Skip site navigation (1)Skip section navigation (2)
Date:      Sun, 12 Jan 1997 18:11:20 +0100
From:      j@uriah.heep.sax.de (J Wunsch)
To:        hackers@freebsd.org
Subject:   Re: DEVFS permissions &c.
Message-ID:  <Mutt.19970112181120.j@uriah.heep.sax.de>
In-Reply-To: <Mutt.19970112003853.davidn@labs.blaze.net.au>; from David Nugent on Jan 12, 1997 00:38:53 %2B0000
References:  <Mutt.19970111201007.j@uriah.heep.sax.de> <16902.853042470@time.cdrom.com> <Mutt.19970112112012.j@uriah.heep.sax.de> <Mutt.19970112003853.davidn@labs.blaze.net.au>

next in thread | previous in thread | raw e-mail | index | archive | help
As David Nugent wrote:

> KISS works better. Any /etc/rc.shutdown has /etc/sysconfig as a
> resource and should use it if relevent.

I don't mind.

> And please - PLEASE - no "runlevels"!

Yeah.  Actually, we do already have runlevels: S, 0, [234] (depending
on the actual SysV vendor), and 6. :-)

But i agree that extending this to more runlevels is useless.  I have
yet to see a single SysV implementation that groks all runlevel
transitions without silly actions like starting up subsystems if you
lower the runlevel etc.

(Shutdown grace period)

> Make it configurable (sysctl var?), but 10 seconds would seem like a

I think the user or machdep facilities of sysctl might come handy
here.  Opinions?

> reasonable default on most systems. I've found that if a news server,
> for example, is particularly busy, it can take up to a minute for
> ctlinnd shutdown to complete. A shutdown 'fastboot' option to bypass
> the shutdown script might also be useful.

There's always `reboot -q'.

> Of course, the 10 seconds wait is a moot point if the script returns
> earlier...

Yes, if the shutdown script exited, init can proceed.

> Naturally, this brings ports in, with /usr/local/etc/rc.d scripts
> again, but you knew that. :-)

Yep.  The scripts there do already get an argument "start" passed from
/etc/rc.

> Perhaps the *existance* of an /etc/rc.shutdown will be enough to
> trigger it.

Yep.  Failure to find the script could be syslogged.

-- 
cheers, J"org

joerg_wunsch@uriah.heep.sax.de -- http://www.sax.de/~joerg/ -- NIC: JW11-RIPE
Never trust an operating system you don't have sources for. ;-)



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?Mutt.19970112181120.j>