Incremental message for PBS backup

[William Edwards]

When creating a backup of a VM not previously backed up to PBS, 'backup was done incrementally' is almost always shown:

INFO: backup is sparse: 2.83 GiB (1%) total zero data
INFO: backup was done incrementally, reused 4.91 GiB (2%)

I take it 'incremental' means space was 'reused' as the backup is sparse. Maybe you could consider not using the word 'incremental', as claiming an only technically incremental backup is incremental can be quite confusing.

 

[ph0x]

PBS splits all backup data in chunks. You probably have a similar VM, which was already backed up, so those chunks could be reused.
And on a side note, therefore your backup was incremental.

 

[William Edwards]

PBS splits all backup data in chunks. You probably have a similar VM, which was already backed up, so those chunks could be reused.
And on a side note, therefore your backup was incremental.

I understand, and while that is technically true, I think the message is confusing from a practical standpoint . See here the first question about this message: https://forum.proxmox.com/threads/first-backup-of-pve-vm.82871/#post-364565

 

[ph0x]

Technically correct is the best form of correct.

 

[oversite]

I understand, and while that is technically true, I think the message is confusing from a practical standpoint . See here the first question about this message: https://forum.proxmox.com/threads/first-backup-of-pve-vm.82871/#post-364565

I don't follow, for me incremental is the perfect word here, how is it confusing you? Incremental means for me the changes between now and the previous backuped data. It's easy to understand and also true. What bothers you here?

 

[ph0x]

That it was the first backup of a specific VM, therefore no data of that VM could have been backuped before.
But taking into account the way the backups are stored I agree with you that this is the correct term.
Might still be a bit misleading for beginners, though.

 

[dcsapak]

just fyi, the sentence:

PBS splits all backup data in chunks. You probably have a similar VM, which was already backed up, so those chunks could be reused.
And on a side note, therefore your backup was incremental.

is not completely correct.. while it may be that the backup shares blocks with already existing ones, for privacy reasons (e.g. untrusted backup server), the client does not actually
know about chunks from different backup groups

the reason that on the first backup something like

INFO: backup is sparse: 2.83 GiB (1%) total zero data
INFO: backup was done incrementally, reused 4.91 GiB (2%)

can happen, is that not only zero chunks are reused, but all
so if a chunk repeats in a single backup, it only gets uploaded once (so here it seems that an addition ~2GiB was from chunks that repeated)

 

[ph0x]

Thanks for clarification!

 

[William Edwards]

Ack all that. My point is: I think it's confusing for 'the average user'.

 

[Darkk]

I too was confused by this behavior and wanted to make sure the backups a new VMs are successful and able to restore without issues. I've already have two more PBS servers as a test and all VMs were able to restore without problems.

The wording is technically correct. At first I was like how is it able to reuse the data if this is first time backing it up. Now it makes total sense as it's part of de-duplication process on the same datastore. It's amazing how it was able to save so much disk space and keeping everything in a restorable state.