NVIDIA
How to get support? Problem with Win10, K2, vGPU and View7
Last summer my customer bought a couple af K2 cards for a small VDI deployment running Windows 10. The POC on View 6.1.1 was successful. After a long period with test my customer decided to move it to production, using the newly released View7. We have now spent 6 months including VMware support trying to get a stable k120q setup. The larger profile the more stable client, but using a k160q is way overkill and too expensive. VMware ended up suggesting latest Nvidia drivers but those give same problems of freezing desktops when running video or having large display areas change. Now I am trying to get support from Nvidia, but cannot find anything but SUMS which my customer does not have. What to do?
Last summer my customer bought a couple af K2 cards for a small VDI deployment running Windows 10. The POC on View 6.1.1 was successful.

After a long period with test my customer decided to move it to production, using the newly released View7.

We have now spent 6 months including VMware support trying to get a stable k120q setup. The larger profile the more stable client, but using a k160q is way overkill and too expensive.

VMware ended up suggesting latest Nvidia drivers but those give same problems of freezing desktops when running video or having large display areas change.

Now I am trying to get support from Nvidia, but cannot find anything but SUMS which my customer does not have. What to do?

#1
Posted 09/12/2016 10:07 AM   
The version history Stable: vSphere 6.0 Horizon View 6.1.1 Nvidia driver package vGPU-346.68-348.27 Unstable: vSphere 6.0 U2 Horizon View 7.0 Nvidia driver package vGPU-361.45.09-362.56 Unstable: vSphere 6.0 U2 w. latest patches as of today Horizon View 7.0 Nvidia driver package vGPU-367.43-369.17
The version history

Stable:
vSphere 6.0
Horizon View 6.1.1
Nvidia driver package vGPU-346.68-348.27

Unstable:
vSphere 6.0 U2
Horizon View 7.0
Nvidia driver package vGPU-361.45.09-362.56

Unstable:
vSphere 6.0 U2 w. latest patches as of today
Horizon View 7.0
Nvidia driver package vGPU-367.43-369.17

#2
Posted 09/12/2016 10:31 AM   
Hi Oletho, This is the problem with every GPU vendor that uses a hardware only model. The support falls on the GPU vendor (i.e. the OEM who sold the card), in turn they can raise a support ticket with NVIDIA. We recognised this gap so GRID 2.0 and up are sold via a software model license with SUMS allowing the user to purchase 24/7 support and direct support with NVIDIA. The legacy hardware only model means your custoemr needs to raise a support ticket with the server OEM who was the seller of the card. You aren't clear as to what "unstable client" actually means/symptoms/ error messages. Our support staff provide goodwill support on these forums if sufficient info is provided. We also make available the self-help KB articles generated by SUMS funding free to all users to search on their symptoms and debug techniques. search on GRD http://nvidia.custhelp.com/app/home/ Since it's VMware version specific it does suggest it could be the right course of action to pursue with VMware. Best wishes, Rachel Best wishes, Rachel
Hi Oletho,

This is the problem with every GPU vendor that uses a hardware only model. The support falls on the GPU vendor (i.e. the OEM who sold the card), in turn they can raise a support ticket with NVIDIA. We recognised this gap so GRID 2.0 and up are sold via a software model license with SUMS allowing the user to purchase 24/7 support and direct support with NVIDIA.

The legacy hardware only model means your custoemr needs to raise a support ticket with the server OEM who was the seller of the card.

You aren't clear as to what "unstable client" actually means/symptoms/ error messages. Our support staff provide goodwill support on these forums if sufficient info is provided. We also make available the self-help KB articles generated by SUMS funding free to all users to search on their symptoms and debug techniques. search on GRD http://nvidia.custhelp.com/app/home/


Since it's VMware version specific it does suggest it could be the right course of action to pursue with VMware.

Best wishes,
Rachel

Best wishes,
Rachel

#3
Posted 09/12/2016 10:57 AM   
You might like to follow this thread and direct VMware support to it if you are using PCoIP and have the video freezing issues: https://gridforums.nvidia.com/default/topic/890/?comment=3231 Again protocol description is a detail we need to provide advice e.g. Blast/PcoIP/ICA/TCP etc... Best wishes, Rachel
You might like to follow this thread and direct VMware support to it if you are using PCoIP and have the video freezing issues: https://gridforums.nvidia.com/default/topic/890/?comment=3231

Again protocol description is a detail we need to provide advice e.g. Blast/PcoIP/ICA/TCP etc...

Best wishes,
Rachel

#4
Posted 09/12/2016 11:00 AM   
Hello Rachel. The video freezing link you refer to match our case exactly (problem #2). Now, I don't know why you point to VMware. Just as the original poster of https://gridforums.nvidia.com/default/topic/890/?comment=3231 we suspect newer Nvidia drivers are to blame. The K2 cards are purchased in HP Proliant servers, I will try to raise the problem with HP. Thanks.
Hello Rachel.

The video freezing link you refer to match our case exactly (problem #2).

Now, I don't know why you point to VMware. Just as the original poster of https://gridforums.nvidia.com/default/topic/890/?comment=3231 we suspect newer Nvidia drivers are to blame.

The K2 cards are purchased in HP Proliant servers, I will try to raise the problem with HP.

Thanks.

#5
Posted 09/12/2016 12:09 PM   
Because the custoemr has found changing the PCoIP setting to what should intuitively be a weightier protocol has resolved the video issue and we haven't seen this with ICA/HDX/Blast I'm still suspicious this in part is worth pursuing on the protocol side too. Hence why I think pursuing with VMware is worthwhile. Configuring PcoIP resolved one issue. The driver roll-back for the second issue I have highlighted to our engineering team. So they will be aware when HP pass it on. Best wishes, Rachel
Because the custoemr has found changing the PCoIP setting to what should intuitively be a weightier protocol has resolved the video issue and we haven't seen this with ICA/HDX/Blast I'm still suspicious this in part is worth pursuing on the protocol side too. Hence why I think pursuing with VMware is worthwhile. Configuring PcoIP resolved one issue.

The driver roll-back for the second issue I have highlighted to our engineering team. So they will be aware when HP pass it on.

Best wishes,
Rachel

#6
Posted 09/12/2016 01:25 PM   
Scroll To Top

Add Reply