gpu-operator icon indicating copy to clipboard operation
gpu-operator copied to clipboard
verified publisher icon NVIDIA

Nvidia-gpu-operator 25.3.3 / Openshift 4.18.21

Open Ali-CTN opened this issue 3 months ago • 1 comments

Hi folks,

We are using 4.18.21 in our Openshift environment. When we installed the Nvidia operator and used Nvidia driver version 535.261.03, there was no problem. We installed the new driver to upgrade the CUDA version. The new versions are: Driver Version: 580.82.07 CUDA Version: 13.0. However, there are currently labels indicating issues with mig operations on my nodes:

nvidia.com/gpu.deploy.nvsm: paused-for-mig-change

and the following logs are present in the nvidia-mig-manager pods;

time="2025-09-12T18:54:04Z" level=fatal msg="Error applying MIG configuration with hooks: error initializing NVML: ERROR_LIBRARY_NOT_FOUND"

Is the operator and driver in this version compatible with the relevant OpenShift version?

thanks for the answers.

Ali-CTN avatar Sep 16 '25 14:09 Ali-CTN