What you want I guess requires a clustered Jboss config (assuming you're using Jboss 3.x). From your description it seems like you're kind of setting up a pseudo cluster without using Jboss's clustering features.
Thanks for reacting. Well, I don't want to cluster now. I want to be in control now.
I have a request that spends 50% of it its cycles in one bean and 50% in other beans. I'd like to put that one bean on another machine. freeing up the processor and see what the effects are when cracking up the load. I think that with using the clustering one is not in control and the systems dispatches the request handling.
I may try clustering later.
Any suggestions on how to do pseudo clustering then?