ansible-role-design

Ansible Role Design

Production-grade role structure patterns derived from analysis of 7 geerlingguy roles.

Standard Directory Structure

Every Ansible role follows this organizational pattern:

role-name/
├── defaults/
│   └── main.yml          # User-configurable defaults (lowest precedence)
├── vars/
│   ├── Debian.yml        # OS-specific internal values
│   └── RedHat.yml
├── tasks/
│   ├── main.yml          # Task router
│   ├── install.yml       # Feature-specific tasks
│   └── configure.yml
├── handlers/
│   └── main.yml          # Event-triggered tasks
├── templates/
│   └── config.conf.j2    # Jinja2 templates
├── files/
│   └── static-file.txt   # Static files
├── meta/
│   └── main.yml          # Role metadata, dependencies
└── README.md             # Documentation

Directory Purposes

Directory	Purpose	Precedence
defaults/	User-overridable values	Lowest
vars/	Internal/OS-specific values	High
tasks/	Ansible tasks	N/A
handlers/	Service restarts, reloads	N/A
templates/	Jinja2 config files	N/A
files/	Static files to copy	N/A
meta/	Galaxy info, dependencies	N/A

When to Omit Directories

Only create directories that are actually needed:

• Omit templates/ if using only lineinfile or copy
• Omit handlers/ if role doesn't manage services
• Omit vars/ if no OS-specific differences
• Omit files/ if no static files to copy

Task Organization

Main Task File as Router

Use tasks/main.yml as a routing file that includes feature-specific files:

# tasks/main.yml
---
- name: Include OS-specific variables
  ansible.builtin.include_vars: "{{ ansible_os_family }}.yml"

- name: Install packages
  ansible.builtin.include_tasks: install.yml

- name: Configure service
  ansible.builtin.include_tasks: configure.yml

- name: Setup users
  ansible.builtin.include_tasks: users.yml
  when: role_users | length > 0

When to Split Tasks

Scenario	Approach
< 30 lines	Keep in main.yml
30-100 lines	Consider splitting
> 100 lines	Definitely split
Optional features	Separate file with when:
OS-specific logic	Separate files per OS

Task File Naming

Use descriptive, feature-based names:

tasks/
├── main.yml              # Router only
├── install.yml           # Package installation
├── configure.yml         # Configuration tasks
├── users.yml             # User management
├── install-Debian.yml    # Debian-specific install
└── install-RedHat.yml    # RedHat-specific install

Variable Organization

defaults/ vs vars/

Location	Purpose	User Override?
defaults/main.yml	User configuration	Yes (easily)
vars/main.yml	Internal constants	Possible but discouraged
vars/Debian.yml	OS-specific values	No (internal)

defaults/main.yml Example

# defaults/main.yml
---
# User-configurable options
docker_edition: "ce"
docker_service_state: started
docker_service_enabled: true
docker_users: []

# Feature toggles
docker_install_compose: true
docker_compose_version: "2.24.0"

vars/Debian.yml Example

# vars/Debian.yml
---
# OS-specific internal values (not for user override)
docker_package_name: docker-ce
docker_service_name: docker
docker_config_path: /etc/docker/daemon.json

Loading OS-Specific Variables

Simple pattern:

- name: Include OS-specific variables
  ansible.builtin.include_vars: "{{ ansible_os_family }}.yml"

Advanced pattern with fallback:

- name: Load OS-specific vars
  ansible.builtin.include_vars: "{{ lookup('first_found', params) }}"
  vars:
    params:
      files:
        - "{{ ansible_distribution }}.yml"
        - "{{ ansible_os_family }}.yml"
        - main.yml
      paths:
        - vars

Variable Naming Convention

Prefix variables with role name:

# Pattern: {role_name}_{feature}_{attribute}

# Examples
docker_edition: "ce"
docker_service_state: started
docker_compose_version: "2.24.0"
docker_users: []

# Grouped by feature
security_ssh_port: 22
security_ssh_password_auth: "no"
security_fail2ban_enabled: true

Benefits

• Prevents conflicts with other roles
• Clear ownership of variables
• Easy to grep across codebase
• Self-documenting

Handler Patterns

Simple Handler Definitions

# handlers/main.yml
---
- name: restart docker
  ansible.builtin.systemd:
    name: docker
    state: restarted

- name: reload nginx
  ansible.builtin.systemd:
    name: nginx
    state: reloaded

Handler Naming

Use lowercase with action + service pattern:

- name: restart ssh      # Not "Restart SSH Service"
- name: reload nginx     # Not "Reload Nginx Config"
- name: reload systemd   # For daemon-reload

Throttled Handlers

For cluster operations, restart one node at a time:

- name: restart pve-cluster
  ansible.builtin.systemd:
    name: pve-cluster
    state: restarted
  throttle: 1

Template Organization

When to Use Templates

Use templates/ when:

• Configuration has conditional content
• Need variable substitution
• Complex multi-line configuration
• Users may need to extend/override

Use lineinfile when:

• Simple single-line changes
• Modifying existing system files

Template Variables

Expose template paths as variables for user override:

# defaults/main.yml
nginx_conf_template: nginx.conf.j2
nginx_vhost_template: vhost.j2

# tasks/configure.yml
- name: Deploy nginx config
  ansible.builtin.template:
    src: "{{ nginx_conf_template }}"
    dest: /etc/nginx/nginx.conf
  notify: reload nginx

Meta Configuration

meta/main.yml Structure

# meta/main.yml
---
galaxy_info:
  author: your_name
  description: Role description
  license: MIT
  min_ansible_version: "2.12"
  platforms:
    - name: Debian
      versions:
        - bullseye
        - bookworm
    - name: Ubuntu
      versions:
        - focal
        - jammy

dependencies:
  - role: common
  - role: geerlingguy.docker
    when: install_docker | default(false)

Role Complexity Scaling

Based on geerlingguy role analysis:

Role Complexity	Directories	Task Files	Examples
Minimal	3-4	1 (main.yml)	pip, git
Standard	5-6	2-4	security, docker
Complex	7+	5-8	postgresql, nginx

Minimal Role

pip/
├── defaults/main.yml
├── tasks/main.yml
├── meta/main.yml
└── README.md

Standard Role

docker/
├── defaults/main.yml
├── vars/{Debian,RedHat}.yml
├── tasks/{main,install,configure}.yml
├── handlers/main.yml
├── meta/main.yml
└── README.md

Complex Role

postgresql/
├── defaults/main.yml
├── vars/{Debian,RedHat,Archlinux}.yml
├── tasks/{main,install,configure,users,databases}.yml
├── handlers/main.yml
├── templates/{postgresql.conf,pg_hba.conf}.j2
├── meta/main.yml
└── README.md

Task Naming Convention

Start task names with action verbs:

# GOOD
- name: Ensure Docker is installed
- name: Configure SSH security settings
- name: Add user to docker group

# BAD
- name: Docker installation
- name: SSH settings
- name: User docker group

File Validation

Validate critical configuration files:

- name: Update SSH configuration
  ansible.builtin.lineinfile:
    path: /etc/ssh/sshd_config
    regexp: "^PermitRootLogin"
    line: "PermitRootLogin no"
    validate: 'sshd -T -f %s'
  notify: restart ssh

- name: Update sudoers
  ansible.builtin.lineinfile:
    path: /etc/sudoers
    line: "{{ user }} ALL=(ALL) NOPASSWD: ALL"
    validate: 'visudo -cf %s'

Documentation

Every role needs a README.md with:

1. Description - What the role does
2. Requirements - Prerequisites
3. Role Variables - All variables with defaults
4. Dependencies - Other roles needed
5. Example Playbook - How to use it

Additional Resources

For detailed role design patterns and techniques, consult:

• references/role-structure-standards.md - Production role structure patterns from geerlingguy analysis
• references/handler-best-practices.md - Handler design, notification patterns, flush strategies
• references/meta-dependencies.md - Role dependencies, Galaxy metadata, platform support
• references/variable-management-patterns.md - Variable naming, scoping, precedence patterns
• references/documentation-templates.md - README templates and documentation standards

Related Skills

• ansible-playbook-design - When to use roles vs playbooks
• ansible-fundamentals - Module selection and naming
• ansible-testing - Role testing with molecule