[ale] 9.10 smart errors

[Mike Harrison]

> SMART may not be as smart as everyone thinks.

In the old days, I used to be good at predicting drive failure.

You could hear them.. as the bearings started to fail
or heads seeked a lot trying to get data off/on.
I could walk the racks of the colo room and
hear the whines and clicks of imminent failure.

Sometimes you'd get days or months of errors
in the log files, seek errors and more.

I've had more than one drive that'd be fine
if you helped spin it up from a cold start with a
pencil eraser on the spindle (back when they
were exposed).

Luckily, they seem to fail a lot less often then they used to.
I haven't had a production machine (< 3 years old) drive fail
in a loooong time. Except for one 2.5" server drive in a strange place
after multiple power failures, lack of AC and other issues,

But when they do fail now, the seem to instantly transmutate
into small bricks. Less warnings, less notice and no chance of
recovery. they no longer spin up or work after a cool-off cycle.

I miss the whine of a failing hard drive bearing, but not much..
not very much at all.

At least I don't park/lock the heads with a little lever anymore.
(Data General Nova III w/ a 5MB HD and others...)