Considering that none of the JBoss Clustering guides even discuss an active-standby cluster, the request of your programmers is atypical.
Have your programmers justified why they want active-standby? That is usually not something the app programmers are concerned with, unless there is some specific reason / bug / feature that does not cluster properly. You should be able to deploy the app according to scalability / uptime requirements that may change after the app is written.
An idle standby node is a waste of resources and will result in less overall capacity. If you are the sys-admin, clarify your role, and make it clear that your job is to deploy and manage that deployment. Their job is to write the app. Your recommendation is for active-active, so there is no downtime, and because it is the standard JBoss clustering approach. It has the advantage of seamless failover / recovery of services, whereas with an active-passive, there will be downtime while the standby is started, plus sessions will not failover.
I cannot claim to be a JBoss expert, but the only clusters I _have_ run were active-active load-balancing. The only ammo I know to give is to refer to the JBoss clustering guide, and note that the recommended configs are all active nodes.
