Ben Munat schrieb:
> Well, long story short, I discovered a stale lock file in /tmp. The
> screwy thing was that the init.d script kept saying "ok" even though
> postgres wasn't actually running. That's a pretty bad bug, if you ask
> me, but I'm not a dev so...
Unfortunately you cannot trust the output of /etc/init.d/<servicename>,
start-stop-daemon plain sucks and lies in your face. I often resort to
ugly hacks using ps and grepping for (hopefully) unique strings and
writing a valid pidfile myself.
There are various solutions (daemontools, supervisor) which run the
started program as a childprocess..., of course the init scripts need to
 be rewritten for those...

cheers
 Paul

-- 
[email protected] mailing list

Reply via email to