Bug#606416: Installation-report: No boot from raid 1 (raid1) partition on debian-squeeze-di-beta1 new install.
Package: installation-reports
TITLE: installation-report: No boot from raid 1 partition on debian-squeeze-di-beta1 new install.
My filename: issue-installation-report_No-boot-from-raid-partition-on-beta1-20101208.utf
Boot method: DVD install.
Image version: Full URL to image downloaded. Only the first DVD 1 of the set is required.
Date: 08 December 2010.
Machine: Motherboard: A bit KT7, Raid, 1 Ghz.
Processor: AMD AThlon K7, 850 Mhz, Socket 'A'
Memory: SDRAM 2 x 512Mb = 1Gb.
Chipset: Via KT 133 ATX, USB 1.0 compliant, UDMA 100, Socket 'A', FSB 200Mhz, AGP 4X.
Updated bios: Award Modular Bios v6.00 PG. 07/11/2002-8363-6A6A-6A6LMA19C-A9 The Latest level.
CMOS Set Up Utility: Advanced BIOS features:
First boot device: CDROM,
Second boot device: ATA100RAID,
Third boot device: ATA100RAID, Boot other device: disabled.
Delay for IDE Initial (Secs) 8.
Intergrated Peripherals: ATA100RAID IDE Controller: Enabled.
Two additional IDE channels provide the capability of adding high performance device(s) to system.
IDE HDD Block Mode: Enabled.
for automatic detection of the optimal number of block read/writes per sector the drive can support.
DASD chip: On-motherboard HPT370 raid controller 2001 www.highpoint-tech.com (hardware raid)
Not used by me for 'hardware raid'. The HPT370 is used by me, only for the fast disk control bus.
Linux Debian MDADM Software raid is set up in the partition manager during 'squeeze-di-Beta1' install.
Onboard Hipoint HPT370 UDMA/ATA100 RAID controller BIOS v01.11.0402
HPT370 BIOS Setting Utility:
The screen button F1 toggles the view ARRAY STATUS shows NO array or devices at this time. Good!
The screen button F1 toggles the view CHANNEL STATUS which shows the real hardware DASD:
DASD: Quantity two: IBM-DTLA-307030 drives, primary master and secondary master,
Mode UDMA5 size 30Gb each, as HDD0 and HDD1 both attached to the HPT370 controller. Good!
HPT370 Bios Setting Utility: Option 7. Set Boot Disk: I have set no disk as Boot Disk here,
as I believe the 'Linux Software Raid Array' is the place to specify /dev/md2 /boot as bootable in partition manager, which I have since set.
IDE Primary master: none.
IDE Primary slave: Pioneer DVD-ROM ATAP (I do not use this.)
IDE Secondary master: LITE-ON LTR-16102B DVD rewriter (I use this all the time.)
IDE Secondary slave: none.
Partitions: df -Tl
df: invalid option --'T'
BusyBox v1.17.1 (Debian 1:1.17.1-4) multi call binary.
usage: df [-Pkmh] [FILESYSTEM]...
df -h
Filesystem Size Used Available Use% Mounted on
tmpfs 505.4m 92.0k 505.3m 0% /dev
/dev/md1 26.6g 473.6m 24.8g 2% /target
/dev/md2 92.0M 17.5M 69.7M 20% /target/boot
/dev/md1 26.6G 473.6M 24.8G 2% /dev/.static/dev
tmpfs 505.4M 92.0K 505.3M 0% /target/dev
Output of lspci -nn
00:00.0 Host Bridge [0600] Via Technologies ... (rev 02)
00:01.0 PCI Bridge [0604] " " ... AGP
07.0 ISA bridge... " " Appollo Super South ... (rev 22)
07.1 IDE ... PIPC Bus Master IDE ... (rev 10)
07.2 USB Controller [0c03]: Via Technologies VT82xxxxx UHCI USB1.1 Controller ... (rev 10)
07.3 " " " " " " USB1.1 Controller ... (rev 10)
07.4 Bridge " " VT82c686 [Appollo Super ACPI] ... (rev 30)
11.0 Ethernet controller ...
00:13.0 Mass storage controller [0180]: HighPoint Technologies, Inc. HPT ... /370/ ... [1103:00041] (rev 03)
01:00.0 VGA compatible controller [0300]: ATI Technologies Inc 3D Rage Pro AGP 1X/2X [1002:47421] (rev 5c)
Base System Installation Checklist:
[O] = OK. [E] = Error
Initial boot: [O]
Detect network card: [O]
Configure network: [N/A]
Detect CD: [O]
Load installer modules: [O]
Detect hard drives: [O]
Partition hard drives: [O]
Install base system: [O] Configure Package Manager: Scanning DVD set disk, Binary-1 only. (yes)
Clock/timezone setup: [O]
User/password setup: [O]
Install tasks: [O]
Install boot loader: [O]
Overall install: [E] Won't auto boot into new installed system.
---
/var/log/syslog excerpt 2010 :-
Dec 8 11:39:18 kernel: [ 486.432151] md: md2 stopped.
Dec 8 11:39:18 kernel: [ 486... ] bind<hdg1>
Dec 8 11:39:18 kernel: [ 486... ] bind<hde1>
Dec 8 11:39:18 kernel: [ 486... ] raid1 personality registered for level 1
Dec 8 11:39:18 kernel: [ 486... ] raid1: raid set md2 active with 2 out of 2 mirrors
Dec 8 11:39:18 kernel: [ 486... ] md2: detected capacity change from 0 to 99602432
Dec 8 11:39:18 md-devices: mdadm: /dev/md/2 has been started with 2 drives.
Dec 8 11:39:18 kernel: [ 486... ] md2:
Dec 8 11:39:18 kernel: unknown partition table <<<=== *** Is this the start of my problems ?? ***
...
Dec 8 11:39:18 kernel: [ 486.432151] md: md0 stopped.
Dec 8 11:39:18 kernel: [ 486... ] bind<hdg2>
Dec 8 11:39:18 kernel: [ 486... ] bind<hde2>
Dec 8 11:39:18 kernel: [ 486... ] raid1 personality registered for level 1
Dec 8 11:39:18 kernel: [ 486... ] raid1: raid set md0 active with 2 out of 2 mirrors
Dec 8 11:39:18 kernel: [ 486... ] md0: detected capacity change from 0 to 99602432
Dec 8 11:39:18 md-devices: mdadm: /dev/md/0 has been started with 2 drives.
Dec 8 11:39:18 kernel: [ 486... ] md0:
Dec 8 11:39:18 kernel: unknown partition table
...
Dec 8 11:39:18 kernel: [ 486.432151] md: md1 stopped.
Dec 8 11:39:18 kernel: [ 486... ] bind<hdg5>
Dec 8 11:39:18 kernel: [ 486... ] bind<hde5>
Dec 8 11:39:18 kernel: [ 486... ] raid1 personality registered for level 1
Dec 8 11:39:18 kernel: [ 486... ] raid1: raid set md1 active with 2 out of 2 mirrors
Dec 8 11:39:18 kernel: [ 486... ] md1: detected capacity change from 0 to 99602432
Dec 8 11:39:18 md-devices: mdadm: /dev/md/1 has been started with 2 drives.
Dec 8 11:39:18 kernel: [ 487... ] md1:
Dec 8 11:39:18 kernel: unknown partition table
...
Dec 8 11:39:19 kernel: [ 487... ] device-mapper: uevent version 1.0.3
Dec 8 11:39:19 kernel: [ 487... ] device-mapper: ioctl: 4.15.0-ioctl (2009-04-01) initialised: dm-devel@redhat.com
Dec 8 11:39:19 kernel: partman: No physical volumes found
Dec 8 11:39:19 kernel: partman: Reading all physical volumes. This may take a while...
Dec 8 11:39:19 kernel: partman-lvm: No volume groups found
Dec 8 11:50:17 kernel: [ 1145... ] Adding 1560556k swap on /dev/md0. Priority:-1 across:1560556k
...
Dec 8 11:50:56 kernel: [ 1184... ] kjournald starting. Commit interval 5 seconds
Dec 8 11:50:56 kernel: [ 1184... ] EXT3 FS on md1, internal journal
Dec 8 11:50:56 kernel: [ 1184... ] EXT3-fs: mounted file system with ordered data mode.
...
Dec 8 11:50:56 kernel: [ 1184... ] kjournald starting. Commit interval 5 seconds
Dec 8 11:50:56 kernel: [ 1184... ] EXT3 FS on md2, internal journal
Dec 8 11:50:56 kernel: [ 1184... ] EXT3-fs: mounted file system with ordered data mode.
...
Dec 8 11:51:59 debootstrap: dpkg: regarding .../base-files_5.9_i386.deb containing base-files, pre-dependancy problem:
Dec 8 11:51:59 debootstrap: base-files pre-depends om awk
Dec 8 11:51:59 debootstrap: awk is not installed.
Dec 8 11:51:59 debootstrap: dpkg: warning: ignoring pre-dependancy problem!
Dec 8 11:51:59 debootstrap: Unpacking base-files (from .../base-files_5.9_i386.deb) ...
...
Dec 8 11:52:01 debootstrap: Selecting previously deselected package base-passwd.
Dec 8 11:52:01 debootstrap: Unpacking base-passwd (from .../base-passwd_3.5.22_i386.deb) ...
Dec 8 11:52:01 debootstrap: dpkg: base-passwd: dependancy problems. but configuring anyway as you requested:
Dec 8 11:52:01 debootstrap: base-passwd depends on libc6 (>= 2.1); however:
Dec 8 11:52:01 debootstrap: Package libc6 is not installed.
Dec 8 11:52:01 debootstrap: Setting up base-passwd (3.5.22) ...
...
Dec 8 12:49:59 os-prober: No volume groups found
Dec 8 12:49:59 os-prober:
Dec 8 12:49:59 os-prober: debug: running /user/lib/os-probes/50mounted-tests on /dev/hde1
/hde2
/hde3
Dec 8 12:49:59 50mounted-tests: debug: /dev/hde3 type not recognised; skipping (Note:This should be my /boot raid md2 )
Dec 8 12:49:59 os-prober: debug: os detected by /usr/lib/os-probes/50mounted-tests
Dec 8 12:49:59 os-prober: debug: running /user/lib/os-probes/50mounted-tests on /dev/hdg5
Dec 8 12:49:59 os-prober: debug: running /user/lib/os-probes/50mounted-tests on /dev/hdg1
Dec 8 12:49:59 os-prober: debug: running /user/lib/os-probes/50mounted-tests on /dev/hdg2
Dec 8 12:49:59 os-prober: debug: running /user/lib/os-probes/50mounted-tests on /dev/hdg3
Dec 8 12:49:59 50mounted-tests: debug: /dev/hdg3 type not recognised; skipping
Dec 8 12:49:59 os-prober: debug: os detected by /usr/lib/os-probes/50mounted-tests
Dec 8 12:49:59 os-prober: debug: running /user/lib/os-probes/50mounted-tests on /dev/hdg5
Dec 8 12:55:21 grub-installer: info: Installing grub on '/dev/hde' (This should be my raid set md master boot record)
Dec 8 12:55:21 grub-installer: info: grub-install supports --no-floppy
Dec 8 12:55:21 grub-installer: info: Running chroot /target grub-install --no-floppy --force "/dev/hde"
Dec 8 12:55:24 grub-installer: Installation finished. No error reported.
Dec 8 12:55:24 grub-installer: info: grub-install ran successfully
...
---
Comments/Problems:
I came across three squeeze-di-beta1 issues, and two wish list items.
I will report them in separate reports.
(Issue 3 of 3.)
Using graphical installer:
The CD or DVD with the following label has been scanned:
Debian GNU/Linux squeeze-di-beta1_Squeeze_-Official Beta i386 DVD Binary-1 20101027-01:46
If you wish to scan another CD or DVD, please insert it now.
Scan another CD or DVD? (No)
Trying to boot from a raid 1 (mirror) set.
I installed the GRUB boot loader to the master boot record.
Finish the installation : Installation complete, so it is time to boot into your new system...
The hardware post and bios progress ok, but when bios hands over to continue the ipl from a disk it just waits.
*** This is where the install fails. It does not boot up into the newly installed system. ***
---
I saw an ARMEL bug description below, followed by it's origin I believe, also below,
that I thought may have roots in my i386 scenario.
http://www.debian.org/devel/debian-installer/errata
"armel: root on RAID boot issue
Please be aware that there is a problem if you choose to have root on RAID that will cause your machine not to start after installing Debian.
Fortunately, there's a easy workaround: when the installer tells you that it's done and asks you whether you want to reboot, choose "Go back" and open a shell in the menu.
Then use nano to edit the file /target/usr/share/initramfs-tools/hooks/flash_kernel_set_root and remove the check that ensures the root device exists.
Type exit, choose "Make the system bootable" and press okay at the "Installation complete" message. ".
Alas, when I started to adapt the ARMEL architecture workaround to maybe my i386 situation, I got lost.
I fancied having a hack at trying to edit :
/bin/nano /target/usr/share/initramfs-tools/hooks/mdadm I may have basically chosen the wrong file.
I got lost in the code, having not been up that particular language learning curve, and not familiar with the 'Principals of Operation', (does such a document exist ?)
I would need some specific 'monkey see, monkey do' help to progress this further.
An excerpt from, what I believe to be the original source of the errata workaround for arm processors, not i386:
http://us.generation-nt.com/bug-558823-dev-disk-by-uuid-not-populated-i-raid-help-168909411.html
" Bug#558823: /dev/disk/by-uuid not populated in d-i with RAID
November 30th, 2009 ... by Martin Michlmayr
Here's a workaround for users who run into this issue: open another
SSH connection to the installer, open a shell and type:
ps aux | grep -i flash_kernel
then kill that PID. Then edit the file
/target/usr/share/initramfs-tools/hooks/flash_kernel_set_root
and remove the check that ensures the root device exists. The
installer will continue. Wait until it reaches
"Installation complete" and hit "Go Back". Then choose "Make
the system bootable" again.
Martin Michlmayr
http://www.cyrius.com/ ".
This is definately not a complaint, I am very very grateful for every ones work.
Regarding Martin's above instructions " remove the check that ensures the root device exists."
Which line are we talking about specifically ? I am looking in /target/usr/share/initramfs-tools/hooks/mdadm .
From the squeeze-di-beta1 Debian-installer/errata at http://www.debian.org/devel/debian-installer/errata
" ensures the root device exists. Type exit, " ( Do I type ' exit ' at the bottom, in the code file? or do I type ' exit ' in the interactive shell command line ? )
For some one like me, not familiar with the coda lingua, and for me to use Martin's patch, work around, I need the instruction to be more explicit at this point.
Just a reflection on me, no one else.
Thanks for any help forthcoming to get my 'raid 1' booting. Regards, McTech.
Reply to: