Add any_errors_fatal to openstack playbooks

[tzumainn]

No description provided.

[tomassedovic]

/ok-to-test

[tomassedovic]

Awesome, thanks! This is not necessary for the localhost jobs, only for the multihost plays, but there's no harm in leaving it there.

Anyway, it works great! Needs a rebase but we can merge it afterwards. Note that some of the plays have now moved under install and they still require any_errors_fatal, too.

[tzumainn]

Thanks for the review! I've rebased and added any_errors_fatal to the install playbook.

[tomassedovic]

So sadly, when I tested this, it didn't seem to work. But there's apparently a few ansible bugs about that, so I'm okay to merge this.

For the record, here's how I'm testing this. I've added the following to the provision playbook, right after the Populate DNS entries task:

  - os_server:
      name: app-node-0.openshift.example.com
      state: absent

It will delete an app server (the name is hardcoded so if your env is different, change this). The play should abort during the TASK [Detecting Operating System from ostree_booted]. In my testing, it continued way after that just skipping tasks related to the app node.

Apparently though, it's possible to enable this option globally in ansible.cfg as well:

any_errors_fatal = True

And that seem so be working fine. Would you mind adding that to the readme? Something along the lines of "It is recommended that you put any_errors_fatal = True which will abort the Ansible execution as soon as any error is encountered."

[tzumainn]

I actually was going to just document the ansible.cfg change and leave it at that, but when I was testing - I did so by removing a nova server after the heat stack was created - the ansible.cfg change did nothing, while my changes worked.

any_errors_fatal only seems to work on the specific tasks I threw it on, so perhaps the reason your test didn't work is because you were expecting it to fail in the playbooks/init/basic_facts.yml playbook?

[tomassedovic]

Interesting. Maybe our Ansible versions are different. The changes in this PR did nothing at all to me -- as if they weren't there. While the ansible.cfg version worked perfectly.

So let's do both.

[tzumainn]

Fair enough! Out of curiosity, what's your ansible version?

[tomassedovic]

@tzumainn: ansible 2.4.1.0

[tzumainn]

@tomassedovic I'm using 2.4.2.0. Now I'm really confused...

[tomassedovic]

Hm, so I updated to 2.4.2.0 but I keep seeing the same thing: ansible.cfg stops the play right after the failure, whereas without it, this patch keeps going for ages afterwards.

[tzumainn]

@tomassedovic Okay, I've added a note to the README. I guess one of the billion options here will work, although it might be a different one for each person.

[tomassedovic]

/lgtm

[tzumainn]

/test gcp

[tzumainn]

/test gcp

[tzumainn]

/test gcp

[tzumainn]

/test install

[tzumainn]

/test containerized

[openshift-merge-robot]

/test all [submit-queue is verifying that this PR is safe to merge]

[tzumainn]

/test install

[openshift-merge-robot]

/test all [submit-queue is verifying that this PR is safe to merge]

[tzumainn]

/test gcp

[tzumainn]

/retest

[tzumainn]

/retest

[openshift-merge-robot]

/test all [submit-queue is verifying that this PR is safe to merge]

[tzumainn]

/retest

[openshift-merge-robot]

/test all [submit-queue is verifying that this PR is safe to merge]

[tzumainn]

/test install

[tzumainn]

/test install

[tzumainn]

/test install

[tzumainn]

/test install

[tzumainn]

/test install

[openshift-merge-robot]

/test all [submit-queue is verifying that this PR is safe to merge]

[openshift-merge-robot]

Automatic merge from submit-queue.

[openshift-ci-robot]

@tzumainn: The following tests failed, say /retest to rerun them all:

Test name	Commit	Details	Rerun command
ci/openshift-jenkins/system-containers	bb20e09	link	/test system-containers
ci/openshift-jenkins/containerized	bb20e09	link	/test containerized
ci/openshift-jenkins/extended_conformance_install_crio	bb20e09	link	/test crio
ci/openshift-jenkins/logging	bb20e09	link	/test logging

Full PR test history. Your PR dashboard. Please help us cut down on flakes by linking to an open issue when you hit one in your PR.

Instructions for interacting with me using PR comments are available here. If you have questions or suggestions related to my behavior, please file an issue against the kubernetes/test-infra repository. I understand the commands that are listed here.