Reverse Incremental Backup

[Khue]

I just ran my first reverse incremental backup. This was my post describing my reasoning and methodology for switching to reverse incremental. Upon my first backup using reverse incremental backup, my expectation was that my vbk file was going to be large, roughly the size of all the virtual machines created in the list. Upon first run, that didn't appear to be the case. My vbks are quite small. I am afraid that if I need to restore from them, I won't be able to. I did offload my old standard incremental backups to tape including the incrementals. I am mostly covered.

Does this seem normal?

[dellock6]

The final size of a backup does not depends on the backup mode used. A full backup done with reversed or forward should be the same, if the content is the same. Since you cannot do the two backups in the same exact moment, small changes in the blocks of the VMs are going to make the final result slightly different, but these differences should be in the order of few MBs (f nothing happens in between like defrags or other activities inside the VMs).
As long as VBR is telling you the reversed completed successfully, I would not worry about its content

[Khue]

I hope your correct. The vbks when doing a regular backup were on the scale of 500gb to 1tb which would make sense as each VM can range between 100 and 200 gbs. Considering there's about 10 VMs in each job I expect the average of each vbk that i have configured to be right around the 500-600 mark.The last vbk was only 60 gigs and the vrbs are much smaller.

[lp@albersdruck.de]

I hope your correct. The vbks when doing a regular backup were on the scale of 500gb to 1tb which would make sense as each VM can range between 100 and 200 gbs. Considering there's about 10 VMs in each job I expect the average of each vbk that i have configured to be right around the 500-600 mark.The last vbk was only 60 gigs and the vrbs are much smaller.

You could start an Instant Recovery from that backup and see if the VMs boot up and everything is how it should be to confirm everything is backed up. And/or use Surebackup to do this.

[veremin]

Hmmm, the situation looks strange, indeed. Have you tried already to restore VMs out of the resulting backup files? Also, I wonder whether there are any exclusions specified in the job (either disks or VMs) that might have resulted in the decreased backup size. Thanks.

[dellock6]

Ah ok, you didn't posted numbers before, now that I see them the difference is indeed quite large.
Check exclusions as Vladimir suggests and also do some restores...

[foggy]

If everything looks good, you can perform a fresh active full to compare the numbers.

[Khue]

Here's the problem I am having with the process foggy.

This is the full backup using regular incremental:


This is the full backup using reverse incremental:


As you can see both are active fulls and both have siginficantly different sizes. There are no modifications to the lists. Both versions of the backup contain the same exclusions and the same virtual machines. Furthermore, it's like this for every list I've changed over to reverse incremental.

[veremin]

Not sure why but the processed values differ dramatically, as if the backup sources were different (100 GB vs 1 TB). I would double check the corresponding exclusions along with other settings, check restore capabilities and open a case, if the latter fails.

Thanks.

[Khue]

It's almost as if the active fulls only did an incremental instead. It's almost as if it expects the old vbks to still be there. I had to dump them off to tape for space. Basically my process was, commit the old vibs, vbms, and vbks to tape then delete all of those files and run an active full with the job set to reverse incremental instead of incremental. My expectation was the first vbk should be roughly the size of all vms in the list (not accounting for compression and dedupe).

[foggy]

My expectation was the first vbk should be roughly the size of all vms in the list (not accounting for compression and dedupe).

Yes, it should be a brand new active full. I recommend contacting support so they could take a closer look at what has happened.

[Khue]

I figured out the issue. I manually deleted the vbm, vbk, and vibs instead of removing the chain from Backups > Disk. Once I removed the chain from that section, the vbks came through as the expected size.

[foggy]

This should not affect active full, which does not look at the existing files at all. Anyway, I'm glad you were able to get back on track, thanks for updating.