Previously, extra restarts were added after setting MySQL root password during setup phase, OpenStack database users were added on nodes where they already existed and cluster status check occurred right before adding Galera to non-bootstrap nodes and success was declared prematurely. As a result, Galera failures occurred on some or all nodes.
With this update, no extra restart is added after setting the root MySQL password. OpenStack database users are created only on the bootstrap node (Galera replicates them to other nodes). This update also ensures that pacemaker has been added before the final 'clustercheck' on each node to confirm the node is a good start. As a result, pacemaker installation is reliable and stable.
Description of problem:
Occasionally, the phase during an HA puppet run where galera is installed, a node will to bring up galera, causing errors for the rest of the puppet run for that node and causing a staypuft deployment to fail.
Comment 3Alexander Chuzhoy
2014-10-07 21:30:32 UTC
Created attachment 944768[details]
messages and mariadb.log from all controllers in the affected setup (bare metal)
Since the problem described in this bug report should be
resolved in a recent advisory, it has been closed with a
resolution of ERRATA.
For information on the advisory, and where to find the updated
files, follow the link below.
If the solution does not work for you, open a new bug report.
https://rhn.redhat.com/errata/RHBA-2014-1800.html