Marc 'Zugschlus' Haber

Lies mal die Debian policy 9.3.2, da wird ganz gut erklärt, was die Mechanismen um ein Initscript herum erwarten. Kurz gefasst: - restart muss restarten - force-reload kann die Konfiguration...

I think that it might be worth going through the motions of tears down and rebuild, that is still quicker than stop, wait for "really dead" and restart. In that...

> @Zugschlus: please check #458 I cannot really comment on the actual change, my C is not fluent enough. I didn't see something blatantly wrong.

> i also think i came up with an easy-to-implement solution for "reloading" on SIGHUP: > i would simply tear down vzlogger like on exit, but then instead of exiting,...

> [wrote this before the previous post but had not submitted:] > even when only internally stopping and restarting the metering thread, we may lose s0 pulses. > and we...

> (i need to add that it seems sad that one can't trust the wrapper to restart vzlogger quickly.) The easy fix would be to have do_stop in the init...

> > > @Zugschlus: please check #458 > > > > > > I cannot really comment on the actual change, my C is not fluent enough. I didn't see...

> > Any no, I cannot help here, I have moved to using a systemd unit for my vzlogger, sorry. > > can you share yours? Sure: ``` [Unit] Description=vzlogger...

> shouldn't it be worth improving the situation by getting systemd setups to use a proper systemd unit instead of the initscript via the generator? > (it seems to me...

> Immer noch das gleiche. Ich sehe keinerlei Meldung, dass ein coredump geschrieben worden wäre :-( Irgendwelche Ideen, wie ich dem Fehler auf die Spur kommen könnte? > > michael@raspy4:~...