[ntp:questions] Re: Make NTP timestamps leap-second-neutral (like GPS time)

David L. Mills mills at udel.edu
Wed Jan 14 02:40:58 UTC 2004


I think you are asking if anyone actually observed NTP behavior at
midnight New Years Eve when running NTPv4 in an operating system with
the leap state machine, with sufficient external comparison to know that
the leap was or was not successful, and sober enough to know all these
things, actually did record and notify the public.

I am probably the only person on the planet stupid enough to do all
these things on scheduled leaps for the last 25 years. I can report that
the latest leap represented by the last second of 1998 using NTPv4 with
the WWV reference driver to deliver the leap warning bit and the kernel
leap state machine did in fact correctly insert the second as intended.
I can speak for no other site, system, version, driver, observer or
degree of intoxication of the latter.


Michel Hack wrote:
> hack at watson.ibm.com (hack) wrote in message news:<btht5j$bbm$2 at news.btv.ibm.com>...
> > In article <bthstv$bbm$1 at news.btv.ibm.com>, hack <hack at watson.ibm.com> wrote:
> >
> > So, here's a simple question:  Was there an NTP jiggle problem after the
> > last LSE in July 1999, and is there likely to be one at the next LSE --
> > or have the algorithms already improved to the point that the pinball
> > effects described so vividly in 1994 have vanished?
> I used google search for "leap second ntp" for a 1-month period around each
> of the last 4 leap seconds.  There were a couple of items about GPS receivers
> not handling the event promptly and some jumpy NTP reactions to such failures;
> there was also a report about NTP v4 misbehaving (when NTP v3 handled events
> apparently well) -- but nothing like the pinball effect described for 1994 and
> its implication that this was "a tradition going back many years".
> Does this mean that the last 3 leap second events were (in the absence of
> misbehaving hardware) perfectly smooth from NTP's point of view?  And has
> NTP version 4 been fixed in this respect (if indeed it was broken)?
> Michel.

More information about the questions mailing list