Skip site navigation (1)Skip section navigation (2)
Date:      Tue, 8 Apr 2014 13:41:33 -0400
From:      "illoai@gmail.com" <illoai@gmail.com>
To:        ml-freebsd-stable <freebsd-stable@freebsd.org>
Subject:   Re: GEOM_CONCAT odd behaviour
Message-ID:  <CAHHBGkpBhUjXvYC9ejNJW8MAEBdpP1XM1-g-U78RspOcB-cXoQ@mail.gmail.com>
In-Reply-To: <20140408123250.GA34915@potato.growveg.org>
References:  <20140408123250.GA34915@potato.growveg.org>

next in thread | previous in thread | raw e-mail | index | archive | help
On 8 April 2014 08:32, John <freebsd-lists@potato.growveg.org> wrote:
> Hello -stable,
>
> I made a concatenated disk with the gconcat program and compiled
> options     GEOM_CONCAT
> into the kernel. When I reboot, this appears in the dmesg:
>
> GEOM_CONCAT: Device data created (id=1971778810).
> GEOM_CONCAT: Disk diskid/DISK-WD-WCC130881133 attached to data.
> GEOM_CONCAT: Disk diskid/DISK-WD-WCC131034794 attached to data.
> GEOM_CONCAT: Disk diskid/DISK-WD-WCC131035892 attached to data.
> GEOM_CONCAT: Device concat/data activated.
> GEOM_CONCAT: Cannot add disk ada3 to data (error=17).
> GEOM_CONCAT: Cannot add disk ada2 to data (error=17).
> GEOM_CONCAT: Cannot add disk ada1 to data (error=17).
>
> I tried to run gconcat status which returned something like
> 'unavailable' (forget the exact error but it suggested the driver wasn't
> loaded) so I loaded the kernel module instead ( geom_concat.ko )
> and now status returns:
>
> [root@ ~]# gconcat status
> Name  Status  Components
> concat/data      UP  diskid/DISK-WD-WCC130881133
>                      diskid/DISK-WD-WCC131034794
>                      diskid/DISK-WD-WCC131035892
>
> ...so why does it work the second way and not the first? Is compiling it
> into the kernel depreciated? This is on PC-BSD 10 but I don't think this
> is specific to that branch. The version is 10.0-RELEASE-p4.

AFIK, geom_* stuff works by magic: the module "tastes"
the last sectors of the disk, & then tries its damnedest to
attach it to an appropriate thingy (concat, stripe, label,
whatnot) based on the metadata.

It looks like it's finding the geom metadata twice & trying
to attach the same disk(s) to the geom_concat twice, &
failing the second time.

-- 
--



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?CAHHBGkpBhUjXvYC9ejNJW8MAEBdpP1XM1-g-U78RspOcB-cXoQ>