[LinuxPPS] Re: strange readings...

[Udo van den Heuvel]

Udo van den Heuvel wrote:
> Udo van den Heuvel wrote:
>> Udo van den Heuvel wrote:
>>> Udo van den Heuvel wrote:
>>>> ntp-4.2.4p0 I have some different readings in this mrtg-made ntp
>>>> statistics plot:
>>>>
>>>> http://pindarots.xs4all.nl/mrtg/ntpstat.html	
>>> Also the pll is kinda high...
>>>
>>> http://pindarots.xs4all.nl/mrtg/ntppll.html
>> Nobody has an idea why the readings change and are so much worse?
>> Is it the LinuxPPS kernel? (aqlso 2.6.20.7 -> 2.6.21.1)
>> Or the ntpd? (4.2.4p0 now)
> 
> I just went back to 2.6.20.7 and ntpd 4.2.4 to see if the graphs get
> back to the original behaviour.
> Does anybody know what changed in between ntp version 4.2.4 and 4.2.4p0?
> If it's nothing big, could Rodolfo explain the big change in behaviour?

http://pindarots.xs4all.nl/mrtg/ntpstat.html now shows the old, more
accurate behaviour. Does anybody here see similar differences in
performance?