drivegroups: replacement op will ignore existing wal/dbs
Since the db/wal is considered "locked/non-avialable" by ceph-volume after the first deployment, the DriveGroup algorithm excludes it from further operations.
We would need a more descriptive label for wals and dbs similar to as requested here: https://tracker.ceph.com/issues/44750
Also, if a db_device field is specified in the drivegroup specs, but the lookup doesn't find a valid db/wal device the deployment should be considered failed
If that works, we're still hitting https://tracker.ceph.com/issues/44749