I just want to run the slack module if my playbook fails. Right now I have it reporting that a deployment starts and finishes successfully.

I learned I can wrap tasks in block/rescue, but I've got 3 roles and each role is broken down to a few includes (like my webserver role has dependencies, configuration, ssl, etc all broken up).

The way we do this is to use the 'include_role' task instead of the 'roles' directive. Since 'include_role' is a task, you can properly wrap it in a block/rescue/always like so:

- name: Run tests
  hosts: nodes
  force_handlers: True
  tasks:
    - block:
        - include_role:
            name: "run-tests"
        - include_role:
            name: "logs"
            tasks_from: "recover"
      rescue:
        - include_role:
            name: "logs"
            tasks_from: "recover"
        - name: Failure detected in run-tests role
          fail:
            msg: "Failure detected in run-tests role"

[–]senorsmile • 2 points 8 months ago 

• permalink
• embed
• save
• parent
• give gold

I would also like to know how to do this; We are using ansible to create a VM shell and deploy Linux/Windows Servers.

If any of the Tasks fail we just want a cleanup script to run to remove the VM, AD accounts and DNS entries We already do this but it has to be done on a role-by-role basis would be nice if i can just wrap this around my entire playbook.