k8s-driver-manager container image is not included as part of the OLM Operator Bundle

[tginer]

The nvcr.io/nvidia/cloud-native/k8s-driver-manager image is required and is not packaged in the Kubernetes Operator Bundle

In an air gapped environment after mirroring with 'oc-mirror' tool the gpu-operator-certified bundle, the k8s-driver-manager image was missing in the offline registry and I had to manually mirror it. Ideally it should be included as part of the bundle package, as it is required and specified in the CSV:

oc get csv -n nvidia-gpu-operator gpu-operator-certified.v24.6.2 -o yaml | grep k8s-driver
                  value: nvcr.io/nvidia/cloud-native/k8s-driver-manager@sha256:740abc3ff657545c10effd5354f09af525200ed9a1b7623f0c2e8c7bd9e4a4e2
  - image: nvcr.io/nvidia/cloud-native/k8s-driver-manager@sha256:740abc3ff657545c10effd5354f09af525200ed9a1b7623f0c2e8c7bd9e4a4e2
    name: k8s-driver-manager-image

In this specific version OCP 4.16.14 and GPU Operaotr 24.6.2, this specific container image was missing:
nvcr.io/nvidia/cloud-native/k8s-driver-manager@sha256:740abc3ff657545c10effd5354f09af525200ed9a1b7623f0c2e8c7bd9e4a4e2
but should be included in every operator version.

[cdesiniotis]

Hi @tginer I do see this image in our 24.6.2 bundle: https://github.com/redhat-openshift-ecosystem/certified-operators/blob/main/operators/gpu-operator-certified/v24.6.2/manifests/gpu-operator-certified.clusterserviceversion.yaml#L227-L228