[ntp:questions] Stick to PPS, even if the prefer server fails
stenn at ntp.org
Thu Mar 26 23:24:57 UTC 2009
There is an apparent shortcoming in the existing PPS model/implementaion
that I am still trying to wrap my head around.
https://support.ntp.org/bugs/show_bug.cgi?id=557 has some information on
this and I am looking for the other thread on that topic.
As I understand it, and I almost certainly have the following wrong in at
least some ways, there are folks out there who want to support multiple PPS
inputs to NTP, and they also want to use a model where ntp's PPS driver
maintains "health state" information about each PPS source. Dave maintains
that the existing infrastructure can already mostly handle this, and while
some work will be needed, the first set of folks are reinventing too much of
What I have been (slowly) working on is figuring out where the two
above-mentioned groups are miscommunicating, and identifying exactly what
needs to be done to come up with a solution.
Harlan Stenn <stenn at ntp.org>
http://ntpforum.isc.org - be a member!
More information about the questions