[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: Bug#244908: is it OK for invoke-rc.d stop not to stop?



On Tue, Apr 20, 2004 at 03:40:30PM -0300, Henrique de Moraes Holschuh wrote:
> Daemons without pidfile control are broken and must be fixed.
> 
> IMHO lack of pidfile control in a daemon is a severity important bug on 
> itself.  If upstream didn't do the job right, fix it yourself... one can get

The pidfiles concept per se is broken in my opinion.  Pidfiles provide
unreliable information, and are subject to races.  I successfully use
service daemons that don't care about pidfiles, can easily be configured
to run in a chroot, and can reliably be controlled through signals.
Such service daemons needn't care about closing possibly open file
descriptors, creating and cleaning up files in /var/run/, the current
process state, and fork and exec.

 http://homepages.tesco.net/~J.deBoynePollard/FGA/unix-daemon-design-mistakes-to-avoid.html

Regards, Gerrit.



Reply to: