Galera databases/operator setup still produce DB locking issues

[BjoernT]

Describe the bug The way Galera is currently deployed still appears to favor locking issues due to either load balancing issues (not active/passive) or issues with maxscale itself.

To Reproduce Steps to reproduce the behavior:

• Keystone DB sync typically encounters that first as issue where the db schema is not fully laid down and upon retires creates SQL errors that the tables are already present
• If the openstack services deploy, rally standard tests do not finish such as OSPC-343

Expected behavior

DB syncs to finish on first pass and rally tests to suceeds

Screenshots

N/A

Additional context N/A