[ntp:questions] NTP Leap Second Trace program updated....
davehart at gmail.com
Wed Mar 25 21:27:42 UTC 2009
On Mar 25, 9:08 pm, Unruh <unruh-s... at physics.ubc.ca> wrote:
> If you have enough servers they should be outvoted.
That's true for -dev (4.2.5) but not -stable (4.2.4) which still has
the old "or" logic where any surviving source having a leap bit set
causes the local leap bit to be set.
> Note that ntp will ignore leapsecond flags except on June 30 and
> Dec 31. While Mar 31 and Sep 30 are also possible leapsecond times, the
> current ntp (well the last I looked at was 4.2.4p4) does not recognize them.
4.2.5 (-dev) will respect a leap insertion at the end of this month.
It will not arm the leap bit until the last 24 hours, so if you're
using -dev, be sure to give David J Taylor's leap second checker a try
soon, or else look at logs on the last day of the month, to avoid a
false leap second insertion.
More information about the questions