Multi Master replication - table locks
Hello all,
We have an older (oracle 7) multi master replication site that has recently decided not to replicate anymore. After some investigation I've found that the execution job (scheduled with DBMS_DEFER_SYS.SCHEDULE_EXECUTION) is getting stuck and holding a lock on one of the replication tables (DEF$_CALLDEST I believe, but I don't have it in front of me). The only way to clear this seems to be to delete the replication queue. This will cause it to unstick for a while (a minute or two) but then it gets wedged again.
Anyone seen anything like this before, or have suggestions about where to start looking?
Thanks,
-Merit