Unable to install the latest GFE version

[S4nfs]

The following Script is downloading the absolete GFE version. I changed the URL inside /install.requirement.ps1 file and still it downloads the v3.13.0.85 Beta Geforce experience. How to solve this.

[acceleration3]

It's supposed to install an "obsolete" version. The newer versions, while being more advanced and feature complete, have more and better checks implemented to detect and block hardware configurations it's not supposed to run on. I have tried to disassemble those versions and patching as many checks as I could but I still haven't gotten it to work. The one the script installs is the last version before all of the checks were implemented.

[S4nfs]

Is it possible that this script works with Nvidia vGPUs ?. I tried with tesla M60 vGPU (no passthrough) installed on my hypervisor 8.1 but the problem occurs when the vGPU tries to get license to break the 14fps capping limit. vGPUs somehow won't support the grid drivers you mentioned in the script. The driver is blocking the nvidia licence manager, However it's working without any other issues rather than it fails to get license which is required to achieve full gpu performance. I also tried with my previously installed vGPU drivers (xenserver vgpu driver) but its showing nvstream API failed while streaming.

One more thing i noticed that the script installs grid drivers from Microsoft, Why not from Nvidia?

[acceleration3]

That driver from Microsoft is the only one that I found that enables this to work under Windows 10 on Azure with an M60 GPU so I recommend that one for maximum compatibility. As for running this under Citrix Hypervisor, it should work fine but you will have to install the optimal driver yourself because the script is optimized for cloud services virtual machines. Remember that you need an audio solution for GameStream to work as well as the monitor you open GeForce Experience on to be connected to the NVIDIA GPU.

[S4nfs]

This is the error on TeslaM60 vGPU which i mentioned earlier. Shield is enabled but i saw it failed to load the complete script after restarting.
Also i mentioned that i can't use grid driver on vGPU.

[acceleration3]

Can you run this command in the virtual machine's CMD:
start "" /D "C:\Program Files\NVIDIA Corporation\NvContainer" "C:\Program Files\NVIDIA Corporation\NvStreamSrv\nvstreamer.exe"
And paste a screenshot or the output here?

[S4nfs]

Nope it's just skipping the content very fast and closes that window (another cmd)

[acceleration3]

My bad, the command should be start /B "" /D "C:\Program Files\NVIDIA Corporation\NvContainer" "C:\Program Files\NVIDIA Corporation\NvStreamSrv\nvstreamer.exe"

[S4nfs]

C:\Users\Administrator>start /B "" /D "C:\Program Files\NVIDIA Corporation\NvContainer" "C:\Program Files\NVIDIA Corporation\NvStreamSrv\nvstreamer.exe"

C:\Users\Administrator>[libprotobuf WARNING ..\src\google\protobuf\descriptor_database.cc:58] File already exists in database: BusMessage.proto
[libprotobuf WARNING ..\src\google\protobuf\descriptor_database.cc:115] Symbol name "BusMessage" conflicts with the existing symbol "BusMessage".
[libprotobuf WARNING ..\src\google\protobuf\message.cc:288] File is already registered: BusMessage.proto
You can run C:\Program Files\NVIDIA Corporation\NvStreamSrv\nvstreamer.exe -help to see all the help options.
#1(I)[2020-12-26 16:52:54,672]=22:22:54={000022D4} Custom Ctrl+C handler successfully set.
#2(I)[2020-12-26 16:52:54,672]=22:22:54={000022D4} PID 8900
#3(I)[2020-12-26 16:52:54,678]=22:22:54={000022D4} Streaming Process ID: 0
#4(I)[2020-12-26 16:52:54,685]=22:22:54={000022D4} P4 Changelist: 23612337
#5(I)[2020-12-26 16:52:54,693]=22:22:54={000022D4} Enumerating network adapters on this system
#6(I)[2020-12-26 16:52:54,708]=22:22:54={000022D4} --- ...: ... / ...
#7(I)[2020-12-26 16:52:54,719]=22:22:54={000022D4} Building head list
#8(I)[2020-12-26 16:52:54,745]=22:22:54={000022D4} Adapter vendor id: 0x10de, device name: \.\DISPLAY3
#9(I)[2020-12-26 16:52:54,756]=22:22:54={000022D4} Creating a Standalone SCI Thread
#0(I)[2020-12-26 16:52:54,764]=22:22:54={000022D4} Received ServerListenForIncomingStreams
#1(I)[2020-12-26 16:52:54,773]=22:22:54={000022D4} Using: OpenSSL 1.0.2n 7 Dec 2017
#2(I)[2020-12-26 16:52:54,933]=22:22:54={000022D4} RND is initialized
#3(I)[2020-12-26 16:52:54,941]=22:22:54={000024C0} Starting initialization of the main server thread
#4(I)[2020-12-26 16:52:54,951]=22:22:54={000024C0} RTSP Server: TCP instance
#5(E)[2020-12-26 16:52:54,958]=22:22:54={000024C0} Failed to start RTSP server on port 48010
#6(I)[2020-12-26 16:52:54,967]=22:22:54={000024C0} Failed to start RTSP tcp server
#7(I)[2020-12-26 16:52:54,975]=22:22:54={000024C0} Allocated 1 ConnectionInfo entries
#8(I)[2020-12-26 16:52:54,983]=22:22:54={000024C0} Completed initialization of the main server thread with state 2
#9(I)[2020-12-26 16:52:54,994]=22:22:54={000022D4} Not sending a message.
#0(I)[2020-12-26 16:52:55,001]=22:22:55={000022D4} SCI Thread creation done
#1(I)[2020-12-26 16:52:55,009]=22:22:55={000022D4} Computing SPS/PPS headers.
#2(I)[2020-12-26 16:52:55,018]=22:22:55={000022D4} Loaded library from '...'
#3(I)[2020-12-26 16:52:55,135]=22:22:55={000022D4} Initialized context for adapter 0: 1366 x 768 @ 60.0 Hz
#4(I)[2020-12-26 16:52:55,144]=22:22:55={000022D4} Driver version is 443.66, branch is r443_62.
#5(I)[2020-12-26 16:52:55,154]=22:22:55={000022D4} System is NOT co-proc.
#6(E)[2020-12-26 16:52:55,197]=22:22:55={000022D4} Failed to initialize CUDA driver API
#7(E)[2020-12-26 16:52:55,205]=22:22:55={000022D4} Failed to initialize adapter context for RTSP SPS/PPS header generation.
#8(I)[2020-12-26 16:52:55,260]=22:22:55={000022D4} Deinitialized context for adapter 0
#9(E)[2020-12-26 16:52:55,269]=22:22:55={000022D4} Error computing SPS/PPS headers.
#0(E)[2020-12-26 16:52:55,277]=22:22:55={000022D4} NvEncodeAPI Not Supported
#1(E)[2020-12-26 16:52:55,284]=22:22:55={000022D4} SCI is not enabled, NvEnc initialization failed
#2(I)[2020-12-26 16:52:55,294]=22:22:55={000022D4} Terminated the SCI thread
#3(I)[2020-12-26 16:52:55,301]=22:22:55={000022D4} Waiting on RTSP handshake to finish
#4(I)[2020-12-26 16:52:55,309]=22:22:55={000022D4} Starting shutdown of the main server thread
#5(I)[2020-12-26 16:52:55,320]=22:22:55={000022D4} Network Event Subscribe: 00007FF63FB390A0 - 00007FF63FCF8990
#6(I)[2020-12-26 16:52:55,331]=22:22:55={000022D4} NATT Initialize: STUN servers count 1. Retransmission period 500, count 5
#7(I)[2020-12-26 16:52:55,346]=22:22:55={000022D4} Network Host Lookup: blocking call
#8(I)[2020-12-26 16:52:55,360]=22:22:55={000022D4} NATT Initialize: use STUN server s1.stun.gamestream.nvidia.com:19308
#9(I)[2020-12-26 16:52:55,373]=22:22:55={000022D4} Network Event Unsubscribe: 00007FF63FCF8990
#0(I)[2020-12-26 16:52:55,384]=22:22:55={000022D4} Starting un-initialization of the main server thread
#1(I)[2020-12-26 16:52:55,395]=22:22:55={000022D4} Completed un-initialization of the main server thread
#2(I)[2020-12-26 16:52:55,407]=22:22:55={000022D4} Completed shutdown of the main server thread
#3(I)[2020-12-26 16:52:55,417]=22:22:55={000022D4} Terminated the main server thread

[acceleration3]

According to this line:
#0(E)[2020-12-26 16:52:55,277]=22:22:55={000022D4} NvEncodeAPI Not Supported
The drivers you are using right now don't support NvEnc. I updated the script a while ago to use jamesstringerparsec's GPU update script if the one from Microsoft doesn't work. Have you tried checking if that installs the correct drivers?

[S4nfs]

I tried the ParsecGPUcloud updater script but not with the same GPU. I'll try that later.

[acceleration3]

Alright, I will leave the issue open for now. Please report back when you've tried.

[S4nfs]

Yes i tried the james parsec script but it seems that it only works with tesla GPUs running without any hypervisor concept. Also they are more likely to brick your VM on some cloud platforms. Cloudgamestream runs fine on all hypervisor GPUs, (i didn't tried it with A100)

Later,I forked the cloudgamestream project but that isn't working somehow. How can i install a custom or a latest GFE version with the same script, i want to know what's blocking the nvidia server.