If you attended my SRM session at VMworld 2014, or one of the VMUG User Conference sessions I’ve presented at, you’ve heard me talk about upgrading SRM and the entire infrastructure.
I stressed the importance of upgrading the Recovery/Target Site’s hosts before upgrading the Protected/Source Site’s hosts.
As you can imagine, upgrading out of order does happen, and I got to see exactly what happens when you’re in that situation. Well, not in a disaster, but in regular day-to-day tasks.
Have you seen this error when trying to protect a VM in SRM, specifically trying to create the placeholder?
Placeholder VM creation error: No hosts with hardware version ’10’ and datastore(s) [datastore01] which are powered on and not in maintenance mode are available
If you have, then your Protected site’s VM level is higher than your Target sites ESXi version. In my specific case, the source was upgraded to 5.5, but the target was only 5.1.
Can you fix this easily?
Depends on your thought of ‘easy’ :D The proper way is to upgrade your Recovery hosts to ESXi 5.5. The ‘easy’ way would be to create a new empty VM at version 9, then attach the vmdks to it, but I neither recommend nor condone that :P
What happens if a disaster occurs while in this state?
None of your vmx-10 VMs will be recovered, you pretty much shot yourself in the foot. The data’s likely replicated, so you could manually import it, but you may have to edit the .vmx file for the port groups you need.
Either way, this is a prime example of why you NEED to upgrade the recovery site first.
As the error states, the cluster that the placeholder is mapped to may be offline as well. I ran into this error and was like DUH, my dr cluster is offline.I think this also speaks to the perils of upgrading VM hardware without an integrated ecosystem strategy.