During a recent test where the VASA provider was stored on a VVOL Datastore to simulate the loss of the VP and the VVOL Datastore going inactive, upon re-deployment and discovery/registration with VSC again, how would one recover the vvols and their metadata? The PE's still exist and the VM's VVOLs exist as well....but nothing works or can be brought back online because the VP thinks there are no vvols - is there something like a database export/import needed? Seems like there has to be something to recover the VP in this case given the incredible dependancy there is on an OVA to stay up for your VVOL environment to function.....
This KB article was sent to me by support as I couldn't really find much in the published docs - I would suggest making it bolder in the docs so people don't hit issues, specifically not having DRS or anything put the VP on a VVOL datastore itself, creating an affitinit rule if needed, anytihng that helps maintain and/or backup the VP OVA:
Virtual volumes: Protecting and Recovering the NetApp VASA Provider
Actually we went through this today and got it recovered.
Like you said, the files exist at the vvol volumes. what we did was to mount those vvol volumes to a Linux server, then find out VASA VM folders and individual vmdk files linked, copies them off, add the VSAS VM back to VC.
We tried to copy the files out via VC datastore browser option, but got permission issue, by going through Linux NFS, we were able to.