]
Dan Berindei commented on ISPN-1523:
------------------------------------
Manik, it's not really incorrect, it's just suboptimal (I think) because each
requestor gets the invalidation command numOwner times.
I'm not sure how big the performance difference would be if only the primary owner
sent the invalidation command. I only noticed it because I had an exception breakpoint set
for TimeoutException and one of the invalidation commands would hit my breakpoint (because
the invalidation command tries to acquire the lock with 0 timeout).
Remote nodes send duplicate invalidation messages
-------------------------------------------------
Key: ISPN-1523
URL:
https://issues.jboss.org/browse/ISPN-1523
Project: Infinispan
Issue Type: Enhancement
Components: Distributed Cache
Reporter: Dan Berindei
Assignee: Pete Muir
I though only the originator should send invalidation messages, but I'm seeing these
messages in the log:
{noformat}
2011-11-11 11:10:27,608 TRACE (OOB-2,Infinispan-Cluster,NodeD-8993)
[org.infinispan.interceptors.DistributionInterceptor] Put occuring on node, requesting
cache invalidation for keys [k1]. Origin of command is remote
2011-11-11 11:10:27,608 TRACE (OOB-3,Infinispan-Cluster,NodeA-31187)
[org.infinispan.interceptors.DistributionInterceptor] Put occuring on node, requesting
cache invalidation for keys [k1]. Origin of command is remote
2011-11-11 11:10:27,608 TRACE (OOB-2,Infinispan-Cluster,NodeD-8993)
[org.infinispan.distribution.L1ManagerImpl] Invalidating L1 caches for keys [k1]
2011-11-11 11:10:27,608 TRACE (OOB-3,Infinispan-Cluster,NodeA-31187)
[org.infinispan.distribution.L1ManagerImpl] Invalidating L1 caches for keys [k1]
{noformat}
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators: