Messed up officially. Linux related.

[pillage2001]

I posted in another thread but please click on thel ink to relate.....

Here's the situation, I have a server running a software RAID and I made a new kernel and updated it. When I rebooted, it showed this:

Uncompressing linux...

...ran out of input data.

-- System halted.


I tried running the old images but it wouldn't run due to the software RAID. It said something about not able to mount the root and went into the kernel panic.

Need steps to work around this problem and I'm running out of ideas. I'm not really linux incline...

Update:

Got to mount the root onto /dev/hda1 but it said it could not open an initial control.

I tried init=1 but gave the same crap. It asked me to use the init= option but it did not budge.

 

[drag]

This stuff is why when you update a kernel you leave the old entry for the old kernel in the boot configuration file, that way if your new kernel doesn't work then you can go back to the old kernel and try again.



So right now your stuck trying to get into your root partition?
You want to edit your lilo.conf and run lilo and reboot, but you can't get into your root partition?


I am not sure what is going on.


Are you using a initrd image?

Initrd is a simple little loopback file system that is put into RAM at boot up time. It's job is to load up kernel modules that you need to get access to the filing system. (like scsi drivers or network drivers if you have a remote root, stuff like that. it's commonly used by distros to provide access to large amounts of different hardware without having to compile support all into the system).

Basicly your goal is here to get into your root filing system mounted from a rescue disk, which having software raid complicates things, chroot to your system, edit your lilo.conf so that it uses your last known working kernel, then run lilo, then exit and unmount everything, then reboot.

Also if your distro has a rescue disk made for this task, then use that it'll make things easy.

This is also why grub is cool, you input your own configurations from a boot-time command line, and get around this bad stuff.

 

[pillage2001]

Originally posted by: drag
This stuff is why when you update a kernel you leave the old entry for the old kernel in the boot configuration file, that way if your new kernel doesn't work then you can go back to the old kernel and try again.



So right now your stuck trying to get into your root partition?
You want to edit your lilo.conf and run lilo and reboot, but you can't get into your root partition?


I am not sure what is going on.


Are you using a initrd image?

Initrd is a simple little loopback file system that is put into RAM at boot up time. It's job is to load up kernel modules that you need to get access to the filing system. (like scsi drivers or network drivers if you have a remote root, stuff like that. it's commonly used by distros to provide access to large amounts of different hardware without having to compile support all into the system).

Basicly your goal is here to get into your root filing system mounted from a rescue disk, which having software raid complicates things, chroot to your system, edit your lilo.conf so that it uses your last known working kernel, then run lilo, then exit and unmount everything, then reboot.

Also if your distro has a rescue disk made for this task, then use that it'll make things easy.

This is also why grub is cool, you input your own configurations from a boot-time command line, and get around this bad stuff.

it's not booting off the RAID. I just got the news and it seems like I just need a rescue disk and do what you suggested??

 

[drag]

What do you mean "it's not booting off of the raid"?

Is this still your kernel that gave out a error when it tries to boot up, or is this your old kernel that was working until you installed the new kernel?

What distro are you using? how is your partitions setup? There is so much I don't know about your setup.

What you need to do is get your old kernel back up and running just like it was before you installed your custom kernel.

you didn't make a extra entry in your boot menu for your old kernel before you tried out the new one, didn't you?

My suggestion is to boot up with a live rescue cdrom or a live linux cd like Knoppix.

I beleive that Knoppix has support for software raid.... Not sure though. I haven't messed around with software raid much before.

So you go and create a /etc/raidtab file. Maybe you can mount one of your disks and copy it over. Then once that is done, you start the raid up, mount it to a directory like /mnt/raid

then you go
cd /mnt/raid/
chroot ./

Chroot changes root. It makes a directory your in act like it's the operating system that your running, so you can then run commands like that.

then you edit your /etc/lilo.conf file and set it up to use the kernel that last worked.
then you run the lilo command
lilo

and then you exit out of the chroot enviroment, unmount the raid array, then reboot. And see if your old kernel still works, which it should.

Don't forget about setting up it's initrd if it had one.


but I dont' know even if you need to do all that, because I am not sure exactly what is going on, but that is generally what you do when your lilo configuration gets screwed up and you don't have a easy to use rescue cdrom thing that your distro setup for you to use to get out of these type of situations.