Subject: VS2000 PS quirk? [Was Re: Anyone running on VS2000 w/small memory?]
To: NetBSD/vax Mailing List <port-vax@NetBSD.ORG>
From: Brian D Chase <bdc@world.std.com>
List: port-vax
Date: 05/25/1998 17:44:24
Okay, after some additional investigation, I think I've determined what
my original problem may have been caused by.  When I first attempted to
netboot the VS2000s, they were exhibiting some strange behavior.  They'd
lock up tight after getting to the login prompt.  Later I found that
they would sometimes do this mid-way into loading the kernel, sometimes
during the power-on tests, and occassionally after a power-cycle the
power-on tests wouldn't even come up.  I was beginning to think the worst
and that I'd a couple of unhappy VS2000s on my hands.

After a while I was able to make some progress. I could boot single-user
and then multi-user on one of the systems, but the other one was still
misbehaving.  For a while I suspected that my shorting of pins 8-9 wasn't
making good contact when dealing with the suspect VS2000.  Maybe I was
just losing contact with the serial console.  But then when I tested the
cable on the working VS2000, it always performed fine. 

I checked everything over, I could only find one difference really, and
that was that I'd put an RD54 back in the (now) working one right when I
was just beginning to troubleshoot the intial problems.  I wasn't using
the drive at all, I basically just hooked it back up in order to get the
machine in its original state.  In retrospect this struck me as odd, since
right after doing that I started making some progress with that machine --
but I didn't make any connection at that time. 

So I grabbed a spare drive I had laying around and plugged it into one of
the suspect VS2000's power connecters.  I didn't even connect the drive
cable (well I couldn't becuase I don't have one for it).  Then I powered
the system up.  It worked perfectly.  I power cycled it again and it still
worked fine.  I disconnect the power from the drive, powered it back up,
now it was broken again.  Power-cycled it... still broken.  Connect the
drive again, now it works.  

Hmmmmm...

I guess the PS really wants something to act as a load on those power
connectors for the disk drives.  I don't know that this is normal behavior
or not, but it seems like I can reproduce the problem.  We used to have a
very large number of VAXstation 2000s at my college, and I'm pretty sure
that we used them as diskless nodes on a VMS cluster with our VAX 6000. I
remember this, because my roommate and I were very disappointed when we
could only find one spare VAXstation 2000 with a disk drive for running
an Ultrix system.  All the other VS2000s we could lay our hands on didn't
have any drives.

Can anybody more familiar with VS2000s or maybe even electrical power
systems shed some light on perhaps why a VS2000 would need a load on the
drive power connectors to operate normally?

--

For the most part, the NetBSD/vax kernel worked just fine on the VS2000s
in only 6Megs (at least once I got them going).  The systems behaved
fairly well when booted to multi-user mode with a standard compliment of
daemons, though there was one exception.  The exception being that when I
included the update daemon in rc.conf, the system would dump to ">>> " if
I tried to telnet into the machine just after booting.  I'm still looking
into this feature, but otherwise things look pretty good.  Well, the
C compiler somtimes dies on a signal 4, but that's a known problem.

-brian.
---
Brian "JARAI" Chase | http://world.std.com/~bdc/ | VAXZilla LIVES!!!