Failure rates

[Paul Lussier]

Benjamin Scott <dragonhawk at iname.com> writes:

> On Apr 24 at 11:43pm, David Ecklein wrote:
>> My bad impression had only been with Samsung hard drives, and that
>> is dated and limited to my own small computer business activity, not
>> "anecdotes". Failure rate was higher than any other brand I had
>> used.
>
>     An interesting phenomenon has been observed when it comes to the
>     public's perception of reliability of commodity brands:  The more
>     units sold, the more units fail, in terms of absolute numbers.
>     Thus, a given brand of something will become popular to the point
>     where it becomes "first choice".  Then, because there are so many
>     of them out there, people begin to see more failures in that
>     brand.  People thus conclude there must be something wrong with
>     that brand.  Opinion drops, and the product falls out of first
>     place, confirming to everyone that there was something wrong.

This very much applies to hard drive manufacturers.  Though, because
there are really on 3 or 4 of them, this phenomenon presents itself in
a round-robin fashion.

Currently I'm seeing the most failures with Maxtor, on average, about
1 drive a week.  I've RMA'ed well over 50 drives in the past 6 months
or so.  That figure may seem high, until you look at the numbers I'm
dealing with: Over 300 systems, many of which 2 drives, 50 or more
have 4 drives.  I probably have over 1000 drives in this place, so in
theory, 1 per week isn't unrealistic, and is probably well below the
advertised MTBF as well.

Ironically, my understanding is that WD and Seagate both have *higher*
failure rates currently, though, since I have far fewer drives from
them, I see significantly fewer failures.
-- 

Seeya,
Paul