Jobs forward incremental backup take long time

[webstyler]

Hello

I have jobs setting with retention 14 days and:
- Incremental
- Synthetic Full backup any wednesday
- Transform previous backup chain into rollback

But, during long time, too long.

How you suggest to change this settings without spend too repository space and perform more fast job?

Thanks

[foggy]

Hi Simon, could you please provide more details - what specifically takes longer than expected, what kind of target storage do you have? Some numbers will also help to understand the issue better.

My first suggestion would be to switch to forever incremental mode, since it takes the same in terms of space and doesn't perform synthetic fulls with transform, which, I guess, is your major concern.

[PetrM]

Hi Simon,

Please take a look at this table which reflects I/O impact of different backup modes: synthetic with transform is the heaviest operation which requires 4 I/O (2x read, 2x write).

Agree with Foggy, forever incremental mode would be the be best option in your case.

Thanks!

[webstyler]

Hi Simon, could you please provide more details - what specifically takes longer than expected, what kind of target storage do you have? Some numbers will also help to understand the issue better.

My first suggestion would be to switch to forever incremental mode, since it takes the same in terms of space and doesn't perform synthetic fulls with transform, which, I guess, is your major concern.

Hi foggy

Yes, sure.

I've several jobs, each with few VM (from 5 to 10 VM x Job)
Size for job: from 700GB to 2,3 TB

Time required for create weekly synthetic full and transform into rollbacks during:
job ~900 GB > 12 hours
job ~2 TB > 42 hours .. !! :O

[webstyler]

Hi Simon,

Please take a look at this table which reflects I/O impact of different backup modes: synthetic with transform is the heaviest operation which requires 4 I/O (2x read, 2x write).

Agree with Foggy, forever incremental mode would be the be best option in your case.

Thanks!

Hi PetrM

Yes, I've check resource linked and also official documentation here https://www.veeambp.com/ and here https://helpcenter.veeam.com/docs/backu ... ml?ver=100

As if I've understand:
- have a synthetic full backup grant faster recovery
- have active backup grant secured backup integrity

This is why we have not setting as default the forever incremental

Right or bad?

Thanks

[foggy]

In this case switching to forever incremental should help since the merge of the oldest increment will occur daily instead of creating the new full from the entire previous chain (plus transformation).

[foggy]

- have a synthetic full backup grant faster recovery

With 14 days retention this is not an issue.

[PetrM]

Constant transformation of full backup makes it fragmented in the long term but you may schedule compact job to defragment it periodically.

Please keep in mind that compact requires space for one more full on the repository.

Thanks!

[webstyler]

Thanks to all

I try to move to forewer incremental and check in the next week time and space on repository

@PetrM : so you suggest to select "defragment and compact full size backup file" ?
If I understand space requide as the same of my previous configuration, with synthetic full + active backup right?
But what about time request for this operation?
Do you suggest also for jobs with 14 days restore points?

[PetrM]

Hello Simon,

If I understand space requide as the same of my previous configuration, with synthetic full + active backup right?

In fact, not exactly the same, compact requires additional space for one more full to create auxiliary temp file that exist on the repository until the end of operation.
Your previous configuration required additional space for the entire inactive chain: full + set of increments.

But what about time request for this operation?

It depends on many factors, for example: full backup file size, disk performance, storage optimization settings etc.
I'd expect that compact would take approximately the same time as required to create synthetic full.

Do you suggest also for jobs with 14 days restore points?

Yes, I'd suggest to run it periodically as long as you have enough free space on the backup repository.

Thanks!