I believe that before processing events coming from a remote node we should acquire a lock on the global configuration lock to make sure we don't mess up with the catalog on that same instance via REST or GUI.
Actually it might be worth (side note) to investigate extending the Config Lock to use Hazelcast distributed locks.
I believe that before processing events coming from a remote node we should acquire a lock on the global configuration lock to make sure we don't mess up with the catalog on that same instance via REST or GUI. Actually it might be worth (side note) to investigate extending the Config Lock to use Hazelcast distributed locks.