Ubuntu 24.04 Image Missing For nvidia-driver-daemonset

[isugimpy]

1. Quick Debug Information

• OS/Version(e.g. RHEL8.6, Ubuntu22.04): Ubuntu24.04
• Kernel Version:6.8.0-1008-aws
• Container Runtime Type/Version(e.g. Containerd, CRI-O, Docker): CRI-O
• K8s Flavor/Version(e.g. K8s, OCP, Rancher, GKE, EKS): K8s
• GPU Operator Version: 24.3.0

2. Issue or feature description

Installed gpu-operator on a cluster hosting Ubuntu 24.04 nodes. Unable to install drivers because nvcr.io/nvidia/driver:550.54.15-ubuntu24.04 does not exist.

3. Steps to reproduce the issue

1. Launch an Ubuntu 24.04 node with a GPU in a cluster.
2. Install gpu-operator.
3. Observe ImagePullBackOff due to image not existing.

[frittentheke]

I just observed the same issue.

While Ubuntu 24.04 is not yet listed as supported OS (https://docs.nvidia.com/datacenter/cloud-native/gpu-operator/latest/platform-support.html#supported-operating-systems-and-kubernetes-platforms) for the Operator, support for Ubuntu 24.04 has come in GRID 17.2 (https://docs.nvidia.com/grid/latest/grid-vgpu-release-notes-ubuntu/index.html#abstract)

But that apparently is NOT the case for AI Enterprise 5.0 (https://docs.nvidia.com/ai-enterprise/5.0/release-notes/index.html // https://docs.nvidia.com/ai-enterprise/5.0/product-support-matrix/index.html#support-matrix__ubuntu), even though there are references over to GRID 17.x.

Also there seem to be no driver containers for ubuntui24.04 being built yet - https://gitlab.com/nvidia/container-images/driver

@shivamerla could you clarify which level of support there currently is (should be) for Ubuntu 24.04?

[isugimpy]

Corresponding issue I opened on the container-images repo. https://gitlab.com/nvidia/container-images/cuda/-/issues/226

[masterkain]

from the linked issue: IMHO do not expect ubuntu:24.04-based images before 2025.

insert sadface here