Kernel panic at boot after successful install on a G3 400Mhz (ppc)

[Florian Vichot]

Hi

I installed Ubuntu on my Mac G3 400Mhz (it has 128Mb RAM), and once the install is finished, I did a
reboot. But after I got past yaboot, linux started to boot, then froze with these lines:

pivot_root: No such file or dir
/sbin/init : 429 : cannot open dev/console: no such file
Kernel Panic: attempted to kill init!

I tried to install Mandrake 9.1 on my mac, before Ubuntu, and I got a very similar problem (something to
do with init, I could retry to install it if you need the exact message).

If anyone can help that 'd be great because I getting hopeless in installing Linux on my mac.
Mac OS X works fine by the way.
Florian

[Matt Zimmerman (mdz)]

Please tell us about the disks in your system. How many do you have? SCSI or
IDE? On which partition did you install?

Do you have any working Linux on the system? We need to collect more
information, and the easiest way to do this is with Linux tools.

[Barry-phease (barry-phease)]

(In reply to comment #1)
> Please tell us about the disks in your system. How many do you have? SCSI or
> IDE? On which partition did you install?
>
> Do you have any working Linux on the system? We need to collect more
> information, and the easiest way to do this is with Linux tools.

I have the same problem. I only have one disk which is IDE and is reported as
/dev/hdc.
The CD is reported as /dev/hda. The standard install created 4 partitions, The
first 2 for booting, 3=ext3, 4=swap.

If you need more information I can chroot from the install disk.

[Matt Zimmerman (mdz)]

If you could send:
- a copy of the /boot/initrd.img which was generated
- lspci output
- dmesg output (from the installer is fine)

that would help.

[Jon Chambers (jon-students)]

Created an attachment (id=965)
Output of dmesg

Hey there.

I'm having exactly the same problem on a similar system. It's a blue/white G3
with a PCI SCSI card (nothing attached), two hard drives on the first IDE
channel (detected second -- see bug #7256), and the CD drive on the second IDE
channel.

Please find the output of dmesg (run from the installation CD at the end of the
installation process) attached.

[Jon Chambers (jon-students)]

Created an attachment (id=966)
lspci output

...and the lspci output...

[Jon Chambers (jon-students)]

(In reply to comment #3)
> If you could send:
> - a copy of the /boot/initrd.img which was generated
...

The image that was generated by my machine weighs in somewhere around 4MB, which
is too large to post here. How else can I get it to wherever it needs to be?

[Jon Chambers (jon-students)]

(In reply to comment #3)
> If you could send:
> - a copy of the /boot/initrd.img which was generated
> ...

Please find a copy of the initrd image generated during my install at
http://ece.olin.edu/~jchambers/ubuntu/initrd_ubuntu_install.img. I can't
guarantee that it will stay there forever, but it should be fine for the active
lifetime of this bug.

[Jon Chambers (jon-students)]

I would like to confirm that the solution proposed in bug #7256 (modifying the
loadmodules file in the initrd image) DOES solve this problem.

[Matt Zimmerman (mdz)]

(In reply to comment #8)
> I would like to confirm that the solution proposed in bug #7256 (modifying the
> loadmodules file in the initrd image) DOES solve this problem.
>

The original bug reported here (bug #9996) looks like a duplicate of bug #7256.
However, your symptoms (as described in bug #7256) are quite different, and so
I'm quite surprised that the solution changes anything for you. I've downloaded
your initrd, but will need to transfer it to a powerpc system to examine it.

This bug has been marked as a duplicate of bug 7256.

[Jon Chambers (jon-students)]

(In reply to comment #9)
> The original bug reported here (bug #9996) looks like a duplicate of bug #7256.
> However, your symptoms (as described in bug #7256) are quite different, and so
> I'm quite surprised that the solution changes anything for you. I've downloaded
> your initrd, but will need to transfer it to a powerpc system to examine it.

Just to be clear as to the sequence of events and to make sure there's no
confusion about which "symptoms" we're talking about:

1) I had the installation problem described here in bug #9996.
2) Noticed the similar problem in bug #7256.
3) Attempted to modify my initrd image to match the changes made to
'loadmodules' as described in bug #7256.
4) Made a mistake in modifying my image and ERRONEOUSLY posted comment #30 in
bug #7256, complaining that the patch wouldn't work for fresh installs.
5) In response to your suggestion that the proc mountpoint was missing from my
image, I checked and discovered that it was, in fact, missing.
6) I repaired the image, adding the missing mountpoints (there were several)
and verified that the modified image would let me boot successfully.
7) Having done so, I posted comment #8 here in bug #9996.

The actual steps I took are described in detail at
http://www.ubuntuforums.org/showthread.php?t=8050. The ramdisk image I sent is
the one produced by the installer and NOT the modified one that I created. I'd
be happy to post that, too, if you'd like.

Thanks!