Backup Error "vma_writer_register_stream 'drive-scsi0' failed" (PVE 6.1-7)

B

bootsie123

Well-Known Member
Dec 29, 2018
53
13
48
Hi! I set up automated backups for my VM templates and during the backup, several templates encountered this error. I tried backing them up manually and also received the same error. For the backup mode, I'm using stop and for the compression, I'm using GZIP. When I switch the mode to snapshot it successfully backs up, but not all the time.

*Note: I'm using a ZFS pool in raid1 and when I back up I'm backing up to a separate mounted HDD (/mnt/data/)

Here is one of the VMs config and the associated error message:

Config
Code: 
agent: 1
bios: ovmf
bootdisk: scsi0
cores: 2
efidisk0: local-zfs:base-118-disk-0,size=128K
lock: backup
machine: q35
memory: 4096
name: Windows-10-v3
net0: virtio=1A:69:22:68:5F:BA,bridge=vmbr0
numa: 1
ostype: win10
scsi0: local-zfs:base-118-disk-1,cache=writeback,discard=on,size=32G
scsihw: virtio-scsi-pci
smbios1: uuid=690a7e8c-1bd6-4220-8e2e-2a9cb36fd7fd
sockets: 2
template: 1
vmgenid: e63eb54c-ef9c-4e61-a199-94f566a065a7

Backup Log
Code: 
INFO: starting new backup job: vzdump 118 --mode snapshot --remove 0 --node vmworld --compress gzip --storage prox_backup
INFO: Starting Backup of VM 118 (qemu)
INFO: Backup started at 2020-02-23 13:01:46
INFO: status = stopped
INFO: update VM 118: -lock backup
INFO: backup mode: stop
INFO: ionice priority: 7
INFO: VM Name: Windows-10-v3
INFO: include disk 'scsi0' 'local-zfs:base-118-disk-1' 32G
INFO: include disk 'efidisk0' 'local-zfs:base-118-disk-0' 128K
INFO: creating archive '/mnt/data/backup/dump/vzdump-qemu-118-2020_02_23-13_01_46.vma.gz'
INFO: starting template backup
INFO: /usr/bin/vma create -v -c /mnt/data/backup/dump/vzdump-qemu-118-2020_02_23-13_01_46.tmp/qemu-server.conf exec:gzip > /mnt/data/backup/dump/vzdump-qemu-118-2020_02_23-13_01_46.vma.dat drive-efidisk0=/dev/zvol/rpool/data/base-118-disk-0 drive-scsi0=/dev/zvol/rpool/data/base-118-disk-1
INFO: ** (process:48111): ERROR **: vma_writer_register_stream 'drive-scsi0' failed
ERROR: Backup of VM 118 failed - command '/usr/bin/vma create -v -c /mnt/data/backup/dump/vzdump-qemu-118-2020_02_23-13_01_46.tmp/qemu-server.conf 'exec:gzip > /mnt/data/backup/dump/vzdump-qemu-118-2020_02_23-13_01_46.vma.dat' 'drive-efidisk0=/dev/zvol/rpool/data/base-118-disk-0' 'drive-scsi0=/dev/zvol/rpool/data/base-118-disk-1'' failed: got signal 5
INFO: Failed at 2020-02-23 13:01:48
INFO: Backup job finished with errors
TASK ERROR: job errors

Let me know what you think the problem might be. Just as a heads up I'm using PVE 6.1-7. If you want any more specifications don't be afraid to ask! Thanks!

EDIT 1: I was using PVE 5.4-13, although I upgraded to PVE 6.1-7 and I'm still experiencing the same issue
EDIT 2: I also tried backing up to the host disk. Same result for some of the VMs
 
Last edited:
Does backing up to any other storage work?
 
Could you run dmesg -w while you execute the backup command? vma_writer_register_stream 'drive-scsi0' failed might happen because of the USB connection.
 
Here's the output from dmesg -w after experiencing the error. Let me know if you want the output from before as well

Code: 
[573759.620615] traps: vma[28290] trap int3 ip:7f8ce1db3261 sp:7ffe8c129c90 error:0 in libglib-2.0.so.0.5000.3[7f8ce1d63000+112000]

I should probably also specify that this drive is connected directly via SATA and not USB
 
Let me know if you want the output from before as well
Click to expand...
That would be great.

Is there any chance that you could upgrade Proxmox VE to version 6? I'd really recommend it as Proxmox VE 5 reaches end of life in a few months. Additionally, we might currently be looking for something that is already resolved in the new version.
 
I could look at upgrading to version 6. How stable is the upgrade process? One concern I have is the use of a new Linux kernel. I'm currently on kernel version 5.4-5 (4.15.18-25) and whenever I try to upgrade to 5.4-13 (4.15.18-53) I run into a few problems with GPU-Passthrough and a few other things.

Here's the full output of dmesg -w before trying to backup

Pastebin Raw: https://pastebin.com/raw/uQCUs94V
 
Last edited:
How stable is the upgrade process?
Click to expand...
Proxmox VE 6 was released on July 16, 2019 and many people have upgraded. I'd say that's pretty stable.

Is the last reboot of your system long ago? If you have a spare system you could try to
  • Install Proxmox VE 5 on it
  • Test if performing a backup from this new system to your problematic drive works
  • Upgrade to Proxmox VE 6 as test for the upgrade of your main system (which I strongly advise to do before Proxmox VE 5 is EOL)
  • Test if performing a backup to your problematic drive works now
 
My last reboot has been 9 days ago. I'll try what you suggested and see how it goes. It's probably a good idea to at least start planning an upgrade to version 6 before 5 reaches end of life
 
  • Like
Reactions: Dominic
UPDATE: I successfully updated to Proxmxo VE 6, although I still get the error :/

Also here is the output from syslog. Not sure if it really adds anything else. The problem also seems to have gotten worse. When I do an automated backup I used to get 2 or 3 VMs erroring out because of this error. Know I get more than half. This is for VMs and VM templates. The problem seems to be more consistent with templates though

Code: 
Mar 10 19:49:19 vmworld qm[19640]: <root@pam> update VM 105: -lock backup
Mar 10 19:49:19 vmworld kernel: traps: vma[19644] trap int3 ip:7f2df51fac75 sp:7ffe7f434b80 error:0 in libglib-2.0.so.0.5800.3[7f2df51c2000+7e000]
Mar 10 19:49:20 vmworld pvedaemon[18620]: ERROR: Backup of VM 105 failed - command '/usr/bin/vma create -v -c /mnt/data/backup/dump/vzdump-qemu-105-2020_03_10-19_49_18.tmp/qemu-server.conf 'exec:gzip --rsyncable > /mnt/data/backup/dump/vzdump-qemu-105-2020_03_10-19_49_18.vma.dat' 'drive-efidisk0=/dev/zvol/rpool/data/base-105-disk-0' 'drive-sata0=/dev/zvol/rpool/data/base-105-disk-1'' failed: got signal 5
 
Last edited:
Hi!

I'm sorry it took so long to come back to you! I just sent a patch to the Proxmox VE developer mailing list which should help track your problem down. The patch is, however, not applied yet. You now have multiple options:
  • Clone the pve-qemu repository, apply the patch and build it yourself.
  • Wait and
    • Use the pvetest repository as this is the first place where the new package will be available.
      • You will be able to download only the relevant package (watch out for dependencies) or
      • add the whole pvetest repository (in a test setup)
    • Wait longer for it to arrive in your current (I assume pve-no-subscription) repository. This might take some time.
Hope that helps!

Best,
Dominic
 
Thanks so much! I’ll let you know what happens with the extra output. Hopefully that will help solve this issue
 
Dominic said:
Hi!

I'm sorry it took so long to come back to you! I just sent a patch to the Proxmox VE developer mailing list which should help track your problem down. The patch is, however, not applied yet. You now have multiple options:
  • Clone the pve-qemu repository, apply the patch and build it yourself.
  • Wait and
    • Use the pvetest repository as this is the first place where the new package will be available.
      • You will be able to download only the relevant package (watch out for dependencies) or
      • add the whole pvetest repository (in a test setup)
    • Wait longer for it to arrive in your current (I assume pve-no-subscription) repository. This might take some time.
Hope that helps!

Best,
Dominic
Click to expand...

I tried working with the pve-qemu repository to patch it and build it myself, although that didn't work out too well. Do you know when the pve-qemu package in the test repository will be updated with the patch?
 
Thanks for applying the patch for me! This makes my life easier. I installed it and overwrote the previous version by doing:

dpkg -i --force-overwrite ./pve-qemu-kvm_4.1.1-4_amd64.deb

and restarted. However, when I backup and encounter the error I don't receive any extra output (I also looked in syslog too). Is there a way I can tell if I'm using the patched version/I am missing something? Thanks!

EDIT: I unpacked the deb file and repacked it with a different version number and it seems to at least be recognized correctly
 
Last edited:
Only one line should have changed. Instead of
vma_writer_register_stream 'drive-scsi0' failed you should get
error for device 'drive-scsi0': vma_writer_register_stream: here an explanation.
 
you also need to shutdown the VM and start it again so that it picks up the new Qemu binary..
 
  • Like
Reactions: Dominic
Hm. It seems I'm not able to reproduce the error with VMs, however, I'm still getting the old error message with my current VM templates. I created a new vm and converted that to a template and tried to back it up a few times and still couldn't reproduce the error. I also made a full clone of a template and then converted that back into a template. I also couldn't reproduce the error with that. Any ideas?
 
Could you please try the following and simply copy all log output into the forum with [spoiler][/spoiler]?

118 is still one of those templates that produces errors and --mode stop works fewer times than snapshot => The first vzdump command is most likely to throw the error?

Code: 
cat /etc/pve/storage.cfg
mount | grep prox_backup
vzdump 118 --mode stop --remove 0 --storage prox_backup --compress gzip
vzdump 118 --mode stop --remove 0 --storage prox_backup --compress 0
vzdump 118 --mode stop --remove 0 --storage A_LOCAL_STORAGE --compress gzip
vzdump 118 --mode stop --remove 0 --storage A_LOCAL_STORAGE --compress 0
qm clone 118 FREE_ID --full 1
qm template FREE_ID
qm config 118
qm config FREE_ID
zfs get all local-zfs/base-118-disk-0
zfs get all local-zfs/base-FREE_ID-disk-0
zfs get all local-zfs/base-118-disk-1
zfs get all local-zfs/base-FREE_ID-disk-1
 
You must log in or register to reply here.

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!

Quick Navigation

Home Get Subscription Wiki Downloads Proxmox Customer Portal About

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!
Community platform by XenForo® © 2010-2024 XenForo Ltd.
Top Bottom
Back