Richard,
Although most of us are dyed-in-the-wool HP 3000 fans, it's hard to argue
with facts.  I have been working with MPE for 21 years, and have been very
impressed by the reliability and stability of it for most of that time.  I
have experienced periods of excessive failures, though, when dealing with an
early patch level of a new OS release, or with a new machine.  It also seems
to me that the overall reliability of MPE has diminished somewhat in recent
years.  If my perception reflects reality, then I assume it has to do with
the increased pace of new features that have been incorporated into MPE.

Clearly, the reliability gap is narrowing.  Operating systems that have
historically been less reliable, such as HPUX, are becoming more reliable as
they mature.  The extra acquisition cost of an MPE system gets harder to
justify if its advantages become more difficult to prove.  This may be the
major causes of MPE's impending abandonment.

The situation you describe with "needing a service pack each week" is one
that concerns me with the upcoming release of 7.5.  Since 7.5 is expected to
be the last release of MPE, I worry that HP won't maintain an adequate staff
of MPE experts in the software lab after the 7.5 release.  There will surely
need to be many patches as early adopters of 7.5 uncover bugs, and I hope HP
will be able to respond to those needs.  I have raised this concern on the
list before, in the hope that someone at CSY would address it.  To date, I
have not seen a response.  Will HP provide real, meaningful support for 7.5
after it's introduction?  Will the software support be "up to par" after the
October, 2003 EOL?  I hope we can receive some meaningful, specific
indication of HP's plans in regard to this issue.

-----Original Message-----
From: Richard Barker [mailto:[log in to unmask]]
Sent: Wednesday, July 31, 2002 12:53 AM
To: [log in to unmask]
Subject: Re: Reliability - myth


Hi

I've calmed down a bit now.

In every case HP were involved and nearly every case it involved a patch to
solve the problem.  One case in particular where a DNS entry on our network
caused the HP to crash, we needed a direct change to a Binary in the
operating system.  The last couple of crashes were because the engineers
investigating the problems switched the console to an LDEV that didn't
exist, after a few hours the buffer filled up and the machine crashed again.
Why does the HP allow you to switch the console to a device that isn't
active, when this can cause such a problem.

I have lots of issues with all of this.  For me the point of paying the
money we did for an HP, is you receive a stable, well tested, reliable
machine with good support.  I don't expect to receive a machine that needs a
service pack every week.  It's all very well saying you need to apply
patches, but to me that shows a big weakness in the OS.

The other problem, is the crashes themselves, surely a machine like this
should be able at least to issue an error message or cope with most
circumstances, not just freeze like a MS OS.

The bottom line for me, is I can't get a good Server and install Linux on it
and I don't believe it will be any less stable than the HP and because of
the huge difference in cost I can afford to have a cluster, which will
eliminate most problems with downtime.

Anyway, thanks to most of you for not 'having a go', obviously with this
being an HP3000 list, I thought I might put a few 'noses out of joint', but
everyone was quite polite.

Regards


Richard

* To join/leave the list, search archives, change list settings, *
* etc., please visit http://raven.utc.edu/archives/hp3000-l.html *