----------------------------------------
MORE UNSCHEDULED DOWNTIME: APRIL 6, 1997
----------------------------------------
At 4:39 PM EDT this afternoon a program being run by a user on the UNIX
host started spawning daughter processes at a tremendous rate.  Within
three minutes (4:41 PM) the runaway program had commandeered every
available process ID, preventing any additional programs from running and
causing several critical programs to quit. 

TRANSLATION:  A program being run by an account holder went into an
endless loop, running thousands of copies of itself.  This made it
impossible for other programs to run, which brought the UNIX host down.

I became aware of the problem at about 5:00 PM EDT when I awoke from a nap
to find the system not responding.  Those of you who read this morning's
System News know why I was taking a nap, and you'd be right to guess that
I was not amused!
 
I came into the office, rebooted the UNIX host at 5:48 PM, and everything
seems fine once again.

The cause of this crash was easy to determine; it was a simple case of
operator error on the part of one of our users.  I still haven't figured
what caused last night's crash, but that will have to wait until tomorrow. 
I'm going back to sleep!
 
---------------------------------------------------------------------
If you have questions about this or about anything else pertaining to
your BCPL Internet Account, please contact the BCPL Help Desk:

      Phone:    410-887-3297
      FAX:      410-887-2091
      E-Mail:   oz@mail.bcpl.lib.md.us