[ntp:questions] ntpdate function

Jacek Igalson (at) jacek.igalsonorange.com
Fri Apr 12 13:54:46 UTC 2013


Uzytkownik "David Woolley" <david at ex.djwhome.demon.invalid> napisal w 
wiadomosci news:kk5mt4$9l0$1 at dont-email.me...
> E-Mail Sent to this address will be added to the BlackLists wrote:
>> Jacek Igalson wrote:
>>
>>> 2. ntpdate can be used with option -p, 1 to 8 sampling.
>>>    Difference of delay is 0.05586 sec to 0.05359 sec.
>>>    Result from ntpq is 28 msec.
>>>    To big to be a statistical explanation
>
> That would be consistent with a factor of two, not a fixed offset.  Cold 
> it be that ntpdate is reporting round trip delay, whereas ntpq is 
> reporting the delay figure actually used to correct for time of flight, 
> which is half that?
>
I check it for the LAN located server with short time trip:
ping => 2msec and ntpdate => 28msec
It seems to be fixed offset


>
>>  ICMP echo may have a lower priority than NTP;
>>
>
> ICMP echo is likely to be handled by the interrupt service routine, with 
> no process scheduling delays. 




More information about the questions mailing list