Difference betwwen TreeCache cluster and JBoss Cluster?

[mcevikce]

Can some one clarify the connection between the TReeCache cluster, Jboss cluster and TOmcat cluster.

As much as I understand cluster-service.xml is configuration for clustering JBoss servers (partitions). We can define which servers should be with in the cluster by the partition name. My question is how to we enforce the same rule with the JBoss TreeCache (TreeCache.xml). There is nothing with in this configuration identifies how and when and between who should the TreeCache replicate. Am I missing something?

Currently we have 2 servers running on 3 partions each. It seem like there is replication happening between these servers. I get classloader errors indicating clasloader not found. Thats because each partion group running its own version of the class. When these cached objects are being replicated between the servers these exceptions occur. I would have imganined with the settings I have there should be no replication between the partitions that are different!

Any help to clrify this will be appriciated.

Thank you.

[jerrygauth]

The three clusters are different logical groups. The Tomcat cluster grouping is handling session replication. The DefaultPartition cluster handles core clustering including HA-JNDI. Any TreeCache related clusters are handling replicated TreeCache instances.

These clusters all use JGroups internally and they are differentiated by cluster name, multicast address and port. Currently separate configurations are required for each cluster. The forthcoming JGroups 2.3 release will allow multiplexing of cluster configurations in cases where cluster groups are able to share configurations (e.g., JGroups stacks).

If you are running multiple partitions in a cluster, you need to distinguish them via the name, address and port settings. This is true for all of the clusters in your topology (e.g., Tomcat, TreeCache).