Lucid live-cd broke raid 0 disk array with intel ich10r controller

[WiLLiTo]

Binary package hint: dmraid

Lucid Lynx live cd brokes raids made with intel ich10r controller

ubuntu@ubuntu:~$ lsb_release -rd
Description: Ubuntu 10.04 LTS
Release: 10.04

ubuntu@ubuntu:~$ apt-cache policy dmraid
dmraid:
  Instalados: 1.0.0.rc16-3ubuntu2
  Candidato: 1.0.0.rc16-3ubuntu2
  Tabla de versión:
 *** 1.0.0.rc16-3ubuntu2 0
        500 http://archive.ubuntu.com/ubuntu/ lucid/main Packages
        100 /var/lib/dpkg/status
ubuntu@ubuntu:~$

ubuntu@ubuntu:~$ sudo dmraid -ay
ERROR: isw: wrong number of devices in RAID set "isw_hfigcaijg_linux" [1/2] on /dev/sda
ERROR: isw: wrong number of devices in RAID set "isw_hfigcaijg_win" [1/2] on /dev/sda
RAID set "isw_hfigcaijg_linux" was not activated
RAID set "isw_hfigcaijg_win" was not activated
ubuntu@ubuntu:~$

The disk utility shows 1 disk with raid flag but the other as a normal disk.

After reboot bios controller shows that one of the raid members now it isn't and can't boot the instaled O.S.

If you need any console output or system log tell me

Sorry for my English

ProblemType: Bug
DistroRelease: Ubuntu 10.04
Package: dmraid 1.0.0.rc16-3ubuntu2
ProcVersionSignature: Ubuntu 2.6.32-21.32-generic 2.6.32.11+drm33.2
Uname: Linux 2.6.32-21-generic x86_64
Architecture: amd64
Date: Fri Apr 30 02:13:46 2010
LiveMediaBuild: Ubuntu 10.04 LTS "Lucid Lynx" - Release amd64 (20100429)
ProcEnviron:
 LANG=es_ES.UTF-8
 SHELL=/bin/bash
SourcePackage: dmraid

[WiLLiTo (victor-gg83)]

After 2 cold reboots (shutdown-wait-turn on) raid is fixed by itself and my other O.S load but if i try to start with the live cd same thing happen again.

Any idea?

[Danny Wood (danwood76)]

This is caused by the fact that you have 2 RAID0 volumes on the same disk set.
I think this has cropped up before but I can't remember the bug.

The normal thing to do is to have an entire RAID0 set and then partition the disk to create your separate portions for your different OS.
If you were to create a single RAID0 set and partition that set it would work fine, I think this is what the other guy did in the other bug.

Did this setup work in previous Linux installs?

[Phillip Susi (psusi)]

The failure to activate the array is because of a bug in dmraid that prevents it from handling arrays with multiple logical raid volumes in them. That does not explain why it causes the disk to be offline after a reboot.

I assume you did a warm boot? After running the livecd, can you try shutting down, and pulling all power from the machine for a minute, then doing a cold boot?

If you only get this problem after a warm boot, then it could be caused by use of the host protected area on the disk. Check your kernel log messages ( /var/log/kern.log ) for mention of the HPA.

[WiLLiTo (victor-gg83)]

Sorry I was away from home at weekend . I'll try it tomorrow. My second option is to backup my data and create one unique raid volume and try it.

Thanks for help. See you tomorrow with my test results.

[WiLLiTo (victor-gg83)]

About the offline disk is the same problem that in Jaunty and Karmic with one or two raid volumes. After reboot always show one of the disk as a offline member.

As you said raid fix itself at shutdown, but i need to shutdown 2 times. First time the raid config is OK, but no OS is detected to load, at second shutdown system is OK and load perfectly.

About the kernel log i found only 1 match to HPA:

May 3 11:28:40 ubuntu kernel: [ 4.069854] ata3.00: HPA unlocked: 586070255 -> 586072368, native 586072368

Attached the log

[WiLLiTo (victor-gg83)]

reply to danwood76

raid arrays with this controller never worked for me (with 1 or 2 volumes). In previous releases Ubuntu mounted the array but when i open gparted to see partitions it drop a message like "Partition on disk "X" is bigger than the disk. Partitions cant exceed disk space" And it shows 2 separate disk (sda and sdb) and the disk array (isw_XXXXXX).

There was possible to write to the disk but only one of the 2 disk are really used (first write to sda and when it runs out os space continue writing to sdb) and install system failed on grub.

With previous releases the "Non raid disk" problem at bios occurred too.

[Phillip Susi (psusi)]

Bingo. For some reason your disk is configured to hide part of itself. The kernel disables this feature, which confuses your bios since the location of the raid signature appears to have changed relative to the end of the disk. Marking as duplicate of #380138.