actually, MS doesn't support active/active clustering. there are ways to do it, it really depends on the application you are running. e.g. if you have 2 nodes and are running SQL, you need 2 instances of SQL running(means more dirves and database planning). in cluster manager you have 1 instance running on node 1, and you fail instance 2 onto node 2 and configure the peference to node 2. MS has their reasons, i.e. N+1 failover, several server having acess to the same storage at the same time may lead to data corruption (linux doesn't have this problem, i think it has to do with MS's op lock on files) another BIG consideration is what is your shared storage? DAS? NAS? Multipathing? if you multipath, you need install the cluster in advanced mode(this is so in 2003 clustering, having done multipathing in 2008 cluster yet) this just looks at the shared LUNs' drive letter instead of ID, PID, etc for failover.
any other questions?
ps. migrating to a new SAN is painful unless you have some built in migration capacity built into the SAN. that then gets into setting the LUN id. server 2008 clustering has tools built into fixing these issues you just need to know them.