[ntp:questions] Clock accuracy stats

David Woolley david at ex.djwhome.demon.co.uk.invalid
Mon Jun 16 07:02:34 UTC 2008


Alan wrote:
>     I installed and setup NTP on my PC as a client, using the server
> time.nist.gov.  I would like to view accuracy statistics.  I tried the

Incidentally, that server is heavily overloaded and is unlikely to be a 
good choice compared with a nearby stratum 2 server.  As a general 
principle leaf nodes shouldn't be using stratum one servers.  For 
non-critical, personal, systems the servers provided by any competent 
ISP should do.

> "ntpq -p" command, but I get the error: "No association ID's
> returned".

I'd assume that it has failed to read the configuration file and has 
started without any servers.  I think it might also happen if you fail 
to resolve the domain names, which might be a service dependency issue.

With regard to the subject, these tools cannot tell you the accuracy of 
the clock, they can only give you information about the quality of the 
time measurements that are being made.  In particular, providing you are 
getting offsets scattered evenly around zero, the clock's variation is 
likely to be a lot less than the offsets you are seeing, but there may 
be a systematic bias in the time that is not reflected in the statistics.




More information about the questions mailing list