[ntp:questions] Re: NTP broadcastclient update times?

Terje Mathisen terje.mathisen at hda.hydro.com
Fri Oct 15 21:53:17 UTC 2004


W. D. wrote:
> Fragment of ntp.log:
> 
> 14 Oct 01:21:19 ntpd[88]: synchronized to 192.168.2.122, stratum=2
> 14 Oct 01:29:55 ntpd[88]: time reset +1.643005 s
> 14 Oct 01:31:07 ntpd[88]: synchronized to 192.168.2.122, stratum=2
> 14 Oct 01:41:44 ntpd[88]: no servers reachable
> 14 Oct 01:49:15 ntpd[88]: synchronized to 192.168.2.122, stratum=2
> 14 Oct 01:55:43 ntpd[88]: time reset -1.314898 s
> 14 Oct 01:56:55 ntpd[88]: synchronized to 192.168.2.122, stratum=2
> 14 Oct 02:18:12 ntpd[88]: time reset +0.485238 s
> 14 Oct 02:19:25 ntpd[88]: synchronized to 192.168.2.122, stratum=2
> 14 Oct 02:53:42 ntpd[88]: time reset +0.638312 s
> 14 Oct 02:54:56 ntpd[88]: synchronized to 192.168.2.122, stratum=2
> 14 Oct 03:18:24 ntpd[88]: time reset -0.400259 s
> 14 Oct 03:19:36 ntpd[88]: synchronized to 192.168.2.122, stratum=2

There are only two likely explanations for this:

a) The cpu crystal is badly wrong, with a static error of much more than 
500 ppm, but the alternating +/- steps seems to be pointing towards some 
other problem.

b) Some other process is setting or otherwise affecting the cpu clock.

Terje
-- 
- <Terje.Mathisen at hda.hydro.com>
"almost all programming can be viewed as an exercise in caching"



More information about the questions mailing list