Operator fails to render a valid daemonset on OCP when using 64K kernel page size.

[mvazquezc]

The relevant code for the version we are using (v24.6.1):

https://github.com/NVIDIA/gpu-operator/blob/release-24.6/controllers/object_controls.go#L2809

The function above will render the "sanitized" string as: 5.14.0-427.37.1.el9.4..64k-rhcos4.16 which is not a valid name.

The resulting DaemonSet name will be nvidia-driver-daemonset-5.14.0-427.37.1.el9.4..64k-rhcos4.16 and API will complain about it when the controller tries to create it:

{"level":"info","ts":"2025-01-16T09:45:22Z","logger":"controllers.ClusterPolicy","msg":"DaemonSet not found, creating","DaemonSet":"nvidia-driver-daemonset","Namespace":"nvidia-gpu-operator","Name":"nvidia-driver-daemonset-5.14.0-427.37.1.el9.4..64k-rhcos4.16"}
.
.
.
{"level":"info","ts":"2025-01-16T09:45:00Z","logger":"controllers.ClusterPolicy","msg":"Couldn't create DaemonSet","DaemonSet":"nvidia-driver-daemonset","Namespace":"nvidia-gpu-operator","Name":"nvidia-driver-daemonset-5.14.0-427.37.1.el9.4..64k-rhcos4.16","Error":"DaemonSet.apps \"nvidia-driver-daemonset-5.14.0-427.37.1.el9.4..64k-rhcos4.16\" is invalid: metadata.name: Invalid value: \"nvidia-driver-daemonset-5.14.0-427.37.1.el9.4..64k-rhcos4.16\": a lowercase RFC 1123 subdomain must consist of lower case alphanumeric characters, '-' or '.', and must start and end with an alphanumeric character (e.g. 'example.com', regex used for validation is '[a-z0-9]([-a-z0-9]*[a-z0-9])?(\\.[a-z0-9]([-a-z0-9]*[a-z0-9])?)*')"}

[tariq1890]

Hey @mvazquezc Thanks for raising this issue and the PR to fix it. For my reference can you share the following node label value of the node running on 64k kernel?

feature.node.kubernetes.io/kernel-version.full

[mvazquezc]

@tariq1890:

feature.node.kubernetes.io/kernel-version.full: 5.14.0-427.37.1.el9_4.aarch64_64k