Unable to mount / (ext error 22) [plug]

[Dean Holland]

Why is it stuff like this only comes to you after a 15 hour shift at work!
:)

I was able to mount the partition successfully as ext2, ran lilo and
rebooted - still got a kernel panic, but this time I saw something I hadn't
last time:

journal inode not found on IDE(0:3)

back into rescue mode, i noticed that when fsck was spamming the console, it
was the same group of dup/bad inodes, and wasn't deleting/cloning them.

tune2fs (re)creating the journal allowed me to mount the partition as ext3,
but I am still getting kernel panics on boot, and fsck still loops on the
same files when scanning for dup blocks.

I can actually get to the partition in rescue mode now, so I'll probably
just recover what I can and format :)

Thanks again!
Dean

----- Original Message ----- 
From: "speedster" <speedster at westnet.com.au>
To: <plug at plug.linux.org.au>
Sent: Tuesday, June 17, 2003 9:16 PM
Subject: Re: Unable to mount / (ext error 22) [plug]


> Whoops, forgot the subject... (using a web-based email atm)
>
> >Quite likely :-(
> >If you did a "fsck -n -f" it'll be fine, but if you did a read/write
> >fsck things might not be so happy. 13 hours strikes me as an
> >amazingly
> >long time... something funny must be going on.
>
> Yeah I thought so too. During the inode pass there was a lot of disk
> activity, but during directory scan there's little to none - and fsck
> can't be interrupted with a ctrl-c (as a last resort). I did stop it
> last night and on starting it this morning it didn't look like it
> actually fixed anything (bar the bad superblock).
>
> >OTOH, it might not be disk geometry that's the problem - in which
> >case
> >all bets are off.
>
> The fact it actually prints the filenames (eg
> /usr/lib/python-2.2blahblah) to me indicates it is reading the
> directory structure properly.
>
> I managed to get a copy of my DNS zone from my secondary ns (bind
> forever!) so the only real loss is my website (not as though that was
> really important though!)
>
> ---- Original Message ----
> From: craig at postnewspapers.com.au
> To: plug at plug.linux.org.au,
> Subject: Re: [plug]
> Date: Tue, 17 Jun 2003 21:01:20 +0800
>
> >> I have a funny feeling that the BIOS detected the wrong translation
> >> type (LARGE) for the disk, instead of using CHS or LBA
> >(unfortunately
> >> I can't remember what it was using on the old P166), and if so -
> >has
> >> fsckíng the drive possibly destroyed the data completely?
> >
> >Quite likely :-(
> >If you did a "fsck -n -f" it'll be fine, but if you did a read/write
> >fsck things might not be so happy. 13 hours strikes me as an
> >amazingly
> >long time... something funny must be going on.
> >
> >> Just want
> >> to know if there's a possible way I could retrieve anything off the
> >> partition before formatting it. (I had my website, DNS zone files,
> >> firewall config, dhcp config etc. on there, no backup)
> >
> >Hard to say. Maybe if you can get the geometry right then mount the
> >partition - otherwise, I can only think "disk editor".
> >
> >OTOH, it might not be disk geometry that's the problem - in which
> >case
> >all bets are off.
> >
> >
> ___________________________________________________
> "All your base are belong to us"
>
> Dean A. Holland (Speedster)
> Ph: 0407 778 027
> email:
> speedster at westnet.com.au
> dholland at student.ecu.edu.au
>