When I set up drive letter for shared drives in Windows 2008 R2 2-node cluster, I accidentally used a drive letter which has already mapped automatically due to the login I used. After I run the cluster validation tests, all the drive letters for shared drives are changed. I fixed this issue by disconnecting the automatic mapped drive and reassigned all the drive letters.
However, I wonder whether CNO is used to perform the failover so such issue will not occur actually. Also, what happen if someone is logged on the passive node when the failover occur and any of the drive letters assigned for the shared drives is already mapped due to that person's login.