[rescue] OT: Linux and USB on Intel

Mike Meredith mike at blackhairy.demon.co.uk
Sun Apr 20 15:36:55 CDT 2003

On Sunday 20 April 2003 19:57, David Passmore wrote:
> On Sun, Apr 20, 2003 at 02:56:19PM -0400, Dave McGuire wrote:
> >    Hmm...I understand your point, and I agree that it's a good point,
> > but I'm not sure that I agree with it entirely.  Windows, in my
> > opinion, has damaged the computer industry (and the science) possibly
> Linux has taught the computer industry that hardware failures are normal,
> that cheap power supplies "just fry", that hardware is cheap, obselete
> quickly, and that it's "ok" to need to replace machines constantly.

There's something funny going on here. Cheap and nasty desktop x86 systems may 
be pretty unreliable, but server x86 systems from decent vendors are a 
different matter.

I'm in favour of using better systems from Sun, SGI, IBM, and HP, but I'll 
happily use x86 *server* hardware when it's appropriate. My objections to 
using x86 hardware for database servers and the like has nothing to do with 
the reliability of x86 hardware.

I don't see large numbers of hardware problems with x86 hardware where I work, 
and there are a reasonable number of the boxes in operation. Admittedly we 
have had severe problems with x86 hardware in the past, before management was 
persuaded that buying cheap gear was a false economy.

And what has Linux got to do with hardware failures ? Surely they will happen 
whatever o/s you're running.

Elsewhere you've said :-
> Are you claiming that x86 hardware in general, has an equal or greater
> mean-time-to-failure than dedicated UNIX hardware from manufacturers like

Ok. Now what about the MTBF on x86 server hardware from HP/Compaq, Dell, and 
the like ?

> that OS concepts that were old twenty years ago are okay in a modern OS.

Ok. So what modern OSs should we be looking at ?

More information about the rescue mailing list