Hello Quentin

Thanks for your hints. Because there is a new "mini.iso" online in the meantime, I 
repeated a basic "out of the box" setup on my hardware (IBM NetVista A40 model 6840-QDG, 
Promise Fastrak TX2300, two Western Digital Caviar Red 2 TB NAS disks).

All partitions are deleted, MBR is zeroed completely, i.e. classic "starting from 
green grass" (bare metal) scenario.

- boot parameters on CD: removed the "vga=788" parameter, added "dmraid=true" 
according [1].
- partitioning: "Guided - Use complete disk", RAID volume detected correctly, option 
"Use separate /usr, /home, /var, ..." (gives a very large /home)
- packet selection as usual, i.e. "SSH server", unselection "Desktop 
environment" in my case
- question about GRUB's location answered with "Yes" into MBR
- "Continue" - finishing installation as usual and reboot
Everything worked flawlessly until this point

BTW small minor bug (offtopic): CD drive drawer stays locked -> sending "eject" 
command like the GParted CD does may helpful - I built a small MS-DOS start floppy tool 
containing a CD-ROM driver and a QBasic tool[2] for that to avoid unnecessary power off/on 
cycles as temporary workaround...

First boot from hard drive: GRUB menu appears correctly but boot fails with the message 
as shown ("Device or resource busy" and BusyBox prompt) in the past.

=> Ctrl+Alt+Del, booting from mini.iso CD, "Advanced options", "Rescue mode" <Tab> 
"dmraid=true" instead "vga=788"
- After successful loading: "/dev/dm-1" as root system, start a shell

*** First important bug: Wrong special device file names in /etc/fstab; there is a extra 
"p" which must be removed, i.e.

# ls -al /dev/mapper
drwxr-xr-x  2 root root      200 Jun 29 14:44 .
drwxr-xr-x 15 root root     3740 Jun 29 14:44 ..
crw------T  1 root root  10, 236 Jun 29 14:44 control
brw-rw----  1 root disk 254,   0 Jun 29 14:44 pdc_ehdiebaf
brw-rw----  1 root disk 254,   1 Jun 29 14:44 pdc_ehdiebaf1
brw-rw----  1 root disk 254,   2 Jun 29 14:44 pdc_ehdiebaf5
brw-rw----  1 root disk 254,   3 Jun 29 14:44 pdc_ehdiebaf6
brw-rw----  1 root disk 254,   4 Jun 29 14:44 pdc_ehdiebaf7
brw-rw----  1 root disk 254,   5 Jun 29 14:44 pdc_ehdiebaf8
brw-rw----  1 root disk 254,   6 Jun 29 14:44 pdc_ehdiebaf9

but

# cat /etc/fstab
...
<file system> <mount point>   <type>  <options>       <dump>  <pass>
/dev/mapper/pdc_ehdiebafp1 /               ext4    errors=remount-ro 0       1
/dev/mapper/pdc_ehdiebafp9 /home           ext4    defaults        0       2
/dev/mapper/pdc_ehdiebafp8 /tmp            ext4    defaults        0       2
/dev/mapper/pdc_ehdiebafp5 /usr            ext4    defaults        0       2
/dev/mapper/pdc_ehdiebafp6 /var            ext4    defaults        0       2
/dev/mapper/pdc_ehdiebafp7 none            swap    sw              0       0
/dev/sr0        /media/cdrom0   udf,iso9660 user,noauto     0       0
/dev/fd0        /media/floppy0  auto    rw,user,noauto  0       0

(look to the extra "p" at the end in all "pdc_ehdiebafpX" entries!) => must be 
changed into

# <file system> <mount point>   <type>  <options>       <dump>  <pass>
/dev/mapper/pdc_ehdiebaf1 /               ext4    errors=remount-ro 0       1
/dev/mapper/pdc_ehdiebaf9 /home           ext4    defaults        0       2
/dev/mapper/pdc_ehdiebaf8 /tmp            ext4    defaults        0       2
/dev/mapper/pdc_ehdiebaf5 /usr            ext4    defaults        0       2
/dev/mapper/pdc_ehdiebaf6 /var            ext4    defaults        0       2
/dev/mapper/pdc_ehdiebaf7 none            swap    sw              0       0
/dev/sr0        /media/cdrom0   udf,iso9660 user,noauto     0       0
/dev/fd0        /media/floppy0  auto    rw,user,noauto  0       0

=> bash
mount -o remount /
mount /dev/mapper/pdc_ehdiebaf9 /home
mount /dev/mapper/pdc_ehdiebaf8 /tmp
mount /dev/mapper/pdc_ehdiebaf5 /usr
mount /dev/mapper/pdc_ehdiebaf6 /var
pico /etc/fstab
(removing the extra "p"s)

Applying UUID disabling (Quentin's hint):

pico /etc/default/grub
=> enabling "GRUB_DISABLE_LINUX_UUID=true" by uncommenting
update-grub
grub-install /dev/mapper/pdc_ehdiebaf

(the second command is also important else you get a mess [and the boot failure 
issue still remains!] with some extra entries in the GRUB boot menu)

=> "exit" twice from rescue mode and reboot - removing CD with the trick (start 
floppy tool) above - et voilĂ  - it successfully boots. :-)

Done several root logins and "shutdown -r now" - it always boots successfully - 
i.e. the problem seems to be fixed. :-)

To the Debian maintainers: Changing the last installation step to allow selection GRUB 
UUID on/off (or automatically disabling UUID in case of dmraid=true before creating GRUB) 
seems to be a useful way to getting rid with this bug. Also don't forget to fix this 
extra "p" issue in /etc/fstab and sending an CD drawer unlock (or even eject) 
command to the CD/DVD drive at the tidy up end stage of installation.

    Andreas

[1] http://wiki.debian.org/DebianInstaller/SataRaid
[2] http://beilagen.dreael.ch/QB/CDSCHUBL.BAS
--
meile.biz IT solutions, Hauptstrasse 63, CH-8242 Hofen SH
PC/Netzwerk-Support, Web-Entwicklung und -Hosting, IT Security
Tel. +41 52 640 04 72 * Fax +41 52 640 04 73 * Mobile +41 79 334 05 67
Postfach 169, CH-8240 Thayngen * i...@meile.biz * http://www.meile.biz

Attachment: smime.p7s
Description: S/MIME cryptographic signature

Reply via email to