Whilst migrating a Windows SQL 2008 MSCS Cluster from 3.5 to 4.1 the Disk Queue Lengths for the RDM's went through the roof.
One of our processes was to change the Multipathing policies to Round Robin from the default fixed policy.
After changing the policy back to Fixed this made the queue lengths show as low again.
I then found this note taken from the following KB:
Note: This policy is not currently supported for Logical Units that are part of a Microsoft Cluster Service (MSCS) virtual machine.
Not sure if anyone has experienced this before?