I have had this problem and it has been driving me crazy. I have performed exactly the same on Jboss 4.2.2.GA. It's seems to be a problem with session replication using JDK 6. You have to start up Jboss with the following:
in the run.conf which helped solve my problem.
Hope this helps - let me know if this fixes it for you...
thanks for the reply.
The error is not thrown anymore, but it seems the session is not replicated at all.
I think this might be another issue. In my setup I've checked the session replication through JMX console and I can see the sessions being replicated correctly. I would double check settings (maybe CacheMode and ClusterConfig UDP settings) in the jboss-service.xml (in \server\\deploy\jboss-web-cluster.sar\META-INF directory).
If that looks ok then on one of the nodes in JMX console go to the the service that is replicating the session data, in my case TomcatClusteringCache service. Check the Members property to make sure all the nodes are appearing.
Hope this helps.
I will do that, thanks.
Also, I'm investigating on other solutions like Terracotta...it's seems more powerful and with better scaling capabilities.
Ok, even after the
flag on the JAVA_OPTS, I got this exception.
15:57:27,073 ERROR [JBossCacheService] loadSession(): id: TLvxyxOX2p5VtvjkEuqjoQ** exception occurred during deserialization
Terracotta worked great, and the implementation took no time. Why this has to be so complicated and undocumented!