It's possible that can happen as well. When this happens,
propagated to the consumer through the master anyway, since that
master is receiving
updates from the other master(s).
I wonder if it is because we had a network blip between sites and there
was an update/lock at the time so when the network came back up and the
master tried to send updates, there was already a stale lock? It's
actually multi-master replication - there are two masters and it's one
of the masters which did this.
Did the replication monitor documentation help at all?
The documentation on the repl-monitor script wasn't detailed enough
perhaps? I tried getting this to work but couldn't. I tried every
setting in the config file I could to no avail. It seems to be
hard-coded to port 389 and I am running SSL (well, TLS).
In any case, I really need something like SNMP alerts when there are
replication problems. I suppose I can look for the string
"NSMMReplicationPlugin" in the error log ...