Bug 1860899

Description of problem:

With RHEL8, "systemd --user" instance spawns when a user slice is created. Upon user slice destruction, a "systemctl --user --force exit" command is executed.
Due to race condition, systemd (PID 1) kills the "systemctl" command above and print a message at NOTICE level, which pollutes the logs.


Version-Release number of selected component (if applicable):

systemd-239-31.el8_2.2


How reproducible:

Almost Always


Steps to Reproduce:
1. Login using ssh as a non-logged in user

  # ssh <user>@vm-rhel8 true

2. Check the journal

  # journalctl -p notice -u user@<userid>.service --follow

Actual results:

Jul 27 13:34:20 vm-rhel8 systemd[1]: user@<userid>.service: Killing process 2891 (systemctl) with signal SIGKILL.


Expected results:

No message, or at least at INFO level max


Additional info:

This seems due to a race condition when closing the last session: systemd(1) still sees the "systemctl --user --force exit" command which makes "systemd --user" exit and wait for it to exit.
A solution may be to use "systemctl --user --force --no-block exit" instead.