Raw LUN maximimums

Just an interesting note that is probably no use to anyone.
Back in ESX 3.5 days, I had a workload request for a fileserver with a 6TB disk. Creating a VM with the OS on a VMFS LUN and then assigning a physical compatibility raw LUN of 6TB worked perfectly.

Later, upgrading to ESX 4.0 in that same environment had no issues with the raw 6T VM. This is where the config maximums stated that 2T-512b was the largest pRaw LUN. Shutting down the VM, storage vMotioning the OS to another datastore, all had no issues.

When I upgraded to ESXi 5.0, then we had issues. I wasn't able to vMotion the VM with the 6TB LUN to a 5.0 host in the cluster. We had to pause the upgrade for a week to find possible solutions and a downtime window for the fileserver.

The fix?

Upgrade one of the existing VMFS3 LUNs to VMFS5. Shut down and power up the VM on ESXi 5.0. The config file for the pRaw disk automatically moved to the VMFS5 LUN and the world started turning again.

So a working but unsupported pRaw disk that went well beyond the maximum supported worked perfectly fine up until it was actually supported (under VMFS5) and needed to be massaged before working in 5.0.

I was reminded of all this when checking out the released 5.1 max guide. As much as a pain it was to memorize the key maximums for a VCP exam, knowing them offhand has been invaluable in conversations with new clients.

So, drink up the 5.1 maximums, you new consultants and VCP wannabes: Cheers