On Sun 2015-05-31T14:28:59 +0000, Poul-Henning Kamp hath writ: > >But they have servers in multiple locations, so if they are running on the > >local time zone they will not leap at the same time... > > Welcome to the (probably painful) world of "heuristic handling of > leap seconds".
And it's not just Microsoft, and not just this next leap. The APNIC APRICOT 2015 lightning talks were mentioned on LEAPSECS back in March. There was a talk about the handling of the 2012 leap second and the desynch that happened among machines running NTP from different servers. Even with NTP the clock behavior in one stratum of client was different from clocks in another stratum because knowing "what time is it" is a measurement process. The slides are at http://www.slideshare.net/apnic/the-leap-second-is-coming-by-tomonori-takada-apricot-2015 The video of the talk is available on the APNIC channel on Youtube: https://www.youtube.com/watch?v=E_AMR3ECIxQ&feature=youtu.be&t=51m Starting at Minute 51. In the case of Microsoft the disclaimer at the bottom of https://technet.microsoft.com/en-us/library/cc773013(v=ws.10).aspx still says W32Time cannot guarantee better than 1 to 2 seconds. UCO/Lick knows that, even with the best available version of NTP, the sloppy (or even lack of) hardware clock on some motherboards means that some Windows systems cannot do better than 1 to 2 seconds. Applications running on networks with these sorts of machines are probably already robust against 1 to 2 second offsets because those happen all the time, but that still means we cannot use the timestamps from Windows for science. According to the best rumors, the failure of Qantas/Amadeus in 2012 was because of the 1 second offset between machines. And in 2012 the financial markets were not running, so as PHK writes, this next leap should be very interesting. -- Steve Allen <[email protected]> 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 [email protected] https://pairlist6.pair.net/mailman/listinfo/leapsecs
