--- Begin Message ---
Package: installation-reports
Version: Etch installer daily build 01/01/07
Greetings. I'm continuing where I left off with bugs #286939, #291928 and
#292221. They were all concerned with installing on an ATARAID system using
the 2.4 ataraid/medley driver pair. If that method is now considered
obsolete (now that d-i defaults to 2.6 kernels using libata), those bugs
could be closed, even though they're not actually solved. I'm not sure of
your policy on such things.
Anyway, I was messing around with Ubuntu recently when I discovered the
udpkg command. I realised that this could be the key to reinstalling Debian
on my machine. And it was.
Instructions for installing Etch onto ATARAID arrays
----------------------------------------------------
Before you start partman (or after it has offered you only sda and sdb which
you don't want), start the shell on the F2 console and do the following four
things:
1. modprobe dm-mod
2. udpkg -i /cdrom/pool/main/d/devmapper/dmsetup-udeb...... (whatever it's
called) (this step may not be necessary, I don't know if dmraid requires it
or not)
3. udpkg -i /cdrom/pool/main/d/dmraid/dmraid-udeb......
4. dmraid -ay
This should find your arrays and create the relevant devices for you in
/dev/mapper. I'm using a Silicon Image controller (SiI3112A), and it worked
for me. If it doesn't, you're no worse off than before I wrote this. If it
does, go back to the main menu and continue with partman. You can do your
partitioning, install the base system, choose a mirror, install more
software, all that stuff. The only slightly freaky thing is that it assumes
that your /dev/mapper devices are LVM2 volume groups, even when they're not.
This may or many not be relevant to what's below:
Outstanding problems with installing Etch onto ATARAID
------------------------------------------------------
There are two main problems remaining:
1. You can't install a boot loader. The log console on F4 says at various
points "root is an LVM volume - can't install GRUB", and your only other
option is installing LILO. This fails with no explanation AFAICT, even
though it worked under 2.4 (see my previous bug reports noted above).
Actually the root is not an LVM volume, but d-i mistakenly thinks that all
your ATARAID devices are LVM volumes. Ubuntu installed GRUB perfectly well
on the same system, which is how I managed to unearth the other problem:
2. Since I had GRUB already installed, I dropped into it and manually gave
it the following info, carefully copied down before rebooting:
root (hd0,3)
kernel /boot/vmlinuz-2.6.18-3-486 root=/dev/mapper/sil_aebdgidebbaj4
initrd /boot/initrd.img-2.6.18-3-486
It booted fine, but stopped at this point:
Begin: Mounting root file system ...
Begin: Running /scripts/local-top ...
Done.
Begin: Waiting for root file system ... ...
And it just stopped. I strongly suspect (though I run out of knowledge at
this point) that the kernel/initrd combo is not properly equipped with the
dm-mod module and the dmraid userspace tool, such that it can't find the
specified root fs. This is supported by screensful of "trying to read beyond
the end of device" error messages, where it's looking for sector 600000000
on /dev/sdb and hitting the end of the device at 300000000 (my medley array
is 2 x 160Gb disks striped, ie. 640000000 sectors or thereabouts). And also
some references to /dev/sd4, which is a dead giveaway that it's mistaken
that for /dev/mapper/sil_......4
Well, that's all for now. If you're interested I kept logs of both my
installation attempts (expert and normal) - you can get them at
http://www.terminalarrogance.com/debian-dmraid-logs.rar
My next attempt is to do a line-by-line comparison of the mkinitrd script
(for which I submitted a successful patch for 2.4) with the Ubuntu one,
which works fine (ie. supports dmraid properly). I may then be able to
finish the install, and supply a patch for others to use. Don't hold your
breath though - it takes me about a year between each set of attempts to
psych myself up for another go (I've been working on this since 2003).
CC
--
No virus found in this outgoing message.
Checked by AVG Free Edition.
Version: 7.5.432 / Virus Database: 268.16.4/615 - Release Date: 03/01/2007
13:34
--- End Message ---
--- Begin Message ---
Source: hw-detect
Source-Version: 1.54
We believe that the bug you reported is fixed in the latest version of
hw-detect, which is due to be installed in the Debian FTP archive:
archdetect_1.54_amd64.udeb
to pool/main/h/hw-detect/archdetect_1.54_amd64.udeb
disk-detect_1.54_all.udeb
to pool/main/h/hw-detect/disk-detect_1.54_all.udeb
ethdetect_1.54_all.udeb
to pool/main/h/hw-detect/ethdetect_1.54_all.udeb
hw-detect_1.54.dsc
to pool/main/h/hw-detect/hw-detect_1.54.dsc
hw-detect_1.54.tar.gz
to pool/main/h/hw-detect/hw-detect_1.54.tar.gz
hw-detect_1.54_amd64.udeb
to pool/main/h/hw-detect/hw-detect_1.54_amd64.udeb
A summary of the changes between this version and the previous one is
attached.
Thank you for reporting the bug, which will now be closed. If you
have further comments please address them to 405484@bugs.debian.org,
and the maintainer will reopen the bug report if appropriate.
Debian distribution maintenance software
pp.
Frans Pop <fjp@debian.org> (supplier of updated hw-detect package)
(This message was generated automatically at their request; if you
believe that there is a problem with it please contact the archive
administrators by mailing ftpmaster@debian.org)
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1
Format: 1.7
Date: Fri, 06 Jul 2007 00:17:19 +0200
Source: hw-detect
Binary: archdetect hw-detect disk-detect ethdetect
Architecture: source amd64 all
Version: 1.54
Distribution: unstable
Urgency: low
Maintainer: Debian Install System Team <debian-boot@lists.debian.org>
Changed-By: Frans Pop <fjp@debian.org>
Description:
archdetect - Hardware architecture detector (udeb)
disk-detect - Detect disk drives (udeb)
ethdetect - Detect network hardware and load kernel drivers for it (udeb)
hw-detect - Detect hardware and load kernel drivers for it (udeb)
Closes: 405484
Changes:
hw-detect (1.54) unstable; urgency=low
.
[ Joey Hess ]
* Remove cruft left over from the days of pcmcia-cs.
- pcmciautils doesn't create a cardmgr.pid, so stop looking for it to
determine if pcmcia is started. Look for the pcmcia_core module being
loaded instead.
- No need for funky redirections when running the init script as it no
longer starts a daemon.
- /var/run/stab was only produced by pcmcia-cs, so the whole
gen_pcmcia_devnames section can no longer run; remove.
- Only modify config.opts in /target if pcmcia_resources is preseeded.
* Remove workaround for long fixed cdebconf bug.
.
[ Frans Pop ]
* disk-detect: detect Serial ATA RAID disks if requested by setting
disk-detect/dmraid/enable at the boot prompt. Closes: #405484.
Thanks to Criss Carr for all the information he has provided, much of
which has been used in implementing the experimental dmraid support
available with this release of hw-detect and other components. Also
thanks to Jérémy Bobbio for his patches and Mark Brown for providing
a test system.
.
[ Updated translations ]
* Punjabi (Gurmukhi) (pa.po) by A S Alam
* Russian (ru.po) by Yuri Kozlov
Files:
d5d852515625e8818990e2ad3399c2b9 914 debian-installer standard hw-detect_1.54.dsc
1ea02ff1164433255bc5527f81c14fa6 139573 debian-installer standard hw-detect_1.54.tar.gz
4dbf8ea04c44f668629918f189563c6a 25036 debian-installer optional ethdetect_1.54_all.udeb
46f7df0008a77e1a5830f046b81ec773 18744 debian-installer optional disk-detect_1.54_all.udeb
77610a44f3ff990c175f50e6ff0ee31f 69406 debian-installer standard hw-detect_1.54_amd64.udeb
7fc8c92836d430dce3009207b5d37564 2270 debian-installer standard archdetect_1.54_amd64.udeb
Package-Type: udeb
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.6 (GNU/Linux)
iD8DBQFGjXJhgm/Kwh6ICoQRAhPPAKCjNwSiuAqnO/NVB3uSNX96eLUlrACg2X30
9O5lpm10eVHXvDwzou5JNX4=
=Wc4C
-----END PGP SIGNATURE-----
--- End Message ---