hard failure or 'checkpoint' type task

Archives Ansible Project
1

I’m running a deployment playbook (stop, rollout, start, post-deploy tasks).

There are times when the whole playbook should abort if any tasks fail (e.g. not all hosts stop).

Is there an easy way to achieve this - what’s happening at the moment is the tasks fail on certain hosts and then the playbook continues using only the non-failed hosts.

I thought of two theoretical approaches - a global ‘fail-hard’ setting where any failure just stops the playbook, or a checkpoint task that only succeeds if all hosts are still in play. I’m sure other approaches are valid.

Is there anything that would solve this currently?

Thanks,
Will

2

There is already a new feature in 1.2 called “any_errors_fatal: True” (an attribute of the play)

I’m fine with there also being an action plugin capable of creating global failures, though currently “fail” will need to fail everything in a play (or a serial batch size) without the any_errors_fatal parameter set.

3

Awesome, any_errors_fatal sounds exactly what I need - I don’t think I need a failure level greater than play.

Thanks!