[ntp:questions] Query about NTP accuracy

David Mills mills at udel.edu
Mon May 25 15:43:10 UTC 2009


Harlan,

The ntpdate program was first written by Mike Petrie and Louie Mamakos 
circa 1984 and so far as I know has not been maintained since. I 
purposely built ntpdate functionality in ntpd some years ago, but that 
has not been popular, probably because that requires a configuration 
file. Properly configured, the time is set within 13 seconds using the 
engineered mitigation and discipline algorithms that ntpdate does not have.

Continued maintenance of ntpdate is not a worthwhile expense; rather, 
those resouces should be put into a command line argument for ntpd that 
could specify one or more servers without requiring a configuration 
file. In this way users would not even know that ntpd was any different 
than ntpdate. That might even be done with a script.

There is precedent in this agenda. We did in fact replace the ntptrace 
program with a script.

Dave

Harlan Stenn wrote:

>ntpdate is broken in many ways, nobody wanted to fix it, and in the interim
>we came up with sntp.
>
>I agree that ntpd -g is usually better than sntp (or ntpdate) to initially
>set the time, but the choice is between "Set the clock well, even if it
>takes a little time" and "Set the clock as quickly as possible and it may be
>wrong."
>  
>




More information about the questions mailing list