|
From: SourceForge.net <no...@so...> - 2005-03-19 14:51:12
|
Read and respond to this message at: https://sourceforge.net/forum/message.php?msg_id=3056475 By: szaka > I am using Mandrake 10.1 Community, kernel 2.6.8.1-10mdk, I do not > know which ntfs module version this kernel uses. Ok, it has: % dmesg | grep NTFS NTFS driver 2.1.15 [Flags: R/O DEBUG MODULE]. > it is bizarre that the BadClusteRs are not marked. Yes, it is indeed strange. Because strings and grep didn't work, you could also try ntfsclone -o - win-full.img | grep BadClusteR One of the error messages below indicates that the read error was detected and the content of the unreadable sector/cluster was filled by 0x02 so the BadClusteR should be there too. What I did: - checked the source tarball you downloaded has the right ntfsclone. It has. - checked the source code for a bug that would result missing this BadClusteR signature but I couldn't find any. - checked bad sector images but they were OK. > I see from my /var/log/messages that there were lot of I/O errors like > the one below during the cloning process. Am I doing something wrong? No, everything happens as it should be. The kernel tries to read the damaged disk surface and it may or may not succeed. > ---- > Mar 17 20:50:00 localhost kernel: hda: dma_intr: status=0x51 { DriveReady > SeekComple > te Error } > Mar 17 20:50:00 localhost kernel: hda: dma_intr: error=0x40 { UncorrectableError> }, > LBAsect=51461830, sector=51461614 > Mar 17 20:50:00 localhost kernel: end_request: I/O error, dev hda, sector > 51461614 > Mar 17 20:50:00 localhost kernel: Buffer I/O error on device hda3, logical block > 3127352 > --- Yep, bad sectors. > here is the error I get when I try to access the /mnt/loop/WINDOWS > directory. > Mar 18 17:01:17 localhost kernel: NTFS-fs error (device loop0): > ntfs_readdir(): Actual VCN (0x202020202020202) of index buffer is > different from expected VCN (0x0). Directory inode 0x1c is corrupt or > driver bug. Very interesting. The 0x202020202020202 clearly indicates that originally the kernel couldn't read the given sector so ntfsclone filled the buffer by this value. How is it possible that you can still see /mnt/windows/WINDOWS using the NTFS driver? You shouldn't be able to do that. Well, one explanation could be that the NTFS driver, the Linux VFS, etc tries harder/repeatedly to read the sector and it succeeds at the 2nd, 3rd, etc attempts. Ntfsclone tries only once. I thought the kernel takes care about these things but perheps this depends on subsystems and kernel versions. I need to check this out. What you could do? Run chkdsk /r then use ntfsclone. The /r option is VERY important. This should result a full list of bad sectors stored in a NTFS file and ntfsclone and the kernel shouldn't report any error unless bad sectors develop too fast. > thanks for your patience in answering my questions It was quite rewarding, we found several problems. I must also lie in bed, having high fever instead of realizing my plans, travelling around in Tunisia so currently I can't really do much anyway ... ______________________________________________________________________ You are receiving this email because you elected to monitor this forum. To stop monitoring this forum, login to SourceForge.net and visit: https://sourceforge.net/forum/unmonitor.php?forum_id=44084 |
