Not booting correctly into SuSE 10.3
 

"Something" happened to my friend's computer, and now he somehow expects me to know how to fix it. Sigh...

The symptoms:
There are two partitions - a Windows XP partition and a larger partition for SuSE 10.3 it seems. He can boot into the Windows, but not into the SuSE anymore. However, he can boot using the Fail-safe method.

I know this is very vague. Please forgive me, but I can't get to the computer at present to check things for myself. Do any of you know what a possible cause of these symptoms could be and how to fix the problem?

Greetingz!
A few possibilities;
1) Your "friend" is running Microsoft products, and isn't really a friend.
2) "Friend" attempted to do something to the Linux parititon (probably while playing around with "Disk Manager" or something)
3) GRUB freaked, needs a tweak.
4) Your "Friend" was actually using openSuSE and ran an update, the kernel was updated, but not GRUB, and *bam*. Failure to boot (Linux).
5) His mom/sister/dad was at "one of those websites" and now the whole thing will need a reload.

Lemme know which one's right.

Without really informational inputs and errors, how is anyone to device what has happened and what went wrong? And did you ask your *friend* what did he do before the system borked? An update? An upgrade? Deleted some files? What? What is to be assumed here? It makes sense to get all the errors you see on the screen and also to get the logs from the system if he can boot into single user mode.

If you don't know what error it is!!!? Take a photo of the error and attach it.

Hi!

I expected some sarcastic replies. Hehe. Yeah, and I understand your frustration.

Okay, I have looked at the computer myself, and this is more or less what it spat out at me:
It asked for the root password to login, which I did and so I have:
Oh yes, it's actually SLES 10 SP2. Don't know whether it makes a huge difference.
My friend has a lot of work related data on the computer.

Does the above-mentioned info help a bit?
It's obvious that there's something wrong with the filesystem. Duh. But I don't know how to run fsck manually or where to go from here?

Thanks!

It appears that the /var file system is located on its own partition and it is not clean. The boot process wants to run the file system checker (fsck) but it is not allowed to do that because /var is a system directory. This is a safety feature in the boot process. The boot scripts require this partition to be checked manually. This is probably not too bad.

When the boot process stops you should see the root interactive prompt. What you need to do is figure out what partition holds the /var file system and run fsck manually.

1) Find the partition that holds /var
2) Run fsck on that partition. The following example shows how to run fsck on the /dev/sda6 partition. Your system may have the /var directory on a different partition.
If fsck successfully finishes then reboot the machine.

If fsck cannot check the file system then there may be a problem with that partition such as having the file system erased. This kind of problem may be unrecoverable unless a backup exists. There is no point in speculating unless the procedure that I outlined fails to repair the system software.

Thanks stress_junkie!

Um, I suppose that should give me some output, so that I know on which partition it is?
Well it doesn't.

Any other idea how I can check it?

Open a terminal and run this command: cat /etc/fstab
That should show you the contents of the fstab file.
If it doesn't....?

gives me:
If I do
I get:
If I do (sda3 or sda4)
I get:
And a paragraph that says that if sda3 is really valid, then the superblock is corrupt.

If I look in /Code or /Work, then it is empty, but I know there should be data there.

Is this very bad news for my friend?

That fstab doesn't look as I had hoped it would look. I was hoping to see something like /dev/sda6 /var ...

Ask the owner of the machine if he/she created any special disk configuration.

Maybe this /var thing is a red herring. I am wondering if the machine has LVM and maybe software RAID. Maybe not. Looking again at the boot messages in reply #5 I see that the root file system is already mounted read-write so it appears to be okay.

I think we need more information about how the machine was set up.

This is probably one of those situations where someone who is very familiar with Linux could boot a live CD and see how the machine was originally configured and determine what needs to be done.

Hi!

Sigh... This is more intricate than you expected. Perhaps I should explain a bit more. My friend happens to be a professor in chemistry, and the sick computer was used for small scale computational chemistry calculations. I am pretty sure that the professor wasn't responsible for the set-up of the computer. So asking him wouldn't help. The person who probably set it up, is not very friendly, but I know that he has something against RAID so I doubt that he would have used it. I vaguely recall that some of the boot-messages said that there wasn't any LVM.

/var is not on it's own partition. It's clumped together with everything else except for /Code and /Work under /.

Hmmm, I have a GParted LiveCD. When I manage to get to the sick computer again I will pop it in and let you know what I see.

Thanks,
;-)

Hi!

Okay, I've been silent because there have been further problems. It doesn't allow me to change the boot order. I go to the advanced BIOS settings and change it, but when I save that and look at the Boot order (on another screen, - I don't know the technical name for it) it still maintains that the harddrive is the first option, and that is what is seen in practice. I don't manage to boot from the GParted LiveCD or the SuSE installation disk.

One of my friends suggested that I take the harddrive out and copy it to somewhere using another computer. I actually just want the data under /Code and /Work otherwise I would've formatted long ago. If I log in as the root as mentioned above, /Code and /Work is empty. Can I believe the computer that it is really empty? I know there should be data there!

Is there a way that I could get into the Linux partitions from the Windows side?

Could you try "Disabling" the other Boot Options? To where "1st Boot Device" is "CDROM" and the others are "DISABLED"?
(Depending on the Make/Model of the Mobo/system, the lingo could be different)

If you login as root, then do the following;
"grep -i code /etc/*fstab" and see what it tells you. Odds are good that /code and /work are different filesystems mounting from somewhere else.

Yes, but I would not recommend you go that route (as you may not get all of the data you're after). Here's what I would suggest;
1) Take a Linux LiveCD (like Knoppix)
2) Walk up to different unsuspecting computer
3) Shut it down safely (or just rip out the power-cord)
4) Disconnect all cables from all hard drives in the unsuspecting system (might want to note what-goes-where beforehand)
5) Rip all hard drives out of the borked system and plug them in (might want to note what-goes-where beforehand)
6) Boot unsuspecting system, go into the BIOS, make sure "Boot From CD-ROM First" is set. (Lingo might be different, but you get the drift)
7) Reboot, insert Linux LiveCD.
8) Login as root, do an "fdisk -l /dev/[hs]d[a-z]|grep ."
9) Go over the output, find out what partitions of what type are where, then mount (read-only!) the partitions you need.
10) Plug in external media such as a USB stick or USB HDD, and copy the data you need.
11) Undo all of what you did.
12) Have a beer.

I could get into the Linux side from the Windows, but there was too much data to copy everything successfully. Eventually I got an external harddrive case from a friend, and copied the files to another computer since the external case allowed me to treat the harddrive as a USB flashdisk. (I guess it's not technically correct to say that, but okay, you get the picture.) Then yesterday my harddrive at home seemed to have crashed and I thought I'd try this fsck method before I start loosening screws, and it worked! So thank you very much! :-D