Random BlackScreens - vGPU VMware Horizon

Hi all,

we have random, rarely black screens in our vdi environment. Only a reboot of the affected VM will help to fix it.
Nvidia Support told us to upgrade vGPU driver from 13.2 to latest 13.4.
We using Nvidia M10 on VMware Horizon with Windows 10 - 20H2 and grid_m10-1b profile and VMware ESX 7.0 U3d

After upgrading we can still see a lot of these events in the nvidia esx logs

+++++++++++++++++
2022-09-09T10:46:31.072Z Er(02) vthread-24234322 - vmiop_log: Assertion Failed at 0x14416cd7:397
2022-09-09T10:46:31.072Z Er(02) vthread-24234322 - vmiop_log: 12 frames returned by backtrace
+++++++++++++++++

Does anybody know if it normal?

Regards,
VM_Master

Hi,
sounds like a known issue. NVES should be aware of this. Upgrading to 13.4 won’t fix it as the root cause is still under investigation. Please contact NVES once again to add your case to the list of affected customers.

regards
Simon

Thanks Simon for this info.

NVES told me that this is fixed in 13.4.
But I can’t find any information in the relase notes about this issue. So why is there no public information about it?

I will talk again with NVES.

Just send me your case ID with PM. Will route your case in the right direction. I’m not aware that this is fixed in 13.4.

Hi Simon,

did you get my PM? Any news regard this issue?

Hi there @VM_Master / @sschaber - are there any updates on this issue?

We are running ESXi 7.0 U3g (20328353) and the latest 13.4 Nvidia vibs / guest drivers.

Last week we had two instances of this. All users on the same graphics card got a black screen - we had to move them to a different host and reboot them. Obviously this is very disruptive!

Hi,

our Eng is working with high priority on this issue. We hope to have a fix soon. Unfortunately this is a very complex issue and therefore requires some time. Please keep your support ticket open or create a new one if already closed to get the information as soon as there is a fix available.

Best regards
Simon

1 Like

Thanks @sschaber - great thanks, good to hear this is being actively worked on. We have a case open and will follow up further there.
Kind regards, Tom

Hi,

good news. Our issue seems to be solved after upgrading to VMware GRID v13.4 (on all esx hosts and also on our VDIs).

And we set this regkey on all VDIs from 1 to 0
HKLM\Software\Policies\VMware, Inc.\VMware Blast\config
PixelProviderForceViddCapture REG_SZ : 0

ESX version: 7.0.3, 20328353
Horizon version: 2111, Agent 8.5.0

We don’t have this random blackscreens anymore. Also the reconnect time is now faster.