NTP Problem

[Wilder Mellotto]

Hi.

I have a problem to syncronize my server with somes NTP server's.
My server is hp-ux 11.0 and I aways syncronized my server with "time.nist.gov" without problems, but now it can't be done.

My /etc/ntp.conf was:
server time.nist.gov minpoll 1 version 3

And this configuration worked very well and I don't understand what's happening

Then I tried to use more servers:
server time.nist.gov minpoll 1 version 3
server utcnist.colorado.edu minpoll 1 version 3
server nist1.datum.com minpoll 1 version 3
server time-a.nist.gov minpoll 1 version 3
server time-b.nist.gov minpoll 1 version 3

Here the output of "ntpq -p"

remote refid st t when poll reach delay offset disp
===================================================
time.nist.gov 0.0.0.0 16 - - 16 0 0.00 0.000 16000.0
india.colorado. 0.0.0.0 16 - - 16 0 0.00 0.000 16000.0
nist1.symmetric 0.0.0.0 16 - - 16 0 0.00 0.000 16000.0
time-a.nist.gov 0.0.0.0 16 - - 16 0 0.00 0.000 16000.0
time-b.nist.gov 0.0.0.0 16 - - 16 0 0.00 0.000 16000.0

[James R. Ferguson]

Hi:

If you server's time differs from the source to which you are trying to synchronize by more than ~1000 seconds, all attempts will fail and the 'xntpd' daemon will die. It that case, you must first adjust your server's time.

Regards!

...JRF...

[Jaime Bolanos Rojas.]

Wilder,

Most likely somebody did a change in the firewall in your company.

Please make sure that the traffic from your ntp server still can go thrught that firewall back and forth.

Regards,

Jaime.

[Wilder Mellotto]

#ntpdate -db time.nist.gov
11 Aug 15:00:29 ntpdate[1102]: ntpdate version=3.5f; Thu Oct 23 13:37:24 PDT 1997 (6)
transmit(192.43.244.18)
receive(192.43.244.18)
transmit(192.43.244.18)
receive(192.43.244.18)
transmit(192.43.244.18)
receive(192.43.244.18)
transmit(192.43.244.18)
receive(192.43.244.18)
transmit(192.43.244.18)
server 192.43.244.18, port 123
stratum 1, precision -18, leap 00, trust 000
refid [ACTS], delay 0.37207, dispersion 0.00587
transmitted 4, in filter 4
reference time: c8874353.32981387 Fri, Aug 11 2006 14:56:35.197
originate timestamp: c8874385.03e0ef99 Fri, Aug 11 2006 14:57:25.015
transmit timestamp: c887443e.9211e000 Fri, Aug 11 2006 15:00:30.570
filter delay: 0.37656 0.37207 0.37698 0.42477
0.00000 0.00000 0.00000 0.00000
filter offset: -185.730 -185.727 -185.731 -185.755
0.000000 0.000000 0.000000 0.000000
delay 0.37207, dispersion 0.00587
offset -185.727542

11 Aug 15:00:30 ntpdate[1102]: step time server 192.43.244.18 offset -185.727542 sec

[Bill Hassell]

It looks like ntpdate got through and jumped the time about 185 seconds. The NIST timeservers are heavily utilizied and may not always be available. You might try using the NTP pool:

http://www.pool.ntp.org/

And in ntp.conf, try just the server names and let the protocol take care of the rest:

server 0.pool.ntp.org
server 1.pool.ntp.org
server 2.pool.ntp.org

When ntpq -p shows no sync, try polling one of the remote servers as in:

ntpq -p time.nist.gov

[Wilder Mellotto]

I edited my ntp.conf

#cat /etc/ntp.conf
server 0.pool.ntp.org
server 1.pool.ntp.org
server 2.pool.ntp.org


Restarted my XNTPD and then
#ntpq -p
remote refid st t when poll reach delay offset disp
=================================================
rapwap.razor.dk 0.0.0.0 16 - - 64 0 0.00 0.000 16000.0
213.222.11.219 0.0.0.0 16 - - 64 0 0.00 0.000 16000.0
203.86.196.5 0.0.0.0 16 - - 64 0 0.00 0.000 16000.0

This result doesn't change, but when I run

#ntpq -p 1.pool.ntp.org
remote refid st t when poll reach delay offset disp
================================================
-172.17.0.12 KMMCMLXXX.dsl.s 2 u 453 1024 377 4.93 4.071 14.84
*ntp1.sp.se .PPS. 1 u 384 1024 377 10.94 -0.226 14.84
+ntp2.sp.se .PPS. 1 u 1032 1024 377 10.69 0.145 14.82
+ntp1.sth.netnod .PPS. 1 u 488 1024 377 1.00 1.560 14.83
+timehost.lysato .GPS. 1 u 406 1024 377 4.92 1.537 14.85

I don't understand why I can't syncronize.

[paolo barila]

Hi,

if possible:

- stop xntpd

- align your clock to server

# ntpdate ntpserver

(your system date will do a jump to correct time)

- start xntpd

[Wilder Mellotto]

Paolo, I already synchronized my clock with ntp server. In ntpq -p output the offset is very short -180~
ntpdate already done.

[paolo barila]

sorry, imo 180 seconds is not short

[Wilder Mellotto]

Paolo, I work with ntp only with the basic configuration, but I already synchronized my servers with 12 minutes of difference. Why 180 seconds is too big?

Now my server is 24 secs and ntpq -p still with same message I posted.


#ntpdate -db 1.pool.ntp.org
16 Aug 11:28:27 ntpdate[3686]: ntpdate version 3.5f: Fri Dec 10 18:32:44 GMT 1999 PHNE_19711
transmit(193.216.51.18)
receive(193.216.51.18)
transmit(193.216.51.18)
receive(193.216.51.18)
transmit(193.216.51.18)
receive(193.216.51.18)
transmit(193.216.51.18)
receive(193.216.51.18)
transmit(193.216.51.18)
server 193.216.51.18, port 123
stratum 2, precision -19, leap 00, trust 000
refid [192.36.143.153], delay 0.43050, dispersion 0.00865
transmitted 4, in filter 4
reference time: c88da786.8e7d45a5 Wed, Aug 16 2006 11:17:42.556
originate timestamp: c88da9f5.486bacf7 Wed, Aug 16 2006 11:28:05.282
transmit timestamp: c88daa0d.60ab3000 Wed, Aug 16 2006 11:28:29.377
filter delay: 0.44553 0.43050 0.51596 0.43973
0.00000 0.00000 0.00000 0.00000
filter offset: -24.3110 -24.3012 -24.3486 -24.3018
0.000000 0.000000 0.000000 0.000000
delay 0.43050, dispersion 0.00865
offset -24.301264

16 Aug 11:28:29 ntpdate[3686]: step time server 193.216.51.18 offset -24.301264 sec

[paolo barila]

I'm just used to sicronize just before starting xntp