Smartd message: What does it mean?

[Charles Curley]

On Sat, Oct 29, 2005 at 12:02:27AM -0700, jdow wrote:
> Do not pass go. Do not collect $100. Do not dilly dally around. Get a
> new drive and move over to it. Your only hope that it is something else
> is the possibility of a flaky ATA cable. Betting on this is like taking
> the bad odds at a craps shoot.

Joanne is correct.

First off, the fact that smartd is reporting N bad sectors does not
mean that you only have N bad sectors. You probably also have a bunch
that the firmware has already either re-allocated (hiding the fact
that they are bad), or recovered (if the defect is small enough) and
re-written.

When a drive is manufactured, it is tested, and a list of bad sectors
is created. These sectors are re-allocated from spares, and the
substitution is utterly transparent to the OS.

When (not if) a new defect occurs, the drive will re-read the sector a
number of times, and try various tricks to recover the data. The
smaller the defect, the greater likelihood of recovering the data. If
the defect is small enough, the drive will simply re-write the sector,
end of discusion.

When (not if) the defect gets large enough, the drive will re-allocate
a substitute sector from a list of spares, and mark the old one as
bad. That list of spares will eventually be exhausted, which can
happen in a matter of minutes.

However, the fact that the drive has enough bad sectors that the
firmware is reporting them to you means that you have more bad sectors
than there were on the drive since it was manufactured. Your drive is
warning you that it is about to die.

Yes, it may take several years to die. On the other hand it could take
15 minutes. 15 seconds.

Now, how much do you want to bet that it will be toward the several
years end of things?

How much is your time worth? I'd rather spend a few bucks on a new
hard drive, copy the data over, and be good to go than try to recover
data from a dieing drive. Been there, done that, got the
T-shirt. Been paid big bucks to do it for other people.

Do as the lady says. Now.

Then start doing regular backups.

-- 

Charles Curley                  /"\    ASCII Ribbon Campaign
Looking for fine software       \ /    Respect for open standards
and/or writing?                  X     No HTML/RTF in email
http://www.charlescurley.com    / \    No M$ Word docs in email

Key fingerprint = CE5C 6645 A45A 64E4 94C0  809C FFF6 4C48 4ECD DFDB
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 189 bytes
Desc: not available
URL: <http://listman.redhat.com/archives/fedora-list/attachments/20051029/31aa3b43/attachment-0001.sig>