[ntp:questions] how will this effect stability

Terje Mathisen terje.mathisen at hda.hydro.com
Wed Feb 14 16:30:03 UTC 2007


marknmbox-88 at yahoo.com wrote:
> I am in an environment where the path to and from a
> time server can vary from milliseconds to a second and
> the return path may vary within the same range.  The
> transit time for each direction will be independent of
> the time for the other.  Thus the path to the server
> may take a few mills but the return path can take a
> second.  The sum of both paths can range from a few
> mills to seconds.
> 
> Will this cause ntpd problems with accuracy or
> stability or will the ntp calculations not be
> effected?

Yes, it will, if the periods of severely one-sided delays last for many 
minutes, i.e. long enough that ntpd can go through 5 or more sync cycles.

The only way you can make this work is by having a much longer history 
buffer inside ntpd, allowing it to discover the actual "real round trip 
time", i.e. when everything works OK.

There is a special option which you can use in your ntp.conf file, which 
specifies this buffer, I'm trying to remember the name...google

http://www.eecis.udel.edu/~mills/ntp/html/miscopt.html

What you want is something like 'tinker huffpuff 86400' to make a 
24-hour buffer in which to look for the minimum round-trip time: All 
packets which are delayed significantly more than this time will be 
checked for consistency (i.e. the resulting time must be within the 
possible interval), but will not directly modify the system 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