Playbooks

Ansible Playbooks offer a repeatable, reusable, simple configuration management and multi-machine deployment system. They are files that allow to define a set of instructions to execute to reach a desired state. Sames as the [ad-hoc]({< ref “ansible/execution/adhoc” >}}) commands, they make use of modules to execute things.

They are written in YAML format and run with the ansible-playbook binary

ansible-playbook -K my_plabook.yml

An example of a simple playbook can be seen below:

- name: Install apache # Name of the whole playbook
  hosts: all # Running on all nodes
  become: true # Same as --become, run this as root
  tasks:
  - name: install apache2 package # Name of the task
    apt: # Using the builtin apt module
      name: apache2 # Name of the package to install
      update_cache: true # Before installing run apt update

More modules, and options can be added to achieve complex behaviors.

Running a playbook

When running a playbook, ansible first always runs a gather_facts task to gain information on the servers. After running, the results will be summarized in the PLAY RECAP, for example:

ok=2    changed=1    unreachable=0    failed=0    skipped=0    rescued=0    ignored=0  

This gives an overview of what the command achieved and wheter or not it succeded

• ok shows the tasks that did not fail (always 1 plus because of the gather_facts tasks)
• changed shows the tasks that actually changed something in the server (Running an install task twice will only change once)
• unreachable will report if the server could not be reached while running the task
• skipped will make sense for conditional tasks (that run depending if something is true)