search

NVIDIA / gpu-operator

NVIDIA GPU Operator creates/configures/manages GPUs atop Kubernetes
Apache License 2.0
1.77k stars 286 forks source link

update Update 0500_configmap.yaml to support L40S vgpu #720

Closed jxdn closed 2 months ago

jxdn commented 4 months ago

this PR to support L40S vgpu profile. latest version didnt support this

https://docs.nvidia.com/grid/latest/grid-vgpu-user-guide/index.html#vgpu-types-nvidia-l40s

cdesiniotis commented 4 months ago

Hi @jxdn thanks for your contribution! Currently, we do development on our Gitlab repository: https://gitlab.com/nvidia/kubernetes/gpu-operator Would you mind opening your PR there?

jxdn commented 4 months ago

@cdesiniotis done.

do you have step by step to update manually on the gpu-operator that run on k8s ?

we need to do it ASAP actually

cdesiniotis commented 4 months ago

@jxdn you can pass a custom vGPU devices configuration to the operator today. You do not need to build and deploy a custom gpu-operator image.

First, create a ConfigMap in the gpu-operator namespace that contains the entries for L40S. Then set vgpuDeviceManager.config.name="<name-of-configmap>" in Clusterpolicy so that the vgpu-device-manager pod uses your custom configuration file.

jxdn commented 4 months ago

Thank you ! @cdesiniotis

image

cdesiniotis commented 2 months ago

Closing as this was addressed in https://github.com/NVIDIA/gpu-operator/pull/799