If you try to restore a proxmox backup from NFS server and receive the above error messages there might be hope.
Most answers from proxmox forum say the backups are broken or there are hardware issues (RAM,CPU). So here is how you can fix your backups, because in our case, hardware and backups were fine.
We assume your backup files are compressed in whatever way. (lzo. gzip…) In our case, it was a vma.lzo-File.
First, try to get the backup files from a remote storage to local storage so extraction is faster.
then decompress the image.
lzop -xF vzdump-qemu-110-2022_10_24-03_26_22.vma.lzo
Note the -F flag. It is to ignore checksums.
If that works, check the vma-file at the end:
vma verify vzdump-qemu-110-2022_10_24-03_26_22.vma
If exit-code is 0, all is good. Then just restore the vma file.
Conclusion:
Looks like proxmox in older versions is not always able to decompress & verify a file stream at the same time and giving up too early. We also saw some kernel traps in ringbuffer (dmesg):
[10943846.988380] traps: vma[10496] trap int3 ip:7f5a2813d3f1 sp:7ffdd0f715b0 error:0 in libglib-2.0.so.0.5000.3[7f5a280ed000+112000]
[10944177.919042] traps: vma[12443] trap int3 ip:7fb4fde6d3f1 sp:7ffea942ebe0 error:0 in libglib-2.0.so.0.5000.3[7fb4fde1d000+112000]
[10950011.679328] traps: vma[13547] trap int3 ip:7fcbaab263f1 sp:7ffd1f02ff90 error:0 in libglib-2.0.so.0.5000.3[7fcbaaad6000+112000]
Decompressing and importing in 2 different steps did the trick for us.