Randall Davis writes:

> But then again, I hate having single points of failure.  Am
> I asking for too much from the HP3000?  These types of solutions are
> available in other platforms.
...
> If I'm wrong, or if there are solutions I haven't thought
> of or don't know about I'd love to hear about them!

While I understand and agree to a point, I would like to caution you
against solutions which *appear* to allow no downtime.

I worked at a site that decided to bring in Tandem equipment for shop floor
tracking.  The hardware was physically imposing, with redundant hardware
for everything (e.g., dual disc drives, CPUs in pairs, etc.).  It was an
eight-CPU machine, in four pairs of 2 CPUs.

Meanwhile, the diminutive-by-comparison Series 70 (2-bay, not 3) sat over
in the corner doing its work.

Problem was, it was not unusual to come in after a weekend to find 7 of the
8 CPUs down!  Since applications were run by specifying both a "primary"
and a "secondary" CPU, if both are down....

Meanwhile, the Series 70 (with its lone, single CPU) sat over
in the corner doing its work.

It was certainly an eye-opening lesson.

--Glenn

P.S.  Your post includes the attachment "rdavis.vcf".