[ntp:questions] Keeping NTP Honest Redux
unruh-spam at physics.ubc.ca
Sat Nov 7 17:05:38 UTC 2009
"David J Taylor" <david-taylor at blueyonder.not-this-bit.nor-this.co.uk.invalid> writes:
>"Evandro Menezes" <evandro at mailinator.com> wrote in message
>news:d9a76148-a9c9-45d5-ac72-4e7588c61cca at m7g2000prd.googlegroups.com...
>> I've come to the conclusion that since NTP doesn't correct for
>> temperature variations, it's long-term compensation is counter-
>> productive. Therefore, instead of letting the poll period increase to
>> the default 1024s, perhaps it'd be best to limit it to 64s. However,
>> this requires that each server has the option maxpoll attached to it.
>> It'd be more convenient to have set this parameter for all servers
>> from a single place, such as the old tinker maxpoll option.
>> So, why was tinker maxpoll removed? Any chance to get it back in?
>What sort of accuracy are you wanting, and what OS are you running? You
>could add a local stratum-1 server with a GPS source if you are after
>really good performance, or you could lock your PCs to one local system
>with 64s poll and have that system sync over the Internet with the default
>automatic poll. Some public servers may consider excessive polling to be
>an attack and disconnect you.
>Here is what you can get with a Garmin GPS 18x LVC puck and Windows-7 (an
>OS not known for its timekeeping). Use FreeBSD and you would likely get
>at least ten times better:
And for a linux system, (but the machine is not heavily loaded at all) see the
"string" graph on www.theory.physics.ubc.ca/chrony/chrony.html. String is
connected to and GPS18LVC and runs ntp. The rest of the machines run chrony off
string. as server. So yes, at least 10 times better than his Windows machine.
This graph shows both the rate and the offset fluctuations.
>A Windows XP system synced to that stratum-1 PC shows these temperature-
>and load-related variations:
More information about the questions