[Bug-ddrescue] Fwd: Verify recovered data

[Zeniff Martineau]

Florian Sedivy wrote:
> Am 04.02.2014 um 03:42 schrieb Zeniff Martineau:
>
>> * Flash drive (source)'s version:
>> ...5b6db6bdfa0c71e3ed952c48fd72249e28e11465c6eb4fba49a41c  ./Windows...
>> * Some, but not all, images made by ddrescue wrote this:
>> ...5b6tb6bdfa0c71e3ed952s48fd72249e28e11465c6eb4fba49a41c  ./Wandowc...
>> (Note the "t" and "s" in the checksum...)
>
>
> You don't have a wrong checksum here, you have a display problem.
>
> As you noticed yourself, there are "t" and "s" in the checksum, which cannot
> be in a hexadecimal hash. Other than those wrong characters the checksums
> are suspiciously similar, which would be too much of a coincidence. The
> conclusion is, those checksums are actually identical, but at least one of
> them is displayed with errors.
>
> Look for problems in your terminal connection or your RAM. You experience
> bit-flips: d-t, s-c and i-a are all pairs with just 1 bit difference. Once
> you eliminated that problem in your terminal you of course have to make sure
> that the actual data you copied was not affected by bit-flips as well.


What should I do to test this?

I don't have much knowledge in this field, sorry. :(

In this case I was using midnight commander's mcdiff utility.

I used mcdiff to swap the panels back and forth, but the wrong
characters followed with the side of the terminal with the copy, not
the original, so I don't feel there is a display issue, unless I
misunderstand your meaning?

I also remounted and copied the same file 20 times, and no 2 file
copies were the same, but they were usually very similar.

The copies varied in different places, but in places where the
variations occurred, they were the same places and with the same
variations characters, at least as far as I could tell.

Would it be a good idea to post here some partial diff utility output?
(I don't want to clutter up the post if it's unhelpful, sorry)

Also, I tried making the ddrescue with 2 different laptops to 2
different destination HDDs with the same results.

The only similarity between the copies that I can think of was that
both computers were running updated versions of Arch Linux with
ddrescue 1.17.

If you have ideas how I can test further for terminal/RAM/etc
problems, that'd be awesome~:)


Thank you and have a great day~:)