Retention Period: Deletion of old files

[heikoh81]

Hi everyone,

I'd like to use Veeam Agent 2.0.0.700 Free to backup my personal computers to my NAS.
Before deploying, I'm testing it.

I have the following setting:
"Keep backups for the past 4 days when the computer was running".


I have done backups on:
Oct 4th, Oct 9th, Oct 10th, Oct 11th and Oct 13th (so I count to 5 backups on 5 different dates).
However, backup files from Oct 4th have not been deleted today, see attached screenshots.


Where's my mistake?
To my understanding of the documentation at https://helpcenter.veeam.com/endpoint/11/retention.html, backup files from Oct 4th should have been deleted today, as those are from 5 backups ago...
Thanks for your answers,

Heiko / Germany

[Dima P.]

Hello Heiko,

It looks like you have two backup chains (.vbk is a full backup and .vib is an incremental backup). Please check if you have periodic Active Full Backup configured under Advanced > Storage tab (on this step of the wizard).

[heikoh81]

Hi Dimitry.

Indeed, periodic Active Full Backup is checked for my testing.
So, how does this affect retention policy? Will old files only be deleted after 4 full backups?

[Dima P.]

If Active Full is on, backup chain is removed only after the last incremental in this backup chain is retired. Please take a look at the article I sent before: it explains retention logic in case you are running periodic full backups.

[heikoh81]

Read it, still don't understand it.
Ran another job today, now I have Restore Points for 5 days. Retention policy still set to 4.

Shouldn't have at least one restore been deleted today?

[vClintWyckoff]

You likely have the Active Full backup set to run once per week - the default is Saturday. The important thing to understand is the backups are a chain - so we cannot delete the previous restore points until a new chain is started and the retention policy has been satisfied. With the scenario of 4 days but the full is set to run every week would result in 7 restore points until a new full is initiated and then 4 incremental backups (have to satisfy the 4 day requirement) would occur before that previous chain is deleted.

So if you only want 4 days you should configure the active full to run on say Saturday and Wednesday for instance.

I hope this explanation helps.

///////

Looking at your screenshot - You have 2 backup chains. The eldest was started on 4/10 and actually only has 3 days worth of restore points in it - so my guess is that you initiated a manual active full backup on 10/10 after the incremental ran? So actually this chain does not even meet the retention policy and now must wait for another active full and 4 more restore points.

[heikoh81]

Thanks. Performed another manual full backup, and now, the old files have been deleted.

Before:


After:


By the way: What's the best strategy for a private environment (just to be protected against Ransomware or Updates gone bad): Endless incremental backups or Full Backups?

[vClintWyckoff]

With your retention set for 4 days you'll always run into this scenario b/c there are only 7 days in the week so one chain will have 3 restore points and the next will contain 4.

By the way: What's the best strategy for a private environment (just to be protected against Ransomware or Updates gone bad): Endless incremental backups or Full Backups?

For this, you definitely want to make sure that your backup target is disconnected when the backup completes or you're sending this off further to another location that's not directly accessible from your workstation. You can achieve this in a few different ways - Veeam Backup & Replication Backup Copy Jor or even something as simple as a robocopy but that method is outside of the Veeam product so your restore points won't automatically be pruned when they expire. Periodic Fulls or Forever Incrementals - either one does not provide any type of "Ransomware" protection with the device readily accessible to the malware. A personal best practice I like is to have periodic fulls at least every month or so this way the chain gets reset in case something goes bad.

Hope this helps.