mailing wrote:
> > > > I would hurry to backup data from the disk before everything is lost. It
> > > > seems like a bad disk to me.
> > > >
> > > OH!!!!! Really?? You mean it is not a single partition problem but physical
>error of whole disk???
> >
> > /dev/sda9 has problems at least. And so far reiserfs does not feel itself
>comfortable enough on such devices as it does not have well tested bad block
>accounting system. The point (afaik) is that when hard drive starts complaining about
>bad sectors -
> > that means that it is unable to remap then and therefore is close enough to start
>getting more and more bad sectors. I am sure that it is not always true, because I
>saw harddrives working for years with few bad sectors, but I hope you got the point
> > reisrefs is on for now.
> >
> > Thanks,
> > vs
> >
> >
> Yes, I just use badblocks to check and it contain many bad sector:
> [root@desktop http]# badblocks /dev/sda9
> 174392
> 174393
> 174394
> 174395
> 174396
> 174397
> 174398
> 174399
> 174400
> 174401
> 174402
> 174403
> 174404
> 174405
> 174406
> 174407
> Can I mark it and keep it unuse at reiserfs?
> --
>
> Carfield Yim, visit my homepage at http://www.carfield.com.hk
It seems that one track cannot be read anymore. Because it is not one block, but a
whole track, I suppose that track has been "overwandered" by other tracks. (Track
wandering happens when writing the same track many times, due to a slight positioning
bias
of the disk head each write.) For some of my drives, such problems could be solved by
rewriting from some tracks before the overwandered track to some tracks after the
overwandered track multiple times.
So in theory, you could create a disk image from blocks 174000 to 17500 (with zeroes
at the unreadable blocks), and try to write it to the position it represents to disk
using "dd".
Do this only if you have a backup, and only if you want to expect other bad blocks.
But for me, this procedure has cured all overwandered tracks so far.
Xu�n.
