In a cluster you do a normal SQL install on one node and do a "Add node" on all the other respective nodes. So, you have 2 nodes, 2 SQL instances. SQL Instance can only reside on ONE node at any given time. If your PRD instance is on Node1 that means all your database files related to that instance is on Node1. If a failover (move from Node1 to Node2) happens that clustered disk which has the database files for PRD will also failover to Node2 and now your PRD instance will be on Node2.