Ensure we only process snapshot clusters if the sending connection has control / owns the object. Update refresh message snapshot id check to use latest logic.