Delay in receiving replicated updates in my ASync Replicated

[spikalan]

My current setup is that I have 10 caches setup in REPL_ASYNC mode with a replication queue trigger on 1000ms interval. On occasion, some cache are not late to receive the changes in the cache. The delay can be significant, in the order of minutes even when the transmitters have ceased to call their put() method.

I noticed that when logging in debug, once the pbcast.STABLE schedules task run, everything is updated. Is there a way to increase the frequency of this scheduled in order to minimize the delay to an acceptable time.

I am user the AOP JDK50 cache running in a SPRING container.

Thanks in advance for your help.

[brian.stansberry]

See "desired_avg_gossip" config attribute on http://www.jboss.org/community/docs/DOC-10902.

For JGroups discussion of what may be the root cause of what you are seeing:

http://www.jboss.org/community/docs/DOC-11244

[spikalan]

I reduced the value from 20000 to 2. I checked and this should ensure that the gossip will be 10000. It seems to have fixed the problem. Now I can see that the caches are more quickly synchronized.

Thanks for the quick reply.