Large incremental backup size for file level backup

[kevin.boddy]

Hi,

I have a customer wanting to do SQL database file backup. We installed the Veeam Agent for Windows and configured a file level backup of the SQL database+log folders.
The CBT driver is installed correctly. The full backup size is 111GB and the daily incremental backup size is 85GB.

This SQL server is also a VM and the VBR backup reports are showing that the incremental size for the same day are 15GB. The SQL server is the only VM in the VBR backup job.

Is it normal for the incremental backup on the Veeam Agent to be so much bigger than the VBR incremental backup with a working CBT driver?
We are trying to get the backup size down on the Veeam Agent incremental backup, is there something we are doing wrong here?

Thanks
Kevin

[HannesK]

Hello,

configured a file level backup

that means every file that is changes will be backed up as a whole.

If you would use "entire computer" or "volume level" backup, then the size would be similar to VM-based backup. Change block tracking (the CBT driver) has impact on performance - not on size.

Best regards,
Hannes

[kevin.boddy]

Hi Hannes,

Thanks for the reply. Is there documentation on how that works exactly?

I went through the documentation I could find and it seems to indicate that the CBT driver is used to retrieve only the data blocks that have changed which should affect the incremental size.
Is that not the case?

https://helpcenter.veeam.com/docs/agent ... tml?ver=50

"2. Veeam Agent reads data from the created VSS snapshot, compresses it and copies it to the target location.

For volume-level backup, Veeam Agent copies data blocks of the whole volume.
For file-level backup, Veeam Agent creates a volume inside the backup file in the target location. The content of the volume in the backup file is synchronized with the volume on the source: Veeam Agent for Microsoft Windows copies only those data that you have selected to back up.

During incremental backup, Veeam Agent uses Changed Block Tracking (CBT) to retrieve only those data blocks that have changed since the previous backup session. To learn more, see Change Block Tracking."

Thanks
Kevin

[HannesK]

Hello,
agree, that documentation is hard to understand. I will talk to tech writers how we can improve it.

The point is that "file backup" backs up "file-based". There is no "block level" backup with "file-based backup" today (version 5). If 1MB of an 1GB file changes, then 1GB will be backed up.

Just in case your next post would be requesting "only back up changed blocks of a large file", then I can say that it is a "classic" feature request, and something we plan to improve.

Best regards,
Hannes

[kevin.boddy]

Thanks Hannes.

I think the documentation should have an exclusion regarding CBT and file level backups and include a section describing what you just mentioned about how the file backup works.

Thanks
Kevin

[HannesK]

also agree on that (and I already asked for that)