There is an article on O'Reilly OnJava written by Bill Burke and Sacha Labourey that is a very good introduction into clustering with JBoss 3.0.
Being new to clustering though, what I don't understand is how database connections are handled. Are all clustered JBoss application servers supposed to talk to one database/database cluster? If you are using a database that doesn't support database clustering, can you have a failover scenario using multiple databases with replication? I don't see how multiple databases using replication will work with clustered JBoss application servers but I thought I would ask anyway.