Understanding NAS backup archiving

[talltim]

I have some daily jobs that backup various CIFS on a NetApp SAN to a physical server added as a backup repository. They are set to Keep all file versions for 28 days
I then have Keep previous file version ticked and set for 3 years, going to an Azure BLOB (Standard, not archive) repository
Am I right in understanding that it should keep 28 backups on the physical server, then archive any older and move them to the Azure repository? Or have I missed something and there needs to be an extra process to move the older backups?

[chris.childerhose]

Yes, it will archive older files past 28 days to Azure. Please see this for retention scenarios - https://helpcenter.veeam.com/docs/backu ... ml?ver=110

[talltim]

I ask because my Backup Repository server filled its disk over Christmas. Looking at the files on it, it seems to have kept all the backups (or large folders with the same dates on them, at least). When I look at the Azure Archive Repository I can see what I assume are the archived backups so that element seems to work. Is Veeam supposed to automatically remove the older backups that it has archived from the Backup Repository as part of the process? Its possible I have missed a setting or that there is a fault, in which case I will log a support call.

[talltim]

This quote from page link by Chris implies that it works automatically, which is what I had assumed, if there is no setting that controls it I could look at I will log a support call

On day 9, the file is removed from the source, file version 9 (denoting the missing file) is added to the backup repository, file version 5 is moved from the backup repository to the archive repository, file versions 2 and 3 are deleted from the archive repository to keep 2 file versions of the deleted file (versions 4 and 5).

[chris.childerhose]

I would suggest a support call to ensure things are operating as they should, as it sounds like something is not based on what you indicated over the holidays.

[Gostev]

Note that the archiving process moves only OLDER versions of files. The last (current) version will always stay in the backup repository, because it is needed for the entire file share recovery.

[talltim]

Ah, with one sentence I think you have made me understand the issue! So a combination of all the backup files will contain 1 copy of each of the files in the file share (plus 28 days of any changed files). This means that (ignoring any compression) the backup of a file share will be at least the same size as the file share. If there is a file that hasn't changed since the first backup, it will be in the first backup file, and this first backup file will get smaller as the files in it are changed and the changed versions will get included in the latest backup, with the previous versions getting archived after 28 days.

[Gostev]

The first two sentences are absolutely correct. The last to sentences relay the idea correctly, but technically NAS backup structure is much more complex than just "backup files" like you have with image-level backups. NAS backup is effectively a proprietary NOSQL database optimized to store large amount of data directly inside of it (whereas normally databases only store "metadata" which refers to some external location storing all the "heavy" stuff the application needs).