Report of slink installation with 2.1.2 disks
Brief summary:
Installation of fresh slink went fairly well. Had big problems with
apt being intolerat for two things: missing .debs in archives and
failing package installation scripts.
Thanks for Jeff Sheinberg providing template for this mail :)
CONFIGURATION:
machine: Abit BH6 mother board, Celeron 300A overclocked to 450MHz
memory: 128MB SDRAM
disk: IBM Deskstart 8GB
partitions: hda1 3GB fat32 or whatever
hda2 2GB same filesystem as hda1
hda3 256MB Linux swap
hda4 linux extented? (see below)
hda5 linux native (see below)
hda6 same as hda5 (see below)
scsi: none
sound: none
video card: Matrox Millennium G200 (8MB)
cd-rom: Mitsumi 32x (IDE/ATAPI)
net card: 3Com 905b
TESTING TARGET:
task: Base system installation (slink)
source: boot-floppies 2.1.2 from Incoming@ftp.lh.umu.se
PREPARATION:
Used dd to create rescue and drivers floppies. Prepared NFS server on
another machine to export directory containing base2_1.tgz and other
boot related files (rescue, drivers, etc.). The machine running NFS
server has also complete mirror of slink and hamm and was later used
to install the rest of the packages with dselect/apt over HTTP.
PROCESS (about the partitions):
Booted with rescue disk. Created partitions for swap and / from free
space intented for Linux, BSDs and other things. This is where things
got interesting.
When I was in dinstall and using cfdisk to create the two partitions,
I never created or saw partition hda4. After rebooting the new system
I noticed that there indeed exists partition hda4, which is listed as
"extented" but only fdisk shows it. However, both cfdisk and fdisk show
that hda5 is "linux native".
This is ok and just as how I thought it should be. But for some reason
dinstall suggested using hda6 as the root partition. I accepted this to
see what would happen.
Below are partition tables as printed by cfdisk and fdisk respectively.
Note that with the "print" command cfdisk gave information about hda4.
The information about hda4 was not shown when creating the partitions.
Partition Table for /dev/hda
---Starting--- ----Ending---- Start Number of
# Flags Head Sect Cyl ID Head Sect Cyl Sector Sectors
-- ----- ---- ---- ---- ---- ---- ---- ---- -------- ---------
1 0x80 1 1 0 0x0C 15 63 1023 63 6144705
2 0x00 15 63 1023 0x0F 15 63 1023 6144768 4096512
3 0x00 15 63 1023 0x82 15 63 1023 10241280 524160
4 0x00 15 63 1023 0x05 15 63 1023 10765440 2048256
5 0x00 15 63 1023 0x83 15 63 1023 63 2048193
Disk /dev/hda: 16 heads, 63 sectors, 16383 cylinders
Units = cylinders of 1008 * 512 bytes
Device Boot Begin Start End Blocks Id System
/dev/hda1 * 1 1 6096 3072352+ c Unknown
/dev/hda2 5120 6097 10160 2048256 f Unknown
/dev/hda3 9216 10161 10680 262080 82 Linux swap
/dev/hda4 10240 10681 12712 1024128 5 Extended
/dev/hda5 10240 10681 12712 1024096+ 83 Linux native
As you can see, /dev/hda6 is not listed but it is mounted as root.
Also, kernel sees it fine during bootup. Here are the partition
check lines from dmesg:
Partition check:
hda: hda1 hda2 < hda5 > hda3 hda4 < hda6 >
This is what mount says:
# mount
/dev/hda6 on / type ext2 (rw,errors=remount-ro)
proc on /proc type proc (rw)
rae:/mnt/vanha/hessu/debian on /mnt type nfs (rw,addr=10.1.1.1)
Can anyone tell how /dev/hda6 can be root even if it is not listed
with fdisk or cfdisk?
Also, there should be free available for more partitions but when I try
to create a new partition with fdisk, it says "No free sectors available".
cfdisk would have let me add a partition (hda6) but I did not try it. I'm
starting to think that my partition table is messed up.
PROCESS (the rest of it):
When installing module for 3c905b, the card was discovered with MAC
address FF:FF:FF:FF:FF:FF. This is a known bug and can be corrected with
a newer driver from Don Beckers web pages. The old driver version
was 0.99e. After I compiled v0.99h of the driver and did rmmod/insmod,
the MAC address was found correctly and it was possible to do the rest
of the installation over NFS.
When the base was installed I created a boot disk and rebooted my new
system.
When the system came up, I did the following:
o created root password
o no for normal user account
o yes for shadow
o log in as root from VC2
o yes for remove pcmcia
o yes for "... choose of several selections..."
o cancel immediately, choose none of the above
o quit dselect without doing anything at this point
Was correctly logged out by the boot script
o log in as root
o start dselect
o choose apt, use local mirror over HTTP
o [U]pdate list of available packages
o [S]elect default selection
o [I]nstall
When running install, two showstoppers happened. First was that
dialog-0.9a-14 was not found. The reason for this was that the
Packages file and my mirror contents were out of sync. The version
of dialog was 0.9a-15 in the mirror and this is why apt refused to
cooperate. However, I found version -14 and copied that in my mirror.
After that apt was able to finish retrieving the packages.
When it was time to configure the packages, pcmica-cs broke the
installation process. Did not proceed further this time. Output of
script is below:
---------------- start of script output --------------------------------
Checking system integrity...ok
The following packages will be REMOVED:
pcmcia-cs
The following NEW packages will be installed:
at man-db libpng2 tetex-base libncurses4-dev manpages doc-debian lynx
mime-support mpack debian-policy mailx bison patch talkd groff xbase locales
ibritish pdksh lsof-2.0.35 liblockfile0 g++ libstdc++2.9-dev
cracklib-runtime binutils procmail libnet-perl libstdc++2.8 wenglish elm-me+
telnetd zlib1g xlib6g dialog dpkg-dev emacsen-common dnsutils dpkg-perl
dpkg-ftp electric-fence sharutils ncurses-term bin86 emacs19 libpam0g
cracklib2 bsdmainutils ispell libg++272 tetex-extra libc6-dev gcc gdb biff
strace netstd perl-suid data-dumper psmisc libpcre1 libpwdb0g m4 file make
cpp mtools ncurses3.4 gpm cpio tetex-bin flex ppp-pam libpam0g-util
nfs-server manpages-dev doc-linux-text info tcsh-i18n rcs cron talk lpr nvi
bc exim less iamerican libident dc perl ed time tcsh
13 packages upgraded, 94 newly installed, 1 to remove and 0 not upgraded.
Need to get 0b/51.4M of archives. After unpacking 131M will be used.
Do you want to continue? [Y/n]
^M^M(Reading database ... 3750 files and directories currently installed.)
Removing pcmcia-cs ...
Shutting down PCMCIA services:start-stop-daemon: stat /sbin/cardmgr: No such file or directory
cardmgrstart-stop-daemon: stat /sbin/cardmgr: No such file or directory
.
dpkg - warning: while removing pcmcia-cs, directory `/etc/pcmcia/cis' not empty so not removed.
dpkg - warning: while removing pcmcia-cs, directory `/etc/pcmcia' not empty so not removed.
dpkg - warning: while removing pcmcia-cs, directory `/etc/apm/suspend.d' not empty so not removed.
dpkg - warning: while removing pcmcia-cs, directory `/etc/apm/resume.d' not empty so not removed.
dpkg - warning: while removing pcmcia-cs, directory `/etc/apm' not empty so not removed.
Error: the current /etc/conf.modules is not automatically generated.
Use "update-modules force" to force (re)generation.
dpkg: error processing pcmcia-cs (--remove):
subprocess post-removal script returned error exit status 1
Errors were encountered while processing:
pcmcia-cs
E: Sub-process returned an error code
Some errors occurred while unpacking. I'm going to configure the
packages that were installed. This may result in duplicate errors
or errors caused by missing dependencies. This is OK, only the errors
above this message are important. Please fix them and run [I]nstall again
Press enter to continue.
installation script returned error exit status 100.
---------------- end of script output ----------------------------------
As noted by Gunnar Isaksson earlier, national keymap (Finnish for me)
was not loaded when the system came up. I believe this has already been
fixed.
Thanks for reading this far :)
// Heikki
--
Heikki Vatiainen * hessu@cs.tut.fi
Tampere University of Technology * Tampere, Finland
--
To UNSUBSCRIBE, email to debian-testing-request@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org
Reply to: