Skip site navigation (1)Skip section navigation (2)
Date:      Tue, 9 Aug 2005 14:59:39 -0700
From:      Brooks Davis <brooks@one-eyed-alien.net>
To:        Jeremie Le Hen <jeremie@le-hen.org>
Cc:        drvince@anonymnet.net, freebsd-current@freebsd.org
Subject:   Re: More into /etc/rc.d/jail
Message-ID:  <20050809215939.GA15004@odin.ac.hmc.edu>
In-Reply-To: <20050809214330.GZ45385@obiwan.tataz.chchile.org>
References:  <N1-uLBXxM-zn8@Safe-mail.net> <20050809214330.GZ45385@obiwan.tataz.chchile.org>

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

--wRRV7LY7NUeQGEoC
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
Content-Transfer-Encoding: quoted-printable

On Tue, Aug 09, 2005 at 11:43:30PM +0200, Jeremie Le Hen wrote:
> Hi,
>=20
> > Hi,
> > I'm using jails inside md devices to limit the disk space each jail can=
 use.  It's working great but I have to start manually all of them at start=
up.  Here's my drill:
> >=20
> > mdconfig -a -t vnode -f ${IMAGE} -u ${ID}
> > fsck_ufs /dev/md${ID}c
> > mount /dev/md${ID}c ${DEST}
> > mount_devfs devfs ${DEST}/dev
> > jail -l -U root ${DEST} ${FQDN} ${IP} /bin/sh /etc/rc
> >=20
> > Therefore, I can't use the /etc/rc.d/jail facility.  So I thought, it w=
ould be good to add *fsck before mounting* and an optional mdconfig beforeh=
and.
> >=20
> > jail_${NAME}_md_device=3D""    # The device to attach or NO
> > jail_${NAME}_image=3D""        # The image file containing the jail, us=
ed with md_device
> > jail_${NAME}_fsck_options=3D"" # Options to pass to fsck
> >=20
> > In fsck_options I could put "-t ufs".  Of course, /dev/md${ID}c must be=
 present before mounting, could happen if the image isn't bsdlabel'ed.
> >=20
> > I'm a terrible coder, I could do it, but I'll need coaching and I've ne=
ver made a patch.  I would gladly hand that to someone else but I also need=
 it to be done, I can't babysit the server forever.  So, how does it sound?
>=20
> This was recently discussed in some thread.  The problem with this
> approach is that file backed md(4) devices are too slow to make it
> standard.  This is why this won't go into FreeBSD, IMHO.

Nonsense.  Integrating support for mounting file backed MD file systems
at boot would be quite logical even if md(4) devices are slow.  In a
large portion of cases, the overhead isn't relevent compared to the
administrative freedom granted by using files instead of partitions.

-- Brooks

--=20
Any statement of the form "X is the one, true Y" is FALSE.
PGP fingerprint 655D 519C 26A7 82E7 2529  9BF0 5D8E 8BE9 F238 1AD4

--wRRV7LY7NUeQGEoC
Content-Type: application/pgp-signature
Content-Disposition: inline

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.1 (GNU/Linux)

iD8DBQFC+SdKXY6L6fI4GtQRAgUSAKDE5OFzG3iRSTfCRsVMbQBrMYduagCgvJyS
hlBwTXd6QUoUDmuHKsuRSPE=
=XZGc
-----END PGP SIGNATURE-----

--wRRV7LY7NUeQGEoC--



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20050809215939.GA15004>