Re: URGENT! Install Base System via network: settings wrong
On Wed, Jan 09, 2002 at 05:50:49PM +0000, Philip Blundell wrote:
> > I then tried a brand new installation, including formatting the partition.
> >In this case, the whole installation, until the base system installation, went
> >smooth. Not a single error and no hack needed. At the base system
> >installation, the error message occured once again.
>
> Okay. First, can you just confirm which version of boot-floppies you were
> using?
>
> > I'd be glad to do testing, but as for debugging I can't promise anything.
> >What do you mean by debugging (I never tested boot floppies of a machine
> >nor am I familiar with the code; Can I run it with a debugger? What is
> >the testing methodology?)
>
> I meant "debugging" in a fairly generic sense, as in figuring out what's wrong.
> A good place to start would be to run the install up to the point where the
> error occurs, then go to a shell prompt and inspect the Release file that
> debootstrap claims is malformed, to see whether it was successfully downloaded
> and contains a valid Components line.
>
> p.
Here are the results of the test, boot-floppies 3.0.18 dated 21.12.2001:
1. After entering the root floppy, and immediately before the first
screen (the Debian 3.0 title), the following error messages occured:
modprobe: modprobe: Can't open dependencies file
/lib/modules/2.2.20/modules.dep (no such file or directory)
I don't know if it has anything to the ``Malformed'' message, I just
write *anything* I met during installation.
2. To make kernel configuration minimal I only installed the IDE/ATAPI floppy
support (I might have skipped that..) and the eepro100.o network module
(for my network card, which is a must).
When trying to install these two, I received the following error message:
/target/usr/sbin/modconf: cannot create: directory nonexistent.
This was followed by a ``successful installation'' message, so I assumed
these errors are not show-stoppers.
3. When I came to the base system installation, chose network, and then
chose my mirror site (ftp://ftp.tau.ac.il/pub/OS/Debian, proxy: none,
port: 8080; proxy & port: didn't change the defaults) came the surprise:
I did NOT get the ``Malformed Release file'' message.
It seemed like the installation is going fine (I didn't get to this
step the previous time I tried to install).
BTW, this time I was working on a brand new unused partition I have
/dev/hda4.
In the previous one (/dev/hda1) I already installed potato from CDs,
and upgraded (network) to woody, so I don't want this erased.
I reformatted my swap partition /dev/hda2, just to ensure nothing is
due to the other installation.
4. Well, I thought to myself, everything is going fine.
But then, another surprise: at_3.1.8-10.deb was corrupt (or so the
installation system said; I guess the installation system is wrong,
or else I couldn't upgrade my potato from that very same site).
I tried to upgrade from the default http://http.us.debian.org/debian
site (immediately after the first attempt has failed; i.e., no boot or
anything in between), but then I received a message that:
base-config_1.33.7_all.deb was corrupt.
The same message was issued from http.uk.debian.org .
Again, I can only assume these messages are wrong and not a problem
in the installation system.
5. When all else fails: I executed a shell.
I tried: find / -name Release .
The only files I found were:
/target/var/lib/apt/lists/debootstrap.invalid_dists_woody_{Release,main_binary-i386_Packages}
I also tried: find / -name '*elease*'
but only the above was found (and /release-notes, which has nothing to do
with this).
6. I manually erased the two files I just mentioned, returned to the
installation system, and tried once again the base system installation.
And the 3rd surprise: ``Malformed Release file''.
I never got this message in the various downloading attempts I had,
but I got it only after *erased* the two files!
7. I executed a shell once again, but this time too: no Release file found.
8. Attempts to continue the installation at this stage fail and result in
the ``Malformed Release file''.
Good luck with the debugging, and let me know if I can be of any help
in the testing!
-- Ariel.
---
Ariel Tankus
School of Computer Science
Tel-Aviv University
arielt@post.tau.ac.il
Reply to: