slow backups after updating to 8.0.0.2030

[tym]

Hi,

just updated to the latest sp (2b?) yesterday. Now our backups run significantly slower. Eg a 1h backup now takes 5:30, and another backup which usually ran at 1 hour is now running since yesterday evening (18 hours, and only 57% finished). Both are incrementl backups, and the processing rates are within normal range. What is this?

Thanks in advance.

Best,
Tim

[Vitaliy S.]

Hi Tim,

What does job bottleneck statistics show for these jobs? Is it the same as you had before the upgrade?

Thanks!

[ursli123]

I can't find out, what chenged between 8.00.2021 and 8.00.2030
Why should I upgrade to Update 2b? Slower Backups?

Regards

Urs

[Vitaliy S.]

Please check out this explanation from Anton's weekly digest:

B&R 8.0 Update 2a was officially released around 10 days ago, but as you may have noticed, B&R update check still does not "see" one. The reason is yet another critical issue we have found in VDDK 6.0 last week, which forced us to rebuild the update code again, now as Update 2b – adding a workaround for this single issue. While it's again a corner case, it may results in B&R quietly creating unrecoverable backups for certain VMs, which previous versions of VDDK had no issues processing (see KB2042 for more information). Sadly, VDDK 6.0 has been one of the worst VDDK releases in terms of the amount of new bugs introduced. More evidence to keep using SureBackup for catching those environment-specific issues!

[BriFar]

Hi Tim,
Can you open the backup log, click on the vm being backed up on the left? This should show the vm's individual job log.
Note the step were the duration is the longest.
I had a customer were the job took 20hrs. 5mins only was the actual backup to disk. The rest was guest file indexing.
I disabled guest file indexing because it wasn't required.
Now the job finishes within minutes.
P.S this was in a VMware environment but I believe it may still be relevant. We do not have update 2b installed.

[tym]

Hi
nevermind. After 1 incremental that took 48 hours, the following incrementals were fine again. I assume b&r had to re-do some thing after the upgrade, indexing or whatever.

[Vitaliy S.]

Check out the amount of data transferred, there could be full backup happening. Thanks for coming back and updating your topic.

[foggy]

This could be due to the driver update. Full VM data is always read after that (however only changes are actually transferred).

[tym]

thanks foggy, that matches what I saw: transfer was only 69 GB from 34,5 TB total.