NVIDIA / gpu-feature-discovery

GPU plugin to the node feature discovery for Kubernetes
Apache License 2.0
292 stars 47 forks source link

cannot generate nvidia.com/cuda.xxx labels on node #75

Closed FLM210 closed 6 months ago

FLM210 commented 6 months ago

I have installed gpu-operator in my cluster, and it appears that all components are running normally, but the nvidia.com/cuda.xxx is missing on a certain node。

image
FLM210 commented 6 months ago

I have found the reason, as gpu-feature-discovery run before the driver installation was completed, resulting in the inability to load NVML library