To: tuhs@tuhs.org
Date: Wed, 22 Jul 2020 13:11:16 -0400 (EDT)
From: norman@oclsc.org (Norman Wilson)
Subject: [TUHS] Bell Labs Sysadmins
Errors-To: tuhs-bounces@minnie.tuhs.org
Sender: "TUHS" <tuhs-bounces@minnie.tuhs.org>

Henry Bent:

  Are there any former Bell Labs sysadmins on this list?  My father was the
  sysadmin for hoh* (Crawford Hill, mostly the radio astronomy folks) in the
  mid-late '80s and early '90s and I would be especially interested to hear
  from hou* (Holmdel, what a building!) folks but also ihn* (Indian Hill) and
  the like.  I'm very curious about what life was like on the ground, so to
  speak.

=====

It is worth pointing out that, like many universities, Bell
Labs had at least two layers of computing and therefore of
sysadmins.  There were official central Comp Centers, which
is the world Henry asks specifically about; but there were
also less-central computing facilities at the divisional and
center and department level.

I never worked for a comp center, but my role in 1127 was
basically that of sysadmin for our center's own systems;
the ones used both for our day-to-day computing (including
that known to the uucp world as research!) and for OS and
other experiments and research and hacking.

There were also two large VAXes called alice and rabbit,
which afforded general-use computing for other departments
in Division 112.  Us 1127 sysadmins (I wasn't the only one
by any means) ran those too; I don't know the full history
but I gather that happened because there was some desire
to run the Research version of UNIX rather than the comp-center
one for that purpose.

One system I had hands in straddled the researcher/comp-center
boundary: 3k, the Cray X-MP/24 bought specifically for
researchers.  It was physically located in the comp center,
but managed jointly: it ran Cray's UNICOS but with
substantial additions from the Research world, including
the stream I/O system (which is rather self-contained so
it was not too hard to graft into a different UNIX) and
Datakit support (using a custom interface board designed
and built by Alan Kaplan and debugged by me and Alan
over several late-night sessions.  (I remember that
the string "feefoefum\n", which is an obscure cultural
reference from one of my previous workplaces, was
particularly effective at shaking out bugs!)
Once UNICOS-a-la-Research was running stably, staff from
the Murray Hill Comp Center looked after day-to-day
operations, with Research involved more for software
support as needed.

Norman Wilson
Toronto ON