On Sat, 31 Dec 2016 20:10:05 -0700, Jack J. Woehr <[email protected]> wrote:
>Paul Gilmartin wrote:
>> I had believed it happened at 23:59:60 UTC regardless of what time zone
>> you're
>> in. So, in America/Denver, 16:59:60.
>
>We have empirical proof of that at one of my clients . Took down a monitoring
>system this evening.
>
But that's not what Walt was discussing.
>Was peacefully spawning services and suddenly the last-started time was in the
>future and it spawned 100 in a second and choked.
>
But why? It should perceive it's ahead of the game and take a (short)
coffee break. (" The last started was 23:59:59.98. So I need to start the
next one at 23:59:59.99. But it's only 23:59:59.01, so I can relax.")
Details? What OS? z/OS? Other (specify)?
So it was slavishly following NTP (or other?) which duplicates the 23:59:59
second
because POSIX doesn't allow it to Do the Right Thing and the "monitoring
ssystem"
did no filtering at all on the NTP (whatever) input.
The world ought to adopt a rational (not yet close to standard) accommodation:
https://developers.google.com/time/smear
... and avoid even the milder disturbance such as:
http://members.iinet.net.au/~nathanael/ntpd/leap-second.html
... which would be tough on z/OS, which has committed to a different path.
-- gil
----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [email protected] with the message: INFO IBM-MAIN
