09.10.2011 01:59, Nicolas Bercher wrote:
> I was simply unable to mount/check/restore the clone
> xfs volume after cloning it with dd.
Most common problem with using dd to close faulty volumes is that people use
conv=noerror and not conv=noerror,sync which results in DD not writing zeroed
blocks to the copied image in place of unreadable source blocks. Also copying
process from faulty source should be iterative, starting with full image copy
using big block size and then trying to retry copying bad block using 512 bytes
size. Using specialized tools like dd_rescue are also of a big help.
I had one case when I had been facing faulty harddrive that had been returning
garbage data when trying to read bad sectors instead of reporting the system
about unreadable block. Key point to dig the data from it was to create ~20
copies of disk image from it and then use statistical methods to determine which
data for each sectors to use.
--
Best regards,
Alexey Loukianov mailto:mooroon2@...
System Engineer, Mob.:+7(926)218-1320
*nix Specialist
Attachments: