Re: etch installation fails on two machines
On Tue, Sep 18, 2007 at 05:33:32PM +0200, das.t wrote:
>> Doug Tutty:
>> > Read the installation manual and submit an instalation report. It goes
>> > in as a bug maintained by the debian-boot team. Once you get the
>> > automated response to your report giving you the bug number assigned,
>> > subscribe to that bug and reply to that bug; it becomes its own
>> > mini-list that is mirrored to debian-boot.
>
> thank you both for your answers. and yes, i actually did mean 96 m instad of
> 96 k ram. ;-)
>
Just for comparison, I'm writing this email via (not from) a P-II with
64 MB ram and an 850 MB hard drive. From it, I ssh to my Athlon to
actually write and send the email. The P-II is basically a slim client.
It used to run Etch, the install proceeding normally. Note that I've
never been able to get Etch's installer to install Grub. Etch ended up
quite slow on the P-II and I have since switched it to OpenBSD and its
just fine. OBSD is also great on my 486 with 32 MB ram.
> well, i finally succeeded. but still this remains quite strange. (faulty cds /
> dvd can be absolutely ruled out, i thrice-checked.)
>
> - i installed some more rum (up to 128 m). installation went past installation
> of base tools (core tools?) which were the problem. (in german the translated
> word is very similar to kernel, which is why i thought of that.) ok, it went
> further, but got stuck when extracting libslang or similar.
> this in fact lets one (strongly) suspect it's the ram, but as the installer
> does support a low memory install i ruled that out.
> - as i used the same hard disk in both computers and had another one lying
> nearby i grabbed that one and - the install succeeded. yay!
> - as soon as i added that other hard disk as a slave into that machine and
> tried a re-install onto the master ... it failed, this time when setting the
> clock.
>
> the hard disk in question has been checked for bad blocks several times, yet
> it is quite old and there might be some other strange issues. windows did not
> complain, though (i was forced to do it ... ;-)
>
> question: is it still worth a bug report? i suspect it will not be possible to
> being reproduced without that special hd.
> but if you think so i will do my best to describe it for the team.
>
I wonder if the disk is having errors that cause several retries. Since
you have a working system on one disk, I would suggest you carefully
remove it and put it away safely before testing the questionable drive
further. If you have both drives on one controller (you mention putting
the other drive in as slave) the bad drive can mess up the controller
and the other drive.
Once you have a safe testbed, try booting up the install CD and go to
the shell. Then try formatting the drive with a badblocks scan while
watching syslog for errors. If you have a LiveCD such as grml, you can
boot that and run wipe on the drive and watch syslog. You could also
download the OBSD install floppy (one image only) and run its shell to
poke at the drive with a totally different OS/kernel.
Finally, my P-II is picky what drives it will tolerate. I almost threw
out a seemingly failing drive only to find that it works perfectly in my
486. Perhaps you're having the same problem.
Doug.
Reply to: