systemd's stricter PIDFile checking breaks EXEC=screen

[undeference]

If you start your server in an existing screen instance, systemd will refuse to monitor it. Currently, systemctl start newrun@whatever will report a timeout or configured resource limit exceeded, but systemctl status -l newrun@whatever will report the correct issue.

In this case, the service systemd is supposed to be monitoring does not belong to the cgroup systemd created for it, so it will always fail. There is no way around that.

You must instead configure unique screen sessions for each newrun-managed server. If the server crashes, the screen session will persist and have to be ended manually before restarting the server.

[undeference]

If the server crashes, the screen session will persist and have to be ended manually before restarting the server.

This is no longer true as of 9843862, which disabled screen guard. Now, when the last window is closed, the screen session should end.