"pcg"@goof.com ( Marc) (A.) (Lehmann ) wrote:
> > > Yes. It should move readable data to good disk and run --rebuild-tree.
> > > Please keep in mind that dd does not copy data properly when input file
> > > contains bad blocks. It looks like it does not make seek on output
> > > device.
> > > I usually run badblocks first and then copy data with dd around bad
> > > blocks manually.
> >
> > Oh, my. Why people do everything the hardest way? Woudln't it be easier for a
> > filesystem developer to simply patch dd (perhaps adding a parameter) instead
> > of going through all that?
>
> What's wrong with dd? Why is a seek on the output device required? I'd
> think that zeroing the block (as dd does) would be more productive than
> letting some random data survive.
>
The problem is that often multiple adjacent blocks are bad. You'll have to detect
them manually. Once you know the bad blocks, just trying to overwrite them usually
does not succeed because the disk wants to seek to that block exactly (which does
not work for the same reason the block is bad). But if the whole track is
rewritten, the bad blocks usually are gone.
I've finally developed a program that eliminates such bad blocks and retains the
data of the blocks which are not bad. If someone needs it, I'm happy to send it.
--
Mit freundlichen Gr��en
Xu�n Baldauf
Medium.net Internet Server Software
