Re: Question about cron behavior with missing directories
Hi,
Steve Greenland wrote:
> (http://bugs.debian.org/40843) because cron exits when it can't find
> /etc/cron.d. He got into this situation because he uses a software
[...]
>
> Originally I agreed with him, but after looking at the code and thinking
> about it a while, I'm not sure I do. My arguments would be thus:
>
> 1. If the spool dir (/var/spool/cron/crontabs) is missing, it logs and
> error and exits.
>
> 2. /etc/cron.d is provided by the cron package (actually, it isn't, but
> *that* will be fixed in the next release). We don't require our packages
> to survive random deletions of their content.
>
> 3. It's going to make for some ugly code (which is probably why Paul
> Vixie made choice #1).
>
> (FWIW, I think 1 and 2 are the compelling reasons...)
>
> So I guess I'm looking for some opinions about this matter.
Hmmm, I agree with you that this is more of a philosophic matter, but:
A) I think customization to the usual setup of UNIX-Systems is a rather
common thing in a heterogenous network.
I've never encountered an /etc/cron.d directory on DEC-UNIX-,
RISC-OS-, SUN-OS-, IBM-AIX-, Linux-Slackware-Systems therefore I
think this is a speciality of this cron. Personally I like behaviour
#1 not very much, but at least this IS standard. We have such a
herterogenous environment here, and like our machines to behave
alike.
B) I rate cron as a service which should work on a system as reliable as
syslog. Therefore I don't like the Idea, of a cron that's just dying
because of a missing (, nonstandard) directory. I see that this is
not really relevant for the guy, sitting on his PC at home, but if
debian should be something for everybody including larger scale
networks.
That's just my humble opinion on this matter (If there exists something
like a humble opinion ;-)
Regards,
Stefan.
--
------------------------------------------------------------
Stefan M. Petters Institute for Realtime-Computersystems -ooo
Arcisstr. 21 TU Muenchen --oooo
80290 Muenchen Tel: ++49/89-2892-3556 ----oo
Germany EMail: petters@lpr.ei.tum.de ---o
------------------------------------------------------------
Reply to: