[ltp] Re: Hard drive going bad ...?

[Richard Neill]

> On Mon, Sep 04, 2006 at 13:00:26 -0400,
>   Paul Michael Reilly <pmr@pajato.com> wrote:
>> The smartd daemon (Fedora Rawhide) is starting to indicate that the
>> hard drive has pending unreadable sectors:
>>
>> Sep  3 08:01:43 roamer smartd[2032]: Device: /dev/hda, 9 Currently unreadable (pending) sectors 
> 
> One important note is that your drive will not remap those sectors to its
> spares until it either gets a good read or you give up on getting the data
> back (signified by writing over the problem sectors). You probably want to
> try to figure out which files these sectors are in so that you know what
> data you have lost. There is a description on how to do this for ext3/2
> file systems at: http://smartmontools.sourceforge.net/BadBlockHowTo.txt
> When you are ready you can write over the sectors using dd.

The real problem when trying to read (especially newer, SATA) drives 
with failing blocks is this: the drive + kernel will keep re-trying for 
almost an eternity to read any given dead sector. Thus it becomes 
impossible to read the good bits!  dd might take a year to complete.

I recommend dd_rescue as a solution. I tried SpinRite - but it didn't 
help at all (although they were very decent about giving me a full refund).

Best wishes,

Richard