Prawdopodobnie mój obraz ucierpiał z powodu dużego obciążenia operacjami I/O mimo, iż miałem Virtio-SCSI: https://forum.proxmox.com/threads/qcow2-corruption-after-snapshot-or-heavy-disk-i-o.32865/

qemu-img check disk.qcow2

ERROR OFLAG_COPIED data cluster: l2_entry=8cfdb0000 refcount=1

43782 errors were found on the image.
Data may be corrupted, or further writes to the image may corrupt it.

424996 leaked clusters were found on the image.
This means waste of disk space, but no harm to data.
524288/524288 = 100.00% allocated, 10.45% fragmented, 0.00% compressed clusters
Image end offset: 41744859136

If -r is specified, qemu-img tries to repair any inconsistencies found during the check. -r leaks repairs only cluster leaks, whereas -r all fixes all kinds of errors, with a higher risk of choosing the wrong fix or hiding corruption that has already occurred.

qemu-img check -r all disk.qcow2

Repairing OFLAG_COPIED data cluster: l2_entry=8000000885b30000 refcount=2
The following inconsistencies were found and repaired:

    424996 leaked clusters
    36838 corruptions

Double checking the fixed image now...
No errors were found on the image.
524288/524288 = 100.00% allocated, 10.45% fragmented, 0.00% compressed clusters
Image end offset: 41744859136

UWAGA! Lepiej zrobić najpierw opcję LEAKS, a potem w ostateczności ALL.

qemu-img check disk.qcow2 

No errors were found on the image.
524288/524288 = 100.00% allocated, 10.45% fragmented, 0.00% compressed clusters
Image end offset: 41744859136

Link do manuala: https://www.qemu.org/documentation/