Date: Sat, 29 Apr 2006 16:55:25 +0200 From: Pawel Jakub Dawidek <pjd@FreeBSD.org> To: Fernan Aguero <fernan@iib.unsam.edu.ar> Cc: freebsd-geom@freebsd.org Subject: Re: panic when loading mirror Message-ID: <20060429145525.GJ4896@garage.freebsd.pl> In-Reply-To: <20060428171226.GG62606@iib.unsam.edu.ar> References: <20060428171226.GG62606@iib.unsam.edu.ar>
next in thread | previous in thread | raw e-mail | index | archive | help
[-- Attachment #1 --] On Fri, Apr 28, 2006 at 02:12:26PM -0300, Fernan Aguero wrote: +> Hi, +> +> I'm playing with gmirror on a box with a pair of spare +> slices before installing a production box. +> +> The box is running 6.1-RC (i386), and has two identical +> disks (ad0, ad2) already sliced and partitioned: +> +> ad0: +> ad0s1 unused slice +> ad0s2 ad0s2b (swap), ad0s2a (/), ad0s2d (/var), ad0s2e (/usr) ... +> +> ad2: +> ad2s1 unused slice +> ad2s2 ad2s2b (swap), ad2s2d (/tmp), ad2s2e (/usr/ports) ... +> +> My idea was to use the unused slices (identical in size) to +> set up a geom mirror. +> +> After reading the man pages and the following guides +> http://www.onlamp.com/pub/a/bsd/2005/11/10/FreeBSD_Basics.html?page=1 +> http://people.freebsd.org/~rse/mirror/ +> this is what I did: +> +> # gmirror label -v -b round-robin gm0s1 ad0s1 ad2s1 +> Metadata value stored on ad0s1. +> Metadata value stored on ad2s1. +> Done. +> +> But then attempting to load the mirror, causes a system +> panic. I'm copying the messages by hand, so excuse any typo. +> Also, I'm copying what I see on the screen ... I can't know +> if there were other errors before the ones I'm transcribing +> (capturing the session with script(1) doesn't help here). +> +> # gmirror load +> WARNING: Expected rawoffset 0, found 63 +> WARNING: Device name truncated! (mirror/gsm0s1ccccccccccccc...c) +> [ this is repeated several times. gsm0s1 is followed above +> by 51 'c's) ] +> +> Fatal double fault: +> eip = 0xc04a7982 +> esp = 0xe2244ff4 +> ebp = 0xe2245038 +> panic: double fault +> +> +> My questions: +> +> did I do something wrong? (maybe the system shouldn't let me +> instead of entering into a panic ...) +> +> Was this caused because I used already sliced and +> partitioned disks? The second example in the guide written +> by rse uses slices ... You should configure both slices and partitions on mirrored provider. To fix it you need to clear the old metadata first. +> Or is this something that is expected to happen and should +> be fixed? This shouldn't act that way and panic your box for sure... -- Pawel Jakub Dawidek http://www.wheel.pl pjd@FreeBSD.org http://www.FreeBSD.org FreeBSD committer Am I Evil? Yes, I Am! [-- Attachment #2 --] -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.2 (FreeBSD) iD8DBQFEU35dForvXbEpPzQRAvmJAKDS2VokgjOAWc+JAQqNlOvg5CZcTwCghsNM KwcZrXdm/nX5RMRqFhcuLL8= =C86k -----END PGP SIGNATURE-----
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20060429145525.GJ4896>