The Ansible website describes it as:

Ansible is a simple automation language that can perfectly describe an IT application infrastructure

First install pip if you have not already installed it

sudo easy_install pip

Then install the latest version of Ansible

sudo pip install ansible

To try the examples of this tutorial please start the vagrant boxes:

vagrant up

A playbook is Ansibles scripting language for orchestrating and configuring servers.

A playbook is in a yaml format making it easy to read and understand.

- hosts: localhost
  tasks:
    - debug: msg='hello world'

We can run ```hello_world.yml` playbook with a simple command

ansible-playbook hello_world.yml

When you run the playbook you will see that it runs all the tasks in the playbook. After all the tasks have been completed it will print out a summary of the results.

PLAY [localhost] ***************************************************************

TASK [setup] *******************************************************************
ok: [localhost]

TASK [debug] *******************************************************************
ok: [localhost] => {
    "msg": "hello world"
}

PLAY RECAP *********************************************************************
localhost                  : ok=2    changed=0    unreachable=0    failed=0  

An inventory describes all your servers in your infrastructure.

An inventory is in the ini format which contains a heading to group a list of servers and then their host names:

[frontend]
192.168.0.32
192.168.0.33
192.168.0.34

[backend]
backend-server-1
backend-server-2

[database]
database-1 ansible_host=192.168.0.128

You can then set the hosts value in your playbook to run all the commands on those servers:

- hosts: backend
  tasks:
    - debug: msg='hello world'

You can then run the playbook by specifying the inventory file to use:

ansible-playbook \
    -i inventory/inventory.ini \
    hello_world_frontend.yml

To use Ansible with Vagrant you need to set ansible_user and ansible_ssh_private_key_file in the inventory configuration

[frontend]
192.168.33.31 ansible_user=vagrant ansible_ssh_private_key_file=.vagrant/machines/machine1/virtualbox/private_key
192.168.33.32 ansible_user=vagrant ansible_ssh_private_key_file=.vagrant/machines/machine2/virtualbox/private_key

To perform tasks that requires sudo you need to pass in the become option.

- hosts: frontend
  become: true

You can run individual tasks with sudo as well by adding it to the task

- hosts: frontend
  tasks:
    - name: Install Nano
      become: true
      yum:
        name: nano
      state: present

We can install a package on all of our servers by setting hosts to all.

- hosts: all
  become: true
  tasks:
    - name: Install Nano
      yum:
        name: nano
        state: present

We can run this example to install a package on all our servers

ansible-playbook \
    -i inventory/inventory.ini \
    install_server_software.yml

You will then see that the server packages are installed on all the servers.

PLAY [all] *********************************************************************

TASK [setup] *******************************************************************
ok: [frontend-2]
ok: [backend-1]
ok: [frontend-1]

TASK [Install Nano] ************************************************************
changed: [frontend-1]
changed: [backend-1]
changed: [frontend-2]

PLAY RECAP *********************************************************************
backend-1                  : ok=2    changed=1    unreachable=0    failed=0   
frontend-1                 : ok=2    changed=1    unreachable=0    failed=0   
frontend-2                 : ok=2    changed=1    unreachable=0    failed=0

We can select groups to run the playbooks in by seeing the host to the group name.

- hosts: frontend
  become: true
  tasks:
    - name: Install HTTPD
      yum:
        name: httpd
        state: present

You can run this with:

ansible-playbook \
    -i inventory/inventory.ini \
    install_http_server.yml 

You will see that it will only install httpd on the frontend servers.

PLAY [frontend] ****************************************************************

TASK [setup] *******************************************************************
ok: [frontend-2]
ok: [frontend-1]

TASK [Install HTTPD] ***********************************************************
changed: [frontend-1]
changed: [frontend-2]

PLAY RECAP *********************************************************************
frontend-1                 : ok=2    changed=1    unreachable=0    failed=0   
frontend-2                 : ok=2    changed=1    unreachable=0    failed=0

Ansible supports variables to make your scripts more portable. We can use the vars to do this:

- hosts: frontend
  become: true
  vars:
    webserver: httpd
    welcome:
      dest: /var/www/html/index.html
      content: 'hello world'
    packages:
      - nano
      - wget

We can use them to install the webserver:

- name: Install Webserver
  yum:
    name: "{{ webserver }}"
    state: present

We can also access variables at different levels:

- name: Insert a welcome page
  copy:
    dest: "{{ welcome.dest }}"
    content: "{{ welcome.content }}"

We can access an array of variables

- name: Install Server Packages
  yum:
    name: "{{ item }}"
    state: present
  with_items: "{{ packages }}"

ansible-playbook \
    -i inventory/inventory.ini \
    configure_server.yml

Now we can view the hello world page running on our webserver.

We can provide configuration by passing a listing of files

- hosts: all
  become: true
  vars_files:
    - vars/common.yml

We can now create vars/common.yml with the following content

webserver: httpd
welcome:
  dest: /var/www/html/index.html
  content: |
    <html>
    <title>Ansible</title>
    <body>
    <h1>Hello World From Ansible</h1>
    </body>
    </html>
packages:
  - nano
  - wget

Handlers enable us to perform operations only on a change.

handlers:
  - name: Start Webserver
    service:
      name: "{{ webserver }}"
      enabled: yes
      state: started

We can call this handler with notify with:

- name: Install Webserver
  yum:
    name: "{{ webserver }}"
    state: present
  notify: Start Webserver

When the playbook is completed the handlers will be run if they are needed. The handlers will be run once so we can call notify as many times as we want.

Roles can be used to organise your playbooks better

- hosts: all
  become: true
  vars_files:
    - vars/common.yml
  roles:
    - common
    - httpd
    - php

We can now structure our project like below:

root
├── build_server.yml
├── inventory
│   └── inventory.ini
├── roles
│   ├── common
│   │   └── tasks
│   │       └── main.yml
│   ├── httpd
│   │   ├── tasks
│   │   │   └── main.yml
│   │   └── handlers
│   │       └── main.yml
│   └── php
│       └── tasks
│           └── main.yml
└── vars
    └── common.yml

In our roles/common/httpd/tasks/main.yml we can add tasks only related to httpd. We would structure our main.yml like below:

- name: Install Webserver
  yum:
    name: "{{ webserver }}"
    state: present
  notify: Start Webserver

- name: Insert a welcome page
  copy:
    dest: "{{ welcome.dest }}"
    content: "{{ welcome.content }}"

We can then update the handler in roles/common/httpd/handlers/main.yml with:

- name: Start Webserver
  service:
    name: "{{ webserver }}"
    state: started

We can then run the ansible playbook

ansible-playbook \                                                                                                                                13:03:26  ☁  master ☂ ➜ ⚡ ✚
    -i inventory/inventory.ini \
    build_server.yml

Our playbook will run each role and we can reuse those roles in other playbooks.

Explain here how Ansible Vault can be used to keep our configuration secure

After the tutorial we can reset the vagrant boxes back to their original state

ansible-playbook \
    -i inventory/inventory.ini \
    reset.yml

Or destroy them with:

vagrant destroy