[ntp:questions] Re: HOWTO prepare ntpd to the leap of a second

Serge Bets serge.bets at NOSPAM.laposte.invalid
Tue Dec 27 13:26:22 UTC 2005


Hello David,

 On Sunday, December 25, 2005 at 1:36:26 +0000, David L. Mills wrote:

> If Autokey is running with a number of servers, only the latest data
> are used. Since old leap epoches don't change, the only thing that can
> happen is the addition of a new epoch at the next opportunity.

Another thing could happen: IERS decision of *no* leap second insertion
at the next opportunity. Result for leap-seconds file is then:

 - Unchanged events table.
 - Unchanged lastmod timestamp.
 - Unchanged filename extension.
 - Expiration timestamp pushed by 6 monthes.

| $ awk '/^#@/{print $2}' leap-seconds.3331497600
| 3360441600

This is currently the NTP timestamp of 28 June 2006. Sometime early
January 2006, IERS would decide to not brake the Earth, and thus no leap
second insertion will be necessary on 30 June. They would announce it in
their bulletin C29. Sometime later in January 2006 NIST would update the
leap-seconds.3331497600 file, with one change only: Expiration date
advanced to 29 December 2006.

The NIST file is absolutely true until expiration. This fact is not
exploited by the NTP daemon, but could perhaps be.


> With Autokey each association is restarted once per day, so the latest
> leapsecond data will be retained from each server. If those servers do
> the same thing, it may take a few days for the data to flow from an
> authoritative source.

OK: Thanks for the confirmation. I had misunderstood your previous
explanation.


Thankfully, Serge.
-- 
Serge point Bets arobase laposte point net




More information about the questions mailing list