[ntp:questions] Interpretation of frequency error / drift_comp

Daniel Kabs daniel.kabs at gmx.de
Thu Feb 23 11:51:58 UTC 2006


I hope this hasn't been covered here too often before but I could not 
research this question as groups.google.com is not responsive currently.
(note how I skillfully shift the blame for asking a dumb question from 
my shoulders to google).

Assume ntpd is configured on a system for a stable time server. After 
one hour, the drift file is written.  I take it, the drift file contains 
the *frequency error* of the local clock.

The *frequency error* of the local clock can also be monitored using 
"ntpq -c rc". According to the implementation documentation of ntpq
"The status, leap, stratum, precision, rootdelay, rootdispersion, refid, 
reftime, poll, offset, and frequency variables are described in RFC-1305 
specification." I could find a table listing "the complete set of system 
variables" on page 15 of http://www.faqs.org/rfc/rfc1305.pdf but I 
failed to find the frequency variable listed there.

I also searched the source code: ntpq links the text "frequency" to the 
symbol CS_DRIFT and ntpd links CS_DRIFT to the double precision variable 
drift_comp. drift_comp is written to the drift file. I searched the RFC 
for "drift" but didn't hit anything relevant.

Is this correct: A positive *frequency error* means that the local clock 
runs slower than the reference clock. So the local clock's frequency has 
to be increased by the *frequency error* in order to run as fast as the 
reference clock. Vice versa for negative values.

Refactor, don't archive! - SamHasler - 28 Aug 2004 - twiki.org

More information about the questions mailing list