[ntp:questions] Leap seconds

David L. Mills mills at udel.edu
Tue Jan 25 02:54:32 UTC 2005


Management frequently has rather amazing procedures when leap seconds 
and spring/fall steps occur. The IBM Sysplex Timer carefully calibrates 
the clock offsets for leap seconds, locl time zone offset and TAI/UTC 
(as specified by the operator). However, and particularly in the fall 
step, I know for a fact that some IBM systems just stop all processing 
during the hour after the switch. I assume these machines are not being 
used for air traffic control, but then all times used in that and the 
maritime community are only in UTC.


Brad Knowles wrote:

> At 1:36 PM -0800 2005-01-24, mayer at gis.net wrote:
>>  NTP is the one that decides how much to change the clock and
>>  tells the O/S. Moving the clock backward by 1 second, which is
>>  in effect what a leap-second does, can have major negative effects
>>  on systems that depend on the time being a monotonically increasing
>>  variable.
>     A leap second would never cause the time to be moved backwards. A 
> leap second shouldn't ever cause the time to be jumped forwards, 
> either.  It should simply result in an extra second that occurs during 
> that minute.

More information about the questions mailing list