[smartmontools-support]When is a dead disk really dead? (question about Reallocated Sector Count)

[ever w. <eve...@ya...>]

Hi folks, I'm a newbie who came across my exact
problem in your archives here:
http://sourceforge.net/mailarchive/message.php?msg_id=11960603

5  Reallocated_Sector_Ct  0x0033  001  001  050 
Pre-fail  Always  FAILING_NOW  1023

Disk = TOSHIBA MK8025GAS (14 months old)

Many thanks to Sebastian for the clear, concise
explanation (attached below). I understand the
situation completely now. But my question is this: Is
it possible to recover dead sectors and raise the
Reallocated_Sector_Ct back above the threshold?

My hard drive failed dismally last week.  I spent the
last 5 days recovering my data (successfully!) and
eventually reformatting & reinstalling my OS. Now
everything seems to work fine except that I continue
to get the "SMART Failure Predicted" error message at
boot up.

I've even run CHKDSK several times w/ no errors found.
Is it possible that my disk was fixed by reformatting?
And if so, could the SMART error message be wrong?

It seems to me that the SMART error reporting is based
on a lifetime history of the drive. So obviously, a
horrendous crash such as mine would be recorded as
having exhausted the Reallocated_Sector_Ct.  But if
those sectors are later recovered, does SMART
replenish the count, or does it still think I've used
up all my reserves? That would explain why I keep
getting the Failure Predicted message even though
performance has returned to normal.

As you can see, I'm terribly confused by all of this.
(Also, I'm desperately wanting to believe that my disk
is still healthy, because my warranty just expired 2
months ago!!)

Regards,
Rick

 On Fri, 3 Jun 2005, Sebastian Vuorinen wrote:
 
 > On Fri, 2005-06-03 at 02:05 +0100, Carlos Silva
wrote:
 > > Now I get this..
 > > [root@localhost ~]# smartctl -a /dev/hda | grep
"Reallocated_Sector_Ct"
 > >   5 Reallocated_Sector_Ct   0x0033   001   001  
050    Pre-fail 
 > > Always   FAILING_NOW 1023
 > 
 > Your harddrive has a number of reserve sectors to
replace damaged
 > sectors. Once the number of reserve sectors left
goes below a limit set
 > by the drive manufacturer SMART considers your
drive as having failed.
 > This is the case here.
 > 
 > The first number (001) is a normalized value
derieved by a calculation
 > specified by the drive manufacturer. Normally it
counts down from 100 as
 > sectors on your drive are replaced with spares.
Remember this is
 > "normalized" and not a one to one mapping. The last
number of the three
 > (050) is the limit set by the manufacturer.
 > 
 > Also of interest is the very last value on the line
(1023). It is the
 > raw value for this item. Usually this should show
the number of actual
 > sectors damaged. Assuming the raw data is correct
your drive has 1023
 > failed sectors. That is a lot, possibly caused by
dropping the drive or
 > serious overheating. What ever the cause, this
drive can not be trusted
 > to store data and if you are still on warranty you
should be able to get
 > a replacement.
 > 
 > You can find more information on the relation of
the "normalized" values
 > to the "raw" values from the smartctl manpage under
the heading "-A, 
 > --attributes".
 > 
 > Sebastian
 > 
 > 


__________________________________________________
Do You Yahoo!?
Tired of spam?  Yahoo! Mail has the best spam protection around 
http://mail.yahoo.com