Issue with datastore usage

[mrbyte]

Hello everyone

I have a PBS where I store backups of a VM with Nextcloud (143 GB) and another with Bitwarden (22 GB). For the past few days, I’ve noticed that the datastore usage has suddenly increased considerably, from around 50-55% to 75-80%, even though I haven’t added any new data:

To try to resolve this, I’ve run the “prune jobs” and “garbage collect jobs” options without success.

I’ve also removed the PBS datastore from the Proxmox node and deleted the disk in the PBS used as a datastore for storing the backups. Then, I re-added the PBS datastore to the Proxmox node. But this didn’t work either. On the first backup I create of the Proxmox node VMs on the PBS datastore, it already occupies 78% of the storage, about 365 GB.

The most curious is that this issue doesn’t occur with a second PBS that backs up VMs from two other different Proxmox nodes; in that setup, datastore usage remains stable at around 30%.

I'm a bit lost here. Any idea what could be happening?

Thank you very much and kind regards.

Albert.

 

[seiji]

Hello mrbyte.

As you know, I am still a beginner with Proxmox PBS. Based on your description,
could there be any duplicate files on PBS1?

Because As you mentioned, There are no issues on PBS2,
However if the same backup files are causing problems on only PBS1,

it might be helpful to search for duplicate files.

The command fdupes -r /mnt/pbs-datastore can be used to search for duplicate files in your datastore directory.
(forexample my datastore is /mnt/pbs-datastore directory. )

https://packages.debian.org/search?keywords=fdupes

fdupes -r /mnt/pbs-datastore

Best regards,

Seiji

 

[mrbyte]

Hello seiji,
Thank you very much for your help...

I've run the command that you suggested:


After I have run a prune job task and a garbage collector and run again the command:


And this is a capture from the garbage collector, only wiped 8 Gb:


The datastore usage is worsening:


I'm going to delete all backups from the datastore and see what happens...


Best regards.

Albert

 

[mrbyte]

Hello!!
I have conducted another test: I started from scratch by setting up a new Proxmox Backup Server and a datastore with the XFS file system instead of ext4, which I had previously used.

Afterwards, I ran the backup job again, and the same issue occurred, more than 70% of the datastore is occupied:



In total, there are about 360 GB of data, even though the VMs occupy 158 GB and 24 GB respectively. It seems like a single backup job is being performed, but in reality, there are two because the data is being duplicated:




I don’t think the issue lies with the VMs themselves. It seems more like a problem with the Proxmox node or the Proxmox Backup Server itself. I’m unsure what further tests I could perform to continue troubleshooting this.

Any idea what could be happening?

Thank you very much and kind regards.

Albert.

 

[mrbyte]

Hello,
Two more screenshots showing the storage used by the VMs:




Best regards.

Albert.