[SOLVED] Backup size of VM is too big

[upcnuwft]

Hello,

I have around 10 Debian VMs running on my Proxmox server.
The backup size for each is around 700-900 MB. But one is around 17 GB.

And I have absolutely no idea why this one VM is so big.

The application running there is Jellyfin.
I already tried to search the error there: https://forum.jellyfin.org/t/reduce-backup-size-of-jellyfin-vm/3687

The VM does only have one drive and if I run df -H it shows me that this drive uses 2,1 GB.
How can the backup be 17 GB then? This is so mysterious... I can't figure it out.

Do you guys have any idea what's going on here?

 

[apoc]

What is the VMS disk size?
Have you enabled trim / discard on the disk?

 

[upcnuwft]

Every of my VMs has a disk size of 32 GB.
And discard is enabled everywhere.

Probably the used size was 17 GB once but now is only 2 GB.
Is it possible that the Proxmox backup job can't detect that it is smaller now?
For most VMs it says: INFO: backup is sparse: 29.46 GiB (92%) total zero data
But for this one: INFO: backup is sparse: 15.65 GiB (48%) total zero data

PS: qemu guest agent is installed, if this matters.

 

[upcnuwft]

Well I just found this one here: https://pve.proxmox.com/wiki/Shrink_Qcow2_Disk_Files
And fstrim -av looks actually very good: /: 27,5 GiB (29552508928 bytes) trimmed on /dev/sda1

INFO: backup is sparse: 28.49 GiB (89%) total zero data

wow... nice. Done. Solved.
How have I missed that?