I have a SQL Server 2008 R2 failover cluster with a passive node and an active node. The servers are physically identical. I have two instances of SQL Server, one set to use all the CPUs in the first 3 NUMA nodes, the other set to use all the CPUs in the 4th NUMA node. This seems to be working okay, even when failover occurs, but is this a bad idea? What would happen in failover if the passive server didn't have as many processors?