From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from lists.gentoo.org (pigeon.gentoo.org [208.92.234.80]) by finch.gentoo.org (Postfix) with ESMTP id ED3D71381F3 for ; Sun, 20 Oct 2013 14:31:22 +0000 (UTC) Received: from pigeon.gentoo.org (localhost [127.0.0.1]) by pigeon.gentoo.org (Postfix) with SMTP id 74FBDE0CA1; Sun, 20 Oct 2013 14:31:16 +0000 (UTC) Received: from smtpq2.tb.mail.iss.as9143.net (smtpq2.tb.mail.iss.as9143.net [212.54.42.165]) by pigeon.gentoo.org (Postfix) with ESMTP id 57BCAE0C5E for ; Sun, 20 Oct 2013 14:31:15 +0000 (UTC) Received: from [212.54.42.135] (helo=smtp4.tb.mail.iss.as9143.net) by smtpq2.tb.mail.iss.as9143.net with esmtp (Exim 4.71) (envelope-from ) id 1VXu2E-0004X9-PE for gentoo-user@lists.gentoo.org; Sun, 20 Oct 2013 16:31:14 +0200 Received: from 54698b76.cm-12-2c.dynamic.ziggo.nl ([84.105.139.118] helo=data.antarean.org) by smtp4.tb.mail.iss.as9143.net with esmtp (Exim 4.71) (envelope-from ) id 1VXu2E-0003XK-6v for gentoo-user@lists.gentoo.org; Sun, 20 Oct 2013 16:31:14 +0200 Received: from [10.20.13.57] (unknown [10.20.13.57]) (using TLSv1 with cipher RC4-MD5 (128/128 bits)) (No client certificate requested) by data.antarean.org (Postfix) with ESMTPSA id 67B584C for ; Sun, 20 Oct 2013 16:30:38 +0200 (CEST) User-Agent: K-9 Mail for Android In-Reply-To: <5263DB18.6020409@hadt.biz> References: <201310150834.21103.michaelkintzios@gmail.com> <201310201054.45575.michaelkintzios@gmail.com> <5263D33E.2050509@hadt.biz> <201310201413.59198.michaelkintzios@gmail.com> <5263DB18.6020409@hadt.biz> Precedence: bulk List-Post: List-Help: List-Unsubscribe: List-Subscribe: List-Id: Gentoo Linux mail X-BeenThere: gentoo-user@lists.gentoo.org Reply-to: gentoo-user@lists.gentoo.org MIME-Version: 1.0 Content-Type: multipart/alternative; boundary="----R3LBTRD8B8CZDSB7B0HH88GU9745N3" Subject: Re: [gentoo-user] Re: [O/T] RAID help - now won't boot From: joost@antarean.org Date: Sun, 20 Oct 2013 16:31:12 +0200 To: gentoo-user@lists.gentoo.org Message-ID: <30a1c0c8-79ea-4606-9961-c4c9e95a6d91@email.android.com> X-Ziggo-spambar: ---- X-Ziggo-spamscore: -4.8 X-Ziggo-spamreport: ALL_TRUSTED=-1,BAYES_00=-1.9,HTML_MESSAGE=0.001,NO_REAL_NAME=0.1,PROLO_TRUST_RDNS=-3,RDNS_DYNAMIC=0.982 X-Ziggo-Spam-Status: No X-Spam-Status: No X-Spam-Flag: No X-Archives-Salt: 3926dcdc-9c5d-456d-9c55-6b94923349d2 X-Archives-Hash: 575fa9d342b80d997728670894067d04 ------R3LBTRD8B8CZDSB7B0HH88GU9745N3 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: quoted-printable Michael Hampicke wrote: >Am 20.10.2013 15:13, schrieb Mick: >> On Sunday 20 Oct 2013 13:57:34 Michael Hampicke wrote: >>> Am 20.10.2013 11:54, schrieb Mick: >>>> Any ideas why the Ubuntu installation won't boot? >>> >>> My guess would be, you cannot boot, because if you install grub in >>> /dev/md0. >>> >>> Upon boot the bios cannot find stage1 of the bootloader, which >normally >>> lies in the MBR (which also houses the partition table). >>=20 >> I see ... so installing the MBR code in the /dev/md0 block device is >further=20 >> down the disk than where BIOS is looking for it and that's why it >errors out? >>=20 > >That would be my guess. Maybe someone more knowledgeable on how mdadm >writes stuff on the disk can jump in and provide additional info. But >I'm pretty sure, if you install grub in md0, it's not in that place on >the disk where the bios is actually looking for. > >>=20 >> It seems to me then that I *have* to create normal partitions on >/dev/sda &=20 >> /dev/sdb, or I would need a different boot drive. Is there another >way to=20 >> overcome this problem.=20 > >Maybe create two mds. md1 (sda1, sdb1) is a small boot partition which >contains stage2+, the kernel and the initramfs. And md2 (sda2, sdb2) >which acts as another block device with partition table, etc... >In this setup you could install grub in the mbr of sda and sdb >(grub-install /dev/sda...) > >A quick google on this subject returned no usable results. But I am off >now until tomorrow. I would suggest trying it by usong the older metadata format. Check the man pages, but I thinl it would be --metadata=3D0.90 (or simila= r) during creation. That might put the metadata at the end, rather then at the front. (Or it'= s the other way round and new metadata does it at the end.) -- Joost Ps. I have never tried it this way (full disk raid for boot device) using= linux software raid. --=20 Sent from my Android device with K-9 Mail. Please excuse my brevity. ------R3LBTRD8B8CZDSB7B0HH88GU9745N3 Content-Type: text/html; charset=utf-8 Content-Transfer-Encoding: quoted-printable
Michael Hampicke <= mh@hadt.biz> wrote:
Am 20.10.2013 15:13, schrieb Mick:
On Sunday 20 Oct 2013 13:57:34 Mich=
ael Hampicke wrote:
Am 20.10.2013 11:54, schrieb Mick:
Any ideas why the Ubuntu installation won't boot?
My guess would be, you cannot boot, because if you install gr=
ub in
/dev/md0.

Upon boot the bios cannot find stage1 of =
the bootloader, which normally
lies in the MBR (which also houses th=
e partition table).

I see ... so installing the MBR cod=
e in the /dev/md0 block device is further 
down the disk than where =
BIOS is looking for it and that's why it errors
out?


That would be my guess. Maybe someone more k=
nowledgeable on how mdadm
writes stuff on the disk can jump in and p=
rovide additional info. But
I'm pretty sure, if you install grub in =
md0, it's not in that place on
the disk where the bios is actually l=
ooking for.


It seems to me then that I *have* to create normal partitions on /de=
v/sda & 
/dev/sdb, or I would need a different boot drive.  Is t=
here another way to 
overcome this problem. 

Maybe=
 create two mds. md1 (sda1, sdb1) is a small boot partition which
co=
ntains stage2+, the kernel and the initramfs. And md2 (sda2, sdb2)
w=
hich acts as another block device with partition table, etc...
In th=
is setup you could install grub in the mbr of sda and sdb
(grub-inst=
all /dev/sda...)

A quick google on this
subject returned no usable results. But I am off
now until tomorrow.=


I would suggest tr= ying it by usong the older metadata format.
Check the man pages, but I thinl it would be --metadata=3D0.90 (or simila= r) during creation.
That might put the metadata at the end, rather then at the front. (Or it&= #39;s the other way round and new metadata does it at the end.)

--
Joost
Ps. I have never tried it this way (full disk raid for boot device) using= linux software raid.
--
Sent from my Android device with K-9 Mail. Please excuse my brevity. ------R3LBTRD8B8CZDSB7B0HH88GU9745N3--