Citrix VDA disconnecting and no reconnect possible

Virtual GPU Forums More vGPU Forums
1

Hi folks

We have a case open for this with NVIDIA and they already acknowledged that this is a bug (internal tracking bug 3445082) since there are more customers with the same problem.

When the issue occurs, the user experiences a sudden disconnect or a black screen within the Citrix session. After the disconnect, the VM itself is not in a hanging state and can be accessed via remote console. After this happens the user has no chance to reconnect to his session again and we have to force restart the Windows 10 VM.

According to NVIDIA support, this only happens to Citrix VMs with 24GB memory or more. I just wanted to see if there is anybody elsa experiencing this problem and maybe we can exchange some information.

Regards, Cedy

2

Hi C3dy

we recognized this behaviour aswell and fount it to be related with the NVIDIA vGPU driver drashing in the guest OS.
Check in the device manager of the affected VM if the GPU driver/device is loaded and without any error.

Regards

1 Like
3

Hi Gormet

Thanks for your answer and it always feels good to not be alone. Yeah the driver is definitly crashing inside the VM at some random point. In the meanwhile NVIDIA informed us that they were able to isolate the issue and provide a fix asap. Will keep you posted.

Regards, Cedy

4

Hello Cedy,

is there anything new related to this issue?

Greetings!

5

Hello Team,

I know, this issue is old but we are facing the similar issue with our physical remotepc . we delivered the desktop via citrix and last year September we upgraded the Nvidia Quadro P1000 driver and VDA to 2203 CU 3 after that we have seen citrix connection automatically hung and disconnected and never reconnect. RDP session also stuck on Welcome screen as soon as we reboot it issue fixed.

We had tried to upgraded the VDA to 2203 cu4 update 1 but no luck. Due to this issue 10 to 15 user randomly get impacted and this issue is mostly happening wih live connected user. our physical device having 64 RAM and 12 CPUs

6

Hello Team,

Can anyone help on this issue. now i can below logs on one of the machine after disabled the IPv6 NIC.

Faulting application name: NVDisplay.Container.exe, version: 1.37.3103.4323, time stamp: 0x621dbda6
Faulting module name: _nvtopps.dll, version: 31.0.15.3141, time stamp: 0x641395ce
Exception code: 0xc0000409
Fault offset: 0x00000000007f7cd1
Faulting process ID: 0x46e4
Faulting application start time: 0x01daa86d7fc28782
Faulting application path: C:\WINDOWS\System32\DriverStore\FileRepository\nv_dispwi.inf_amd64_c7d30625dd117952\Display.NvContainer\NVDisplay.Container.exe

7

Hello everyone we have a similar problem that the VM’s Server OS Win2019 with 128 GB RAM and nvidia_a16-16q and Citrix VDA 2311 on ESX8.0.2 after a vMotion Process, the users fly out of their sessions and can no longer reconnect.the Vm’s can also no longer be connected via RDP. The only remedy is to power off the affected VMs.
If vMotion is switched off, this problem does not occur.
The current Nvidia driver NVIDIA-GRID-vSphere-8.0-550.54.16-550.54.15-551.78 is currently being used.