[LUNI] Re: Testing IDE drives

[Martin Maney]

On Sun, Sep 02, 2007 at 03:17:37PM -0500, David Gibbs wrote:
> I've had the same feelings about SMART for quite a while ... I've
> *NEVER* had SMART predict a true drive failure ... and, when SMART does
> report a impending failure, normal diagnostics (even the manufacturers)
> report no problems.

Yeah, I think it's safe to say that many drives give *no* advance
notice - they just up and quit.  That certainly describes all the drive
failures I've had personal dealings with.

And that should really have been my point.  You can look at what SMART
reports, you can run this test or that test, and in the end all you can
really say is that the drive was (or wasn't, but that's the simple
case) working when the last test completed.  Tomorrow it could be belly
up.  If that makes you feel kind of stupid to be keeping anything you
care even a little bit about on any single drive without a known
working backup in place... welcome to the club.

-- 
The dualist evades the frame problem - but only because
dualism draws the veil of mystery and obfuscation
over all the tough how-questions  -- Daniel C. Dennett