Path: utzoo!utgpu!news-server.csri.toronto.edu!rpi!zaphod.mps.ohio-state.edu!samsung!umich!terminator!pisa.citi.umich.edu!rees
From: rees@pisa.citi.umich.edu (Jim Rees)
Newsgroups: comp.protocols.time.ntp
Subject: Re: ntpd -> AIX/370 on a 3090
Message-ID: <50a37873.1bc5b@pisa.citi.umich.edu>
Date: 28 Mar 91 18:32:43 GMT
References: <1513@ucl-cs.uucp> <1991Mar27.180231.1269@unixg.ubc.ca>
Sender: usenet@terminator.cc.umich.edu (usenet news)
Reply-To: rees@citi.umich.edu (Jim Rees)
Organization: University of Michigan IFS Project
Lines: 12

In article <1991Mar27.180231.1269@unixg.ubc.ca>, whaley@elbo.ucs.ubc.ca (Paul Whaley) writes:

  There's not much point in trying until IBM fixes the bugs in adjtime. Both the
  unix ntp implementations (ntp from louie.udel.edu and xntp) compile fairly
  easily on AIX/370, but the errors in adjtime defeat the timekeeping - we had 
  our clock 90 minutes slow at one point. While awaiting IBM's fix, we run
  ntpdate every hour.

We find that the clock on our 3090 is very stable, and once a day is plenty.
Don't forget to give ntpdate the '-b' flag.  Also note that if you have
several AIX systems on a single machine with VM, each one keeps a separate
clock so each one needs ntpdate.