We are attempting to put together a Cluster of JBOSS ESB nodes (on AIX) with the some of our heavy lifting services spread across nodes for better load balancing and throughput.
An example being that one of the services which does some heavy I/O
(computes checksums on large files on the disk) has its queue/listener
clustered so when the current node is already performing that service,
the message is get sent to another node thereby improving our throughput.
The cluster appears to be working as expected only when all the nodes
happened to be on the same physical machine. When we spread the nodes
across machines (all nodes obviously have same oracle schema as juddi,
message store, etc.), the message never gets sent to nodes running on
other machines. I can send our queue definition, JBM config, jboss-esb
and/or other files if they can help to figure out if we are missing any
in our setup.
We have been struggling for past few days as to why the setup won't work
across machines. Any working sample you can point to which shows JMS
clustering across machines can help us tremendously.