[LEAPSECS] USNO NTP incident of 2012-11-19

[mike cook]

Le 26 nov. 2012 à 01:26, Steve Allen a écrit :


> On Sun 2012-11-25T19:59:14 +0000, Matsakis, Demetrios hath writ:

>> And I suppose many on this list will have even more to say ...

> 


Like you Steve, I think the silence is basically due to what you are complaining about, that is the lack of available technical analysis of the incident. I couldn't find any. 


> I think this list is largely silent while wondering whether the ITU-R

> is capable of doing anything besides the non-decisions during this year.


Any news from the WP7A junket to Ecuador?  


> 

> Thank you for the report on the USNO incident.  My concern upon

> reading this is that the notice on the web page

> http://tycho.usno.navy.mil/ntp.html

> looks as if it is a temporary modification which will, in short order,

> simply vanish from view.

> 


There is a lot of anecdotal evidence of its effects on the web. The USNO response in that link is frankly insufficient, but what it clearly indicates is that even the time lords don't have robust procedures in place to assure vital time dissemination. The big issue is of course is that incorrect data was being propagated and not just no data.  One would  hope that they have plugged that hole.


> Behind many arguments about leap seconds are references to previous

> events and decisions which, upon inspection, turn out not to be

> mentioned in the available documentation.  Among the effects of

> that are

> => in the future, sysadmins will not have access to documents which

> can warn about the severity of the effects and the importance of

> robust configurations


The fact that it caused so much grief is due to the lack of understanding some sysadmins have of the configuring NTP. Those affected will have probably fixed their setups, but new boys on the block won't have the history to prevent them falling into the same traps.   
NTP could itself do a sanity check at startup and simply kill itself if there is not enough redundancy ( need a non default override) and also log any descent into the danger zone of 2 clocks. 


> => similarly, delegates to subsequent ITU-R meetings about the

> broadcast time scale will lack basis for making their decisions

> 

> My question:  Is there a published description of this event which

> will be citeable for the sake of posterity?

> 

> --

> Steve Allen                 <sla at ucolick.org>                WGS-84 (GPS)

> UCO/Lick Observatory--ISB   Natural Sciences II, Room 165    Lat  +36.99855

> 1156 High Street            Voice: +1 831 459 3046           Lng -122.06015

> Santa Cruz, CA 95064        http://www.ucolick.org/~sla/     Hgt +250 m

> _______________________________________________

> LEAPSECS mailing list

> LEAPSECS at leapsecond.com

> http://six.pairlist.net/mailman/listinfo/leapsecs