On Fri, May 17, 2013 at 08:36:30PM +0000, Scott Weikart wrote:
> > It's statistically possible that if the two drive were produced on
> > the same assembly line via the same process that they would fail
> > at around the same time.
>
> This happened to me: two disks failed within 2 days of each other.
>
Statistics-shmatistics.
The airplane "multiple correllated failure" story goes like this.
So they were flying a Ford tri-motor from Miami to Havana (this was before Fidel)
and the idiot light lights up on the first engine. They shut it down,
continue flying. The idiot light lights up on the second engine,
they start getting worried and turn back. Soon, the idiot light lights up
on the third (last) engine and they relax - no way all 3 engines
can fail for real - must be a "computer glitch", but they keep flying
back (on two engines). Then both working engines seize-up (stop turning)
one after the other and now they are in trouble, getting ready for a swim.
Luckily they remember the very first engine, manage to restart it,
fly home on one engine and land. The last engine seizes-up as they taxi
to the gate. Investigation shows that all 3 engines were missing
engine oil plugs - removed for inspection, "forgot to put them back"
before flying. Every engine had no oil, every engine failed.
So the impossible can happen quite simply. (Also read up on Gimli glider
as ObCanadianContents or if you think Ford tri-motors belong with abacuses).
The equivalent correllated failure for disk drives would be - computer was dropped,
damaging both disks at the same time; power supply glitch, damaging both disks
at the same time; the same batch of "bad capacitors" used in both disks; the shipping box
containing both disks was dropped; both disks have the same bad firmware; cooling
fan stopped and both disks were damaged by overheating at the same time; etc.
--
Konstantin Olchanski
Data Acquisition Systems: The Bytes Must Flow!
Email: olchansk-at-triumf-dot-ca
Snail mail: 4004 Wesbrook Mall, TRIUMF, Vancouver, B.C., V6T 2A3, Canada
|
