GFS Retention Policy

[foggy]

Ok, doing so will allow to have Sunday's data in the backup copy job restore point.

[SpiderPig]

Thank you kind sir.

[SilentDeath]

Hello,

I've setum a job containing only one VM which is a virtual fileserver.
This job runs daily at 07:00 PM and will hold 7 restorepoints.
I make use of Incremental jobs and create an active full backup every saturday.

Now, this customer asked some GFS implementation.
We want to achieve: 4 full, weekly back-ups and 12 monthly back-ups.
There is no offsite storage needed, the restore points will be stored on the same server.

So the first idea was to work with back-up copy jobs which works somewhat other than I expected.
While creating the back-up copy job, I assume I have to select 16 restore points.
I don't really get what the sync interval does?
Now, in my copy job I selected 4 weekly restore points and 12 monthly that need to run on Sunday.
I've set this up on 10 may and checked today the results.
I expected in this folder: 2 full back-ups: from 15 and 22 May.
The result however, is one full backup from 10 May, the day I created the copy job and the rest are all incrementals based on that full.

I recently started diving in Veeam powershell and it looks like I can achieve my wishes far more easy with this.
I was thinking of creating a script that runns every sunday (full backup is on saturday) which copies the active full backup to another folder.
The script checks if not more than 4 restore points are in that folder. If yes, before deleting a restore point copy if to a monthly folder.

[PTide]

Hi,

I don't really get what the sync interval does?

Basically it specifies how often the backup copy job will synchronize primary backup chain with the secondary backup chain. What's the value of sync interval in your backup copy job?

While creating the back-up copy job, I assume I have to select 16 restore points.

You can select any number of restore points, GFS will just keep weeklies and monthlies.

Thanks

[SilentDeath]

The value is set to sync every day at 12:00h.

The problem is, I don't want a secundary back-up chain.
I only want my copy job to store the active full restore point, I do'nt care about the incrementals being synced or not.

[PTide]

The result however, is one full backup from 10 May, the day I created the copy job and the rest are all incrementals based on that full.

GFS kicks in at the moment when the retention condition is met (16 restore points in your case). You can set your backup copy job retention to 2 restore point and leave sync interval as it is now. That will make backup copy job to copy changes daily (so you will have the most recent restore point safe) and start creating GFS restore points sooner.

Thanks.

[Dja]

We have been with Veeam now for 6 months or so, previously a Commvault shop. We are trying to achieve the simple extended protection policies that we had in place with Commvault and previous vendors. That was to create a secondary copy of Monthly and Yearly jobs in order to extend the retention period. In our Commvault environment, all longer retention was to tape. We did start out trying to do this with tape but abandoned that pretty quickly having witnessed Veeam's lack of capability in that sphere.

We have also now significantly minimised the workloads we are seeking to protect for longer than the 30 days that we hold on the primary disk repository. Limited now to Exchange, Sql and File. 3 Jobs.

At present we are running active fulls at end each week and incremental daily. Backup is across the FC fabric so all pretty quick.

What we want to do with longer retention is to create secondary copies of the weekly jobs (last Friday of the month or last Friday of year) to another disk target in separate data centre. We may eventually offload to a cloud disk service but at present will be deduped to disk. The combined total for the 3 jobs is ~10TB for an active full and will be able to get several years on disk deduped, given the current rate of change.

I have just got off my second Veeam support call regarding how I might do this, but it seems there is just no capability to wait for my end of month to complete and then copy to secondary location.

We obviously need to rethink this completely (rather than think in Commvault or other vendor terms) but not sure where to start. I am surprised that anyone can run a secondary copy in the way I describe, we are not attempting to do anything complex here, on the contrary the logic is very simple. ie. Wait until job finished, send it somewhere else, apply different protection policy.

I appreciate any advice from you experienced Veeam folk.

Any Veeam people reading, would appreciate advice as whether this simple capability is on the dev road-map.

[PTide]

Hi,

It seems that backup copy job is what you need. Please review the thread and feel free to ask questions.

Thanks

[Dja]

Right, the obvious choice. But the support team advises me that what we are trying to do just can't be done with the options available. My observations and experimentation back that up. Powershell was one (unsupported ) suggestion. If you have a magic panacea for this I would be very pleased to hear it. Thanks in advance.

[Shestakov]

Could you specify what exactly can be done here? It looks like you need to create a backup copy job with set monthly GFS retention.
By the way, what was your support case number?
Thanks!

[WellMark]

I'm relatively new to Veeam, and I'm a little confused on configuring my Monthly backups.

I have a machine being backed up nightly, with a 90 day retention.

I want to have a backup go to a remote DR site on the last day of every month, for the last 15 months.

Using a Backup Copy Job:

1st screen: Copy every? If I only want once a month what do I choose here?

3rd screen: Restore points to keep? 15 for the last 15 months?

Check the Keep the following restore points for archival purposes. Monthly backup: 15 Last day of the month?

Any good/best practice settings under Advanced?


Thank you!!

[Shestakov]

Hello Mark,
Backup Copy job has both simple and GFS retention.
If you want only 15 monthly points you should set restore points to keep =2, monthly backup = 15.
The logic is described in this topic and User Guide. Please get familiar and ask additional questions if have any.
Thanks!

[Kdoel]

Case #01846068

Good day Forum,
I have a customer that I am already backing up (FF with 7 incrementals) every day with a synthetic FB on Saturday. They have just requested that we provide them with year's worth of Backup retention. I am interested to know what the "Best Practice" is to achieve 52 backups (weekly) or 12 (Monthly) backups for this customer. Can I use Backup Copy Job to copy off the Synthetic Full every week to another Source(not Tape)?

[neilmacneil]

Hello,

VBR version: v9 u2 (upgraded from v1 yesterday to see if it fixed the issue)

I have a case open for this #01869502.

All my backup copy jobs are experiencing this behavior. The problem is that my weekly's are being created the day before they are scheduled. For example:

Backup Job = Veeam
Schedule: daily at 5AM
Finished this morning at 5:14 AM

Backup Copy= Veeam - Backup Copy
scheduled sync: 5:10AM
retention: 14 days
GFS: 4w/4m
weekly Saturday's 5:10AM
monthly's Last Saturday
Finished this morning at 5:28 AM

My backup copy job did a weekly this morning even though it was scheduled for saturday. This is a problem because last weekend it ended up doing 2 fulls, 1 weekly full a day before and 1 monthly (last saturday). I was expecting the monthly and weekly to occur on the same day and end up with just 1 full restore point.

Anyone know what might be causing this? Since it appears to be happening on all my backup copy jobs i'm thinking about just setting my weekly 1 day ahead of schedule :/

Thanks,

-Neil

[foggy]

Neil, GFS restore points are offloaded when the full backup of the regular chain reaches the specified day, not exactly on that day. On Friday, your regular full reached previous Saturday and was offloaded as GFS.

[vdb421]

Hello,

We are currently using Veeam to backup 10 VMs to a first NAS with an active full on the week-end and forward incrementals on the week, with a retention of 14 days.

These backups are sent to a second NAS for yearly archival purpose with a backup copy job.

This job was initially set with these settings:

Restore points to keep - 365
Weekly backup - 4
Monthly backup - 4
Quarterly backup - 4
Yearly backup - 1

There are actually 340 points restore points available but only 1 full and incrementals points in the explorer.

Ideally, we want to keep backups for a year, with the 4 last weeks full backups, and monthly / quartlerly backups for the rest.

These settings are corrects?

Second, i want to reduce the restore points number to create these points before 1 complete year of backup.
How many restore points must be configured with these settings?

Thanks !

[foggy]

Kevin, GFS restore points will start to be created once the job reaches the specified regular retention (365 restore points).

[michaelp]

Good morning,

I am trying to get the GG-GFS (GreatGrandfather - Grandfather - Father - Son ) set up but i have no idea where to tell it to do so. I saw some policies on the backup copy job but i would like to do this on the regular backup jobs. Please assist as we are going to be running some restore testing soon and dont want to be missing too many cycles of the GG-GFS.

Thanks
Michael


PS my point of reference was: https://helpcenter.veeam.com/backup/vsp ... y_gfs.html

[Shestakov]

Hello Michael,
Indeed GFS retention works with backup copy jobs only since it`s a historical retention for secondary job. Could you explain why you need GFS on the primary backup?
Thanks!

[doggatas]

Hi All,

I think I have enough understanding of the GFS policy and most efficient way achieving our backup requirements. However, it's also possible I have no idea; hence this post.

Firstly, our backup requirements are as follows:

• Ability to restore from any of the previous 31 days (1 Full Backup + 30 increments)
• Ability to restore from the last day of the month from the previous 12 months (12 full backups)
• Ability to restore from the last day of the year from the previous 7 years (7 full backups)

To achieve this efficiently:

• Regular forward incremental backup job with 1 restore point (full backup on a Saturday or Sunday) targeting "Weekly Repository" (maximum of 7 restore points 1 VBK + 6 Incs)
• And a Backup Copy job as outlined below, focusing on the Job and Target setting pages

Backup Copy job
Job: Copy every one day
Target:

• Backup Reposity: "GFS Repository"
  Restore Points to Keep: 31
  Keep the following restore points for archival purposes: Ticked
  Weekly: 0
  Monthly: 12
  Quarterly: 0
  Yearly: 7

Any input/advice is appreciated and thanks in advance.

David

[doggatas]

Unable to edit the post above.

I realise there might be 2 full backups and 60 increments, given the 31 restore points on the backup copy job(does it work that way to for the backup copy jobs?). I think it might be better to have 31 restore points on the backup job with a weekly full and rollback option ticked. The backup folder will have; 1 vbk, 6 vibs and 24 vrbs.

Now on the backup copy job, can I set the restore points to keep at 2?
Does this mean there will be always synthesise on the copy job after the first two runs?
What gets copied to the target repository of the copy job when the weekly job runs?
Will Veeam use whatever is there to create the full backup for the monthly or yearly on that given day?

I will be testing this on a smaller scale before final configurations, but it takes time as I have to wait for the days to tick over.

[foggy]

Both approaches are viable. Just keep in mind that setting job retention to just a single restore point (regular job in the first approach) is not recommended, as you might find yourself without any valid restore point at all.

Now the questions:

I realise there might be 2 full backups and 60 increments, given the 31 restore points on the backup copy job(does it work that way to for the backup copy jobs?).

Unless the "Read from source..." check box is selected, backup copy job will have the exact specified number of restore points in the chain.

Now on the backup copy job, can I set the restore points to keep at 2?

Yes.

Does this mean there will be always synthesise on the copy job after the first two runs?

Yes.

What gets copied to the target repository of the copy job when the weekly job runs?
Will Veeam use whatever is there to create the full backup for the monthly or yearly on that given day?

Please review the corresponding user guide section for explanation on how GFS fulls are created.

[doggatas]

Appreciate the reply.

Going to go with the 31 restore points on a "Monthly" 31-day repo and the 2 restore points on the "GFS" repo.

It's such a shame 9.5 is not generally available as this is for a new environment, which is almost ready for live workloads. With all the synthesis going on, the 9.5/2016-ReFS would have been ideal. Not to worry, I have my Server 2016 servers ready to go and it shouldn't be too much work to migrate the backup jobs over.

Regards,
David

p.s. I LOVE VEEAM!

[Shestakov]

David,

Going to go with the 31 restore points on a "Monthly" 31-day repo and the 2 restore points on the "GFS" repo.

That confused me a bit. To meet your initial goal you need to set Restore points to keep = 31, Monthly backup = 12, Yearly backup = 7.

By the way, v9.5 is already around the corner.

[doggatas]

That confused me a bit. To meet your initial goal you need to set Restore points to keep = 31, Monthly backup = 12, Yearly backup = 7.

Both approaches are viable. Just keep in mind that setting job retention to just a single restore point (regular job in the first approach) is not recommended, as you might find yourself without any valid restore point at all.

Sorry about that, I'll try to explain.

Based on the advice from foggy (thanks BTW) rather than have the 31 Restore Points (as I said I might do in my first post on the topic) on the Backup Copy job, I will have the 31 Restore points on the actual Backup Job, which will reside on a Repository for the month worth of restore points (the 31).

I will then configure the Backup Copy job to only hold 2 Restore points, but using GFS i'll set it to have 12 Monthly restore points and 7 Yearly restore points; set to generate on the Last day of the month and Last day of December. So, at the end of the first 12 months (my calculations and for the sake of theory, we started the schedule in Janurary), the Repository holding the Backup Copy job will have 1 Full backup and a VIB (for the 2 restore points) and an additional 12 Full backup files. The last full backup file being tagged as an end of month and the first yearly.


Regards,
David

[Shestakov]

Understood. The plan should work fine!

[adr67]

Hi,

I've set up backup copy job with GFS retention, with everyday copy start at 0 am :

Restore points to keep - 2
Weekly backup - 0 saturday at 1 am
Monthly backup - 12 first saturday of the month
Quarterly backup - 0 first saturday of the quarter
Yearly backup - 5 first saturday of the year


I already have the 2 regulars backup copy points (1 vbk and 1 vib) on the repository, but GFS (monthly and yearly) never start.
What wrong with my config ?
Thanks

[veremin]

Wait till the last restore point (.vbk) reaches Saturday, then, first weekly point will be created. Thanks.

[forumit]

Hi,

What are the correct GFS settings for the below retention?

-Daily Cycle with 7 days retention
-Weekly cycle with 4 weeks retention
-Monthly cycle with 12 months retention
-Yearly cycle with 7 yrs retention

So far the Backup Copy Job is configured as:

Restore points to keep: 14
Weekly: 4 (Sunday 10:00 PM)
Monthly: 12 (First Monday of each month)
Quarterly: 1 (First Wednesday of each quarter)
Yearly: 7 (First Friday of each year)


Thanks much appreciated.

[Shestakov]

Hi,
You just need to set the retention you want to achieve, i.e.
Restore points to keep: 7
Weekly: 4
Monthly: 12
Yearly: 7
As for the backup day, if you set the same day for weekly/monthly/yearly, the backups will not be duplicated, but same file is to be marked as Weekly and monthly, so you save the repository space.
Thanks!