Arithmetic Overflow Error / Abnormal transfer rates

[ferrus]

Case ref: 04269924

We had an issue in the Veeam B&R console, when we started using WAN acceleration.
As the jobs were running, we'd receive frequent errors in the console about Arithmetic Overflow Errors.
At the same time, the transfer rates of the jobs would show wildly extravagant speeds - into several hundred thousand GBps.

I opened a support case, and after some investigation we updated the infrastructure to v10a, and the issue went away.

It's been running perfectly for about 8 weeks now.
But last week I reinstalled the Veeam B&R server, to migrate the OS to Windows 2019, and the issue has returned.

My current workload means I probably won't be in a position to open a support case for a while now, for a simple GUI/console bug.
If this was a known WAN acceleration issue resolved in 10a however - it appears to have returned for us, with Windows 2019.

[PetrM]

Hello,

Actually, we don't know if it's a simple GUI issue or in your particular case there is some complicated low-level logic behind which does not behave as it's intended to do.

I recommend to re-open the support case or to raise a new one and to ask our support engineers to perform a deeper investigation.

Thanks!

[ferrus]

Okay, thanks.
I was just wondering if it had been observed somewhere else, or if a bugfix was planned for the next CU.
When my workload clears, I'll open/reopen a case to look into it.

[Princes]

Has anyone seen any of their Copy Jobs reporting unfeasibly high Processing Rates, e.g. 1679700.8GB/s and similar? Began happening after our upgrade from v9.5 Patch 4 to v10a, and we're now on v10a Build 10.0.1.4854?

[foggy]

Hi James, split the post as it is not related to the original discussion. As for the issue you're seeing, please contact technical support - might be a UI glitch or something.

[ferrus]

Yeah - had the same bug, with similar versions.

Only seemed to affect our Backup Copy Jobs that used WAN Acceleration.
Unfortunately, I removed WAN Acceleration a couple of days ago, following a WAN link upgrade - but could roll it back for testing.
The issue was still present in the GUI, every time I clicked on one of the affected jobs. It didn't appear to affect the actual job or performance - only the GUI.

Here's the original post (mods - feel free to merge them and edit this post):
veeam-backup-replication-f2/arithmetic- ... 69961.html

[foggy]

Thanks - the case notes are telling though that originally you've seen the issue on version 10 P2 and updating to 10a resolved the console error.

[ferrus]

It did for a while, but a subsequent re-install following an OS upgrade brought the error back.
It was like that for a further 3 months - until I switched from WAN Acceleration to Direct a couple of days ago.

[foggy]

Then it seems somewhat consistent and we'd definitely like to investigate.