Hello everyone!
I have the very similar issue on virtual environment one of our customer.
I have tested this configuration:
1. one blade server on westmere architecture: CPU Intel Xeon E5649.
2. one blade server on sandy bridge architecture: CPU Ontel Xeon E5-2440.
3. they both are in one HA-cluster without EVC.
4. When I try to vMotion test VM from Westmere to Sandy -> everything is ok and without problems.
5. When I try to vMotion test VM in backward, from Sandy to Westamere -> host on Sandy architecture became unresponsible until Mgmt network restarted from console at vMotion progress 65%. But VM has migrated very well.
6. I created turned on EVC feature on this cluster but it wasn't a better decision. The result was the same.
7. And one more interesting thing: when I shutdown my guest OS the host with Sandy architecture has gone unresponsible too )))
Very strange situation.
Anybody knows, can be two hosts Westmere and Sandy bridge in one cluster?