Bug#406483: System fails to boot after installation
On Friday 12 January 2007 17:43, you wrote:
> > A simple failure to reboot is not a critical issue, especially not if
> > it only happens for one user.
>
> Generally right, but it cause unusable all system for all users on
> server.
It is still only one system which is not even really broken, just
misconfigured. The fact that _you_ cannot reboot, does not mean that the
installer is completely broken for all hardware and should not be
released. It works perfectly for almost all others and this probably _is_
a known issue that _can_ be worked around.
> > OK, so we have a successful installation. Good.
>
> i maked it to now about 6 times :( all the same :(
That does not surprise me. If you do the same installation, you get the
same result.
The problem _could_ be that the RC1 installer is using the 2.6.17 kernel
while the installed system will have the 2.6.18 kernel if you do a floppy
installation.
However, the fact that you have the same problem if you install from the
full CD image makes that unlikely.
> this can wait over 4 hour! it's not good procedure to restore server
> after crash.
Well, 4 hours is not the amount of time I was thinking of. More like 5 or
10 minutes.
> > - give us _all_ and the _exact_ messages that are visible on the
> > monitor when the boot "hangs".
>
> ugh. i must rewrite it by hand :( I'll do it.
We really do need those messages to be able to help you.
You can also try booting over serial console. Then you should be able to
just copy&paste the messages.
> I think it's very strange, that install sarge and next upgrade to etch
> is ok.
That is somewhat strange.
I assume you are using an initramfs-tools initrd and the grub bootloader.
You can add additional boot parameters in grub by editing the menu item
that is used to boot Debian.
Try adding the parameter 'rootdelay=10'.
During the boot the initrd will go through certain stages and print
messages about the scripts its running. Based on that you can ask it
to "break" at a certain point so you can investigate if anything is wrong
from a debug shell.
For that, add the boot parameter 'break=XXXX', where XXXX is one of top,
modules, premount, mount, bottom or init. These are scripts that are
executed in the order they are listed. Try to break as late as possible.
See 'man initramfs-tools' for additional information.
Good luck,
FJP
Reply to: