We’re just getting into deploying our XenDesktop VDA setup. We’ve got a set of machines running with the M10-0B (512 MB VRAM) profile and working just fine. However for some machines we need a bit more and so I’m trying to create a set of machines running the M10-1Q (1 GB VRAM) profile. When I create the VDAs with this profile those machines never get into Windows and reboot over and over. I tried the other 1 GB VRAM profiles as those do the same. If I set it back to 512 MB they again work fine. I’m pretty new to this whole vGPU thing, so I’d appreciate any guidance I can get on this. Here’s some information about what we’re running.
OS: Windows Server 2016
Citrix XenDesktop Version: 7.15 CU2
Citrix App Layering Version: 4.13
Driver Version 23.21.13.9103
Let me know if there’s any other relevant information that would help. Thanks!
Are you using PVS or MCS to deploy your VMs? If yes, which vGPU profile does your Master image have attached … Make sure your VMs have the same vGPU profile as your Master. When you change vGPU profiles, the OS will see it as a different GPU, and this typically requires a reboot. If you have non-persistent VMs, this could be what’s happening …
As you’re using a 512MB profile, you won’t be using NVEnc on any of those VMs. NVEnc only works with 1GB profiles and above. Also, unless you need Quadro features, a 1B profile will be cheaper on the licenses.
This was exactly the issue. We are running these non-persistent so that makes a lot of sense. I created a new Platform layer in PVS and made a new machine template from that using the M10-1B GPU profile. I also had to uninstall/reinstall the driver in Device Manager before it would work for some reason. But when I created the new VMs from that template they working great. Thanks for pointing me in the right direction!