Reiserfs corruption problem.

Reiserfs corruption problem.

[Jean-Michel Bonnefond]

Hi all,

I have a 40Go ide disk that is mounted with one big partition containing all 
my clients datas on a linux debian woody. I was working as usual on the 
server when I suddenly couldn't no more list the contents of the root mount 
point (/part) but could continue working and looking into the files all under 
the root, (/part/distrib) and can cd into all the sub-directories. I just 
can't access to the root partition contents with an error message that I 
unfortunately haven't keep.

I decide to reboot the server in order to plug another ide disk to copy all 
data to this disk but, as soon as I umount the disk, I was unable to remount 
it. It replies me :
mount: is not a repertory (traduction of the french error message).

The reiserfsck do reply :

root@interne:~# reiserfsck /dev/hdc1
<--------reiserfsck 3.6.4, 2002-------->

Will read-only check consistency of the filesystem on /dev/hdc1
Will put log info to 'stdout'

Do you want to run this program?[N/Yes] (note need to type Yes if you do):Yes

******************************************************
* Warning: The dma on your hard drive is turned off. *
* This may really slow down the fsck process.        *
******************************************************
###########
reiserfsck --check started at Tue Mar  1 15:19:46 2004
###########
Replaying journal..
0 transactions replayed
Checking internal tree../  1 (of   2)bad_internal: vpf-10320: block 8384, 
items 5 and 6: The wrong order of items: [2 2310 0x52a2001 IND (1)], [0 2311 
0x69db001 IND (1)]
 the problem in the internal node occured (8384), whole subtree is skipped
finished
Comparing bitmaps..vpf-10640: The on-disk and the correct bitmaps differs.
Bad nodes were found, Semantic pass skipped
1 found corruptions can be fixed only during --rebuild-tree
###########
reiserfsck finished at Tue Mar  2 10:20:00 2004
###########
root@interne:~#

So I does a dd_rescue of the disk to another one wich hadn't notice any 
hardware problem, all the physical disk can be acceseed well. 
and then I launch a --rebuild-tree.

The first stage seem to pass correctly but the second pass loop infinitely on 
message :

do_pass_2: The block (level=28869, nr_item=905, free_space=49895 rdkey), 
marked as a leaf on the first two passe, is not a leaf! Will be skipped.


So do you have any suggestion on what I can do now and how I can acces again 
to my datas, even temporarily to copy them eslewhere? 

Thanks for your help.
Jean-Michel.

Re: Reiserfs corruption problem.

[Vitaly Fertman]

On Tuesday 02 March 2004 12:35, Jean-Michel Bonnefond wrote:
> Hi all,
>
> I have a 40Go ide disk that is mounted with one big partition containing
> all my clients datas on a linux debian woody. I was working as usual on the
> server when I suddenly couldn't no more list the contents of the root mount
> point (/part) but could continue working and looking into the files all
> under the root, (/part/distrib) and can cd into all the sub-directories. I
> just can't access to the root partition contents with an error message that
> I unfortunately haven't keep.
>
> I decide to reboot the server in order to plug another ide disk to copy all
> data to this disk but, as soon as I umount the disk, I was unable to
> remount it. It replies me :
> mount: is not a repertory (traduction of the french error message).
>
> The reiserfsck do reply :
>
> root@interne:~# reiserfsck /dev/hdc1
> <--------reiserfsck 3.6.4, 2002-------->
>
> Will read-only check consistency of the filesystem on /dev/hdc1
> Will put log info to 'stdout'
>
> Do you want to run this program?[N/Yes] (note need to type Yes if you
> do):Yes
>
> ******************************************************
> * Warning: The dma on your hard drive is turned off. *
> * This may really slow down the fsck process.        *
> ******************************************************
> ###########
> reiserfsck --check started at Tue Mar  1 15:19:46 2004
> ###########
> Replaying journal..
> 0 transactions replayed
> Checking internal tree../  1 (of   2)bad_internal: vpf-10320: block 8384,
> items 5 and 6: The wrong order of items: [2 2310 0x52a2001 IND (1)], [0
> 2311 0x69db001 IND (1)]
>  the problem in the internal node occured (8384), whole subtree is skipped
> finished
> Comparing bitmaps..vpf-10640: The on-disk and the correct bitmaps differs.
> Bad nodes were found, Semantic pass skipped
> 1 found corruptions can be fixed only during --rebuild-tree
> ###########
> reiserfsck finished at Tue Mar  2 10:20:00 2004
> ###########
> root@interne:~#
>
> So I does a dd_rescue of the disk to another one wich hadn't notice any
> hardware problem, all the physical disk can be acceseed well.
> and then I launch a --rebuild-tree.
>
> The first stage seem to pass correctly but the second pass loop infinitely
> on message :
>
> do_pass_2: The block (level=28869, nr_item=905, free_space=49895 rdkey),
> marked as a leaf on the first two passe, is not a leaf! Will be skipped.
>
>
> So do you have any suggestion on what I can do now and how I can acces
> again to my datas, even temporarily to copy them eslewhere?

Obtain the latest reiserfsprogs from our ftp site and try it again please.

-- 
Thanks,
Vitaly Fertman