Re: time synch between two systems: HOWTO?

[Bob Arendt <rda rincon com>]

Charles A. Crayne wrote:
> On Fri, 31 Oct 2003 16:28:59 -0700
> Bob Arendt <rda rincon com> wrote:
>
> :You can use ntptrace to see your timing chain:
>
> And if I get:
>
> localhost.localdomain: stratum 16, offset 0.000058, synch distance 1.10818
> 0.0.0.0: *Not /synchronized*
>
>  what does that tell me that I need to do?
I suspect you're not sync'd to a server.  Use ntpq to check.
If you do a "locate nptq" you should be able to find some docs.
For me it was in:   /usr/share/doc/ntp-4.1.2/ntpq.htm
(also look at /usr/share/doc/ntp-4.1.2/debug.htm)

Example:
===================
% ntpq

ntpq> peer
     remote           refid      st t when poll reach   delay   offset  jitter
==============================================================================
*ahab.rinconres. seismo.cmr.gov   3 u  663 1024  377    0.632  -51.586   4.498
 LOCAL(0)        LOCAL(0)        10 l   34   64  377    0.000    0.000   0.015

ntpq> assoc
ind assID status  conf reach auth condition  last_event cnt
===========================================================
  1 11068  9634   yes   yes  none  sys.peer   reachable  3
  2 11069  9034   yes   yes  none    reject   reachable  3

ntpq> rv 11068
status=9634 reach, conf, sel_sys.peer, 3 events, event_reach,
srcadr=ahab.rinconres.com, srcport=123, dstadr=194.9.200.146,
dstport=123, leap=00, stratum=3, precision=-11, rootdelay=271.667,
rootdispersion=166.779, refid=seismo.cmr.gov, reach=377, unreach=0,
hmode=3, pmode=4, hpoll=10, ppoll=10, flash=00 ok, keyid=0,
offset=-51.586, delay=0.632, dispersion=15.339, jitter=4.498,
reftime=c34e54e0.5d952000  Sat, Nov  1 2003  8:50:56.365,
org=c34e5519.d2195000  Sat, Nov  1 2003  8:51:53.820,
rec=c34e5519.df62c775  Sat, Nov  1 2003  8:51:53.872,
xmt=c34e5519.df395c42  Sat, Nov  1 2003  8:51:53.871,
filtdelay=     0.63    0.61    0.59    0.61    0.68    0.68    0.64    0.58,
filtoffset=  -51.59  -56.08  -56.47  -58.28  -58.36  -59.46  -34.54   26.55,
filtdisp=      0.50   15.88   31.27   46.64   62.00   77.38   92.77  108.16

ntpq> exit
===================

The "peer" command shows who I'm talking to, and how often.  For instance,
ahab is polling seismo every 1024 seconds.  The delay, offset, and jitter
numbers (in msec) make me suspect it's sync'd.  It shows all the time
servers that ntpd is aware of.

The "assoc" command gives me the connection info.  If you have any other
server, it "reachable" probably isn't a "yes".  It's rejecting you for
some reason.  The "condition" will show how the servers are being used.
After restarting ntpd, you will probably see "insane" in this column.
This is good - you're getting updates, but ntpd doesn't have 8 samples yet.
At startup , it samples every 64 seconds, so after a little over 8 minutes
you'll start getting sync info.

For the impatient, you can use the "assID" from as given by "assoc" to
examine the state using the "rv" command.  The primary thing of interest
is the table of 8 samples at the bottom.  At the start, this will have
bogus numbers (thus it's insane).  As samples accumulate, you'll see the
the table values fill in and shift.

Bottom line - you're probably not connecting to a time server.
I did a google on "ntp servers by region", and it gives a lot of
leads.  You might start at:
http://www.eecis.udel.edu/~mills/ntp/servers.html

I usually use the IP address of the server on the server entry.
Excluding comments, my /etc/ntp.conf looks like:

% grep -v ^# /etc/ntp.conf

server  194.9.200.2     # ahab - inside our firewall

server  127.127.1.0     # local host
fudge   127.127.1.0 stratum 10

driftfile /etc/ntp.drift
authenticate no


Hope this helps.  I out of town for while.
-Bob Arendt