Hi, I would like to know if enabling a GPU in a VM using iommu/passthrough (a 1:1 setup, where the host is unable to use the gpu and it becomes available to only one VM) is in the realm of “vGPU” , which are the compatible nvidia cards, and if i need a “licensing server” for that License Server User Guide :: NVIDIA Virtual GPU Software License Server Documentation
I’m trying to run a few VMS (qemu - ubuntu 20.04) on a Host (ubuntu 20.04), and setup a kubernetes cluster on those vms, and use nvidia-device-plugin to manage the gpu resources.
The host has a NVIDIA Quadro RTX P4000 card and i want to use it on a single vm (gpu passthrough).
I did manage to configure the passthrough and install the nvidia-device-plugin so the graphics card is visible from the VM but when i run a container on it it freezes completely.
thanks in advance