Mark Griebling wrote:
> Seagate man, Seagate!
>
> Later,
>
> At 07:38 AM 6/8/00 +0100, Ted Wager wrote:
>
> Hi...
> I am buying a new h/drive for my machine.It has a Maxtor in at the moment but
> I need another drive...I know Ican read the h/ware howtos but they are not so
> up to date......Any info on recommended drives or ones to steer clear of other
> than Western Digital would be welcome...
>
> Regards Ted
>
> Ted Wager......Mandrake linux
> g3tpi.ampr.org 44.131.147.8
If your current drive has performed well and you plan to keep it installed, stay
with your brand. The timing requirements are likely to be much more compatible.
My own personal considerations these days run to IBM and Quantum because the data
I have collected suggests a better MTBF in those brand names. It is however
dangerous to assume that MTBF by brand name calculations are valid--there are too
many models and technical approaches within a single technology to say that
figuring backwards from data points will produce the correct curve.
Seagate is good for certain chipsets, but I have observed signal reflections on
i810, VIA MVP3, and MVP4, Intel 430 TX, and ALi Aladdin V under Pentium code. The
problem is highly pronounced with TX chipsets, resulting in corrupt data that
appears to write OK and is totally unreadable when the TX chipset is clocked at 75
MHz (as for a Cyrix processor). I did get "lost interrupt" errors on a very
modern i810 with a Celeron 400 and a Barracuda 10.2G. Windows, FreeBSD, and such
loaded fine, but Stampede, Enoch and Mandrake could not even install. (All were
processor optimised) Mandrake 486 could run this with only occasional reports of
disk errors.
Civileme
