From: Guillaume Filion <gfk@logidac.com>
To: linux-raid@vger.kernel.org
Subject: Re: Crooked raid
Date: Wed, 16 Nov 2005 11:39:15 -0500 [thread overview]
Message-ID: <437B60B3.5030601@logidac.com> (raw)
In-Reply-To: <200511161541.jAGFfBjZ027253@cichlid.com>
Andrew Burgess a écrit :
>>I did this, rebooted and the system wouldn't reboot. Yikes! I was
>>however able to boot with giving root=/dev/hdc2 to the kernel.
>
> I didn't realize we were talking about your root filesystem else I would have
> been more cautious (or maybe just kept my mouth shut). Glad you got it to boot!
Hey no problem there. I prefer having someone trying to help me and
having problems than getting no help at all.
> Do you recall why it didn't want to boot?
I don't remember the exact errors, I you need them, I can get them
tonight when I get back home. From memory, I got the first error very
early in the boot process -- before trying to load any md stuff -- the
error was something like:
Unable to boot from "0900"
Then I tried with an older kernel. This older kernel is bigger and I
suspect that it has the raid stuff compiled in. With this kernel, md
loaded and then I got this error:
Can't find Superblock on disk..
I was finally able to boot with the older kernel by specifying boot=/dev/hd2
It looks to me that it's still trying to mount md0 from hdc rather than
hdc2.
> What are the kernel command line args?
I don't think there's any. append= is commented out in lilo.conf
> Did you check for a bootup script somewhere screwing things up?
> egrep -i 'raid|mdadm' /etc/rc.d/* /etc/rc.d/init.d/* /etc/*
There's /etc/init.d/mdadm-raid that starts "/sbin/mdadm -A -s -a" if
/etc/mdadm/mdadm.conf exists -- and it does, see below.
There's /etc/init.d/raid2 which would start raids from /etc/raidtab but
there's no raidtab and /etc/default/raid2 says to disable this.
Other than that I have this in /etc/modules.conf :
root@ali:~# egrep -i 'raid|mdadm' /etc/modules.conf
### update-modules: start processing /etc/modutils/raidtools2
alias md-personality-2 raid0
alias md-personality-3 raid1
alias md-personality-4 raid5
### update-modules: end processing /etc/modutils/raidtools2
> And did you already say that you don't have a mdadm.conf file?
I have a /etc/mdadm/mdadm.conf file:
DEVICE /dev/hdc2 /dev/hdg2
ARRAY /dev/md0 level=raid1 num-devices=2
UUID=b013e39b:ec629293:98df4657:97255939
> You might need to pick which mirror (hdc2 or hdg2) you trust more as your root
> filesystem (since they may be different now) and then start over and follow the
> HowTo for making a normal root filesystem into a raided one.
>
> But if you can't find out why the system is still looking at hdc then it might
> all happen again...
>
>
>>Here's the relevant part of dmesg:
>>device-mapper: 4.1.1-ioctl (2004-04-07) initialised: dm-devel@redhat.com
>>md: can not import ide/host0/bus1/target0/lun0/part2, has active inodes!
>
> I guess this is because you booted from it as hdc2 so its busy and mdadm
> won't try to use it?
Yeah, that makes sense.
> Do you know why the system sometimes says ide/host0/bus1/target0/lun0/part2 and
> sometimes says hd[cg]2 ? Its confusing...
No, that's something that confuses me too. I installed devfsd sometime
in the past but deinstalled it because I didn't need it.
> Any other raid superblocks around?
> for dev in /dev/hd? /dev/hd??; do mdadm -E $dev; done
root@ali:~# for dev in /dev/hd? /dev/hd??; do mdadm -E $dev; done 2>&1 |
egrep -v 'cannot|small'
mdadm: No super block found on /dev/hdc (Expected magic a92b4efc, got
00000000)
mdadm: No super block found on /dev/hde (Expected magic a92b4efc, got
69686766)
mdadm: No super block found on /dev/hdg (Expected magic a92b4efc, got
00000000)
mdadm: No super block found on /dev/hdc1 (Expected magic a92b4efc, got
ffffffff)
/dev/hdc2:
Magic : a92b4efc
Version : 00.90.00
UUID : b013e39b:ec629293:98df4657:97255939
Creation Time : Wed Dec 29 21:32:26 2004
Raid Level : raid1
Raid Devices : 2
Total Devices : 3
Preferred Minor : 0
Update Time : Tue Nov 15 16:01:25 2005
State : clean
Active Devices : 2
Working Devices : 2
Failed Devices : 1
Spare Devices : 0
Checksum : ff92c98b - correct
Events : 0.103
Number Major Minor RaidDevice State
this 1 22 2 1 active sync /dev/hdc2
0 0 34 2 0 active sync /dev/hdg2
1 1 22 2 1 active sync /dev/hdc2
mdadm: No super block found on /dev/hde1 (Expected magic a92b4efc, got
00000000)
mdadm: No super block found on /dev/hdg1 (Expected magic a92b4efc, got
c8938b73)
/dev/hdg2:
Magic : a92b4efc
Version : 00.90.00
UUID : b013e39b:ec629293:98df4657:97255939
Creation Time : Wed Dec 29 21:32:26 2004
Raid Level : raid1
Raid Devices : 2
Total Devices : 2
Preferred Minor : 0
Update Time : Tue Nov 15 17:29:57 2005
State : active
Active Devices : 1
Working Devices : 1
Failed Devices : 1
Spare Devices : 0
Checksum : ff92de3d - correct
Events : 0.104
Number Major Minor RaidDevice State
this 0 34 2 0 active sync /dev/hdg2
0 0 34 2 0 active sync /dev/hdg2
1 1 0 0 1 faulty removed
> And any raid autodetect partitions around?
> fdisk -l | grep raid
fdisk -l doesn't output anything.
One thing that might be a clue about the problem is the warnings that I
get when I run lilo -v:
root@ali:~# lilo -v
LILO version 22.6.1, Copyright (C) 1992-1998 Werner Almesberger
Development beyond version 21 Copyright (C) 1999-2004 John Coffman
Released 17-Nov-2004, and compiled at 20:01:15 on Sep 29 2005
Debian GNU/Linux
Reading boot sector from /dev/hde
Warning: '/proc/partitions' does not match '/dev' directory structure.
Name change: '/dev/ide/host2/bus1/target0/lun0/disc' -> '/dev/hdg'
Name change: '/dev/ide/host2/bus1/target0/lun0/part1' -> '/dev/hdg1'
Warning: Kernel & BIOS return differing head/sector geometries for
device 0x81
Kernel: 23989 cylinders, 16 heads, 63 sectors
BIOS: 1024 cylinders, 255 heads, 63 sectors
Name change: '/dev/ide/host2/bus1/target0/lun0/part2' -> '/dev/hdg2'
Name change: '/dev/ide/host2/bus0/target0/lun0/disc' -> '/dev/hde'
Name change: '/dev/ide/host2/bus0/target0/lun0/part1' -> '/dev/hde1'
Name change: '/dev/ide/host0/bus1/target0/lun0/disc' -> '/dev/hdc'
Name change: '/dev/ide/host0/bus1/target0/lun0/part1' -> '/dev/hdc1'
Name change: '/dev/ide/host0/bus1/target0/lun0/part2' -> '/dev/hdc2'
Name change: '/dev/md/0' -> '/dev/md0'
/boot/boot.1600 exists - no master disk volume ID record backup copy made.
Backup copy of master disk volume ID record in /boot/boot.2200
...
In case you need to see it, /proc/partitions looks like this:
root@ali:~# cat /proc/partitions
major minor #blocks name rio rmerge rsect ruse wio wmerge wsect
wuse running use aveq
34 0 78150744 ide/host2/bus1/target0/lun0/disc 38 240 688 190 3
0 12 0 -166 22379007 12182291
34 1 642568 ide/host2/bus1/target0/lun0/part1 9 12 168 60 0 0
0 0 0 60 60
34 2 77508144 ide/host2/bus1/target0/lun0/part2 14 127 288 80 1
0 8 0 0 80 80
33 0 251392 ide/host2/bus0/target0/lun0/disc 400 713 2226 760
4330 3059 14778 146790 0 128510 147550
33 1 251632 ide/host2/bus0/target0/lun0/part1 387 626 2026
720 4328 3059 14774 146710 0 128390 147430
22 0 78150744 ide/host0/bus1/target0/lun0/disc 183212 108424
2332636 2782770 503378 1212603 13756060 27924377 -167 22375017 20476641
22 1 642568 ide/host0/bus1/target0/lun0/part1 62 104 1328 400
287 2195 21104 13580 0 2430 13980
22 2 77508144 ide/host0/bus1/target0/lun0/part2 183135 108219
2331076 2782170 503089 1210408 13734952 27910797 0 2151990 30704027
9 0 77508032 md/0 0 0 0 0 0 0 0 0 0 0 0
Thanks,
GFK's
--
Guillaume Filion, ing. jr
Logidac Tech., Beaumont, Québec, Canada - http://logidac.com/
PGP Key and more: http://guillaume.filion.org/
-
To unsubscribe from this list: send the line "unsubscribe linux-raid" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
next prev parent reply other threads:[~2005-11-16 16:39 UTC|newest]
Thread overview: 10+ messages / expand[flat|nested] mbox.gz Atom feed top
2005-11-16 15:41 Crooked raid Andrew Burgess
2005-11-16 16:39 ` Guillaume Filion [this message]
2005-11-16 23:55 ` Guillaume Filion
2005-11-21 3:50 ` Crooked raid [solved] Guillaume Filion
-- strict thread matches above, loose matches on Subject: below --
2005-11-16 16:07 Crooked raid Andrew Burgess
2005-11-15 18:44 Andrew Burgess
2005-11-15 23:24 ` Guillaume Filion
2005-11-15 15:20 Guillaume Filion
2005-11-15 23:07 ` Neil Brown
2005-11-16 13:47 ` Guillaume Filion
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=437B60B3.5030601@logidac.com \
--to=gfk@logidac.com \
--cc=linux-raid@vger.kernel.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).