[ntp:questions] Re: Linux NTP server problems serving time.

Danny Mayer mayer at gis.net
Sat Jul 16 18:24:02 UTC 2005


Maarten Wiltink wrote:
> "Danny Mayer" <mayer at gis.net> wrote in message
> news:42D86196.5050105 at gis.net...
> [...]
> 
>>Rule #1 of NTP is that you need at least 3 servers. You only specified
>>one. It's not going to be able to work this way.
> 
> 
> Nonsense. I would use a stronger word if you were someone else.
> 
> NTP disciplines and synchronises a clock. It does this based on
> information about other clocks. 

Right, but you have to have valid information first.

If it has information about only
> one other clock, it will simply track it. But this does _not_ break
> the chain to a reference clock.
> 

It's not a matter of breaking a chain, it's a matter of having a valid 
reference in the first place.

> With a single server, you don't really know how far out it is, but
> with three servers you don't really know if you have one falseticker
> or two colluding ones either.
> 

That's always true. However, based on lack of information to the 
contrary you need to trust some NTP server or refclock.

Danny

> Groetjes,
> Maarten Wiltink
> 
> 
> _______________________________________________
> questions mailing list
> questions at lists.ntp.isc.org
> https://lists.ntp.isc.org/mailman/listinfo/questions
> 




More information about the questions mailing list