mounting 'drive-scsi1.img.fidx/part/2' failed: all mounts failed or no supported file system (400)

[complexplaster27]

Hi All,

I'm not sure what is going on here but when I try to back up our File Server which is a Win Server 2019 VM, it has 5 NTFS drives on C (scsi0), D (scsi1), E (scsi2), F (scsi3) and G (scsi4).

When I try to perform a File Restore within the Proxmox GUI on the servers backup tab and try to expand the large storage directory on D drive partition 2 I get this mount failure error:



However I can expand some of the other drives without any issue like the E drive partition 2 which is 3.17 TB, the F drive which is 200 GB and the G drive which is 550 GB but all the partition 1 drives throw the same error as above:



Here's the full list of mounting failures:

mounting 'drive-scsi0.img.fidx/part/2' failed: all mounts failed or no supported file system (400)
mounting 'drive-scsi0.img.fidx/part/3' failed: all mounts failed or no supported file system (400)
mounting 'drive-scsi1.img.fidx/part/1' failed: all mounts failed or no supported file system (400)
mounting 'drive-scsi1.img.fidx/part/2' failed: all mounts failed or no supported file system (400)
mounting 'drive-scsi2.img.fidx/part/1' failed: all mounts failed or no supported file system (400)
mounting 'drive-scsi3.img.fidx/part/1' failed: all mounts failed or no supported file system (400)
mounting 'drive-scsi4.img.fidx/part/1' failed: all mounts failed or no supported file system (400)

I have no clue what the issue is, however this is our main File server and I need to be able to restore individual files when user requests for them and I really don't want to use another file backup solution for this. The backup works fine when I initiate it so any tips on what I can do to troubleshoot this would be fantastic as I would love to get this fixed.

I'm on Virtual Environment 8.1.10 and Backup Server 3.1-5, if any more info is required I'm happy to post and reply.

 

[fiona]

Hi,
you could give the following build a try: https://forum.proxmox.com/threads/file-restore-test.137420/post-630034
The change is already in git, but the package is not yet rolled out to repositories yet.

 

[complexplaster27]

Hey @fiona,

Thanks I installed this on my 3 node cluster however I'm still experiencing the same issue with the exact same partitions.

First of all, should I delete the backups and start fresh? If I don't have to delete them should I just take another snapshot backup?

I did notice that another server that I could replicate the issue on is now resolved and I can browse the partition that I was not able to previously.

Here's a screenshot of it as I was previously unable to access Partition 2 on scsi1 on this VM.


So it seems I'm getting a 50/50 success rate. If you need any further info let me know and I can provide!

 

[fiona]

First of all, should I delete the backups and start fresh? If I don't have to delete them should I just take another snapshot backup?

That should not make a difference for file restore, the new build for the proxmox-file-restore does not influence how backups are created.

I did notice that another server that I could replicate the issue on is now resolved and I can browse the partition that I was not able to previously.

Here's a screenshot of it as I was previously unable to access Partition 2 on scsi1 on this VM.

View attachment 66416

So it seems I'm getting a 50/50 success rate. If you need any further info let me know and I can provide!

Does success depend on the backup you select or the node on which you access it?

 

[complexplaster27]

That should not make a difference for file restore, the new build for the proxmox-file-restore does not influence how backups are created.



Does success depend on the backup you select or the node on which you access it?

Previously, every backup selected would fail on that particular partition, I will try migrate the VM to the node where the other machine is failing to see if this impacts the backups. I'll let you know of the outcome.

 

[complexplaster27]

Hmm Interesting @fiona, so the VM was on Node B and I migrated it to Node A and expanding the main drive I wanted worked:



However I still get the same error for scsi0 Partition 2 and 3:



I'm noticing that the expansion thinks about it for a little while say 20 seconds and then fails, as well it does seem that every Partition 1 is failing with that same error still. I'm thinking it's failing on mounting the EFI system partition and the Recovery Partition as neither of them are formatted as NTFS.

As well I'm not sure what Partition 1 would be on any of the other drives as this is what it looks like in Windows:

 

[fiona]

Hmm Interesting @fiona, so the VM was on Node B and I migrated it to Node A and expanding the main drive I wanted worked:

Are you sure you have the same version installed on both nodes?

View attachment 66420

However I still get the same error for scsi0 Partition 2 and 3:

View attachment 66421

I'm noticing that the expansion thinks about it for a little while say 20 seconds and then fails, as well it does seem that every Partition 1 is failing with that same error still.

Please share the log for this, which should be in /var/log/proxmox-backup/file-restore

I'm thinking it's failing on mounting the EFI system partition and the Recovery Partition as neither of them are formatted as NTFS.

Yes, it might very well be that mounting those is just not supported.

 

[complexplaster27]

Are you sure you have the same version installed on both nodes?

It's a 3 node cluster and yeah I can confirm that the following packages are installed on all 3:

proxmox-backup-restore-image/now 0.6.1~test1 amd64 [installed,local]
proxmox-backup-restore-image-debug/now 0.6.1~test1 amd64 [installed,local]

Please share the log for this, which should be in /var/log/proxmox-backup/file-restore


Yes, it might very well be that mounting those is just not supported.

Checked the qemu.log and I believe my previous assumption was right, NTFS3, ext4, ext2, FAT and VFS tries to mount the non NTFS partitions and fails.

[2024-04-16T11:44:52.114Z ERROR proxmox_rest_server::rest] GET /api2/json/list?path=ZHJpdmUtc2NzaTEuaW1nLmZpZHgvcGFydC8x: 400 Bad Request: [client 0.0.0.0:807] mounting 'drive-scsi1.img.fidx/part/1' failed: all
 mounts failed or no supported file system
EXT4-fs (vdc1): VFS: Can't find ext4 filesystem
EXT4-fs (vdc1): VFS: Can't find ext4 filesystem
EXT2-fs (vdc1): error: can't find an ext2 filesystem on dev vdc1.
FAT-fs (vdc1): bogus number of reserved sectors
FAT-fs (vdc1): bogus number of reserved sectors
VFS: could not find a valid V7 on vdc1.
ntfs3: vdc1: Primary boot signature is not NTFS.
ntfs3: vdc1: Alternative boot signature is not NTFS.

And I suspect as this server is old the disks were formatted in MBR and not GPT partition layouts so PBS might be backing up the MBR of each disk and dumping it as Partition 1 which can't be mounted. Maybe it's worth not even showing these partitions as I doubt anyone will want to restore files from the EFI partition anyway and if Partiton 1 is the MBR headers you wouldn't be able to explore that anyway in a file restoration program, is there a way to hide it so only Data disks show up in the file restore GUI?