we have a JBOSS 4.x application server in our setup which is running on a MSCS (win2k3) 3 node cluster. there is another application instance also running on the same cluster. i do not know who has implemented this.
when i go thru JBOSS documentation i find the best practice to cluster JBoss is to use its own clustering mechanism. apart from a HA perspective what could be the reason the implementer has used MSCS to cluster JBOSS.