Weird behavior with CentOS 8.2 container

[koalillo]

Hi,

After running user cron jobs, the user's systemd service goes failed:

$ sudo systemctl status user@1000.service

● user@1000.service - User Manager for UID 1000
   Loaded: loaded (/usr/lib/systemd/system/user@.service; static; vendor preset: disabled)
   Active: failed (Result: timeout) since Tue 2020-06-16 11:02:04 UTC; 25min ago
  Process: 1202 ExecStart=/usr/lib/systemd/systemd --user (code=exited, status=0/SUCCESS)
 Main PID: 1202 (code=exited, status=0/SUCCESS)
   Status: "Startup finished in 27ms."

Jun 16 11:00:04 system.example.com systemd[1202]: Stopped target Paths.
Jun 16 11:00:04 system.example.com systemd[1202]: Stopped target Timers.
Jun 16 11:00:04 system.example.com systemd[1202]: Stopped Mark boot as successful after the user session has run 2 minutes.
Jun 16 11:00:04 system.example.com systemd[1202]: Stopped target Sockets.
Jun 16 11:00:04 system.example.com systemd[1202]: Closed D-Bus User Message Bus Socket.
Jun 16 11:00:04 system.example.com systemd[1202]: Reached target Shutdown.
Jun 16 11:00:04 system.example.com systemd[1202]: Starting Exit the Session...
Jun 16 11:02:04 system.example.com systemd[1]: user@1000.service: Processes still around after final SIGKILL. Entering failed mode.
Jun 16 11:02:04 system.example.com systemd[1]: user@1000.service: Failed with result 'timeout'.
Jun 16 11:02:04 system.example.com systemd[1]: Stopped User Manager for UID 1000.

After the cronjob runs, status prints out briefly:

Jun 17 18:51:01 systemdtest2.h2.int.pdp7.net systemd[1]: user@1001.service: Killing process 1187 (systemctl) with signal SIGKILL.

But no such process with pid 1187 exists, neither inside the LXC container nor outside in the Proxmox host.

Thoughts?

I've filed:

https://bugzilla.redhat.com/show_bug.cgi?id=1847437

, but this looks more like a LXC/Proxmox issue.

Cheers,

Álex

 

[koalillo]

Filed https://bugzilla.proxmox.com/show_bug.cgi?id=2807

 

[absolutesantaja]

I'm also seeing some systemd instability in CentOS 8.2 containers. In my case it's systemd-tmpfiles-setup that's not behaving well but I've had several occasions where systemctl just quits working altogether returning a dbus error. Are you using an unprivileged container or a privileged one?