chapman:~ # systemctl start dhcrelay chapman:~ # systemctl status dhcrelay|grep Active: Active: active (running) since Tue 2018-01-09 18:20:49 CET; 5s ago chapman:~ # grep PID /usr/lib/systemd/system/dhcrelay.service chapman:~ # ls /var/run/dhcrelay.pid /var/run/dhcrelay.pid chapman:~ # systemctl stop dhcrelay chapman:~ # systemctl status dhcrelay|grep Active: Active: inactive (dead) chapman:~ # ls /var/run/dhcrelay.pid ls: cannot access '/var/run/dhcrelay.pid': No such file or directory

can you please add more details why PIDfile is required on my machine its run without any issue http://pastebin.nue.suse.com/18727

Revoking my LGTM review

I agree, that the IPv4 relay agent starts on my virtual test machines just fine, whereas on a physical server while serving 46 vlan interfaces distributed over several physical interfaces and managed by a cluster agent the service is started and terminated again at once.

Please see http://susepaste.org/71360090

In the succeeding start (with PIDFile statement) there is the log line from systemd stating the unavailability of the PID file at that very moment, it seems dhcrelay is slower than systemd in this startup condition. The cluster itself is a 4 node cluster in 2/2 pairs, each master/slave. The relay only runs on one pair, and only on the master.

@mtomaschewski, @rmax: review reminder