[SOLVED] Restore failed: unable to parse raw blob - wrong magic

Rotzbub

Member
Sep 18, 2020
11
6
8
30
PVE 7.4-16
PBS 2.4-2

What I did is a basic restore in PVE:​

  • Go to Datacenter
  • select the Backup-Storage 'disk'
  • From BackupsI select one
  • click on Restore
  • Restore: VM I select the Storage (to where it will be restored) - click on Restore
Taskviewer - Logs 'task-pve-qmrestore...' (attached):
Code:
...
progress 19% (read 612032839680 bytes, zeroes = 6% (40319844352 bytes), duration 9827 sec)
progress 20% (read 644245094400 bytes, zeroes = 6% (42316333056 bytes), duration 10366 sec)
progress 21% (read 676457349120 bytes, zeroes = 6% (44304433152 bytes), duration 10904 sec)
restore failed: unable to parse raw blob - wrong magic
temporary volume 'tank:vm-100-disk-2' sucessfuly removed
temporary volume 'tank:vm-100-disk-0' sucessfuly removed
temporary volume 'tank:vm-100-disk-1' sucessfuly removed
error before or during data restore, some or all disks were not completely restored. VM 100 state is NOT cleaned up.
TASK ERROR: command '/usr/bin/pbs-restore --repository root@pam@127.0.0.1:diskname vm/111/2023-07-12T03:15:04Z drive-scsi1.img.fidx /dev/zvol/tank/vm-100-disk-2 --verbose --format raw --keyfile /etc/pve/priv/storage/diskname.enc --skip-zero' failed: exit code 255


What I tried also:​

  • proxmox-backup-debug recover index drive-scsi1.img.fidx /mnt/datastore/diskname/.chunks --ignore-missing-chunks --keyfile /etc/pve/priv/storage/diskname.enc
    got same message:
    Code:
    Error: unable to parse raw blob - wrong magic
  • checked diskpace/inodes
  • later backup (same vm) has the same error on the same percentage! - differnt backup (another vm) works
  • file Restore is working

I only need one drive from the backup. Maybe there is another way to restore this to mount it to a vm ?

Thanks in advance!
 

Attachments

  • task-pve-qmrestore-2023-07-18T10 45 26Z.log
    12.8 KB · Views: 1
can you try to do

Code:
proxmox-backup-debug inspect file <file>
for every file in the backup directory on pbs and post the outputs ?
 
Thanks for the quick reply:

I shortened the chunks output...
Code:
# root@akroma:/mnt/datastore/diskname/vm/111/2023-07-14T03:15:06Z# ls -al
total 674782888
drwxr-xr-x  2 backup backup         4096 Jul 19 12:48 .
drwxr-xr-x 13 backup backup         4096 Jul 15 04:30 ..
-rw-r--r--  1 backup backup         2664 Jul 14 11:31 client.log.blob
-rw-r--r--  1 backup backup         4128 Jul 14 11:31 drive-efidisk0.img.fidx
-rw-r--r--  1 backup backup       135168 Jul 14 11:31 drive-scsi0.img.fidx
-rw-r--r--  1 root root 690952863744 Jul 18 11:31 drive-scsi1.img
-rw-r--r--  1 backup backup     24580096 Jul 14 11:31 drive-scsi1.img.fidx
-rw-r--r--  1 backup backup          621 Jul 14 11:31 index.json.blob
-rw-r--r--  1 backup backup          394 Jul 14 05:15 qemu-server.conf.blob

# proxmox-backup-debug inspect file client.log.blob
size: 2664
encryption: encrypt

# proxmox-backup-debug inspect file drive-efidisk0.img.fidx
size: 131072
creation time: Fri Jul 14 05:15:07 2023
chunks:
  "2795d9fadf103a4452b46a2870ee3dc6adca333c23e36319d208797a1a008629"
 
# proxmox-backup-debug inspect file drive-scsi0.img.fidx
size: 17179869184
creation time: Fri Jul 14 05:15:07 2023
chunks:
  "b8d7a99ec43da6f3562336c8ef7205d72d80de2dbddabeee82ffb3ed55870dd8"
  "7de95139467f45d34af5154548157c0f7d15cae0f144a69d5d7b71433468eb65"
  "554e3c765b41495c6fb3e832ce8e43d6248559348a95d2a416229bce316faf3b"
  ...
 
# proxmox-backup-debug inspect file drive-scsi1.img
Error: Only .blob, .fidx and .didx files may be inspected

# proxmox-backup-debug inspect file drive-scsi1.img.fidx | more
size: 3221225472000
creation time: Fri Jul 14 05:15:08 2023
chunks:
  "892b521b9d4094178df2d07a8a835ef84f7a847e362dafeee18e43b3e4bafa8a"
  "9e2588a91d847e592aa1861359a2db38df76862cdf5fce83dc819a7d4a249075"
  "11116e72fd74a72e95cbaeadde1e30b2d121ac1b925f356b43bdac645f4df494"
  "c351af23f4d04d7ea1ea636574d7acc215106c7752dec356982efb5e07721995"
  "1473b8e9c6699b5af8ff31cecc7fbdf43d17e8e8fdaea3e38e9a2d0d59fc34bc"
  "37a5a3c7e40ecd609f58462770b73cb6da4c06113de6d8bc44d95c99188558fa"
  "e390924675501e912723a75344401a3a6f77e9a770d55ab70d61f2e425cf3d2b"
  "9751bea447ba214723d32a8c33a6049a8c6a4a22093d36cbd8f961aa222b13ca"
  "6e13d2f0e8e279dea9d0afa857915344ccf57827f3d1a3f3597de902508fc92a"
  "650b39ceb103a8b3e86adb38513f77f5fab4a18c805bd985794b824a96ec999f"
  "f8a942a6819be2cf889ed3831fa0f27919691e5129903943d37bb35161c96e43"
  "57f85b79e7d2f6193be77df527eec8eb9ba45b5b762bf68cad094baa39058830"
  "4b584d9bcdb1d09c89b4d489780b6ae775c982638b02c97710d037fa7b4ca459"
  "0274be21c1f48b3fc1245f35903ebc2c802b22f5e314375238769b59f20aae85"
  "d0419e00db88a2a41d3c27a46b63efe27e64fd1ad23b9b8a70e583661b938089"
  "b51acdf1909ab94c99bd5be99d2c77bdc2d749458566fa04efd4bb86d77c1206"
  "dde15043dba5d9e4fcc07bb3c2ab3e43d7e72d03433265bf889187713324f30a"
  "409c11caa72d612becea542d2f36fdaa510b530168e41ec15c253d60d92eae99"
  ...

# proxmox-backup-debug inspect file index.json.blob
size: 621
encryption: none

# proxmox-backup-debug inspect file qemu-server.conf.blob
size: 394
encryption: encrypt
 
Last edited:
mhmmm.. why is there an .img file in there?

can you try to move that somewhere else? maybe it's trying to read that and fails
 
true that,

here are some other ls outputs from that datastore - one of the other vm (112) i restored sucessfully
and i tried again with the one from 2023-07-12 - logs attached - same error
Code:
root@pve:/mnt/datastore/diskname/vm/111/2023-07-14T03:15:06Z# # ls -al ../../112/2023-07-14T09:31:25Z/
total 8284
drwxr-xr-x  2 backup backup    4096 Jul 14 13:04 .
drwxr-xr-x 10 backup backup    4096 Jul 15 04:30 ..
-rw-r--r--  1 backup backup    2673 Jul 14 13:04 client.log.blob
-rw-r--r--  1 backup backup  266240 Jul 14 13:04 drive-scsi0.img.fidx
-rw-r--r--  1 backup backup 8196096 Jul 14 13:04 drive-scsi1.img.fidx
-rw-r--r--  1 backup backup     570 Jul 14 13:04 index.json.blob
-rw-r--r--  1 backup backup     377 Jul 14 11:31 qemu-server.conf.blob
# ls -al ../2023-06-30T03:15:02Z/
total 24164
drwxr-xr-x  2 backup backup     4096 Jun 30 11:08 .
drwxr-xr-x 13 backup backup     4096 Jul 15 04:30 ..
-rw-r--r--  1 backup backup     2649 Jun 30 11:08 client.log.blob
-rw-r--r--  1 backup backup     4128 Jun 30 11:08 drive-efidisk0.img.fidx
-rw-r--r--  1 backup backup   135168 Jun 30 11:08 drive-scsi0.img.fidx
-rw-r--r--  1 backup backup 24580096 Jun 30 11:08 drive-scsi1.img.fidx
-rw-r--r--  1 backup backup      623 Jun 30 11:08 index.json.blob
-rw-r--r--  1 backup backup      394 Jun 30 05:15 qemu-server.conf.blob
# ls -al ../2022-12-30T04:15:01Z/
total 24164
drwxr-xr-x  2 backup backup     4096 Dec 30  2022 .
drwxr-xr-x 13 backup backup     4096 Jul 15 04:30 ..
-rw-r--r--  1 backup backup     2620 Dec 30  2022 client.log.blob
-rw-r--r--  1 backup backup     4128 Dec 30  2022 drive-efidisk0.img.fidx
-rw-r--r--  1 backup backup   135168 Dec 30  2022 drive-scsi0.img.fidx
-rw-r--r--  1 backup backup 24580096 Dec 30  2022 drive-scsi1.img.fidx
-rw-r--r--  1 backup backup      622 Dec 30  2022 index.json.blob
-rw-r--r--  1 backup backup      390 Dec 30  2022 qemu-server.conf.blob
Z# ls -al ../2023-07-12T03:15:04Z/
total 24164
drwxr-xr-x  2 backup backup     4096 Jul 12 11:04 .
drwxr-xr-x 13 backup backup     4096 Jul 15 04:30 ..
-rw-r--r--  1 backup backup     2695 Jul 12 11:04 client.log.blob
-rw-r--r--  1 backup backup     4128 Jul 12 11:04 drive-efidisk0.img.fidx
-rw-r--r--  1 backup backup   135168 Jul 12 11:04 drive-scsi0.img.fidx
-rw-r--r--  1 backup backup 24580096 Jul 12 11:04 drive-scsi1.img.fidx
-rw-r--r--  1 backup backup      618 Jul 12 11:04 index.json.blob
-rw-r--r--  1 backup backup      394 Jul 12 05:15 qemu-server.conf.blob
 

Attachments

  • task-pve-qmrestore-2023-07-19T10 51 45Z.log
    12.8 KB · Views: 2
  • pbs-debug_of_111_2023-07-12.txt
    2.8 KB · Views: 3
is there anything in the pbs journal/syslog during a restore that might indicate the problem?
 
ok a colleague reminded me that this also can happen if the encrypted chunks are corrupt

you can check how many chunks are impacted with the proxmox-backup-debug recover command but with the '--ignore-corrupt-chunks' parameter (this replaces corrupt chunks with zeroes)

did you run a verify recently? this should have caught that and marked the chunk as corrupted
 
Yess,Thanks ! The Error is due to a corrupt chunk it seems:

Code:
root@pve:/mnt/datastore/diskname/vm/111/2023-07-12T03:15:04Z# proxmox-backup-debug recover index drive-scsi1.img.fidx /mnt/datastore/diskname/.chunks --ignore-missing-chunks --keyfile /etc/pve/priv/storage/disk8tb1.enc --ignore-corrupt-chunks --output-path /tank/drive-scsi1.img
WARN: chunk "/mnt/datastore/diskname/.chunks/0257/0257bcf3021ea63daf7a272ba5b8e698ca188b65f9680d780b0fa06e0f7259f2" is corrupt - unable to parse raw blob - wrong magic
WARN: replacing output file 690952863744..690957058048 with '\0'

I did a proxmox-backup-debug recover index but only with --ignore-missing-chunks not --ignore-corrupt-chunks. Also thats why there was an .img file in the folder.
 
  • I will now import the created .img - qm importdisk 111 drive-scsi1.img tank --format raw
  • And run a (in PBS) a verification of the disk to solve this in the future
  • As I have another PBS backup of this vm, maybe I will restore from there or can crosscheck to fix this corrupted chunk - but that would be another thread (will link if I found something)
 

About

The Proxmox community has been around for many years and offers help and support for Proxmox VE, Proxmox Backup Server, and Proxmox Mail Gateway.
We think our community is one of the best thanks to people like you!

Get your subscription!

The Proxmox team works very hard to make sure you are running the best software and getting stable updates and security enhancements, as well as quick enterprise support. Tens of thousands of happy customers have a Proxmox subscription. Get yours easily in our online shop.

Buy now!