CurrentPendingSector

[Chris Murphy]

On Wed, Jan 28, 2015 at 2:06 PM, Patrick Dupre <pdupre at gmx.com> wrote:
> Hello,
>
> I get:
> Device: /dev/sdc [SAT], 1 Currently unreadable (pending) sectors
>
> How can I manage this issue?
>


You can write over the affected sector, and if it's a genuinely bad
sector (persistent write failure) the LBA will be mapped to a reserve
sector. And at that point it'll continue to work. There's not much
evidence a one off bad sector is a big problem, other than of course
the data that was on it is lost. But sector failures tend to come in
groups and once multiple bad sectors start to happen, there's a lot of
evidence this is probably a drive pre-fail indication.

The procedure to write over the affected sector depends on whether the
drive is 512n or 512e, which can be determined with:

parted /dev/sdc u s p

You'll see a line for logical and physical sector size in bytes. If
it's 512/512 then it's a 512n drive and you can write over that single
sector using dd with a count of 1 and the default bs, and the LBA
value is used as the dd seek value.

If it's 512/4096 then it's a 512e drive and you'll need to define
bs=4096 and you'll need to divide the bad sector LBA value by 4096 to
get the proper seek= value, and use a count of 1. (If you try to use
the default bs and count of 1 (or even 8) the drive actually tries to
read the sector which will result in a read error. It won't even
attempt to write and thus the problem can't be fixed this way.)

-- 
Chris Murphy