Hi Simon,
[email protected] wrote:
>> For me, the issue still is to identify whether there is
>> a bug in ntfsclone or if the image got somewhat
>> corrupted for another reason.
>>
>> Did you initially compress the image ? If you could
>> decompress it, the corruption probably occurred
>> earlier than compression, otherwise you would not
>> be able to decompress.
>>
> No I did not compress the image.
> As far as I can remember, I just attached the harddisk to a blank machine and
> booted the "SuSE Rescue System" via Network. Then i mounted an NFSv3 share
> to /mnt/data (nolock,tcp,hard,rw,async).
> Afterwards I think I used this command line to invoke ntfsclone:
> ntfsclone -s /dev/sda1 -O /mnt/data/sda1.ntfs
>
Nothing wrong there.
> ntfsclone did not report anything strange, that would have caught my attention
> during the process. I used ntfsclone before and everything seemed as usual.
>
> Btw. the NTFS fileystem was created by, or at least used by "Windows Vista".
> (I'm not sure if Vista created the filesystem or some OEM with a prebuild
> setup)
> ntfsclone reports "NTFS volume version: 3.1"
>
I do not think the issue is related to Vista : I have
backup'ed and restored a Vista system partition with
no problem.
> Also I don't know yet, whether the source NTFS filesystem had errors.
> As far as i know the machine was powered down correctly before the image was
> taken. (I don't have access to the machine any more and the harddrive got
> formated - thats why I needed the data ;-) )
>
>
>>> To me these errors look a bit as if the variable, that holds the code to
>>> write in ntfsclone, doesn't get initialized all the time for code 0x1
>>> blocks, because most of the time ntfsclone just wrote a 0x0 were a 0x1
>>> would belong.
>>>
>> This looks like a bug in ntfsclone, though such
>> error pattern has not been reported so far.
>>
>> If this error pattern is caused by a bad ntfsclone,
>> the good news is that an auto-repair can be built.
>>
> Indeed!
> Your check() function provided the core part of that program.
> It works exceptionally well I have to admit....
>
>
>>> But in a few cases there were other values (0xC8) there. But this is just
>>> a wild guess. And of course the image was created with the old version of
>>> ntfsclone that was shipped with the openSuSE 12.1 x64 DVD. Maybe it is
>>> already gone in the newest version....
>>>
>> Can you be more precise about the ntfsclone version
>> used to build the image ? ("ntfsclone -h" will display it)
>> There has been a major port in 2010 from a dead
>> project, and some code cleaning was done at the
>> time.
>>
> On the openSuSE 12.1 x64 Rescue System:
> ntfsclone v2011.4.12 (libntfs-3g)
>
This version was released by ntfs-3g, and if it is
buggy, the bug is probably still present (unless
caused by a patch specific to openSuSe).
> On my desktop:
> ntfsclone v2012.1.15 (libntfs-3g)
> (with your patch applied to it!)
>
This patch was intended to locate a possible
corruption, it does not fix anything.
>>> I could now try to create a new ntfsclone-image from the raw-image and
>>> see if the errors persist?!
>>>
>> Yes, please, try to create an image again with the
>> original version, and if you got a corrupted image,
>> try again with a recent version.
>>
>>
> OK. I do that now. I will report back as soon as I know more....
> I will also run "chkdsk.exe" against a copy of the raw-image, to see whether
> it had errors.
>
Thanks.
Jean-Pierre
------------------------------------------------------------------------------
Everyone hates slow websites. So do we.
Make your web apps faster with AppDynamics
Download AppDynamics Lite for free today:
http://p.sf.net/sfu/appdyn_sfd2d_oct
_______________________________________________
ntfs-3g-devel mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/ntfs-3g-devel
