Skip site navigation (1)Skip section navigation (2)
Date:      Fri, 8 Nov 2002 19:35:06 -0800
From:      Mike Makonnen <mtm@matrix.identd.net>
To:        current@freebsd.org
Subject:   Re: rc.d and sysctl.conf
Message-ID:  <20021109033506.GB6608@matrix.identd.net>
In-Reply-To: <200210302236.OAA26726@windsor.research.att.com>
References:  <200210302236.OAA26726@windsor.research.att.com>

next in thread | previous in thread | raw e-mail | index | archive | help

--CUfgB8w4ZwR/yMy5
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
Content-Transfer-Encoding: quoted-printable

On Wed, Oct 30, 2002 at 02:36:01PM -0800, Bill Fenner wrote:
[ snip ]
>
> The update to the /etc/rc.d infrastructure keeps the ability to run
> twice, but does not actually run it twice.  I started creating an
> /etc/rc.d/sysctl-last that would run "/etc/rc.d/sysctl lastload",
> but realized that I didn't know how to say where the first/second
> call should go.  To strictly follow /etc/rc.d, I could change the
> existing /etc/rc.d/sysctl to say "BEFORE: serial" and add "BEFORE:
> securelevel" to sysctl-last, but I'm not sure this is appropriate given
> the meta-checkpoints that we have.

One of the hard things while I was doing the porting was deciding whether
something in /etc/rc was there because it *must* run before the commands
that were after it and after the commands that came before it. Since there
haven't been any complaints in that regard I don't think the current order
has broken anything. The general
rule is to put something in REQUIRE and/or BEFORE only if it is necessary
that some script be run before or after the current script. So, if the
sysctls *must* be set before SERIAL, it should be in the BEFORE line.
Otherwise, I would leave it as is and run `/etc/rc.d/sysctl lastload' in
/etc/rc.d/securelevel just before rasing the securelevel (Please see the
attached patch).

>
> (It also raises the question of if /etc/rc.d/securelevel actually
> runs at the right time.  /etc/rc puts it almost at the absolute end,
> while rcorder sticks it somewhere in the middle -- number 67 of 102
> on my system.)

We wanted to keep the differences between our scripts and NetBSD's to
a minimum. So, if it turns out we've broken something because of where
rcorder puts the securelevel script, then we'll have to modify the
BEFORE line of the affected script.

Cheers,
Mike

--=20
GPG Key: http://www.identd.net/~mtm/gpg.key
pub  1024D/7D39509A 2002-10-08 Mike Makonnen <mtm@identd.net>
Key fingerprint =3D 5491 488A 0445 2DCC 777B  1F03 F3AB F9F8 7D39 509A

--CUfgB8w4ZwR/yMy5
Content-Type: application/pgp-signature
Content-Disposition: inline

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.1 (FreeBSD)

iD8DBQE9zIJo86v5+H05UJoRAj6NAKC8EB4iQ/SzHVzvAW1q8ikwfScMpQCfWePo
f2OAlcN4aTii6H3vbGxjKCw=
=wIKO
-----END PGP SIGNATURE-----

--CUfgB8w4ZwR/yMy5--

To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-current" in the body of the message




Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20021109033506.GB6608>