Looks like systemd does not execute the statements in status) case of the init script
at all, but just checks if the daemon process exists. My '/etc/init.d/<my own
daemon> status' did much more, i.e., it checked if the daemon was actually able to
do some real work.
So far I have had no luck in finding the answer from the web.
You will find the answer here:
*
http://homepage.ntlworld.com./jonathan.deboynepollard/FGA/unix-daemon-readiness-protocol-problems.html
* https://freedesktop.org/software/systemd/man/sd_notify.html
* https://freedesktop.org/software/systemd/man/systemd-notify.html
*
https://lists.freedesktop.org/archives/systemd-devel/2014-April/018797.html
If you want to make your daemon interoperate with systemd's status
mechanism to the extent of having custom status reports, you have to
modify your daemon to send readiness notification messages through a
socket to the systemd service manager. That way, not only will
"/etc/init.d/jh status" report your custom statuses, so too will
"systemctl status jh" and (possibly) so too will (some) GUI
administration tools.
To do this, you must write a service unit for your service. Sticking
with a van Smoorenbug rc script and relying upon the
systemd-sysv-generator to write a compatibility service unit on the fly
to encapsulate it won't work, for several reasons. For starters, the
generator doesn't generate Type=notify service units and uses
RemainAfterExit=true (which prevents detection of service abends).
