bug: reiserfsck reports "neither new nor old reiserfs format found"

bug: reiserfsck reports "neither new nor old reiserfs format found"

[Moritz Franosch]

Hello,

I think I've just lost a reiserfs partition and reiserfsck did not
work. Here is my bugreport, as the man page of reiserfsck says.


The system is Intel 486, 66 MHz, 16 MB RAM.
/dev/hda1 boot partition with ext2
/dev/hda2 root partition with reiserfs

I've installed the system with Suse 8.0, the system booted (many
times), then hang and I switched off.


The next boot, the kernel paniced with the message

read_super_block: can't find a reiserfs filesystem on (dev 03:02, block 64, size 1024)


Thus I've downloaded reiserfsprogs-3.6.3.tar.gz and did

tar -xzf reiserfsprogs-3.6.3.tar.gz
cd reiserfsprogs-3.6.3
./configure
make
cd fsck
gcc --static -g -O2 -Wall  -o reiserfsck  main.o pass0.o pass1.o pass2.o semantic_check.o semantic_rebuild.o pass4.o lost+found.o ubitmap.o uobjectid.o ustree.o ufile.o check.o check_tree.o info.o super.o ../lib/libmisc.a ../reiserfscore/libcore.a

to get a statically liked version of reiserfsck (unser Suse Linux 8.0,
gcc 2.95.3, Suse's default 2.4.18-4GB kernel).


I copied reiserfsck via floppies to the ext2 filesystem on /dev/hda1
(using a rescue system, Suse's did not work because of too few
memory).


I booted the Suse bootdisk (Suse kernel 2.4.18-4GB), loaded the
reiserfs module, went to the console, mounted /dev/hda1 and executed
reiserfsck as

reiserfsck --check --logfile check.log /dev/hda2

It reported (I've to type from the screen):

reiserfs_open: neither new nor old reiserfs format found on /dev/hda2
reiserfsck: could not open filesystem on "/dev/hda2"

"echo $?" reported 0 as exit status, check.log has size 0.


According to the man page, reportig 0 as an exit status in this case
is a bug:
man reiserfsck says:
  2.  Run reiserfsck --check --logfile check.log /dev/hda1.  If
  reiserfsck --check exits with status 0 it means no errors were
  discovered.


reiserfsck --rebuild-tree --logfile rebuild.log /dev/hda2

reiserfs_open: neither new nor old reiserfs format found on /dev/hda2
reiserfsck: could not open filesystem on "/dev/hda2"

rebuild.log has size 0.


Trying to mount /dev/hda2 with

mount /dev/hda2 /mnt2

gives the following kernel messages on the console:

...
VFS: Can't find ext2 filesystem on dev ide0(3,2)
read_super_block: can't find a reiserfs filesystem on (dev 03:02, block 64, size 1024)
read_super_block: can't find a reiserfs filesystem on (dev 03:02, block 8, size 1024)
...


I think reiserfsck should do _something_.
man reiserfsck says:
  6. If the --rebuild-tree step fails or does not recover what you
  expected, please submit this as a bug report. Try to provide as much
  information as possible and we will try to help solve the problem.

The --rebuild-tree step failed and did recover absoulutely nothing, so
I report this as a bug.


I suspect the magic number of the reiserfs filesystem is overwritten
and reiserfsck gives up. But there really _is_ a reiserfs filesystem
on that partition. I'll save the first 1.4 MB of /dev/hda2 on a floppy
for further analysis, if required.

Please CC me, as I'm not on the reiserfs-list@namesys.com.


Thank you,

Moritz



-- 
Dipl.-Phys. Moritz Franosch
http://Franosch.org

Re: bug: reiserfsck reports "neither new nor old reiserfs format found"

[Vitaly Fertman]

Hi, 

> I've installed the system with Suse 8.0, the system booted (many
> times), then hang and I switched off.
>
> The next boot, the kernel paniced with the message
>
> read_super_block: can't find a reiserfs filesystem on (dev 03:02, block 64,
> size 1024)
>
> Thus I've downloaded reiserfsprogs-3.6.3.tar.gz and did

Please take the last 3.6.4 version of reiserfsprogs.

> reiserfsck --check --logfile check.log /dev/hda2
>
> It reported (I've to type from the screen):
>
> reiserfs_open: neither new nor old reiserfs format found on /dev/hda2
> reiserfsck: could not open filesystem on "/dev/hda2"

Your super block got corrupted somehow or you have your partition table 
changed so that the blocknumber of the superblock was changed.

To rebuild the corrupted superblock you can use reiserfsck --rebuild-sb.

> "echo $?" reported 0 as exit status, check.log has size 0.
>
> According to the man page, reportig 0 as an exit status in this case
> is a bug:
> man reiserfsck says:
>   2.  Run reiserfsck --check --logfile check.log /dev/hda1.  If
>   reiserfsck --check exits with status 0 it means no errors were
>   discovered.

We decided to not provide a special exit code for corrupted super block
to avoid an automation of super block recovery. This is a difficult case 
and should be done manually. 

>
> Thank you,
>
> Moritz

-- 

Thanks,
Vitaly Fertman

Re: bug: reiserfsck reports "neither new nor old reiserfs format found"

[Moritz Franosch]

Vitaly Fertman <vitaly@namesys.com> writes:

> > Thus I've downloaded reiserfsprogs-3.6.3.tar.gz and did
> 
> Please take the last 3.6.4 version of reiserfsprogs.

I've the new version now.

> > reiserfsck --check --logfile check.log /dev/hda2
> >
> > It reported (I've to type from the screen):
> >
> > reiserfs_open: neither new nor old reiserfs format found on /dev/hda2
> > reiserfsck: could not open filesystem on "/dev/hda2"
> 
> Your super block got corrupted somehow or you have your partition table 
> changed so that the blocknumber of the superblock was changed.
> 
> To rebuild the corrupted superblock you can use reiserfsck --rebuild-sb.

Thank you very much for that hint. It's documented in 'man
reiserfsck', but I've overlooked it. I've tried --rebuild-sb but it
reported that it couldn't find the journal. Then I've done

reiserfstune --make-journal-standard

and also experimented with --no-journal-available in various
combinations with --rebuild-sb and --rebuild-tree but I've had no
success.

Thank you very much for your quick reply.


I've never had a single problem I could blame reiserfs for, the only
partition under Linux I've lost so far was an ext2.

But I've the feeling that e2fsck is much simpler: Just execute it and
answer everything with "yes", you don't understand it anyway. Then
your filesystem is either gone or still there and you can look in
lost+found for important data. No need to experiment with
--do-this-and-that --without-doing-that.


Thank you very much again,

Moritz


-- 
Dipl.-Phys. Moritz Franosch
http://Franosch.org

Re: bug: reiserfsck reports "neither new nor old reiserfs format found"

[Hans Reiser]

Vitaly Fertman wrote:

>Hi, 
>
>  
>
>>I've installed the system with Suse 8.0, the system booted (many
>>times), then hang and I switched off.
>>
>>The next boot, the kernel paniced with the message
>>
>>read_super_block: can't find a reiserfs filesystem on (dev 03:02, block 64,
>>size 1024)
>>
>>Thus I've downloaded reiserfsprogs-3.6.3.tar.gz and did
>>    
>>
>
>Please take the last 3.6.4 version of reiserfsprogs.
>
>  
>
>>reiserfsck --check --logfile check.log /dev/hda2
>>
>>It reported (I've to type from the screen):
>>
>>reiserfs_open: neither new nor old reiserfs format found on /dev/hda2
>>reiserfsck: could not open filesystem on "/dev/hda2"
>>    
>>
>
>Your super block got corrupted somehow or you have your partition table 
>changed so that the blocknumber of the superblock was changed.
>
>To rebuild the corrupted superblock you can use reiserfsck --rebuild-sb.
>
>  
>
>>"echo $?" reported 0 as exit status, check.log has size 0.
>>
>>According to the man page, reportig 0 as an exit status in this case
>>is a bug:
>>man reiserfsck says:
>>  2.  Run reiserfsck --check --logfile check.log /dev/hda1.  If
>>  reiserfsck --check exits with status 0 it means no errors were
>>  discovered.
>>    
>>
>
>We decided to not provide a special exit code for corrupted super block
>to avoid an automation of super block recovery. This is a difficult case 
>and should be done manually. 
>
Tell users that when fsck runs and finds a bad super.

>
>  
>
>>Thank you,
>>
>>Moritz
>>    
>>
>
>  
>