Hi Heiko,
On Thu, 14 Aug 2003, Heiko wrote:
> 1) I've run ntfsmeta -di8 on my partition and got :
>
> VCN LCN Run length
> 0 LCN_HOLE abbeec
> abbeec LCN_ENOENT 0 (runlist end)
So NTFS didn't mark any cluster being faulty one.
> 2) I've also had the nerve to experiment once more to find out if the
> ntfs-driver freezes and found out that the rattling disk noice stops after a
> few minutes (never wanted to wait that long and didn't expect it to stop
> anyways) and the drive can be accessed after that again (with the same
> results if getting bad reads).
>
> Complete log aftter executing ls on my root directory after mounting :
>
> NTFS driver 2.1.0a [Flags: R/O MODULE].
This driver is old and has several bugs, e.g. it can very easily hang.
Anyway this has nothing to do with your problem, I just thought I mention.
> NTFS volume version 3.0.
> hde: dma_intr: status=0x51 { DriveReady SeekComplete Error }
> hde: dma_intr: error=0x40 { UncorrectableError }, LBAsect=46568583,
> sector=46568520
> end_request: I/O error, dev 21:01 (hde), sector 46568520
This is the IDE (disk) driver. It can't read the disk.
> NTFS-fs error (device 21:01): ntfs_end_buffer_async_read(): Buffer I/O error,
> logical block 46568520.
The hardware error propagated to the NTFS driver.
> hde: dma_intr: status=0x51 { DriveReady SeekComplete Error }
> hde: dma_intr: error=0x40 { UncorrectableError }, LBAsect=46568584,
> sector=46568521
> end_request: I/O error, dev 21:01 (hde), sector 46568521
Next hardware error, and this repeats further on.
> 3) Experimenting with dd again on a memorized single bad sector gave a good
> read ! A mass read of 10 sectors gave a bad read. But I know I've got a bad
> read when I tested it with this single sector days ago ! Testing some more
> and I got the following :
>
> - reading more than 1 sector heightens the chance of a bad read considerably
> but only in critical regions like 46550000-46615000
> - every sector can probably be read doing single dd reads (count=1)
> - the more sectors are read in sequence the longer the freeze
Interesting. Maybe you should let this know to the IDE driver maintainer.
Also your IDE chipset, disk and kernel version and preferable you should
also try the latest kernel, maybe they improved the situation. However if
Windows had the same problem, I doubt much more could be done ...
> I'm not sure my drive is damaged at all. Could it be the controller ? But why
> does it happen always in the middle drive region ? Do you have any idea ?
Well, maybe the controller, maybe the IDE cable, maybe the hot weather, etc.
> Can you write the tool ntfscat you mentioned ? I know now all my other files I
> want to save but can't access via the damaged directory tree.
Ok. But unfortunately I couldn't do it now because, as a bad habit, the
ntfsprogs build was broken again in most development environments:
% ./autogen.sh
processing .
Running aclocal ...
Running autoheader...
/usr/bin/m4: configure.in: No such file or directory
Running automake --gnu ...
configure.ac: 5: required file `./[config.h].in' not found
Running autoconf ...
autoconf: configure.in: No such file or directory
**Error**: autoconf failed.
and at present I don't have time to play with this.
Szaka
|