So much for being on schedule. I have been insanely busy with VMware and SAN projects, and some of it not so much fun.
It would seem that here lately there have been a rash of issues with environments where a LUN is either failed over manually or automatically and at least one ESX 3.5 host ends up not being happy about it. From outright host crashes to SCSI lock issues, this appears to be a bubbling problem. Hopefully the 3 times I have run into it in the past 2 months is only a very, very, strange coincedence. I know that there are some current patch issues, but the events that have occured don't seem to fit the bill.
As for SRM, it seems to be rather tempermental in getting it to cooperate with your chosen storage and replication. In scanning the discussion groups I have seen mixed information that seems to follow this theory. With my latest project, I had to reconfigure the replication to work out the best balance of schedule and resources and need to configure the protection groups. I have had issues thus far with getting this step done, so we will see.