qa/suites/orch/cephadm/smb: retry CTDB healthy checks via workunit

Replace one-shot ctdb status greps with smb.workunit ctdb_healthy so
tests wait for all three nodes OK after orch reports running. Apply the
same pattern to the ctdb_res suites, and make ctdb_healthy check every
cluster_id so multi-cluster ports2c covers both CTDB groups.

Fixes: https://tracker.ceph.com/issues/79326
Signed-off-by: Yael Azulay <yazulay@redhat.com>