Hi all, is there some way under Linux except monitoring "nvidia-smi -q | grep License" to know when vGPU license is active? There is nothing in man nvidia-gridd. I have one service that depends on CUDA so it needs to be (re)started when vGPU becomes licensed. Thank you.
The License is automatically allocated and active when the VM boots and stays active until the VM is rebooted / powered off. If your CUDA dependant service can have a delayed start, then it shouldn’t need to be restarted.
Thank you for your reply. But how can I determine a moment when license is active on my VM? There may be networking problems and license will be active not very quickly after VM start…
Your VM will experience poor performance and reduced functionality if a license has not been allocated. Frame rate limitations and reduced resolution will be the main things that are noticeable.
Which version of Linux are you running?