Date: Sat, 12 May 2001 10:18:47 +0200 From: Gerhard Sittig <Gerhard.Sittig@gmx.net> To: cvs-all@freebsd.org Subject: Re: $EDITOR $CONF && kill -HUP $PID (was: cvs commit: doc/en_US.ISO_8859-1/books/handbook/advanced-networking chapter.sgml) Message-ID: <20010512101846.G253@speedy.gsinet> In-Reply-To: <Pine.NEB.3.96L.1010511165523.30751C-100000@fledge.watson.org>; from rwatson@freebsd.org on Fri, May 11, 2001 at 04:59:55PM -0400 References: <20010511195046.B1672@freebie.demon.nl> <Pine.NEB.3.96L.1010511165523.30751C-100000@fledge.watson.org>
next in thread | previous in thread | raw e-mail | index | archive | help
[ the subject looks this funny to avoid religious wars on editors
and stuff :) ]
On Fri, May 11, 2001 at 16:59 -0400, Robert Watson wrote:
>
> On Fri, 11 May 2001, Wilko Bulte wrote:
> >
> > Would this be something for FreeBSD? I, for one, tend to
> > forget this kill -1 thingy.
>
> Maybe it's just me, but I hate it when I discover "automagic"
> things like that. As the administrator, I know when the
> contents of the file /etc/exports will be consistent on disk,
> and I indicate that to the daemon by sending it a signal.
Would something like a "vicrontab(1)" - or whatever it could be
named - wrapper be appropriate? This idea does a great job on
regenerating password databases from their text representation
already.
Another hack(?) could be some Makefile style thing, much like
there is in /etc/mail or next to your DNS database. But I
wouldn't know which criterion describes best the fact that a
running daemon hasn't seen a new config file yet ("[ $CONF -nt
$PID ] && kill -HUP `cat $PID`"? but this would require the
daemon to touch or rewrite the pidfile upon reload -- or the
Makefile had to keep track).
> I'm all for making restarting daemons easier -- you can imagine
> a daemonctl or something that speaks "restart" for a bunch of
> daemons, but automagic can sometimes hurt you more than it
> helps you.
Starting / stopping / hupping / 'stat'ing / etc daemons would be
easier if /etc/rc granularity was finer and the pieces would look
like the /usr/local/etc/rc.d scripts. But past discussions
usually turned into "that's a Linux thing and therefor cannot be
good for us". :> I understand that splitting logically oriented
groups is not the problem but the wrapper to keep the sequence
(would be nice: while obeying dependencies) is. And there have
been concerns that repeated "source"ing of config files slows
operation down (although I estimate that we talk about single
seconds - if at all - once in a boot sequence; plus code has been
sketched to flag an already read file by a variable).
virtually yours 82D1 9B9C 01DC 4FB4 D7B4 61BE 3F49 4F77 72DE DA76
Gerhard Sittig true | mail -s "get gpg key" Gerhard.Sittig@gmx.net
--
If you don't understand or are scared by any of the above
ask your parents or an adult to help you.
To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe cvs-all" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20010512101846.G253>