As you are aware, OpenJPA L2 caches communicate to their counterparts running in other remote processes via RemoteCommitProvider. Out-of-box providers communicate via TCP sockets (do not remmeber whether JMS protocol is provided out of the box, but one did exist for OpenJPA's earlier incarnations) . However, configuring such synchronization manually (yes, you have to figure out the ip address/port numbers for the sockets) in a clustered environment can be daunting as number of participating nodes grow. The usable solution was to piggyback on a protocol that is already operational in the cluster (say, for heartbeats). One such provider was developed for BEA Weblogic -- but I do not know its current status or availability [1]. It will be good to see an equivalent one for WAS environment :)
The other option is to use a distributed L2 cache that manages its own synchronization such as Coherence or Extreme Scale or EHCache. There are several available -- for example [2] (WARNING: self-plug). [1] http://openjpa.208410.n2.nabble.com/jira-Created-OPENJPA-943-How-to-configure-cache-synchronization-in-the-case-of-multiple-jvm-in-same-e-td2387993.html#a2394663 [2] http://ppoddar.blogspot.com/2009/10/l2-cache-in-jpa-20-and-openjpa-plug-in.html ----- Pinaki -- View this message in context: http://openjpa.208410.n2.nabble.com/Creating-a-custom-L2-cache-tp5295541p5321875.html Sent from the OpenJPA Users mailing list archive at Nabble.com.
