Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Path: utzoo!watmath!clyde!cbatt!ucbvax!GRINNELL.MAILNET!McGuire_Ed
From: McGuire_Ed@GRINNELL.MAILNET.UUCP
Newsgroups: mod.computers.vax
Subject: (none)
Message-ID: <8702070627.AA16796@ucbvax.Berkeley.EDU>
Date: Fri, 6-Feb-87 13:24:00 EST
Article-I.D.: ucbvax.8702070627.AA16796
Posted: Fri Feb  6 13:24:00 1987
Date-Received: Sun, 8-Feb-87 05:27:15 EST
Sender: daemon@ucbvax.BERKELEY.EDU
Organization: The ARPA Internet
Lines: 12
Approved: info-vax@sri-kl.arpa

I hope I have the outline of this story right.  The person who told it to me
will probably see this posting, and may wish to set me straight and remind me
of the details.

The story is that there was a period of time during VMS cluster development
when there was an occasional crash of one node of a test cluster. The
developers eventually discovered that every node of the cluster was using the
same swap file.  Occasionally a process would swap when memory got tight, and
the process actually got INSWAPPED ON A DIFFERENT NODE.  This process was the
one listed as "current" in the crash dump.  The only problem was that the
process had I/O channels assigned that were inconsistent with the I/O database,
etc., which which made VMS rather paranoid.