No. My issue has been I've been thinking (and my group too) from a O/S level, and not instance level. It makes a lot more sense when I get my mind off of the Windows Clustering.

I think where my manager is getting the load balancing is that you can have two nodes, and a instance on both with each of it's respected apps hitting each instance. So Application 1 is hitting instance on node A, and application 2 is hitting instance on node B. My thought is it wouldn't make a difference with it being shared disk.

Management wants two node cluster that will automatically failover with as little to no downtime as possible. So by having just one instance on Node A, something goes wrong on it, instance fails over to Node B.

Does Node B need to have an active database instance running on it for it to be online? Or can it be running without an instance until a failure?