[SOLVED] boot problem after update to current

kristizz · 02-03-2012, 01:13 PM

Hello. I updated a 13.37 install to current. Well, not quite, I actually just changed /etc/slackpkg/mirror to current and did slackpkg update *kernel and edit lilo accordingly. I have an initrd in boot but didn-t do a new one because the required modules at boot haven't changed. Anyway, when I boot I get:

loading linuxebda is big;kernel setup stack overlaps lilo second stage.

Obviously not good.
Can some one tell me what this means?

ponce · 02-03-2012, 02:12 PM

you should have made the new initrd.gz image, because the one you got cointains modules built for the old kernel, so the new one cannot load them.

try booting with a boot disk, chroot in your root partition, create the initrd for the new kernel and run "lilo".

I think the message you got is because you probably forgot to run lilo again after updating the kernel: the new one is located differently on the disk and lilo cannot find it, looking for the position of the old one that's not around anymore.
The positions of the kernels it has to look for are updated each time you run the "lilo" command, accordingly on what's in /etc/lilo.conf.

kristizz · 02-04-2012, 04:37 AM

Quote:

you should have made the new initrd.gz image, because the one you got cointains modules built for the old kernel, so the new one cannot load them.

try booting with a boot disk, chroot in your root partition, create the initrd for the new kernel and run "lilo".

I think the message you got is because you probably forgot to run lilo again after updating the kernel: the new one is located differently on the disk and lilo cannot find it, looking for the position of the old one that's not around anymore.
The positions of the kernels it has to look for are updated each time you run the "lilo" command, accordingly on what's in /etc/lilo.conf.

Thanks for help. I didn't write my question very well. I should have said that I ran lilo before rebooting and kept getting errors. I ran sh /usr/share/mkinitrd/mkinitrd_command_generator.sh -i to build my initrd, but got this:

Code:

ash-4.1# sh /usr/share/mkinitrd/mkinitrd_command_generator.sh -i
Modules for kernel 2.6.37.6 aren't installed.

I did a clean install and upgraded to current (only took 20 mins!), but still get the mkinitrd_command_generator_looking.sh looking for 2.6.37.6 modules, and not seeing the 3.2.2 modules that are definitely there.

I'm going try and understand how the script works to see if I can do anything with it, but the basic problem is that it obviously can't accurately read the contents of /lib/modules.

I'm also going to look at writing a mkinitrd command manually to force it to use the right modules.

ponce · 02-04-2012, 05:18 AM

Code:

mkinitrd -c -k 3.2.2 -f ext4 -r /dev/sda7 -m ext4 -u -o /boot/initrd.gz

if you're using ext4 and don't need special hardware modules, this should do (substitute /dev/sda7 with your root device)

you can also do a

Code:

/usr/share/mkinitrd/mkinitrd_command_generator.sh -k 3.2.2

kristizz · 02-06-2012, 03:49 AM

Quote:

Originally Posted by ponce

Code:

mkinitrd -c -k 3.2.2 -f ext4 -r /dev/sda7 -m ext4 -u -o /boot/initrd.gz

if you're using ext4 and don't need special hardware modules, this should do (substitute /dev/sda7 with your root device)

you can also do a

Code:

/usr/share/mkinitrd/mkinitrd_command_generator.sh -k 3.2.2

All sorted. Luckily I had kept a copy of the mkinitrd command that I had generated when I did the original install. I just changed the target kernel and it's fine.