[ntp:questions] Re: HOWTO prepare ntpd to the leap of a second
David L. Mills
mills at udel.edu
Sun Dec 25 01:51:30 UTC 2005
From the survivors of the clustering algorithm the protocol runs an
agreement algorithm. The simplest one implemented lights the host leap
bit if any survivor leap bit is lit. You should be able to work out the
many scenarios that can occur.
Harlan Stenn wrote:
> I'm now wondering how ntpd will handle the "problem" of the case where it
> gets the time from a leap-unaware refclock *during* the leap second (given
> that other leap-aware sourced of time are available).
> I suspect it will simply be thrown out as a bad sample.
> I also suspect I am not awake enough to be contemplating such things...
More information about the questions