I am new to the use of Linux and I appear to have run into serious trouble.

I downloaded and installed a new kernel (-2.6.9). Now laptop will not even boot into linux. It's got a load of stuff displayed ending with:

reiderfs: disagrees about version fo symbol struct_module
insmod: error inserting '/lib/modules/2.6.6-7.95-smp/kernel/fs/reiserfs.ko' : -1 Invalid module format Waiting for device /dev/hda to appear: ok
rootfs: major=3 minor=4 devn=772
warning: can't open /etc/mtab: No such file or directory
Kernel panic - not syncing: VFS: Unable to mount root fs on unknown-block(3,4)

This very scared new user is desperate for help!c

Hopefully, grub still references the old kernel and you can boot that. It appears the reiserfs modules is no good. You could try the make modules and make modules_install again or recompile the kernel. Double check that reiser is included as a modules or built in. I've also found that if you remove support for Ext2 and Ext3 even if your partitions are not formated with them, suse you give you headaches. I've found its generally good to include support for all three file systems regardless of which file system you have formated the hard drives with.

Thanks.
how do I make modules or recompile when the machine won't even boot?

You are not alone - the good news is that there is a solution - which is just as well because I am deeped p***ed off with SuSE. I have just suffered the same under SuSE 9.2. I'm just about to post to suse-linux-e

First - don't do anything drastic and do it all very slowly. If you do it right, the most wrong thing that will happen is that you will reset many of your upgrades. For example, I am now running KDE 3.3.0 instead of 3.3.2


The process is very similar to the fresh install/upgrade - take your original SUSE install disc (DVD) s (CDROM) and boot from them

choose installation
then when you have chosen language/keyboard
AND THIS IS THE SCIENCE BIT - DO NOT CHOOSE INSTALL

first choose repair system - just in case - choose fully automatic, it will take a while, and when asked assume bootloader isn't wonky

This bit is slightly hazy because I cannot remember the exact names on the menu options - but nothing can go wrong until you press the go button

then choose to replace/upgrade software packages (can't remember exactly what it is called - but you'll spot it)
(it will first show you the ones it can't do anything to - don't worry ignore that)
then get the list of all software packages installed up on the screen - option on LHS of menu
when you've fone that you'll see two columns "available version" and "installed version"

find the kernel you'll see that the kernel available is older than the installed version - but click on the check box until a tick appears - this will downgrade your kernel

If the non-GPL kernel module is installed do the same

Now the bit I got wrong was to leave everything else alone - the net result was that everything got reinstalled. -all 1693 packages - but at least I'm up and running and giving you a clue

It might be (and I am guessing here, but it seems reasonable based on my experience) that for all the other packages you need to set the check box to "do not install" - or you could avoid the risk and be assured that just doing what I did got it working.

You will lose nVidia
You will not lose your data

Finall run suse watcher with manually select patches and upgrade everything except the f***ing kernel

Merry Christmas SUSE

You could try the failsafe option, but that is why I said "Hopefully, grub still references the old kernel ". Always keep 1 previous kernel to boot in case of emergencies. The CD probably allows you to boot to a shell.
If you can to that,

mount /dev/hda1 /mnt
cd /mnt
update-env

If I remembered the command correctly, that will mount your hard drive. change to the directory and create a shell that thinks the root directory is /mnt

Then you can play with your system just as it was booted to the command line

Hi,

I installed SUSE 9.2 on an IBM thinkpad x40, all went fine but the ipw2100 driver didn't work with WPA, it's ok as I knew this is not only SUSE's problem.

Then I ran online update in Yast, saw the new kernel and the description mentioned some fixes in ipw2100 + WPA , I installed the update and reboot and encountered the same kernel panic problem.

Now my thinkpad is unusable. Yes maybe it's not hard to fix, but a major problem like this with kernel update is just... unacceptable. At least I think it should create a reference to the old kernel in grub.

And to be fair, other than this problem, I am quite impressed with SUSE 9.2, the default KDE theme is beautiful, suspend to disk works (although it take too long to recover), and changing system settings in Yast is very convenience. (Afterall laziness is the reason I want to try SUSE/Mandrake )

if you have the install CD, choose repair which can allow you to install the stock kernel and should restore your system to working order, however you will be running the stock kernel again.

All you need to do is run the repair option from the install CD, the only repair that needs to be done is to the grub boot loader so uncheck the other options, it then boots just fine into the newly installed kernel.

I did this yesterday on four PC's all running flawlessly on the new kernel.

I ran into this "disagree about struct_module" version thing myself on a SuSE 9.2 Pro system. After some head scratching, I traced the problem down to a misconfiguration in the /boot directory on my machine. What appears to have happened is that the links in /boot weren't properly set up after a kernel update (which was performed using YaST's 'update system' capabilities).

The specific problem was that vmlinuz was soft linked to vmlinuz-2.6.8-24.10-bigsmp, while initrd was left pointing to a different version - it wasn't pointing at initrd-2.6.8-24.10-bigsmp as it should have been. I had to boot up using disc1 of the SuSE 9.2 installation media, select Install, then select "boot existing installation" in order to gain shell access to the filesystem. Fixing the initrd symlink took care of the problem for me.

HTH

This makes sense. The Reiserfs module is in SuSE's "initrd".

I have this similar problem this morning, after 64bit 9.1 to suse9.2 online update of new kernel, system hang at boot cause my / (root) filesytem was full in additon to the warning message.
Like a dummy, I assumed my filesystem was currupt and cleaned up some error files, efsck my /dev/hde1 device and my have made things worst. In doing some research some people are saying that the kernel-um (user mode kernel) has a problem that this created a link in /boot from user mode initrd to /boot/initrd and the wrong kernel is being booted. This will be the last time I use that online update utility.. Thank goodness for snapshot backups.