Retention Policy on Disk

[LeisJohn]

Good morning,
We just started using Veeam in January of this year. I'm required to keep the 12-Monthly Full 8-Weekly Full with daily incremental. I need some help understanding how I setup GFS policy.

Our backup infrastructure contains the following
1 - ExaGrid 7000
1 - ExaGrid 13000
7 - Veeam Proxy

[veremin]

Hi, John,

I believe you have already created backup jobs, right?

Then, just create backup copy job, specifying existing backup jobs as a "source" for it. In the settings of backup copy job select "Keep backups for archival purposes" option and specify below the number different GFS restore points you want to keep - 7 daily, 8 Weekly, 12 yearly.

I'd also recommend reviewing the blog post that describes ultimate backup infrastructure; might be helpful for you.

Thanks.

[LeisJohn]

v.Eremin-
Correct I have already created my backup jobs. If I use the Backup Copy Job will I not be using twice the space on my ExaGrid?

[veremin]

The backup copy job doesn't copy backup file as a whole, but rather synthetically creates required restore points in "target" location from VM data in source backup repositories. So, you will definitely have to allocate some space for restore points created by backup copy job. Though, the actual amount of it depends on how frequently you run the backup copy job, amount of changes made between these intervals, and many other factors. Thus, only the tests are likely to give you the final numbers here.

Please, also bear in mind that GFS restore points are stored as full backups.

It might be worth using different device as a repository for backup copy job in order not to have all eggs in one basket (both backup and backup copy data).

Thanks.

[LeisJohn]

v.Eremin-
My issue is that I need 4 weeks of backups available 4-Friday full & 4-Mon-Thurs Incremental in my case that's a lot of space and now I need space for separate full backups if I want to use GFS. I really wanted to keep the monthly full on the ExaGrid since they would Deduplicate. Is there a way for me to manually delete Weekly chains and update the database in order to meet my goal?

[foggy]

All restore points falling out of the specified retention policy will be deleted automatically. With your initial requirements (8 weekly and 12 monthly backups), you need space for 20 fulls plus the required number of incrementals. You can either implement GFS either using backup copy jobs or simply copying the full backup once a month to some other folder on the storage.

[veremin]

Is there a way for me to manually delete Weekly chains and update the database in order to meet my goal?

Can you elaborate on it a little bit? As mentioned above, the number of restore points kept on disk is controlled by retention policy. Besides it, there is no way to delete existing backup chain and start it anew. Thanks.

[LeisJohn]

Foggy-
That's great but I'm still using more space then needed since ExaGrid is Deduplicating the backups especially the full. With dedup I have ample space to hold my required I just don't have the extra space that it seem Veeam requires in order to use GFS policy. The full backups aren't the problem when it comes to space it's the daily incremental, because they don't dedup very much if at all. So Veeam doesn't have any way for me to do a manual delete and rescan the repository and cleanup the database. Most other software allows you to run an inventory and then catalog it. I may have to revert back to my old ways and put the monthlies off to tape and only keep say 65 restore points on disk but, this causes time issues if I need to restore from 3 months ago. This is also a huge waste of my ExaGrid appliance as I see it.

[foggy]

With dedup I have ample space to hold my required I just don't have the extra space that it seem Veeam requires in order to use GFS policy. The full backups aren't the problem when it comes to space it's the daily incremental, because they don't dedup very much if at all.

Do you mean increments created by the backup copy job here? You can minimize the number of those by setting the number of restore points to keep to the minimum "2" value. This will still allow for GFS archiving of the required fulls, while requiring additional space for just one incremental file.

[LeisJohn]

Foggy-
I'm required to keep 4 weeks of incremental backup which means 4 or 5 full.

[foggy]

But we are talking about backup copy job here, aren't we? As you've already said, you need "4 weeks of backups available 4-Friday full & 4-Mon-Thurs Incremental" + "space for separate full backups if I want to use GFS". That's correct and my point was to eliminate the need for additional increments created by the backup copy job (that would implement your GFS strategy), as they are not deduped well.

[dellock6]

John,
maybe the misunderstanding here is about the mix of a regular retention and the GFS. Let me suggest another setup, and then tell me if it's ok for you:

4 weeks of incrementals: these are the first level of backups and restores, so it's better to create this via regular backup jobs, if you have enough space (maybe the exagrid 7000?)

GFS rotation: has Alexandere has explained you, in order to have GFS, backupcopy also needs some regular restore points in its job. Those restore points are NOT the ones created by the regular backup job, maybe this is the reason of the misunderstanding.
Anyway, since you already have thow 4 weeks of restore points, you can have only the minimum amount of restore points in the backupcopy job (2 points as already explained), since the only use for these is to create the GFS rotation.

Hope it's understandable.

Luca.

[LeisJohn]

Luca-
Please excuse me but, I just really don't understand it. I'm going to contact Veeam support and try to get someone to work with me so that I understand exactly what I need to do. I haven't seen any paper that explains it or draws it out. Thanks for your help and input anyway.

[foggy]

Probably having someone to talk with directly will be more effective than forum correspondence. However, the main point is that you can achieve what you're after (4-Friday full & 4-Mon-Thurs Incremental with monthly fulls) using a daily regular backup job set to keep 20 restore points and a monthly backup copy job with retention set to "2", i.e. without the need for additional space (except only for a single increment). Another, probably more straightforward approach is just a monthly manual/scripted offload of the full backup to another folder on the same storage.

[LeisJohn]

foggy-
That's for your input. I just need to get an understanding one on one exactly how this works when your talking about 13tb full and 2tn incremental. thanks again.

[foggy]

Sure, though feel free to ask here if you need some additional clarification.