Hyper-V Windows Server 2012 issue with Ubuntu Server 18.04 with through physical disks?

Question

linux

Good day.
The problem is as follows:
There is a Windows Server 2012 with the Hyper-v role
In Hyper-V, Ubuntu Server 18.04 is spinning with forwarding of physical drives
In Ubuntu, the disks are configured and mounted, there are entries in fstab
When the host machine is rebooted, the disks in Ubuntu Server are castling, that is: there are three disks sdb (RAID 1 3TB) sdc (RAID 1 3TB
) sdd (RAID5 6TB), backup is performed on sdd, after a reboot, the disks remain with the same identifiers, but the sdd (backup) and sdc disks change logically
backup failure errors because there are no specified directories.
There is an assumption that the problem is somewhere at the level of disk initialization at the start of the virtual machine, but I did not find how to find out or check it for sure.
I would really appreciate any help in this matter.

Answer 1

Try mounting by UUID

Answer 2

Apparently, the backup was done by copying byte-to-byte using dd? or rsync file copy?
First you need to determine which disk was the original disk and which one was copied, and determine which files were not copied, first comparing based on the names, sizes and time of the last modification, and if no differences were found, repeat the process but already with a content check
. Comparison can be carried out with using the diff utility, in principle it is more than enough, but working with its output can be inconvenient
Personally, to compare directories, I collect two lists of files in text form (relative path to the file + size + timestamp time + hash, count md5sum) in turn for the source directory and for the directory with a copy, then I sort the files (for example, using sort by name) and compare them already with the help of diff - it turns out a beautiful and convenient list of differences. Repeat the operation twice, first without counting the hashes - a quick comparison, and then, if necessary, repeat but having already collected the hashes. The script that does this takes five lines.
Since the files are placed on different physical devices, hash counting can be carried out simultaneously for the source and the archive copy (hashing itself does not waste the processor, only the disk), usually this process goes at the maximum speed of the disks, so it will not take much time (a few hours )