helm instal gpu-operator was in Init stage for a long time

[JShuang7711]

gpu-operator nvidia-driver-daemonset-5qcw4 0/1 Init:0/1 0 173m
$ kubectl describe pod nvidia-driver-daemonset-5qcw4 -n gpu-operator
subscription-config-0:
Type: HostPath (bare host directory volume)
Path: /etc/pki/entitlement
HostPathType: Directory
subscription-config-1:
Type: HostPath (bare host directory volume)
Path: /etc/yum.repos.d/redhat.repo
HostPathType: File
subscription-config-2:
Type: HostPath (bare host directory volume)
Path: /etc/rhsm
HostPathType: Directory
kube-api-access-f6dtc:
Type: Projected (a volume that contains injected data from multiple sources)
TokenExpirationSeconds: 3607
ConfigMapName: kube-root-ca.crt
ConfigMapOptional:
DownwardAPI: true
QoS Class: BestEffort
Node-Selectors: nvidia.com/gpu.deploy.driver=true
Tolerations: node.kubernetes.io/disk-pressure:NoSchedule op=Exists
node.kubernetes.io/memory-pressure:NoSchedule op=Exists
node.kubernetes.io/not-ready:NoExecute op=Exists
node.kubernetes.io/pid-pressure:NoSchedule op=Exists
node.kubernetes.io/unreachable:NoExecute op=Exists
node.kubernetes.io/unschedulable:NoSchedule op=Exists
nvidia.com/gpu:NoSchedule op=Exists
Events:
Type Reason Age From Message

---

Warning FailedMount 2m32s (x94 over 177m) kubelet MountVolume.SetUp failed for volume "subscription-config-1" : hostPath type check failed: /etc/yum.repos.d/redhat.repo is not a file

[JShuang7711]

@shivamerla can you help to check this issue.

[gaurav-cloudera]

we are also hitting the same issue while deploying the nvidia gpu operator. I see the code we have a map based on the os type and we are iterating through it to configure the mount points.

gpu-operator/internal/state/driver_volumes.go

Line 59 in dfbb6c5

var SubscriptionPathMap = map[string]MountPathToVolumeSource{

I am not sure if the file is expected to be available with the OS installation or not. Can anyone please help on this.

Helm command used:
helm install --wait --generate-name -n gpu-operator --create-namespace nvidia/gpu-operator --version=v25.3.0 --set mig.strategy=single

`Events:
Type Reason Age From Message

---

Normal Scheduled 114s yunikorn Successfully assigned gpu-operator/nvidia-driver-daemonset-sq5d7 to
Warning FailedMount 50s (x8 over 114s) kubelet MountVolume.SetUp failed for volume "subscription-config-1" : hostPath type check failed: /etc/yum.repos.d/redhat.repo is not a file
gmalik@Gauravs-MacBook-Pro default %`

[gaurav-cloudera]

@shivamerla could you please share any pointers on this.

[github-actions]

This issue is stale because it has been open 90 days with no activity. This issue will be closed in 30 days unless new comments are made or the stale label is removed. To skip these checks, apply the "lifecycle/frozen" label.

[cdesiniotis]

@JShuang7711 can you please provide more information about your cluster (k8s, container runtime, etc.) and how you are installing GPU Operator (version, values, etc.)?