I'm about to begin deployment of a new SQL Server 2008 R2 cluster. This is our main production cluster, and will host approx. 200 different databases. But, as I've learned, SQL Server 2008 uses the tempdb different than earlier versions of SQL Server. I've read BOL regarding distributing tempdb over numerous files for each CPU. (our new cluster will have 32 CPUs on each node). Any one with any pointers how I can monitor tempdb usage and scale the files correctly?
I'm also looking into to have configure the cluster in an active/active configuration, with prod databases on one node, and test on the second node. Any one with thoughts about that kind of configuration?