[ntp:questions] False Reports of Leap-Seconds

David E. Ross nobody at nowhere.not
Mon Jan 26 00:00:34 UTC 2009


On 1/25/2009 11:45 AM, David Woolley wrote:
> David E. Ross wrote:
> 
>> NTP servers that report leap-seconds within a month are buggy and should
>> be fixed.  I have attempted to send E-mail messages to the operators of
> 
> NTP servers need to start announcing more than a day in advance, so I'm 
> not sure what you mean by within a month.  The reason for this is the 
> time that it can take to propagate from stratum 1 to stratum 15.
> 
>> buggy servers that I have detected.  This is not always successful
>> because listed E-mail addresses are sometimes obsolete.

I believe the two-bit leap indicator in the NTP message indicates that a
leap-second will occur at the end of the current day.  At least, that's
how my NTP client (SocketWatch by RoboMagic) is interpreting the
message.  That's also per RFC 4330, the latest specification that I can
find for the protocol.

Leap-seconds are, of course, announced by the IERS well more than a
month in advance (announced almost six months in advance for the 31
December 2008 leap-second).  That announcement is repeated in the United
States by the USNO.

-- 

David E. Ross
<http://www.rossde.com/>.

Don't ask "Why is there road rage?"  Instead, ask
"Why NOT Road Rage?" or "Why Is There No Such
Thing as Fast Enough?"
<http://www.rossde.com/roadrage.html>




More information about the questions mailing list