[ntp:questions] Re: Correcting my time servers clock drift on Alpha ES40s / Tru64

Richard B. Gilbert rgilbert88 at comcast.net
Sun Oct 23 15:41:05 UTC 2005

Tom Smith wrote:

> David L. Mills wrote:
>> Note the prefix of your log messages. It is not possible for the 
>> kernel to simulate that. I conclude the source has been changed. I 
>> have no problem at all with that, but the only way I can be 
>> completely confident with help is with the unabridged version that 
>> leaves here.
> Dave,
> The prefix on log messages comes strictly from the name of the executable
> (and its PID), not from the bits within it. It is used to show the source
> of the message. The version is shown in the log at the time (x)ntpd 
> starts:
> 20 Jul 17:40:44 xntpd[764]: ntpd 4.0.98a Wed Oct 16 17:36:24 EDT 2002
> To the best of my knowledge, all versions of (x)ntpd announce their 
> actual
> versions in this way.
> None of this has anything to do with the original question, of course,
> which was, essentially, "Why do my clocks that have no fixed reference
> not keep accurate time?".
> -Tom

Your clocks do not keep accurate time because they are not accurate!  
They drift!  The typical computer clock keeps time less well than the 
typical cheap wristwatch.   The wristwatch is designed to keep time and 
people will not buy it if it does not keep time.   The computer is not 
designed as a time keeping device.  Time keeping is an afterthought.

No configuration of ntpd that does not include an external source of 
accurate time is likely to help you.  That external source of accurate 
time may be an internet connection, a GPS receiver, a dial-up phone call 
to Automated Computer Time Service (ACTS) operated by the National 
Institute of Standards and Technology (NIST), or a radio receiver 
equipped to receive and decode the time signals broadcast by WWV or 
WWVH  (2.5, 5, 10, 15, and 20 MHz) or WWVB (60 KHz).  All three of these 
stations are operated by NIST.   If you are rich, you can even purchase 
your own atomic clock, have it calibrated by NIST and interface it to 
your computers. 

Asking two inaccurate clocks to synchronize to each other may, if you 
are lucky, cause them to keep the same time but in no way will it cause 
them to keep accurate time!

More information about the questions mailing list