disk error

[Erik Trulsson]

On Sat, Feb 16, 2008 at 07:30:37PM +0100, Peter Boosten wrote:
> Brian A. Seklecki wrote:
>> On Sat, 2008-02-16 at 17:59 +0100, Peter Boosten wrote:
>>> Hi all,
>>> 
>>> Just found these messages in my logfile. Is it something to worry about?
>>> I've never seen them before upgrading to 6.3.
>>> 
>>> ra kernel: ad0: TIMEOUT - READ_DMA48 retrying (1 retry left) LBA=281550271
>>> ra kernel: ad0: FAILURE - READ_DMA48 status=51<READY,DSC,ERROR> 
>>> error=4<ABORTED> LBA=281550271
>> 
>> Yea -- normally that means a bad sector(*), and where there's one,
>> there's bound to be more.  Failed drive eventually.
>> 
>> I would pull this server from rotation and run a full surface sector
>> scan on it (download an ISO of "Hiran's Boot CD")
>> 
> 
> Brian, thanks for your answer (and sugggestion).
> 
> Isn't a drive supposed to mark a bad sector as bad and ignore it (that is: 
> not use it anymore)?

The drive can only remap bad sectors when you write to them.  When you read
from a bad sector the drive does not know what data was supposed to be there
and thus can only return an error or return garbage data.  Returning an
error (which is what disks do) is a much better choice.





-- 
<Insert your favourite quote here.>
Erik Trulsson
ertr1013 at student.uu.se