Multimaster asynchronous serial/parallel replication propagation
I have oracle 10g and 2masters with parallel propagation (PUSH ...delay_seconds=>0, parallelism=>1) configured. From time to time replication stops (PUSH job locks some resources) and after some time DB gets unsynchronized. To start it back I do kill-9 p00x. I did not determined a reason yet, but my support advice me to change PUSH job to parallelism=0 (serial propagation) to avoid using p00x processes. I must also say that DB is small one.
Was this right decision? Does anybody know if there is a bug in p00x ? Or perhaps someone knows what is a most critical task in parallel propagation that can lock resources forever.