Moving from reverse incremental to SOBR with capacity tier

[entropyiv]

Hi Everyone,
I have been reading about SOBR and GFS settings for a couple weeks now and I haven't been able to get my head around it well enough to figure out how to get from A to B. I was hoping someone could tell me what the best path would be to move from my current setup to something that will let me use S3 compatible storage for long term retention.

First, here is my current setup:
Veeam B&R 11, license Enterprise Plus
I have a two simple repositories. One for machine backups and one for the Veeam config.
All machine jobs are set up as reverse incremental. (This made the most sense at the time to ensure that any restore from tape would only require a single file.)
I have a tape job that runs an active full backup every Sunday and gets all backup files. I keep those for four weeks.
I manually run that same active full backup tape job on the first business day of the month, and that goes off site.
I keep the Monthly tapes for 12 months, and the December EOM (first business day of January) tapes for 10 years.

My goal would be to only store the Monthly and Yearly full backups files in the S3, and continue to use my tape jobs for the on site weekly.
I know I can convert my main simple repository to SOBR to add the capacity tier S3, but what is the correct path to achieve getting only the Monthly and Yearly fulls into the S3 storage?

Thank you in advance for any insight!

[Brad.Barker]

Hello Mike!

There are a few things that you'd want to do in order to achieve what you're looking to do.

First, you'd want to move from Reverse Incremental Backup to Forward Incremental with Weekly full backups. Synthetic full backups tend to be the better option for the weekly full backups, but the actual best option between synthetic full backup vs active full backup depends on the initial storage target. If you're repository initial repository is Linux with XFS or Windows with ReFS though, you'll definitely want to lean towards synthetic full backups for example. The other aspects of moving away from Reverse Incremental Backups is that they're deprecated in the current release of Veeam Backup and Replication. I would recommend taking a look at the retention behavior documentation for forward incremental so you can see how the backups will behave compared to what reverse incremental backups: https://helpcenter.veeam.com/archive/ba ... thods.html

Second, you'd then build out your SOBR with the capacity tier added to it. You'd de-select the "Copy" option and then select the "Move" option when configuring the Capacity Tier option. The "copy" option creates full redundancy between the local repository and the S3 compatible repository, which would include the daily restore points that you're looking not to send to the S3 storage. The "Move" option comes with a configuration option for a number of days before offloading data. Here you'd want to set your move operation to be more than what your local daily retention is, so that backups are only moved when they fall into your GFS retention for the monthly backups. Take a look at the "Move" operations functionality here: https://helpcenter.veeam.com/archive/ba ... _move.html

Finally, for your tape jobs, as long as your regular backups run their synthetic full backup on Saturday night and the tape jobs is Sunday morning, or the synthetic full backup run Sunday before your tape jobs run, then there shouldn't be any issues with getting the full backups to tape though this would only capture the full backup. The real question I'd have with the weekly tape backups is whether you're trying to take the incremental backups as well. If you are, then you may want to actually run the tape job the day before your synthetic full backup if you only want to run the tape job once a week.

Hope this helps!

[entropyiv]

Thank you for your response! I will study it and see if I can make something work.

[entropyiv]

I have spent some time playing with one of my backup jobs and I have a few more questions.

Since I am going to continue my once a week tape job, can I turn off the weekly option in the GFS settings?
Basically the end goal for all my jobs would be:

• The full (synthetic?) backups to occur on Saturday, starting a new backup chain for incrementals.

• Retention period is set for 7 days. So at any time the repository would have 1 week of backups, 1 "full" and the incrementals.

• On Sunday, this gets written to tape

• GFS enabled for monthly and yearlys only keeps one of the full backups per month and year

• Monthly and yearly gets moved to the capacity tier S3 when created (or shortly after).

Does that look right?

With the forward incrementals, wouldn't this mean that since I want a weeks worth of backups on hand, that I in fact would need two "full" backups and their respective incremental files per machine to cover the 7 days retention? This is the reason the reverse incremental worked so well for us. The tape just needs to be a snapshot of that week. I haven't looked yet but I am guessing there is no option to have the tape job get only the most recent full backups that get generated on the previous day (Saturday)?

Again, thank you for all the insight!

[Brad.Barker]

Hello Mike!

If you don't need more weekly Full backups than the previous week and current week, then you don't need the weekly GFS restore points enabled.

Your read on forward incremental is correct (if I'm reading it right). Effectively, you have your previous week, and the current week. So minimum of 7 backups (your previous week), this then builds up to 14 and then once the 14th point is created drops back down to 7.

You can have the tape job only take the full backup if that's what you're looking to do. There should be a checkbox during the Tape Job configuration to uncheck taking incremental backups to tape, and then when the tape job runs on Sunday it'll capture the full backup created on Saturday.

Let me know if you have any other questions!

[entropyiv]

OK, so at any given time I would have to have disk space for 3 full backups I think?
Week 1 - full and its daily incrementals
Week 2 - full and its daily incrementals
Week 3 - full gets created, at which point week 1 full and week 1 incrementals get removed.

Is that correct?
If so, does that mean that if I carefully time my tape backups (and exclude incrementals), I would still need the space for two full backups on each tape?

[Brad.Barker]

If you have your retention set to 7 it's only 2 weeks (though having working space is not a bad idea).

The day the 14th restore point is created on Friday, Retention activities will see that all 7 of the previous weeks full backups are outside of retention, and then remove them as part of that retention clean up dropping you back down to 7. The only real thing I could think of that could lead to you having additional full backups kept is how you would configure your monthly and yearly tiering activities to your S3 Compatible target. If you set your "Move" operation to like 30 days or something like that, you'd see that GFS point sit locally for a bit before it moves for example. If you have your Move set to 7 and your retention set to 7 with GFS monthly backups you should only ever have between 7 to 14 restore points locally under normal circumstances.