Urgent help needed - lvm thin pool problems (no disk space problem)

User Support
1

Hello dear qubes community - i am using qubes for quite some time and i have a certified laptop with 1TB of disk space.

Following problem:
I use a linux compatible wifi dongle, sometimes when i put it into sys usb the whole system freezes. Today the tlsame happened again i usually turn the computer off and restart, i did the same with just keeping the power buttons pressed.
After reboot no vms start and i have the following error with every vm:

Thin pool qubes_dom0-vm–pool-tpool (253:9) transaction id is 10690, while expected 10691
Failed to suspend qubes_dom0/vm-pool with queued messages.
Failed to update pool qubes_dom0/vm-pool

I unfortunately have importabt personal data on it and the last backup was over a month ago.
I am only using 30% of the disk space so i do not think it is a memory issue because that was the only thing i found when searching thr qubes forum.

I am worried that if i turn the computer off and on again i will be unable to get into the system.
If somebody could help me and solve this issue i would also make a 100 dollar donation.

I would like to:

  1. backup everything (how when i can start no vms?)
  2. Have the backups recovered on another laptop with qubes. If the backup is working then reinstall qubes and restore the backup

Thanks in advance!

2

Hi @Ruben12, I’m sorry you’ve encountered such a problem.

For me, depending on the importance of the data on the disk, I would
start by taking a full image of the disk - attach a usb storage device,
boot from a live iso, and then use dd to get the image.

Almost anything you do has the potential to trash those volumes, which
is why a disk image is an essential.

At least a plausible approach would be to try to fix the error.
You could try this, as root in dom0:

  1. Take a backup of the volume group configuration.
    vgcfgbackup qubes_dom0 -f /home/user/VG_backup
  2. Edit the new file, and look at the section under logical_volumes->vm-pool->segment1
    change the transaction_id to 10691.
  3. Save the edited file.
  4. Read the edited data back in:
    vgcfgrestore qubes_dom0 -f /home/user/VG_backup
  5. Activate the edited volume:
    lvchange -ay qubes_dom0

Another approach would be to try to repair the pool:

  1. Deactivate all qube volumes in the thin pool.
    lvchange -an /dev/qubes_dom0/vm*
  2. Try a repair on the thin pool:
    lvconvert --repair qubes_dom0/vm-pool
  3. Reactivate the qube volumes.

Note that it may be possible for you to extract the data directly
from the qube volumes.
In dom0, try mounting an image directly:
sudo mount -o ro /dev/qubes_dom0/vm-work-private /mnt
You may then be able to copy out the important data to another device.

Using these methods you do at least stand a chance of getting your data
back. I would take a disk image, try to extract the data manually, try
to repair the pool, then try to “fix” the error, in that order.
If you need help at any stage please follow up.

I never presume to speak for the Qubes team.
When I comment in the Forum I speak for myself.

4 Likes