Unable to run NVIDIA containers

[muety]

Issue Description

I installed Podman on Ubuntu 22.04 from the official sources. In addition, I installed the NVIDIA Container Toolkit as described here and generated a CDI spec as described here. Running nvidia-ctk cdi list properly gives the following output:

INFO[0000] Found 2 CDI devices                          
nvidia.com/gpu=0
nvidia.com/gpu=all

However, when trying to run

podman run --rm --device=nvidia.com/gpu=0 ubuntu nvidia-smi -L

I'm getting the following error.

Error: stat nvidia.com/gpu=0: no such file or directory

The above NVIDIA docs mention that the --device argument is only available from Podman 4.1.0, while I'm on 3.4.4, so maybe that's an issue? If yes, what is the correct way of running GPU-dependent containers with my Podman version?

Steps to reproduce the issue

See above.

Describe the results you received

Error: stat nvidia.com/gpu=0: no such file or directory

Describe the results you expected

Actual nvidia-smi output.

podman info output

host:
  arch: amd64
  buildahVersion: 1.23.1
  cgroupControllers:
  - memory
  - pids
  cgroupManager: systemd
  cgroupVersion: v2
  conmon:
    package: 'conmon: /usr/bin/conmon'
    path: /usr/bin/conmon
    version: 'conmon version 2.0.25, commit: unknown'
  cpus: 48
  distribution:
    codename: jammy
    distribution: ubuntu
    version: "22.04"
  eventLogger: journald
  hostname: ufo
  idMappings:
    gidmap:
    - container_id: 0
      host_id: 44350
      size: 1
    - container_id: 1
      host_id: 1000000
      size: 1024
    uidmap:
    - container_id: 0
      host_id: 246740
      size: 1
    - container_id: 1
      host_id: 1000000
      size: 1024
  kernel: 6.2.0-34-generic
  linkmode: dynamic
  logDriver: journald
  memFree: 128253014016
  memTotal: 134896099328
  ociRuntime:
    name: crun
    package: 'crun: /usr/bin/crun'
    path: /usr/bin/crun
    version: |-
      crun version 0.17
      commit: 0e9229ae34caaebcb86f1fde18de3acaf18c6d9a
      spec: 1.0.0
      +SYSTEMD +SELINUX +APPARMOR +CAP +SECCOMP +EBPF +YAJL
  os: linux
  remoteSocket:
    path: /run/user/246740/podman/podman.sock
  security:
    apparmorEnabled: false
    capabilities: CAP_CHOWN,CAP_DAC_OVERRIDE,CAP_FOWNER,CAP_FSETID,CAP_KILL,CAP_NET_BIND_SERVICE,CAP_SETFCAP,CAP_SETGID,CAP_SETPCAP,CAP_SETUID,CAP_SYS_CHROOT
    rootless: true
    seccompEnabled: true
    seccompProfilePath: /usr/share/containers/seccomp.json
    selinuxEnabled: false
  serviceIsRemote: false
  slirp4netns:
    executable: /usr/bin/slirp4netns
    package: 'slirp4netns: /usr/bin/slirp4netns'
    version: |-
      slirp4netns version 1.0.1
      commit: 6a7b16babc95b6a3056b33fb45b74a6f62262dd4
      libslirp: 4.6.1
  swapFree: 2147479552
  swapTotal: 2147479552
  uptime: 4h 10m 47.75s (Approximately 0.17 days)
plugins:
  log:
  - k8s-file
  - none
  - journald
  network:
  - bridge
  - macvlan
  volume:
  - local
registries: {}
store:
  configFile: /home/ws/pu0288/.config/containers/storage.conf
  containerStore:
    number: 0
    paused: 0
    running: 0
    stopped: 0
  graphDriverName: overlay
  graphOptions:
    overlay.mount_program:
      Executable: /usr/bin/fuse-overlayfs
      Package: 'fuse-overlayfs: /usr/bin/fuse-overlayfs'
      Version: |-
        fusermount3 version: 3.10.5
        fuse-overlayfs: version 1.7.1
        FUSE library version 3.10.5
        using FUSE kernel interface version 7.31
  graphRoot: /var/containers/pu0288
  graphStatus:
    Backing Filesystem: extfs
    Native Overlay Diff: "false"
    Supports d_type: "true"
    Using metacopy: "false"
  imageStore:
    number: 5
  runRoot: /run/user/246740/containers
  volumePath: /var/containers/pu0288/volumes
version:
  APIVersion: 3.4.4
  Built: 0
  BuiltTime: Thu Jan  1 01:00:00 1970
  GitCommit: ""
  GoVersion: go1.18.1
  OsArch: linux/amd64
  Version: 3.4.4

Podman in a container

No

Privileged Or Rootless

Rootless

Upstream Latest Release

No

Additional environment details

Additional environment details

Additional information

Additional information like issue happens only occasionally or issue happens with a particular architecture or on a particular setting

[Luap99]

You need to use a newer podman, 4.6 or 4.7 is strongly recommend.
We do not support older versions upstream.

[muety]

What is the recommended way to get Podman on Ubuntu 22.04? Since we have a whole pool of workstations to manage, I'd preferably wouldn't want to have to compile it myself.