This may be better suited for one of the other forums, but it doesn't quite fit there, so...
These days, its a reasonable assumption that most commercial applications are going to be deployed into a clustered environment - for failover if nothing else, but most of them benefit from increased scalability as well. I'd love to get some feedback about best practicies for Seam applications.
By that, I mean - assuming that you're deploying on JBoss AS, how do you set your clusters up? What's your load balancing strategy? Are you replicating the entire session (probably, but...)? Any issues with conversational failover?
Basically, some kind of message that says, "The team at JBoss has thought about it, and assuming that money is not a problem, here's how we would set up a deployment cluster to run a large-scale Seam app, assuming that you're using Seam, JMS, EJB3 entities, statefull and stateless beans, et cetera. If I'm wishing, it would include a "minimum," setup, and then a "to scale, add ___ and change ___ setting appropriately" section.
Even something saying, "Wait until JBoss AS 5.0 is released because then its all integrated and you do X" would be cool, too.
Thanks! If such recommendations exist out there, my google-fu is inadequate to find them and I'd appreciate a pointer.
I should add that I have read through (or at least searched through) the EJB3, Cluster, and Seam forums before asking this - a lot of the information seems to be old, or possibly incorrect, and in at least one part I saw a recent update from Gavin saying to read the docs rather than looking through old posts :) but I'm not entirely sure which set of docs he's referring to.