Lost weekly GFS after v11 upgrade

[karsten123]

Sorry, if it was answered already.

I had a backup copy job with 30 rps and 4 weekly GFS points in VBR v9.5.4. So i had 34 rps altogether.
After the upgrade to v11, veeam deleted all weeklies and left 30 rps with 1 full and the incrementals.

I did noticed the part in the upgrade checklist for v11a, but i dont really know what to do, tbh.

What is the right way to manage this?

Tia Karsten

[foggy]

Hi Karsten, this looks similar to the issue where monthly GFS points were deleted during the upgrade in case the latest full backup was corrupted. But your case is a bit different - please contact our technical support for a closer investigation.

[Gostev]

BTW, does 30 restore point not cover the entire 4 weeks period already? 4 weeks = 28 restore points in case of daily backups.

[karsten123]

In fact it is not that tragic, because i have insider protection for this customer.

What is the right way to prepare for this and transfer the GFS to the v11 method?

The part of the upgrade checklist is not that clear to me.

Do you have backup copy jobs with synthetic GFS full backups? Before the upgrade, make sure that all GFS candidates (incremental restore points created on days when GFS was scheduled and that are expected to be transformed into full GFS restore points) are already transformed into GFS restore points. To force the backup copy job to transform all GFS candidates, you can temporarily decrease the short-term retention to a value less than the number of restore points between the latest restore point and the most recent GFS candidate and then wait till all the candidates are transformed.
Before Veeam Backup & Replication version 11, Veeam Backup & Replication created GFS candidates on days when GFS was scheduled and only then transformed them into full GFS restore points according to the short-term retention. For more information on how restore points were transformed, see Synthetic Weekly Full Backups. Starting from Veeam Backup & Replication version 11, Veeam Backup & Replication creates GFS restore points according to a new schedule and creates them right on the scheduled days. For more information, see Changes in GFS Retention. After the upgrade, Veeam Backup & Replication no longer transforms previous GFS candidates into full GFS restore points. This means, that all GFS candidates lose their GFS status, they become regular incremental restore points and are deleted according to the short-term retention policy.

Is it worth to set to 60 RPs, to cover the 2 months, wait for the GFS points to create and than reduce to 30 afterwards?

Thx Karsten

[karsten123]

Can someone please give me some insights? I still don‘t get it.

[Gostev]

It seems you're suggesting exactly the opposite retention change (increasing) instead of what the quote says (reducing). According to the explanation above, you should temporarily reduce RPs to just a few days to get weekly GFS created. This makes sense according from how I remember GFS retention worked back then. I admit it's not a very clean process since you will not have those 30 daily backups after the retention change, so it might be best to copy those backups aside in case you need to restore while the new restore points are being created following the upgrade.

[karsten123]

That makes sense.
But indeed. It is not a smooth transition experience.

Thank you Anton

[JacquesvZ]

Hey

This post caught my eye in "THE WORD FROM GOSTEV" mailer.

Currently running 10a with multiple "secondary" backup copy repositories. All the backup copy jobs are set with GFS of:

- Last 2 restore points
- Last 3 weeklies
- Last 6 monthlies

Does this mean Veeam will purge my weeklies and monthlies if I upgrade because my retention is only 2 days? Or am I miss understanding?

Thanks

[Dima P.]

Hello folks,

Pre v11: GFS retention logic was working based on number of restore points, so oldest GFS restore point is not removed by retention unless you got required amount of GFS restore points. With synthetic full backup GFS incremental restore point is marked as GFS but full is created only after such increment is merged into very first full backup (then GFS full backup is kept away from the regular backup chain).

v11 and later: every GFS restore point has an expiration mark on it, so GFS restore point is removed according to the retention at the given day without any delay (moreover, system retention job can process GFS full backups meaning that you do not actually have to run a backup job). GFS full backups are created immediately at the set date.

For backup copy jobs with synthetic full backup GFS there is a trick to make sure that all GFS restore points are actually synthesized into full backups: Help Center - Upgrading to Veeam Backup & Replication 11 or 11a (the one that was quoted in the previous post).

JacquesvZ,
Are you using synthetic full backups for backup copy GFS?

[sps-tech]

I'm lost on this one as well. Can we maybe get a concrete example of a "before", a "what you do to prepare", and what that looks like after?

I have a GFS Backup Copy job that is set to 8 RP, then full backups for 4 weekly, 6 monthly, 5 quarterly, etc. The source backup job runs daily. I don't particularly care a lot if I would lose my daily GFS copies, but I would care a great deal if I lost any of the weekly/monthly/quarterly fulls. So I would need to reduce my RP to less than 7 (so another weekly would get created), or increase to 8+4+6+5+whatever to be sure it keeps ALL my restore points??

[Gostev]

Reduce your RP to less days that what has passed since the day of the week when your weekly full is scheduled. The goal is to get put a job into the situation when the restore point that needs to become a GFS full falls out of short-term retention and normally needs to be deleted by the retention policy. Because it fits the GFS schedule, it will be set aside and seal it as weekly GFS full instead.

In any case, with your short-term retention policy (what you call RP) being so short you already have all GFS backups were already set aside and sealed long ago, except most recent weekly GFS backup. So in any case you're only risking that one weekly full only. The OP's situation is unique in that his RP range of 30 days also covers his entire weekly GFS retention period of 4 weeks.

[eubenitez]

Hello,

Same question over here. I don't really full understand what I have to do previously to upgrade V&R.
For example I have the following backup copy job

Restore points 30
Keep the following restore points as full backups for archival purposes

Weekly backup --> 4 --> Saturday
Monthly backup --> 12 --> first Sunday of the month
Quarterly backup--> 0
Yearly backup --> 5 --> First Sunday of the year

As far as I read should I reduce every weekly monthly and yearly backup restore points?. and the 30 ?

Thanks in advance.

[veremin]

Just to be sure - you are upgrading from 10a to 11a, correct?

[eubenitez]

Correct veremin