The problem with that, especially with NetApp arrays, is that you can't just arbitruarly use round robin and let ESX work it out.. I've had several discussions with the ESX storage folks, and I keep trying to get through to them the problem here is that NetApp has proixy/non-proxy paths when you use clustered filers. ESX does not (I repeat does NOT) natively recognize that one path should be better than another (for Windows, we have to use NetApp's MPIO ware), and as such, some of our LUN's go down a proxy path unless manually changed.. If we don't go through and clean all of htat up, we end up getting notifications that we have fcp partner mismatch errors and tickets auto get generated... It's very annoying. 

For ESX, as the subject implies, there is a utility kit (NetApp brands this an FCP Attach Kit) that has a tool that can configure multipathing properly for LUN's to insure this situation doesn't occur... Currently, with ESXi, not so much. That's a part of why we aren't making a jump to ESXi in a larger way.. If we couldn't use those tools, life would NOT be fun..