Comment 6 for bug 1841620

Revision history for this message
OpenStack Infra (hudson-openstack) wrote : Fix merged to config (master)

Reviewed: https://review.opendev.org/679796
Committed: https://git.openstack.org/cgit/starlingx/config/commit/?id=86abe22b804af6ea569295265adb39c8d15d04f7
Submitter: Zuul
Branch: master

commit 86abe22b804af6ea569295265adb39c8d15d04f7
Author: Stefan Dinescu <email address hidden>
Date: Tue Sep 3 16:31:36 2019 +0300

    Retry puppet runtime apply in case of missing info

    There are cases where sysinv-agents gets runtime puppet
    apply requests before it gets the host_uuid from the
    sysinv-conductor.

    This causes the received request to be dropped because the
    agent thinks the command is not for itself.

    The fix for this is to expand the retry logic of the
    config_apply_runtime_manifest method to retry if any
    of the needed inventory info is needed.

    Tests done:
    - install of an SX setup and 10 lock-unlocks
    - install of a 2+2 setup and 2 lock-unlocks on each node

    Change-Id: I130e73881bd572398072aff7c46b3b7f165ca6b9
    Closes-bug: 1841620
    Signed-off-by: Stefan Dinescu <email address hidden>