reiserfsck problem

[svar]

I had some file corruption and could (not even as root)
even ls -l some files. I tried

reiserfsck --rebuild-tree /dev/vg0/varnew

and after a long time of printing
[fccc000] etc to the screen it stopped

I rebooted and it says that fsck.reiserfs on /dev/sdc6 (i.e. /)
failed(status ff) Run it manually

root is mounted read only
use bash# mount -n -o reount,rw /

then I tried fsck -t reiserfs /dev/sdc6
and it always says this contais a mounted fs,
that fsck.reiserfs on /dev/sdc6 (i.e. /)
failed(status ff) Run it manually

and run it manually
fsck.

what should I do?

Thanks, svar

 

[elgato1906]

Don't mount the file system first. just run fsck on the partition while it's un mounted.

 

[svar]

Hmm. Yes, butthis is root. How do I unmount root?

 

[elgato1906]

boot from a rescue disk. What distro are you using?

 

[ifincham]

Hi,

Not too sure about if there are differences with reiserfsck but certainly ordinarily a filesystem check is done with the filesystem mounted as read-only. Fairly obviously, this is to prevent any process adding to damage that may already be there. The usual way you do this is boot to maintenance / single user level (or init to the same runlevel if already running). Then run fsck. For redhat you'd want run-level 1 which you can reach either by booting as follows (do control-x to exit the lilo graphical boot screen) :

boot:linux 1

Otherwise you can do 'init 1'. Suse would be similar except you need run-level 'S' instead of '1'.

Also, as far as I recall, 'fsck.reiserfs' is just a symbolic link to reiserfsck so all invocations of fsck referring to reiserfs end up running reiserfsck. Online man pages are here -->

http://www.namesys.com/manp.html

.

Regards

 

[svar]

Well, I just noticed that the rescue disk is unusable.
I tried init S and init 1 but I get the same error
-says I need to run it manually.
Does this mean I need to reinstall?

Thanks, svar

 

[ifincham]

Hi,

Not necessarily... If you get to single user mode and it tells you to run fsck then thats what you would do. The filesystem should be mounted as read-only (just enter 'mount' to find out). If it is mounted r/w then just do a read-only remount :

mount /dev/scd6 -o ro,remount /

Then run the actual fsck...

fsck.reiserfs /dev/sdc6

If fsck (reiserfsck) can't fix it then it does look like a re-install I'm afraid.

Regards

 

[svar]

Hmm. I do
mount /dev/scd6 -o ro,remount /
and there is no complaint, but when I do
mount

it says /dev/sdc6 on / type reiserfs(rw)

proc on /proc type proc(rw)

and then of course fsck fails saying this is a mounted system

any ideas??