Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Path: utzoo!utgpu!water!watmath!clyde!cbosgd!ihnp4!ptsfa!ames!ucbcad!ucbvax!CITHEX.CALTECH.EDU!carl
From: carl@CITHEX.CALTECH.EDU.UUCP
Newsgroups: comp.os.vms
Subject: Re: publishing patch (david bolthouse)
Message-ID: <870613150632.05t@CitHex.Caltech.Edu>
Date: Sat, 13-Jun-87 18:14:51 EDT
Article-I.D.: CitHex.870613150632.05t
Posted: Sat Jun 13 18:14:51 1987
Date-Received: Tue, 16-Jun-87 01:43:47 EDT
Sender: usenet@ucbvax.BERKELEY.EDU
Distribution: world
Organization: The ARPA Internet
Lines: 16


 > On another tack, I just can't imagine a computer which is not available for
 > a  reboot  (i.e.   at worst 15 minutes) at some scheduled time, let's say 3
 > o'clock on a nice Saturday morning.  So please  save  us  this  destructive
 > moan about your users!

Let me give your imagination some help.  Picture the following:
   A group of physicists spends years designing an experiment to be run on  an
   accelerator someplace (CERN, SLAC, FERMILAB, wherever).  They then wait for
   another extended period of time to get  the  experiment  scheduled  on  the
   accelerator.   They start the experiment, and have a VAX doing on-line data
   collection, with some  on-line  analysis  to  select  only  the  events  of
   interest.  You,  as  a system manager, then tell them that they're going to
   lose 15 minutes of data so that you can perform an elective reboot  of  the
   system.  I guarantee, they won't be pleased.
Yes, a system can always be rebooted, but not always at a reasonable cost.