as individually installable content.
Variables from vars/ and defaults/ are imported into play scope unless you disable it via the public option in import_role/include_role.

You can add other YAML files in some directories, but they won’t be used by default. They can be included/imported directly or specified when using include_role/import_role. For example, you can place platform-specific tasks in separate files and refer to them in the tasks/main.yml file:

# roles/example/tasks/main.yml
- name: Install the correct web server for RHEL
  import_tasks: redhat.yml
  when: ansible_facts['os_family']|lower == 'redhat'

- name: Install the correct web server for Debian
  import_tasks: debian.yml
  when: ansible_facts['os_family']|lower == 'debian'

# roles/example/tasks/redhat.yml
- name: Install web server
  ansible.builtin.yum:
    name: "httpd"
    state: present

# roles/example/tasks/debian.yml
- name: Install web server
  ansible.builtin.apt:
    name: "apache2"
    state: present

Or call those tasks directly when loading the role, which bypasses the main.yml files:

- name: include apt tasks
  include_role:
      name: package_manager_bootstrap
      tasks_from: apt.yml
  when: ansible_facts['os_family'] == 'Debian'

Directories defaults and vars may also include nested directories. If your variables file is a directory, Ansible reads all variables files and directories inside in alphabetical order. If a nested directory contains variables files as well as directories, Ansible reads the directories first. Below is an example of a vars/main directory:

roles/
    common/          # this hierarchy represents a "role"
        vars/
            main/    #  <-- variables associated with this role
                first_nested_directory/
                    first_variables_file.yml
                second_nested_directory/
                    second_variables_file.yml
                third_variables_file.yml

Storing and finding roles 

By default, Ansible looks for roles in the following locations:

in collections, if you are using them
in a directory called roles/, relative to the playbook file
in the configured roles_path. The default search path is ~/.ansible/roles:/usr/share/ansible/roles:/etc/ansible/roles.
in the directory where the playbook file is located

If you store your roles in a different location, set the roles_path configuration option so Ansible can find your roles. Checking shared roles into a single location makes them easier to use in multiple playbooks. See Configuring Ansible for details about managing settings in ansible.cfg.

Alternatively, you can call a role with a fully qualified path:

---
- hosts: webservers
  roles:
    - role: '/path/to/my/roles/common'

Using roles 

You can use roles in the following ways:

at the play level with the roles option: This is the classic way of using roles in a play.
at the tasks level with include_role: You can reuse roles dynamically anywhere in the tasks section of a play using include_role.
at the tasks level with import_role: You can reuse roles statically anywhere in the tasks section of a play using import_role.
as a dependency of another role (see the dependencies keyword in meta/main.yml in this same page).

Using roles at the play level 

The classic (original) way to use roles is with the roles option for a given play:

---
- hosts: webservers
  roles:
    - common
    - webservers

When you use the roles option at the play level, each role ‘x’ looks for a main.yml (also main.yaml and main) in the following directories:

roles/x/tasks/
roles/x/handlers/
roles/x/vars/
roles/x/defaults/
roles/x/meta/
Any copy, script, template or include tasks (in the role) can reference files in roles/x/{files,templates,tasks}/ (dir depends on task) without having to path them relatively or absolutely.

Note

vars and defaults can also match to a directory of the same name and Ansible will process all the files contained in that directory. See Role directory structure for more details.

Note

If you use include_role/import_role, you can specify a custom file name instead of main. The meta directory is an exception because it does not allow for customization.

When you use the roles option at the play level, Ansible treats the roles as static imports and processes them during playbook parsing. Ansible executes each play in this order:

Any pre_tasks defined in the play.
Any handlers triggered by pre_tasks.
Each role listed in roles:, in the order listed. Any role dependencies defined in the role’s meta/main.yml run first, subject to tag filtering and conditionals. See Using role dependencies for more details.
Any tasks defined in the play.
Any handlers triggered by the roles or tasks.
Any post_tasks defined in the play.
Any handlers triggered by post_tasks.

Note

If using tags with tasks in a role, be sure to also tag your pre_tasks, post_tasks, and role dependencies and pass those along as well, especially if the pre/post tasks and role dependencies are used for monitoring outage window control or load balancing. See Tags for details on adding and using tags.

You can pass other keywords to the roles option:

---
- hosts: webservers
  roles:
    - common
    - role: foo_app_instance
      vars:
        dir: '/opt/a'
        app_port: 5000
      tags: typeA
    - role: foo_app_instance
      vars:
        dir: '/opt/b'
        app_port: 5001
      tags: typeB

When you add a tag to the role option, Ansible applies the tag to ALL tasks within the role.

Note

Prior to ansible-core 2.15, vars: within the roles: section of a playbook are added to the play variables, making them available to all tasks within the play before and after the role. This behavior can be changed by DEFAULT_PRIVATE_ROLE_VARS. On more recent versions, vars: do not leak into the play’s variable scope.

Including roles: dynamic reuse 

You can reuse roles dynamically anywhere in the tasks section of a play using include_role. While roles added in a roles section run before any other tasks in a play, included roles run in the order they are defined. If there are other tasks before an include_role task, the other tasks will run first.

To include a role:

---
- hosts: webservers
  tasks:
    - name: Print a message
      ansible.builtin.debug:
        msg: "this task runs before the example role"

    - name: Include the example role
      include_role:
        name: example

    - name: Print a message
      ansible.builtin.debug:
        msg: "this task runs after the example role"

You can pass other keywords, including variables and tags, when including roles:

---
- hosts: webservers
  tasks:
    - name: Include the foo_app_instance role
      include_role:
        name: foo_app_instance
      vars:
        dir: '/opt/a'
        app_port: 5000
      tags: typeA
  ...

When you add a tag to an include_role task, Ansible applies the tag only to the include itself. This means you can pass --tags to run only selected tasks from the role, if those tasks themselves have the same tag as the include statement. See Selectively running tagged tasks in reusable files for details.

You can conditionally include a role:

---
- hosts: webservers
  tasks:
    - name: Include the some_role role
      include_role:
        name: some_role
      when: "ansible_facts['os_family'] == 'RedHat'"

Importing roles: static reuse 

You can reuse roles statically anywhere in the tasks section of a play using import_role. The behavior is the same as using the roles keyword. For example:

---
- hosts: webservers
  tasks:
    - name: Print a message
      ansible.builtin.debug:
        msg: "before we run our role"

    - name: Import the example role
      import_role:
        name: example

    - name: Print a message
      ansible.builtin.debug:
        msg: "after we ran our role"

You can pass other keywords, including variables and tags when importing roles:

---
- hosts: webservers
  tasks:
    - name: Import the foo_app_instance role
      import_role:
        name: foo_app_instance
      vars:
        dir: '/opt/a'
        app_port: 5000
  ...

When you add a tag to an import_role statement, Ansible applies the tag to all tasks within the role. See Tag inheritance: adding tags to multiple tasks for details.

Role argument validation 

Beginning with version 2.11, you may choose to enable role argument validation based on an argument specification. This specification is defined in the meta/argument_specs.yml file (or with the .yaml file extension). When this argument specification is defined, a new task is inserted at the beginning of role execution that will validate the parameters supplied for the role against the specification. If the parameters fail validation, the role will fail execution.

Note

Ansible also supports role specifications defined in the role meta/main.yml file, as well. However, any role that defines the specs within this file will not work on versions below 2.11. For this reason, we recommend using the meta/argument_specs.yml file to maintain backward compatibility.

Note

When role argument validation is used on a role that has defined dependencies, then validation on those dependencies will run before the dependent role, even if argument validation fails for the dependent role.

Note

Ansible tags the inserted role argument validation task with always. If the role is statically imported this task runs unless you use the --skip-tags flag.

Specification format 

The role argument specification must be defined in a top-level argument_specs block within the role meta/argument_specs.yml file. All fields are lowercase.

entry-point-name:

The name of the role entry point.
This should be main in the case of an unspecified entry point.
This will be the base name of the tasks file to execute, with no .yml or .yaml file extension.

short_description:

A short, one-line description of the entry point. Ideally, it is a phrase and not a sentence.
The short_description is displayed by ansible-doc -t role -l.
It also becomes part of the title for the role page in the documentation.
The short description should always be a string and never a list, and should not end in a period.

description:

A longer description that may contain multiple lines.
This can be a single string or a list of strings. In case this is a list of strings, every list
element is a new paragraph.

version_added:

The version of the role when the entrypoint was added.
This is a string, and not a float, for example, version_added: '2.1'.
In collections, this must be the collection version the entrypoint was added to. For example, version_added: 1.0.0.

author:

Name of the entry point authors.
This can be a single string or a list of strings. Use one list entry per author. If there is only a single author, use a string or a one-element list.

options:

Options are often called “parameters” or “arguments”. This section defines those options.
For each role option (argument), you may include:

option-name:

The name of the option/argument.

description:

Detailed explanation of what this option does. It should be written in full sentences.
This can be a single string or a list of strings. In case this is a list of strings, every list element is a new paragraph.

version_added:

Only needed if this option was added after the initial role/entry point release. In other words, this is greater than the top level version_added field.
This is a string, and not a float, for example, version_added: '2.1'.
In collections, this must be the collection version the option was added to. For example, version_added: 1.0.0.

type:

The data type of the option. See Argument spec for allowed values for type. The default is str.
If an option is of type list, elements should be specified.

required:

Only needed if true.
If missing, the option is not required.

default:

If required is false/missing, default may be specified (assumed null if missing).
Ensure that the default value in the docs matches the default value in the code. The actual default for the role variable will always come from the role defaults (as defined in Role directory structure).
The default field must not be listed as part of the description unless it requires additional information or conditions.
If the option is a boolean value, you should use true/false if you want to be compatible with ansible-lint.

choices:

List of option values.
Should be absent if empty.

elements:

Specifies the data type for list elements when the type is list.

options:

If this option takes a dict or list of dicts, you can define the structure here.

Sample specification 

# roles/myapp/meta/argument_specs.yml
---
argument_specs:
  # roles/myapp/tasks/main.yml entry point
  main:
    short_description: Main entry point for the myapp role
    description:
      - This is the main entrypoint for the C(myapp) role.
      - Here we can describe what this entrypoint does in lengthy words.
      - Every new list item is a new paragraph. You can have multiple sentences
        per paragraph.
    author:
      - Daniel Ziegenberg
    options:
      myapp_int:
        type: "int"
        required: false
        default: 42
        description:
          - "The integer value, defaulting to 42."
          - "This is a second paragraph."

      myapp_str:
        type: "str"
        required: true
        description: "The string value"

      myapp_list:
        type: "list"
        elements: "str"
        required: true
        description: "A list of string values."
        version_added: 1.3.0

      myapp_list_with_dicts:
        type: "list"
        elements: "dict"
        required: false
        default:
          - myapp_food_kind: "meat"
            myapp_food_boiling_required: true
            myapp_food_preparation_time: 60
          - myapp_food_kind: "fruits"
            myapp_food_preparation_time: 5
        description: "A list of dicts with a defined structure and with default a value."
        options:
          myapp_food_kind:
            type: "str"
            choices:
              - "vegetables"
              - "fruits"
              - "grains"
              - "meat"
            required: false
            description: "A string value with a limited list of allowed choices."

          myapp_food_boiling_required:
            type: "bool"
            required: false
            default: false
            description: "Whether the kind of food requires boiling before consumption."

          myapp_food_preparation_time:
            type: int
            required: true
            description: "Time to prepare a dish in minutes."

      myapp_dict_with_suboptions:
        type: "dict"
        required: false
        default:
          myapp_host: "bar.foo"
          myapp_exclude_host: true
          myapp_path: "/etc/myapp"
        description: "A dict with a defined structure and default values."
        options:
          myapp_host:
            type: "str"
            choices:
              - "foo.bar"
              - "bar.foo"
              - "ansible.foo.bar"
            required: true
            description: "A string value with a limited list of allowed choices."

          myapp_exclude_host:
            type: "bool"
            required: true
            description: "A boolean value."

          myapp_path:
            type: "path"
            required: true
            description: "A path value."

          original_name:
            type: list
            elements: "str"
            required: false
            description: "An optional list of string values."

  # roles/myapp/tasks/alternate.yml entry point
  alternate:
    short_description: Alternate entry point for the myapp role
    description:
      - This is the alternate entrypoint for the C(myapp) role.
    version_added: 1.2.0
    options:
      myapp_int:
        type: "int"
        required: false
        default: 1024
        description: "The integer value, defaulting to 1024."

Running a role multiple times in one play 

Ansible only executes each role once in a play, even if you define it multiple times unless the parameters defined on the role are different for each definition. For example, Ansible only runs the role foo once in a play like this:

---
- hosts: webservers
  roles:
    - foo
    - bar
    - foo

You have two options to force Ansible to run a role more than once.

Passing different parameters 

If you pass different parameters in each role definition, Ansible runs the role more than once. Providing different variable values is not the same as passing different role parameters. You must use the roles keyword for this behavior, since import_role and include_role do not accept role parameters.

This play runs the foo role twice:

---
- hosts: webservers
  roles:
    - { role: foo, message: "first" }
    - { role: foo, message: "second" }

This syntax also runs the foo role twice;

---
- hosts: webservers
  roles:
    - role: foo
      message: "first"
    - role: foo
      message: "second"

In these examples, Ansible runs foo twice because each role definition has different parameters.

Using `allow_duplicates: true`

Add allow_duplicates: true to the meta/main.yml file for the role:

# playbook.yml
---
- hosts: webservers
  roles:
    - foo
    - foo

# roles/foo/meta/main.yml
---
allow_duplicates: true

In this example, Ansible runs foo twice because we have explicitly enabled it to do so.

Using role dependencies 

Role dependencies let you automatically pull in other roles when using a role.

Role dependencies are prerequisites, not true dependencies. The roles do not have a parent/child relationship. Ansible loads all listed roles, runs the roles listed under dependencies first, then runs the role that lists them. The play object is the parent of all roles, including roles called by a dependencies list.

Role dependencies are stored in the meta/main.yml file within the role directory. This file should contain a list of roles and parameters to insert before the specified role. For example:

# roles/myapp/meta/main.yml
---
dependencies:
  - role: common
    vars:
      some_parameter: 3
  - role: apache
    vars:
      apache_port: 80
  - role: postgres
    vars:
      dbname: blarg
      other_parameter: 12

Ansible always executes roles listed in dependencies before the role that lists them. Ansible executes this pattern recursively when you use the roles keyword. For example, if you list role foo under roles:, role foo lists role bar under dependencies in its meta/main.yml file, and role bar lists role baz under dependencies in its meta/main.yml, Ansible executes baz, then bar, then foo.

Running role dependencies multiple times in one play 

Ansible treats duplicate role dependencies like duplicate roles listed under roles:: Ansible only executes role dependencies once, even if defined multiple times, unless the parameters, tags, or when clause defined on the role are different for each definition. If two roles in a play both list a third role as a dependency, Ansible only runs that role dependency once, unless you pass different parameters, tags, when clause, or use allow_duplicates: true in the role you want to run multiple times. See Galaxy role dependencies for more details.

Note

Role deduplication does not consult the invocation signature of parent roles. Additionally, when using vars: instead of role params, there is a side effect of changing variable scoping. Using vars: results in those variables being scoped at the play level. In the below example, using vars: would cause n to be defined as 4 throughout the entire play, including roles called before it.

In addition to the above, users should be aware that role de-duplication occurs before variable evaluation. This means that Lazy Evaluation may make seemingly different role invocations equivalently the same, preventing the role from running more than once.

For example, a role named car depends on a role named wheel as follows:

---
dependencies:
  - role: wheel
    n: 1
  - role: wheel
    n: 2
  - role: wheel
    n: 3
  - role: wheel
    n: 4

And the wheel role depends on two roles: tire and brake. The meta/main.yml for wheel would then contain the following:

---
dependencies:
  - role: tire
  - role: brake

And the meta/main.yml for tire and brake would contain the following:

---
allow_duplicates: true

The resulting order of execution would be as follows:

tire(n=1)
brake(n=1)
wheel(n=1)
tire(n=2)
brake(n=2)
wheel(n=2)
...
car

To use allow_duplicates: true with role dependencies, you must specify it for the role listed under dependencies, not for the role that lists it. In the example above, allow_duplicates: true appears in the meta/main.yml of the tire and brake roles. The wheel role does not require allow_duplicates: true, because each instance defined by car uses different parameter values.

Note

See Using Variables for details on how Ansible chooses among variable values defined in different places (variable inheritance and scope). Also, deduplication happens ONLY at the play level, so multiple plays in the same playbook may rerun the roles.

Embedding modules and plugins in roles 

Note

This applies only to standalone roles. Roles in collections do not support plugin embedding; they must use the collection’s plugins structure to distribute plugins.

If you write a custom module (see Should you develop a module?) or a plugin (see Developing plugins), you might wish to distribute it as part of a role. For example, if you write a module that helps configure your company’s internal software, and you want other people in your organization to use this module, but do not want to tell everyone how to configure their Ansible library path, you can include the module in your internal_config role.

To add a module or a plugin to a role: Alongside the ‘tasks’ and ‘handlers’ structure of a role, add a directory named ‘library’ and then include the module directly inside the ‘library’ directory.

Assuming you had this:

roles/
    my_custom_modules/
        library/
            module1
            module2

The module will be usable in the role itself, as well as any roles that are called after this role, as follows:

---
- hosts: webservers
  roles:
    - my_custom_modules
    - some_other_role_using_my_custom_modules
    - yet_another_role_using_my_custom_modules

If necessary, you can also embed a module in a role to modify a module in Ansible’s core distribution. For example, you can use the development version of a particular module before it is released in production releases by copying the module and embedding the copy in a role. Use this approach with caution, as API signatures may change in core components, and this workaround is not guaranteed to work.

The same mechanism can be used to embed and distribute plugins in a role, using the same schema. For example, for a filter plugin:

roles/
    my_custom_filter/
        filter_plugins
            filter1
            filter2

These filters can then be used in a Jinja template in any role called after ‘my_custom_filter’.