1

I had some containers running smoothly in Centos 8.2 with podman. After upgrading to Centos 8.5, I found several of the containers failing to run. In the command line, I see this:

$ podman run --name=digikam-test lscr.io/linuxserver/digikam
Error: OCI runtime error: container_linux.go:380: starting container process caused: process_linux.go:545: container init caused: Running hook #0:: error running hook: exit status 1, stdout: , stderr:

In journalctl -xe, I see this:

Dec 27 06:55:36 oci-systemd-hook[2869749]: systemdhook <debug>: c4e81974f5be: :memory:/user.slice/user-1000.slice/session-4.scope
Dec 27 06:55:36 oci-systemd-hook[2869749]: systemdhook <debug>: c4e81974f5be: Found cgroup
Dec 27 06:55:36 oci-systemd-hook[2869749]: systemdhook <debug>: c4e81974f5be: PATH: /user.slice/user-1000.slice/session-4.scope
Dec 27 06:55:36 oci-systemd-hook[2869749]: systemdhook <debug>: c4e81974f5be: SUBSYSTEM_PATH: /sys/fs/cgroup/memory/user.slice/user-1000.slice/session-4.scope
Dec 27 06:55:36 oci-systemd-hook[2869749]: systemdhook <debug>: c4e81974f5be: memory path: /sys/fs/cgroup/memory/user.slice/user-1000.slice/session-4.scope/memory.limit_in_bytes
Dec 27 06:55:36 oci-systemd-hook[2869749]: systemdhook <debug>: c4e81974f5be: LIMIT: 9223372036854771712
Dec 27 06:55:36 oci-systemd-hook[2869749]: systemdhook <debug>: c4e81974f5be: Limit in bytes: 9223372036854771712
Dec 27 06:55:36 oci-systemd-hook[2869749]: systemdhook <error>: c4e81974f5be: Failed to mkdir journal dir: /var/log/journal/c4e81974f5be: Permission denied

I disabled selinux with setenforce 0 while trying to solve this issue, but it does not change the error message. After updating to Centos 8.5, I've migrated the OS to Rocky Linux 8.5. The error message remained the same. podman version is 3.3.1. I have at least one container that runs correctly after the 8.2->8.5->Rocky migration. That one container is not sourced from linuxserver.io.

Can anyone help me figure out what the problem is here with the inability for the container to create a journal dir? Thanks for any help!

sysadmin1138
  • 131,083
  • 18
  • 173
  • 296
molsqueak
  • 11
  • 1
  • i think the most issues might be that it could be that the cgroup changed fron 1 to 2 so cgroup2 should be used? – djdomi Jan 01 '22 at 16:57
  • It appears that cgroup v1 is in use. `$ podman info host: arch: amd64 buildahVersion: 1.22.3 cgroupControllers: [] cgroupManager: cgroupfs cgroupVersion: v1 ` – molsqueak Jan 02 '22 at 14:56

1 Answers1

0

Are you running it as root? Most likely it is running as a non-privileged user and cannot create the directory.

tilleyc
  • 914
  • 4
  • 11
  • I am running it as my normal user as I had done in Centos 8.2 when things were working. I had originally ran them this way because it appeared to be best practice with podman. I will look into giving my user account permission to create directories in /var/log/journal – molsqueak Dec 31 '21 at 14:43