Just had a scary one… After setting up Windows Server 2012 R2 to back up all its Hyper-V VMs through the host’s nightly Windows Backup I found that only one of the three Windows Server 2008 R2 VM clients was being backed up. The other two were failing, one with a disk CRC error, the other for no listed reason. Based on some searching, I suspected a problem with disk ID clashes–all the VMs were from the same template .vhdx file, so they all had the same “unique” ID. Oops. This is actually supposed to be handled okay by the current version of Windows Backup on a Hyper-V host, but I figured I’d eliminate it as a potential issue anyway. So, I had run DISKPART UNIQUEID DISK to reset all the Ids on all the partitions (virtual disks) in the pool of VMs to unique values. This morning I checked the logs and saw the backups up the two VMs again failed.
Next I tried a chkdsk on all the volumes on the affected VMs. However, after scheduling the chkdsks and rebooting the VMs I got the dreaded black screen of “no boot device can be found.” This happened on both the VMs I had tried (the third I hadn’t attempted to reboot yet), for which I, of course, had no backup. Oh no. Unlike Windows 7, Windows Server 2008 R2 lacks any automatic repair tool. Fortunately, a quick Google search turned up a Tom’s Hardware thread that indicated that the “repair” command prompt available from the original install DVD would allow the repair of the boot record with this command:
That command actually will scan all the mounted partitions and offer to add any Windows installations to the boot loader. In my case, there was only one installation offered per VM and by selecting it I was back in business. The scheduled disk checks even kicked in and now all the machines are up and running. Whether this actually fixes the backup issue we’ll see tomorrow…