To Reproduce Steps to reproduce the behavior:

1. Upgrade RHOSP 16.2 to 17.1
2. During the `overcloud upgrade run ...` command, the galera container fails to start after the DB is upgraded
3. After a while, a different step in the upgrade process brings the container up and the upgrade seems to finish without error

Expected behavior

• <your text here>

Bug impact

• The time that MariaDB is down is proportional to the number of compute nodes in the environment. The more computes, the longer it takes to complete the steps before it reaches the point in which MariaDB is started again (which I believe is step 4)

Known workaround

• Unknown

Additional context

• • rhn-engineering-dciabrin , rhn-engineering-lbezdick , rhn-support-lmiccini working on it