I've enabled trace flag 1800 on all replicas now and I'm still receiving this error. I did not restart the SQL service though. I place the trace flag in the start up parameters and issues a DBCC TRACEON (1800, -1) command. I believe this should do it. Would you agree that restarting the service should not be necessary?