NVIDIA / gpu-operator

NVIDIA GPU Operator creates, configures, and manages GPUs in Kubernetes
https://docs.nvidia.com/datacenter/cloud-native/gpu-operator/latest/index.html
Apache License 2.0
1.8k stars 293 forks source link

Add support to GPU operator for RHEL 9 #1035

Open kimminw00 opened 1 week ago

kimminw00 commented 1 week ago

We are requesting support to GPU operator for Red Hat Enterprise Linux (RHEL) 9. Our organization is planning to upgrade to RHEL 9(or Rocky9) and we need GPU Operator to be compatible with this version. Currently, GPU Operator only supports up to RHEL 8. We believe that supporting RHEL 9(or Rocky9) will benefit not only our organization but also other users who are planning to upgrade to the latest version of RHEL.

fabiendupont commented 1 week ago

@kimminw00, what makes you say that only RHEL 8 is supported? Are you running vanilla Kubernetes on RHEL 9? If yes, do you expect the operator to deploy the drivers or do you install them as RPMs on RHEL 9?

kimminw00 commented 1 week ago

We initially believed that RHEL 9 was not supported, based on the information provided on the NVIDIA website. However, we are currently planning to upgrade our Linux version and would like to confirm compatibility. Our expectation is that the operator will handle driver deployment.

francisguillier commented 1 week ago

We support upstream K8s with RHEL8 only for the moment (GPU operator will deploy the driver container as well as the rest of the operands).

We intend to support RHEL9 in CY 2025 but we have not set the exact release vehicle yet (we are still sorting out the priorities of other features we need to implement).