Path: utzoo!utgpu!jarvis.csri.toronto.edu!cs.utexas.edu!tut.cis.ohio-state.edu!rutgers!netnews.upenn.edu!scotty.dccs.upenn.edu!litwack
From: litwack@scotty.dccs.upenn.edu (Mark Litwack)
Newsgroups: comp.sys.dec
Subject: Re: DS3100 Page faulting bug question
Keywords: pmax decstation 3100
Message-ID: <17738@netnews.upenn.edu>
Date: 6 Dec 89 04:41:42 GMT
References: <7205@pt.cs.cmu.edu>
Sender: news@netnews.upenn.edu
Reply-To: litwack@scotty.dccs.upenn.edu (Mark Litwack)
Organization: University of Pennsylvania
Lines: 52

In article <7205@pt.cs.cmu.edu> te07@edrc.cmu.edu (Thomas Epperly) writes:
>A couple months ago several people on the bboard posted a complaint about
>the page faulting bug on the DECstation 3100.  Process are swapped out and
>running (RW) which cases to the machine to go crazy.  I was very impressed
>when someone from DEC searched out my phone number and called me to ask
>some questions.  I am beginning to wonder what has become of this bug.  My
>machine still has to be reboot every couple of days because of it.  Has DEC
>reproduced it, debugged it and fixed it yet?  When can we expect a bug fix?

Unfortunately, nothing seems to be fixed yet.

We got DEC to lend us 12meg of memory so that our DS3100 would be up
to 24meg (fully populated) while they worked on the problem.  Since
the extra memory was installed, we didn't have a lock up or crash in
45 days.  The segmentation faults when trying to start emacs also
disappeared.

Last week we finally received a patch from DEC claiming that the
problem was solved.  I installed the patch, removed the extra 12meg,
rebooted, and the system locked up twice within a few days.  The emacs
segmentation faults were also back.  Oh well.

I called DEC back and now they claim that my system configuration
isn't supported: they say I need 16meg for a server instead of 12meg.
Ok, fine.  I brought it up to 16meg this evening.  I have my doubts
that this will solve anything because it locked up before with 16meg,
although admittedly it was without the patch.

I would have let everyone know immediately if the problem had been
fixed.  Good thing I tested it first.

It's no surprise that DEC called you.  I sent every article about this
that was posted to the person who was handling my bug report.  Hope
you don't mind, but they claimed they never heard of it before and
they were leaning dangerously close to saying that it was something
specific to my network and that I had done something wrong.

DEC has tried to close my open bug report 4 times now and actually
succeeded once because I was on vacation and they talked to someone
else.  They seem desperate to get the call closed because probably
it's beginning to look bad.

At any rate, today I had a 5-way conference call with the Ultrix
support engineers and customer service.  One engineer that I talked to
seemed to understand the problem and was genuinely interested in
finding a solution.  Perhaps I've found the right person to talk to.

I'll post if there's anything of interest that comes out of this.

-mark
University of Pennsylvania
Data Communications and Computing Services