Subject: Re: /var/cron: hierarchy design arguments
To: Curt Sampson <firstname.lastname@example.org>
From: Bill Studenmund <skippy@macro.Stanford.EDU>
Date: 04/13/1999 13:29:31
On Thu, 8 Apr 1999, Curt Sampson wrote:
> Miles Nordin, in a great posting he made a couple of days ago,
> argued for moving crontabs by doing the following:
> o Dividing files into classes
> o Showing where in the hierarchy these classes of files fit
> o Explaining the purpose he's trying to achieve with this division
> o Showing how right now many files are in the right place for this
> scheme, but cron isn't.
> I've not heard anything like this from the other side, however.
> Can someone who believes that /var/cron is in its correct place
> right now please explain how he divides up files, which ones go in
> /etc, /var, etc. and show where cron fits in this? And also, unless
> everything else is in the correct place already, explain what other
> files are not in the correct places and should be moved.
> I'm just not seeing the general plan that makes people want to put
> crontabs in /var, rather than /etc.
The general plan, as I see it, relies on a couple of things, most all of
which involve the fact that (AFAICT) /var is much more likely to be its
own partition while /etc needs to be part of the root partition.
Think about diskless machines. They each need their own /var partitions,
but certainly could share /.
Given that, I think that cron files belong in /var. They certainly should
be machine specific, and of the two places to put system config info (/etc
or /var), for diskless machines, /var is the machine-specific one.
While I agree that the definitions in hier are wrong/out of date, I don't
think that's a reason to go moving stuff. :-) We should just fix hier(7).