Proxmox 5.3 - Tesla P40 vGPU issues

[dominiaz]

Proxmox 6.2 with Nvidia Driver: 440.87 works stable.

 

[janosch1337]

Hey dominiaz,

do you actually working with the nvidia vgpus on proxmox? and is it still working stable?

the cards are made for rendering / machine learning / video trancoding and so on but not for gaming. do you think it's possible to use this card in a vm for "gamestreaming VM"

 

[dominiaz]

Card works very very stable on Proxmox 5.4.

I have some issues on 6.x.

I think that gamestreaming VM will be working very good.

Why didn`t you passtrough "normal" gamer card to VM?

 

[avttd]

Hello!
Do you want to say that I can take Tesla T4 and give my Windows virtual machines a vgpu profile (for example: 8 virtual machines with 2 GB of video memory?) ???

And I just need to buy nvidia licenses for the Nvidia vDWS?

 

[dominiaz]

Hello!
Do you want to say that I can take Tesla T4 and give my Windows virtual machines a vgpu profile (for example: 8 virtual machines with 2 GB of video memory?) ???

And I just need to buy nvidia licenses for the Nvidia vDWS?

Yes, exactly.
Just install proxmox 5.4 because its the most stable version for vgpu.

 

[avttd]

Yes, exactly.
Just install proxmox 5.4 because its the most stable version for vgpu.

thank you very much!

 

[jsterr]

Yes, exactly.
Just install proxmox 5.4 because its the most stable version for vgpu.

Can you share some details howto start with Tesla T4. There is a lot of demand for vgpu support for proxmox and no real good guide oder documentation howto start

• Do I still need to buy Grid Licences?
• What software needs to be installed on proxmox hosts or does it work out of the box?
• Are there special drivers that need to be used?
• How can I provide a vm a vgpu?
• What problems occure on Proxmox 6.x?

Sorry for the many questions. We are thinking about buying a Tesla T4, but want to collect as many information before buying.
Thank you very much, would really appreciate some details how you managed it.

Edit: I see there is a little information in this thread how you managed it, is this still correct? Anything that should be mentioned for someone who wants to get startet with Nvidia Tesla T4?

 

[Krony]

Instruction step-by-step:

1) Enable iommu on proxmox host: https://pve.proxmox.com/wiki/Pci_passthrough
2) Download NVIDIA vGPU for Linux KVM (NVIDIA-Linux-x86_64-430.67-vgpu-kvm.run or newer) from https://nvid.nvidia.com/dashboard/
3) Execute code on proxmox host

chmod +x NVIDIA-Linux-x86_64-430.67-vgpu-kvm.run
apt install gcc make pve-headers-`uname -r`
./NVIDIA-Linux-x86_64-430.67-vgpu-kvm.run
mkdir /usr/lib64
ln -s /usr/lib/x86_64-linux-gnu/libnvidia-ml.so /usr/lib64/libnvidia-ml.so
ln -s /usr/lib/x86_64-linux-gnu/libnvidia-vgxcfg.so /usr/lib64/libnvidia-vgxcfg.so
ln -s /usr/lib/x86_64-linux-gnu/libnvidia-vgpu.so /usr/lib64/libnvidia-vgpu.so
ln -s /usr/lib/x86_64-linux-gnu/libnvidia-cfg.so /usr/lib64/libnvidia-cfg.so

4) Reboot proxmox host
5) Add it all vm with diffrent uuid 'args: -uuid 00000000-0000-0000-0000-000000000100' to vm file /etc/pve/local/qemu-server/100.conf

Hey,
I'm running into the same issue, but im using the nvidia-vgpu-ubuntu-510_510.73.06_amd64.deb installer.

Error:

root@pve:~# qm start 100
swtpm_setup: Not overwriting existing state file.
kvm: -device vfio-pci,sysfsdev=/sys/bus/pci/devices/0000:01:00.4/00000000-0000-0000-0000-000000000100,id=hostpci0,bus=ich9-pcie-port-1,addr=0x0: warning: vfio 00000000-0000-0000-0000-000000000100: Could not enable error recovery for the device
kvm: -device ide-cd,bus=ide.0,unit=1,drive=drive-ide1,id=ide1,bootindex=103: Can't create IDE unit 1, bus supports only 1 units
stopping swtpm instance (pid 1720) due to QEMU startup error
start failed: QEMU exited with code 1

This fix didnt help 5) Add it all vm with diffrent uuid 'args: -uuid 00000000-0000-0000-0000-000000000100' to vm file /etc/pve/local/qemu-server/100.conf

anyone have another idea?

 

[Neobin]

kvm: -device ide-cd,bus=ide.0,unit=1,drive=drive-ide1,id=ide1,bootindex=103: Can't create IDE unit 1, bus supports only 1 units

https://forum.proxmox.com/threads/pci-gpu-passthrough.83429/#post-366893