Bugzilla – Bug 1221278
trying fedora39 container on Leap 15.6 Beta fails on inappropriate ioctl for device
Last modified: 2024-07-18 12:02:32 UTC
lkocman@localhost:~> rpm -qa | egrep "docker|podman|distrobox" docker-24.0.5_ce-150600.200.49.x86_64 distrobox-1.6.0.1-bp156.1.1.noarch podman-4.7.2-150600.1.40.x86_64 docker-bash-completion-24.0.5_ce-150600.200.49.noarch docker-buildx-0.11.2-bp156.1.9.x86_64 docker-rootless-extras-24.0.5_ce-150600.200.49.noarch distrobox-bash-completion-1.6.0.1-bp156.1.1.noarch lkocman@localhost:~> distrobox-create --name fedora39 --image fedora:39 Creating 'fedora39' using image fedora:39 [ OK ] Distrobox 'fedora39' successfully created. To enter, run: distrobox enter fedora39 lkocman@localhost:~> distrobox-enter fedora39 Starting container... /usr/bin/distrobox-enter: line 567: /home/lkocman/.cache/.fedora39.fifo: No such file or directory /usr/bin/distrobox-enter: fork: retry: Resource temporarily unavailable /usr/bin/distrobox-enter: fork: retry: Resource temporarily unavailable /usr/bin/distrobox-enter: fork: Interrupted system call /usr/bin/distrobox-enter: fork: retry: Resource temporarily unavailable /usr/bin/distrobox-enter: fork: retry: Resource temporarily unavailable /usr/bin/distrobox-enter: fork: Interrupted system call /usr/bin/distrobox-enter: fork: retry: Resource temporarily unavailable bash: fork: retry: Resource temporarily unavailable (in another tab) lkocman@localhost:~> distrobox-enter fedora39 Error: OCI runtime error: runc: exec failed: unable to start container process: inappropriate ioctl for device lkocman@localhost:~>
Added to https://en.opensuse.org/openSUSE:Known_bugs_15.6#Containers_and_Virtualization
This is quite a popular usecase for distrobox https://itsfoss.com/distrobox/
The problem seems to affect only Podman. The issue is not reproducible if the container manager is Docker. The problem is also not reproducible on Tumbleweed, but it is present on Leap 15.6. If both Docker and Podman are available on the machine, distrobox will use Podman by default. Setting DBX_CONTAINER_MANAGER=docker overrides the auto-select mechanism. In Podman, the issue seems to be present only in rootless mode. Running Distrobox with real root (--root argument) solves the issue as well. In rootless mode, it seems that Distrobox is stuck at "Starting container" phase. Enabling verbose log shows that Distrobox is stuck in the following command: podman --log-level debug logs -f fedora39 The container itself does not seem to produce any logs and follow (-f) keeps the create command waiting forever. For now, the best workarounds are using Docker or rootfull Podman.
Can you still reproduce this issue? I tested again and everything seems to be working, previously I had issues trying to get the container up and running. ~> distrobox ls ID | NAME | STATUS | IMAGE 952b56a393a6 | fedora39 | Up 5 minutes | registry.fedoraproject.org/fedora:39 ~> distrobox enter fedora39 alexandre@fedora39 ~]$ cat /etc/os-release NAME="Fedora Linux" VERSION="39 (Container Image)"