[ntp:questions] Using ntpdate -b SERVER shortly after SERVER boots

Donald Murray, P.Eng. donaldm314 at gmail.com
Mon Feb 12 19:40:02 UTC 2007


On 12 Feb 2007 14:12:49 GMT, Steve Kostecke <kostecke at ntp.isc.org> wrote:
> On 2007-02-10, Donald Murray, P.Eng. <donaldm314 at gmail.com> wrote:
>
> > Okay, I've added iburst to the SERVER as follows:
> > server  127.127.1.0 iburst      # local clock
>
> iburst has _NO_ effect on ref-clocks (including 127.127.1.x).

Ah. Okay, iburst is gone again. :-)

> > Unless anyone has some other suggestions,
>
> You could use 'server 127.127.1.0 minpoll 2' to reduce the default poll
> interval to 16 seconds. So the initial sync time would be reduced to 3 *
> 16 seconds, or 48 seconds. On my ntp-4.2.2 test system this consistently
> produced an intial "LocalCLK sync" of between 50 and 53 seconds.
>
> I don't believe that you'll do better without patching ntpd.
>
> BTW: The use of minpoll _may_ have some side effects.

Yeah, I think we'll stick with the hack I've already described. The SERVER
will be able to tell us its current time much sooner than ntp (even though
ntp would be a more accurate approach).

>
> > it looks like I'm going to have to use /bin/date...
> >
> > More details on power for the curious. These are drilling rigs and all
> > "power" comes from generators.
>
> If you can access these systems to modify the configuration files you
> should be able to drop in a modified ntpd.

You've probably got a point. We could modify ntpd and ship it with
our next release. For the moment, we'll just use /bin/date if ntpd is not
ready to serve time.



More information about the questions mailing list