Yep, that's possible, but I just wanted to rule out any more simple explanations first. I was particularly worried when you said the VR server couldn't ping itself but maybe I misinterpreted what you were saying.
In any case, you can check to see if this is your problem by looking for a line like this in the esx.conf of the hosts that you upgraded:
/net/vmkernelnic/child[0001]/tags/4 = "true"
If so, delete the line from the esx.conf and reboot the host. Then, when the host is back, double check that the nic assigned to the vmkernel has a reachable IP address.
I believe the "child[0001]" part may be different on some setups so if you don't find the exact line I specified above, check around for lines with a different child[] part.