On Thu, Mar 10, 2016 at 12:06 PM, Dimitri Maziuk <
dmaziuk AT bmrb.wisc DOT edu> wrote:
>
> On 03/10/2016 01:23 PM, compdoc wrote:
> >>> You don't have to guess about failing hard drives. You only have to
> >>> read the SMART info from each of your drives. That will tell you.
> >>
> >> That is not entirely true. "Desktop" drives will spend a lot of time
> > trying to relocate data from failing sectors before they consider themselves
> > failing.
> >
> >
> > It is completely true.
>
> BS. I've thrown out dozens of dead drives whose SMART report and
> self-tests were "healthy as an ox". I repeat, if you have a low-end
> "desktop" drive, "only reading SMART info" is not enough: they come with
> a well-known failure mode that produces false negative on the SMART test.
For anybody interested, Google studied
failure trends in large disk populations and one of their conclusions was:
...we find that failure prediction models based on SMART parameters alone are likely to be severely limited in their prediction accuracy, given that
a large fraction of our failed drives have shown no SMART error signals whatsoever.
In other words. SMART says there's a problem, it's very likely to fail catastrophically within a short time period. SMART says there's no problem, doesn't mean a thing.