Skip to content

Unable to deploy pod that uses gpu-device-plugin, failed to generate spec: failed to mkdir #2296

Description

@UntouchedWagons

Describe the bug

Kubernetes pods that want to use an intel i915 gpu fail to be created, the following error is provided by Kubernetes:

Warning Failed 6s (x13 over 2m40s) kubelet spec.containers{vainfo}: (combined from similar events): Error: failed to generate container "79bcfdf507c2346c290f11ea5ac0381704751ce3f902d8b082fd75d8f791163f" spec: failed to generate spec: failed to mkdir "/dev/dri/by-path/pci-0000:00:02.0-platform-simple-framebuffer.0-card": mkdir /dev/dri/by-path/pci-0000:00:02.0-platform-simple-framebuffer.0-card: file exists

To Reproduce

Deploy single-node cluster on Ubuntu 26.04 with Intel GPU installed. Install Cert-manager, node feature discovery, intel-device-plugins-operator and intel-device-plugins-gpu helm charts then apply this vainfo pod.

Expected behavior

The vainfo pod to start and list supported encode/decode profiles in its stdout.

Screenshots

Not sure what could be added

System (please complete the following information):

  • OS version: Ubuntu 26.04 Server
  • Kernel version: 7.0.0-22-generic
  • Device plugins version: intel-device-plugins-operator (0.36.0), intel-device-plugins-gpu (0.36.0)
  • Hardware info: MinisForum MS-01 with Intel Core i5-12600H

Additional context

This identical deployment works perfectly on Debian 13.

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Type

    No type

    Fields

    No fields configured for issues without a type.

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions