Comment 8 for bug 1548271

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

Reviewed: https://review.openstack.org/296712
Committed: https://git.openstack.org/cgit/openstack/fuel-library/commit/?id=0318216c0539d7d0b7ed38df51c11711af70dbcf
Submitter: Jenkins
Branch: master

commit 0318216c0539d7d0b7ed38df51c11711af70dbcf
Author: Alex Schultz <email address hidden>
Date: Wed Mar 23 13:22:58 2016 -0600

    Update keystone db task to wait for all databases

    This change updates the keystone db task to wait for the database tasks
    on all nodes to complete before getting run. I believe that this is
    causing the galera cluster to become corrupt because the following
    condition occurs:

    - primary-db comes up
    - secondary-db comes up and syncs from primary-db
    - teriary-db comes up and syncs from secondary-db while keystone task
      runs on primary-db causing cluster to become desynced
    - primary-db ends up trying to sync from the cluster and everything
      falls apart

    The idea is to wait until the cluster has completely formed before
    attempting to start openstack deployment. Additionally, this change
    removes the one-by-one on the secondary database tasks which was added
    as an attempt to address the cluster creation issues.

    Change-Id: I4d62728922bb3a74576c419a934c1afd9187a35d
    Closes-Bug: #1548271