From owner-freebsd-current@FreeBSD.ORG Mon Aug 2 15:35:58 2004 Return-Path: Delivered-To: freebsd-current@freebsd.org Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125]) by hub.freebsd.org (Postfix) with ESMTP id 5E39C16A4CE; Mon, 2 Aug 2004 15:35:58 +0000 (GMT) Received: from gromit.dlib.vt.edu (gromit.dlib.vt.edu [128.173.49.29]) by mx1.FreeBSD.org (Postfix) with ESMTP id F1E6743D68; Mon, 2 Aug 2004 15:35:57 +0000 (GMT) (envelope-from paul@gromit.dlib.vt.edu) Received: from hawkwind.Chelsea-Ct.Org (pool-151-199-91-61.roa.east.verizon.net [151.199.91.61]) by gromit.dlib.vt.edu (8.12.11/8.12.11) with ESMTP id i72FZs4P053847 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=OK); Mon, 2 Aug 2004 11:35:55 -0400 (EDT) (envelope-from paul@gromit.dlib.vt.edu) Received: from [192.168.1.25] (zappa [192.168.1.25])i72FZmBW006347; Mon, 2 Aug 2004 11:35:48 -0400 (EDT) From: Paul Mather To: Lukas Ertl In-Reply-To: <20040729215647.F625@korben.in.tern> References: <20040726191107.B10DD16A513@hub.freebsd.org> <1091130224.99074.55.camel@zappa.Chelsea-Ct.Org> <20040729215647.F625@korben.in.tern> Content-Type: text/plain Message-Id: <1091460946.46029.22.camel@zappa.Chelsea-Ct.Org> Mime-Version: 1.0 X-Mailer: Ximian Evolution 1.4.6 Date: Mon, 02 Aug 2004 11:35:47 -0400 Content-Transfer-Encoding: 7bit cc: freebsd-current@FreeBSD.org Subject: Re: Vinum status X-BeenThere: freebsd-current@freebsd.org X-Mailman-Version: 2.1.1 Precedence: list List-Id: Discussions about the use of FreeBSD-current List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Mon, 02 Aug 2004 15:35:58 -0000 On Thu, 2004-07-29 at 15:57, Lukas Ertl wrote: > On Thu, 29 Jul 2004, Paul Mather wrote: > > > Does this imply that root-on-gvinum is working now? Last time I tried > > it (shortly after gvinum was committed to the tree), it worked except > > that only half of the drives, plexes, and volumes in my Vinum RAID 1 > > mirror were detected when the gvinum module read the configuration from > > the drives during boot. > > I still didn't have a chance to build a root-on-gvinum setup, but I'd be > happy if you could test it. It's quite likely that the errors were > related. I rebuilt my kernel+world as part of the GCC 3.4.2 import, and re-attached my two root-on-vinum test drives. I dumped over the current system to the vinum volumes and edited /etc/fstab for an all-vinum setup. I'm happy to report that the current geom_vinum kernel module successfully detects all plexes during boot when loaded from /boot/loader.conf. So, I now have a working root-on-gvinum setup running (including swap-on-gvinum). Thanks!! My only question is that although writes go to both disks in my mirrored setup, reads are only done from one disk of the mirror. In my case, all reads are always to drive ad2 of my ad0+ad2 mirror. This implies to me a default "readpol prefer " on my defined volumes rather than the "readpol round" that appears to be the default under vinum. (At least on my 4.10-STABLE system, both drives are used for reads, implying a "readpol round" default behaviour.) Does gvinum support "readpol round" or is it just that its default differs from the old vinum? If the latter, is there a way to change the read policy on a volume without deleting the old volume or otherwise destroying its contents? The vinum configuration I'm using is an exact duplicate of a 5.1-RELEASE system I created. That system (since upgraded to 5.2.1-RELEASE-p9) also sends reads to both drives using vinum. Again, great job on the gvinum work! (Will you link the kernel module into the normal kernel build? I forgot it wasn't the first time I built my kernel, and had to go back and build manually the geom_vinum module when I couldn't find it to load when at the boot loader prompt. Ditto for including gvinum in the regular userland build process.) Cheers, Paul. -- e-mail: paul@gromit.dlib.vt.edu "Without music to decorate it, time is just a bunch of boring production deadlines or dates by which bills must be paid." --- Frank Vincent Zappa