Comment 3 for bug 1488962

My first try to verify the current state was with the ports.conf trick, but that failed me (as it worked).

So I looked for other ways to reproduce.
The linked suse bug report used kill to recreate - and this worked for me.

Ok here we have our simplified test (not even needs to change the conf now) and the verification that it is still an issue in Xenial.

reset the conf , restart apache
killall -9 apache2
systemctl status apache2
● apache2.service - LSB: Apache2 web server
   Loaded: loaded (/etc/init.d/apache2; bad; vendor preset: enabled)
   Active: active (exited) since Wed 2016-04-06 16:05:43 UTC; 48s ago
     Docs: man:systemd-sysv-generator(8)
  Process: 1109 ExecStop=/etc/init.d/apache2 stop (code=exited, status=0/SUCCESS)
  Process: 1152 ExecStart=/etc/init.d/apache2 start (code=exited, status=0/SUCCESS)

Apr 06 16:05:43 xenial-tgt-tests systemd[1]: Starting LSB: Apache2 web server...
Apr 06 16:05:43 xenial-tgt-tests apache2[1152]: * Starting web server apache2
Apr 06 16:05:43 xenial-tgt-tests apache2[1152]: AH00558: apache2: Could not reliably determine the server's fully qualified domain name, using 127
Apr 06 16:05:43 xenial-tgt-tests apache2[1152]: *
Apr 06 16:05:43 xenial-tgt-tests systemd[1]: Started LSB: Apache2 web server.

That said - the behavior is not limited to apache - e.g. also nscd.
But on the other hand others (probably with real own service files) work - tested on openssh-server.
The current assumption is - as also mentioned in the initial report that it applies to all that come via systemd-sysv-generator.