I am using fedora 18 beta for some automated tests.
Sometimes it is unclear when the daemon has been restarted by systemd.
In the old sysv world, we had a restart wrapper that took an option in
the config file to say restart=no.
Obviously, I can just set Restart=no in the service file, but we end up
running "make install" and forgetting to change this.
So looking through the systemctl man page, I see:
disable [NAME...]
Disables one or more units. This removes all symlinks to the
specified unit files from the
unit configuration directory, and hence undoes the changes made by
enable. Note however that
this removes all symlinks to the unit files (i.e. including manual
additions), not just those
actually created by enable. This call implicitly reloads the systemd
daemon configuration
after completing the disabling of the units. Note that this command
does not implicitly stop
the units that is being disabled. If this is desired an additional
stop command should be
executed afterwards.
However, that _seems_ to be a lie?
First I test the systmd restarting:
[root@west ~]# pidof pluto
1762
[root@west ~]# killall -9 pluto
[root@west ~]# pidof pluto
1895 1893
[root@west ~]# pidof pluto
1893
[root@west ~]# killall -9 pluto
[root@west ~]# pidof pluto
1992
So the daemon restarts all the time. So let's tell systemd to disable
it, while leaving the service running:
[root@west ~]# systemctl disable ipsec.service
[root@west ~]# pidof pluto
1992
[root@west ~]# killall -9 pluto
[root@west ~]# pidof pluto
2095
So I guess the man page is either lying or the disasble command is not
working as expected.
[root@west ~]# rpm -q systemd
systemd-44-23.fc17.x86_64
In case you decide to fix the man page, how could I get the service to
not restart without editing the systemd service file?
Paul
_______________________________________________
systemd-devel mailing list
[email protected]
http://lists.freedesktop.org/mailman/listinfo/systemd-devel
