Date: Mon, 8 Nov 2004 14:06:49 +0100 From: Michael Riexinger <michael.riexinger@de.clara.net> To: Pawel Jakub Dawidek <pjd@freebsd.org> Cc: freebsd-geom@freebsd.org Subject: Re: gmirror problems Message-ID: <200411081406.50208.michael.riexinger@de.clara.net> In-Reply-To: <20041108115506.GB8120@darkness.comp.waw.pl> References: <200411081140.57792.michael.riexinger@de.clara.net> <200411081247.39449.michael.riexinger@de.clara.net> <20041108115506.GB8120@darkness.comp.waw.pl>
next in thread | previous in thread | raw e-mail | index | archive | help
On Monday 08 November 2004 12:55, Pawel Jakub Dawidek wrote: > On Mon, Nov 08, 2004 at 12:47:38PM +0100, Michael Riexinger wrote: > +> On Monday 08 November 2004 12:05, Pawel Jakub Dawidek wrote: > +> > On Mon, Nov 08, 2004 at 11:40:57AM +0100, Michael Riexinger > wrote: +> > +> Hi, > +> > +> > +> > +> I set up FreeBSD 5.3 with gmirror for my 2 ide disks ad0 and > ad1. +> > So the +> rootfs is included in the mirror. I wanted to > test what +> > happens when +> one disk fails, so I removed ad1. > Unfortunately after +> > a reboot the +> system didn't boot anymore > with the error: +> > +> GEOM_MIRROR: Force device mirror0 start due > to timeout +> > +> and > +> > +> GEOM_MIRROR: Device mirror0 destroyed > +> > > +> > Could you add: > +> > > +> > kern.geom.mirror.debug=2 > +> > > +> > to your /boot/loader.conf and try to boot without ad1 again. > +> > It should start mirror with only one component (ad0) connected > after +> > this timeout and now it looks like it can't find any > component. +> Ok, done that: > +> > +> the messages are: > +> GEOM_MIRROR[2]: Tasting ad0. > +> magic: GEOM::MIRROR > +> version: 1 > +> name: mirror0 > +> did: 1626481947 > +> all: 2 > +> syncid: 1 > +> priority: 0 > +> slice: 4096 > +> balance: load > +> mediasize: 20485785088 > +> vectorsize: 512 > +> syncoffset: 1297612800 > +> mflags: NONE > +> dflags: DIRTY SYNCHONIZING > > Ok, everything is clear now. You disconnected ad1, but ad0 was marked > for synchronization. It that case gmirror cannot find any valid > component, because the only one it can find is stale. > So: > 1. You should test it when mirror is in complete state (then you can > remove any component). Ok, thanks, done that. Now both ad0 and ad1 are in dirty state. What exactly means dirty? I found nothing about it in the manpage. Kind regards, Michael Riexinger systems engineer -- claranet gmbh internet service provider tel +49 (0) 69 - 40 80 18 - 300 email: michael.riexinger@de.clara.net http://www.claranet.de/
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?200411081406.50208.michael.riexinger>