G
George Potemkin
Guest
Earlier mentioned knowledgebase article "Clarification of the -pica replication enabled database startup parameter" knowledgebase.progress.com/.../P104638 says: "There is a PICA queue on both the source and target databases" Should the -pica parameters be the same on source and target databases or should we tune them individually? These databases are running on a different hardware and under a different load. I guess that in the customer's case I posted above the performance degradation on source db was caused by slowness of replication agent on target db. Theory: replication server was spending a lot of time waiting for the acknowledgments from replication agent that in turn resulted in the filled pica queue on source db => source's APW was waiting for pica access => the updates on source db was very slow. Is it a reasonable explanation? Will we also get the negative impact on source database if the pica queue is filled on target database for any reasons? Will it cause the filled pica queue also on source db as well?
Continue reading...
Continue reading...