Race sometimes causes successful deployments to return failed

[Steve Baker]

These lines[1] have a race condition where the fetched execution is state:SUCCESS but the last payload is status:RUNNING. This leads to the status:SUCCESS payload never being returned to the caller, and the deployment failing even though it succeeded.

Even though triggering this seems to be rare, it does happen. And the same issue appears to go back to Queens.

[1] https://opendev.org/openstack/python-tripleoclient/src/branch/master/tripleoclient/workflows/base.py#L94-L95

[OpenStack Infra (hudson-openstack)]

Reviewed: https://review.opendev.org/679873
Committed: https://git.openstack.org/cgit/openstack/python-tripleoclient/commit/?id=b5b5cab61da98c8bcf2c4e52a6d8ce0108dcfc64
Submitter: Zuul
Branch: master

commit b5b5cab61da98c8bcf2c4e52a6d8ce0108dcfc64
Author: Steve Baker <email address hidden>
Date: Wed Sep 4 10:09:04 2019 +1200

    Fix race in execution finishing

    An execution state can go from RUNNING to SUCCESS between fetching the
    last message from the websocket and polling the execution state. This
    means the SUCCESS payload is never returned and the overcloud
    deployment fails at the end with no indication as to why.

    This change turns the output of the execution into the last payload,
    allowing the calling SUCCESS logic to run.

    Change-Id: Ic22021ba9a2717de199629e361c656e2f562fb38
    Closes-Bug: #1842987

[OpenStack Infra (hudson-openstack)]

Fix proposed to branch: stable/stein
Review: https://review.opendev.org/681243

[OpenStack Infra (hudson-openstack)]

Reviewed: https://review.opendev.org/681243
Committed: https://git.openstack.org/cgit/openstack/python-tripleoclient/commit/?id=b0c714e3226b585751988cefdbed9a84e4df0599
Submitter: Zuul
Branch: stable/stein

commit b0c714e3226b585751988cefdbed9a84e4df0599
Author: Steve Baker <email address hidden>
Date: Wed Sep 4 10:09:04 2019 +1200

    Fix race in execution finishing

    An execution state can go from RUNNING to SUCCESS between fetching the
    last message from the websocket and polling the execution state. This
    means the SUCCESS payload is never returned and the overcloud
    deployment fails at the end with no indication as to why.

    This change turns the output of the execution into the last payload,
    allowing the calling SUCCESS logic to run.

    Change-Id: Ic22021ba9a2717de199629e361c656e2f562fb38
    Closes-Bug: #1842987
    (cherry picked from commit b5b5cab61da98c8bcf2c4e52a6d8ce0108dcfc64)

[OpenStack Infra (hudson-openstack)]

This issue was fixed in the openstack/python-tripleoclient 12.2.0 release.

[OpenStack Infra (hudson-openstack)]

This issue was fixed in the openstack/python-tripleoclient 11.5.2 release.