37

I'm working on several Ansible playbooks to spin up a new server instance. There are approximately 15 different playbooks I need to run in a specific order to successfully spin up a server.

My initial thought was to write a shell script that executes ansible-playbook playbook_name.yml and duplicate it one entry for each playbook I need to run.

Is there a smarter/better way to do this using a master playbook and if so what would it look like (examples are appreciated).

I could write one monolithic playbook that does it all but there are some plays that run as root first then as a sudo user later.

Improve this question
2
  • 2
    use include in your main playbook docs.ansible.com/ansible/playbooks_roles.html
    – c4f4t0r
    Jan 21, 2016 at 20:13
  • 1
    To handle the case of running as root then as a sudo user, you can use the block feature - put the become: part at end of each block. You might need to create a new play to switch connection user from root to sudo user though.
    – RichVel
    Oct 3, 2016 at 6:48

3 Answers 3

Reset to default
31

Build many sub-playbooks and aggregate them via include statements.

- include: playbook-one.yml
- include: playbook-two.yml

If your playbooks must run in order and if all of them are mandatory, build a main playbook and include files with tasks. A playbook should always be a closed process.

Improve this answer
4
  • 6
    could you elaborate on the last part of your answer? what do you mean by "A playbook should always be a closed process." ?
    – Mike Vella
    Dec 1, 2016 at 13:31
  • 1
    Is there a way to specify all playbooks using regular expression? For Example: - include : books/*.yml
    – Chenna V
    Jan 20, 2017 at 21:04
  • 13
    Looks like include is deprecated. docs.ansible.com/ansible/latest/playbooks_reuse.html I think import_playbook: foo is the right way to go, but I'm not super experienced.
    – Andrew
    Sep 12, 2017 at 18:48
  • 2
    @MikeVella "A playbook should always be a closed process" essentially means that each playbook should be unto itself and should be independently complete. Nothing in playbook-a should be reliant or necessary from playbook-b. Keep it all as together as possible.
    – jnovack
    Aug 10, 2020 at 20:32
34

For newer versions of Ansible, you can build many sub-playbooks and aggregate them via import_playbook statements:

---
- import_playbook: A-systemd-networkd.yml
- import_playbook: B-fail2ban-ssh.yml
- import_playbook: C-enable-watchdog.yml
Improve this answer
1
  • By any chance, do you know how to continue with B if A fails? A way other than setting ignore_errors: yes for every task in A ?
    – Tag Wint
    Dec 28, 2020 at 20:31
3

From: https://docs.ansible.com/ansible/latest/collections/ansible/builtin/import_playbook_module.html

- hosts: localhost
  tasks:
    - debug:
        msg: play1

- name: Include a play after another play
  import_playbook: otherplays.yaml
Improve this answer

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .