[olug] Superblock corruption on ext3

[Bradshaw Terence]

Hello again, all...

My screensaver problem is gone.  Along with the rest of my former Linux
installation.

I can't recall exactly what I was doing when my system crapped out, but
crap out it did.

My system:

Mainboard:  Biostar M7VIB (built in audio, but disabled)
CPU:  AMD Athlon XP 1700+
256MB DDR memory
Maxtor ~13GB hard drive (3 partitions, root ~4GB ext3, ~500MB swap, home
~8GB ext3)
Nvidia GeForce2 MX400 (manufactured by PNY)
Soundblaster Live audio
Linksys 10/100 ethernet card
Mandrake 9.0 exclusively on the machine - no Windows to goof things up.

The sequence of events: I used Mandrake Update to look for bugfixes and
security updates, and there were some security updates, which I downloaded
and installed.  Some time after that, I tried to fire up Armagetron...
nothing happened.  So, I fired up a console and tried to run glxgears.  It
reported a problem with glx permissions being too tight.  Per Nvidia
documentation, I modified a security file to remove a <dev> line, issues a
chmod and a chown to make the nvidia stuff usable again, and glxgears ran
like a champ, as did Armagetron.

For the life of me, I can't recall what I was doing when the system went
south.  I just remember waiting longer than expected for something to
happen, the drive was working like mad, and then the display became
totally garbled.  <CTRL><ALT><BACKSPACE> did nothing.  <CTRL><ALT><DELETE>
did nothing.  I got out my laptop, and was able to SSH into my Linux box. 
I did a "shutdown -r now" and called it a night.  When I went to use my
machine this morning I was greeted with a message about a corrupt
superblock, that the filesystem was not a proper ext2 filesystem...
instruction of using an alternate superblock, etc.  I couldn't get fsck to
run against the drive.  The instructions to retrieve an alternate
superblock did not work for me.  A command I found on Google to find
alternate superblocks also did not work.  It just kept saying that the
device did not contain a valid ext2 filesystem.

Since I had nothing critical on my system, I decided to reinstall from the
ground up.

Several questions:

First, what is likely to be causing this?  The updates I installed?  Drive
failure?  BIOS settings?  Hardware/Kernel misconfiguration?  Other
software on the system?  Crappy hardware?  (It's a cheap motherboard with
VIA chipsets).  I'm willing to drop money into this box to make it more
Linux friendly.  I will not go back to Windows on this machine.

Second, why do I need to use fsck anyhow (that's what the error message
was saying to do)?  I thought the journaling of ext3 was supposed to get
you around the fsck usage.

Third, why was it insistent that I should have an ext2 filesystem on the
device, when I had an ext3 filesystem?

Last, how should I handle this problem in the future (shoud it happen at
all)?  A reinstall is a stupid way to handle it, but I just wanted it
running again.  In the long run, I want to know how to handle it the right
way.

I've had the same superblock corruption issue arise in Mandrake 8.1 and
8.2 as well.  I never had this trouble when I was running SuSE, but it was
on different hardware, so I'm not sure it's a fair comparision.

Sorry it's a long post, but the questions are many and the problem is
frustrating.

I'm looking forward to meeting some of you at the next OLUG meeting.


Terry