[ntp:questions] Time going backwards

Steve Kostecke kostecke at ntp.org
Thu Aug 14 16:00:28 UTC 2008

On 2008-08-14, Unruh <unruh-spam at physics.ubc.ca> wrote:
> Paul.Croome at softwareag.com writes:
>>There's also the theoretical possibility of a negative leapsecond,
>>but of course that's intentional.
> It is not supposed to step backwards then It is supposed to take a second
> seconds to move forward at a glacial pace. -- ie, if during that time the
> clock is read, it is supposed to give a later time than the previous
> reading (by a microsecond or so), but otherwise it is stopped.
> I guess it is true that if the time is out by more than 125ms is it it does
> step.

The default step threshold is 128ms.

This threshold may be changed with the 'tinker step N' command.

Stable release documentation:


Dev release documentation:


Steve Kostecke <kostecke at ntp.org>
NTP Public Services Project - http://support.ntp.org/

More information about the questions mailing list