Pool Showing Less Space

[NAS0_DATA1]

I got a pool with 2-3TB HDDs in a mirrored RAID. I use it to backup my Proxmox VMs to a NFS share. Recently I got a notification on TrueNAS showing space usage is at 73%. Pool shows 1.99 TiB (2.18 TB) being used and only 662GB available. Thought this was odd since in Proxmox it shows only 462GB being used. I have a daily snapshot task enabled with a 2 week retention policy set. Deduplication is set to OFF. What’s eating up the space in this pool? Why the difference in used/available space reporting between TrueNAS & Proxmox?

 

[chuck32]

I don't follow the math here, why is proxmoxx showing only 1.13 GB?

I have a daily snapshot task enabled with a 2 week retention policy set.

None the less I assume this is the reason. The VM backups in proxmox are stored as .zst files, each can eat up several Gb.

When you create a snapshop most likely most of your VMs will have changed, resulting in new zst files. These are now different from the other files and thus are stored additionally for 2 weeks.

This is not an incremental backup, i.e.

01.01.23 - VM1 - 3 Gb
02.01.23 - VM1 - 2.99 Gb
03.01.23 - VM1 - 3.01Gb
...

with daily changes storing a 3 Gb VM dump can use as much as 42 Gb of space.

* assuming your VM dumps work the same way my dumps worked when I played around a year ago with proxmox.

 

[NAS0_DATA1]

I don't follow the math here, why is proxmoxx showing only 1.13 GB?

None the less I assume this is the reason. The VM backups in proxmox are stored as .zst files, each can eat up several Gb.

When you create a snapshop most likely most of your VMs will have changed, resulting in new zst files. These are now different from the other files and thus are stored additionally for 2 weeks.

This is not an incremental backup, i.e.

01.01.23 - VM1 - 3 Gb
02.01.23 - VM1 - 2.99 Gb
03.01.23 - VM1 - 3.01Gb
...

with daily changes storing a 3 Gb VM dump can use as much as 42 Gb of space.

* assuming your VM dumps work the same way my dumps worked when I played around a year ago with proxmox.

I think you meant 1.18TB, not 1.13GB, but yes, same here, not sure why Proxmox shows that when the pool is 3TB. Thereby my question "Why the difference in used/available space reporting between TrueNAS & Proxmox?".

Thanks for the feedback about the snapshots. I guess I'll have to play around with that. In Proxmox I have the backup retention policy set to "keep the last 2". I do weekly backups on my VMs, so it's keeping 2 weeks worth of dumps. As mentioned originally, in TrueNAS I have the snapshot retention policy set to 2 weeks. Shouldn't TrueNAS be deleting snapshots older than 2 weeks?

 

[chuck32]

Shouldn't TrueNAS be deleting snapshots older than 2 weeks?

At least for me in the prior version the used space display was broken and often showed N/A.

But you go navigate to the datasets tab and from there to manage snapshots.
There you can see all snapshots, how much the snapshot uses and how much size is referenced and when (if) they are automatically destroyed.

If you change your retention policy, I'm pretty sure the old snapshots will still keep their original retention policy (on mobile, can't check right now). You would need to destroy them manually.

Edit: On a side note I don't use proxmoxx anymore but I changed the way I backup VMs on truenas. Every week I replicate the VMs to another pool, only keeping that single snapshot. But some of my VMs already have docker volumes on my data pool so some applications also live outside of the VM.
Backups of the VMs themselves are done inside the VM, i.e. creating backups of all the docker containers etc.