For the Eagan machines, add these 4 lines to /etc/ntp.conf
   server 163.231.100.252
   server 163.231.100.253
   server 163.231.101.253
   server 163.231.101.253
uncomment xntp from /etc/rc.tcpip (or chrctcp -a xntpd)
and startsrc -s xntpd

=========================================================================

  Some note on NTP (Network Time Protocol)           September 22, 2003

See http://www.eecis.udel.edu/~mills/ntp/servers.html for a brief
intro to Public NTP Servers.  At the bottom of that page, is this link
(http://www.eecis.udel.edu/~mills/ntp/clock2a.html) to a list of 173
"Public NTP Secondary (stratum 2) Time Servers".

Among them are these two located in San Jose,
   US CA clock.sjc.he.net (216.218.254.202)
   US CA ntp1.sf-bay.org (207.126.97.57)
On the day I checked, the first one, clock.sjc.he.net, seemed to be the
closest (11 hops via traceroute versus 12) and quickest (9 milliseconds
versus 18), so I chose that one for AFS1 (which had been using
nist1.sjc.certifiedtime.com, but this apparently no longer exists).

To change the NTP server on afs1, you actually have to modify the 
runntp lines in the /usr/afs/local/BosConfig file.  When I started,
these lines were
   bnode simple runntp 1
   parm /usr/afs/bin/runntp nist1.sjc.certifiedtime.com
   end
And these two processes were running,
    root 16274 10098   0   Sep 21      -  0:00 /usr/afs/bin/runntp nist1.sjc.certifiedtime.com 
    root 18340 16274   0   Sep 21      -  0:00 ntpd -f -c /tmp/ntp.conf

To change this, you first delete, then redefine it via bos commands.
   bos stop afs1 runntp -wait -localauth
Both the ntp & runntp processes are now no longer running.  Then
   bos delete afs1 runntp -localauth
And now those lines are removed from /usr/afs/local/BosConfig.  Finally,
   bos create afs1 runntp simple -cmd "/usr/afs/bin/runntp clock.sjc.he.net" -localauth

=========================================================================


=========================================================================