Folks, and especially people who know the ntpd code,
We are now nearly halfway through August, and I'm wondering where we
stand on the issue of bogus leap seconds. We had one at the beginning
of this month for unknown reasons - but they seem like they could have
stemmed from an ntpd bug: http://bugzilla.ntp.org/show_bug.cgi?id=2246
("sys_leap is sticky").
Questions in what I think are priority order:
1. How can administrators ensure that they will not experience another
bogus leap second at the end of August? Will proper installation of a
current leap second file suffice?? Closely related to this is: Can
someone who knows or can read the ntpd code, and/or who is clear on the
intent (and hopefully the implementation) of the code, answer this
question from Martin a previous thread:
- from which sources [does ntpd accept] a leap warning under which
conditions: from
a single upstream server (earlier NTP versions) or only from a majority
(current NTP versions), from refclocks, from a leap second file, and *which
priorities come into effect if there are upstream servers *and* one or more
refclocks *and* a leap second file, or another combination of those*.
2. When can the answers to the above questions be clearly documented,
presumably at or near
http://support.ntp.org/bin/view/Support/ConfiguringNTP ?
3. What is the current status of the stratum 1 servers vis-a-vis the
leap-second bits? I don't know a good way to survey the servers to see
what the current status is; at last check in an earlier threads
(sometime this month), a substantial fraction of those servers were
still advertising a (bogus) leap second. Can anyone check and/or
provide a script that can be run to acquire that info?
4. What are the next steps to tracking down the issue(s) leading to the
bogus leap second at the end of July 2012? Is someone who is familiar
with the ntpd code willing/able to address that?
Thanks very much,
--Jeff
_______________________________________________
questions mailing list
questions@lists.ntp.org
http://lists.ntp.org/listinfo/questions
