I’m trying to do a recovery with Windows backup on server 2016. I’s a hyper_v server with three VMs.

Backup was setup to backup the whole server. Now I’m looking at recover items and there is no D: volume (where the VMs are stored). Just an option tor restore Hyper-V. If I check that, I get a check box for Hyper-V and a sub-box for Host Component.

Is this what I need to recover the three VMs?

8 Spice ups

All I need to recover is the VMs. I’m looking at the list of what the default setting for items backed up and it includes Bare metal recover, Sytem State EFI part, Local C: drive, Recovery, EX2016 (VM), DC01 (VM), SRV01 (VM), and Host Component.

All of the VMs and virtual drives were on the D: partition. Nothing else.

Did you check all of the backups to verify the information?

Why no, I did not, but when I did, I found that the last backup did not contain all of the data that the previous ones did. The D: drive was there and I restored it.

The back up was listed as successful, but it was missing information. It’s a scheduled backup, I’m not sure how it got messed up.

Wait… Did you have a conversation with yourself or do multiple people share that same login? …

Well, nobody else seemed to be available, so I had to help myself. :slight_smile:

Its was 2am on Sunday night, not many people will be around ??

What are you backing up and what are you trying to restore ??

There is a very big difference between a VM and the files on a host…being able to backup and verify that you successfully backup vHDD files does not mean they can be restored to VMs…or that the VMs when booted up will not give weird OS errors or even BSOD (as the VMs were not properly quiescence or backup as VM).

Yeah, I had to redo a raid array and they don’t like it when I take the whole network offline during a working day. It was only 10am to 9pm EST for me. I guess everybody else was out enjoying the day.

I had raid issues with the Hyper-V host. Intel recommended redoing the raid array. The host runs on a raid 1 drive (C:) and the all of the VM files are on a raid 10 array (D:). Both drives are backed up at the host level. C: drive was operating just fine. It was the D: that needed to be replaced.

The initial issue was that I opened the last backup set and didn’t see the D: listed in the recoverable file set. Just the C:, system state and Hyper-V. That was a rush I never want to experience again. I didn’t know if I was looking at missing data or whether Server 2016 backup did things differently than pervious versions (this was the first time I’ve restored anything in 16). I’m thinking maybe Windows backup handled VMs differently that regular data. Hence the first post because my only option was to restore Hyper-V. But that did not list the three servers, just a check box for Hyper-V and a sub check box for Host Component. I wanted to see if someone could steer me in the right direction. So after I posted, I started looking to see how a backup of the whole server could leave out a whole drive. The details list stated that everything should have been backed up, including the three VMs. At the point I looked at the previous backup and everything was there. Why the last back excluded stuff I don’t know. When I got to the office this morning, I checked the back up and it had run with the exact same message as the previous backups. No indication that it skipped stuff.

So to make a long story short, I ran the recovery from Friday night, restored the entire D: drive, and everything started right up. Domain is up, email is flowing and I know in morning if the apps are working. So much for a relaxing weekend.

So the 3 VMs included a Domain Controller ? Then I guessed you were very very lucky…

I would never have recovered DCs or even bothered to backup DCs as it will give you more errors and issues especially if you have applications and DBs running 24/7.

Then also never to rely on Windows backup as it cannot differentiate VM data vs files…if your VMs were powered on, you may get BSOD or other issues (unless you powered off the VMs before backup).

Then most of the time, we would create OBR instead RAID 1 with a RAID 10…

And have at least 2 hosts so that we can move (or migrate or vMotion) VMs to other hosts when we need to down either hosts.

This was kind of a hail Mary sort of thing. The VMs were shut down before being backed up. The VMs them selves are backed up by ShadowProtect client on each VM to a local server the replicated to the cloud. That process stopped when the raid card failed. SC tech told me (after I sent him the diagnostic logs) that they didn’t support DFS and it would have to be disabled. !?). So those backups were days old. For insurance, I removed the old array intact, and replaced all of the drives with new ones. I could revert back to those if I needed to.