[ntp:questions] Excessive NTP configuration of a stratum 2 server
Richard B. Gilbert
rgilbert88 at comcast.net
Wed Jan 17 16:58:49 UTC 2007
shoppa at trailing-edge.com wrote:
> Eric wrote:
>>For the group's amusement, below are the results of an NTPQ "peer" command to a
>>us.pool.ntp.org server (located in Florida). It also claimed an offset of -0.220 and
>>a root dispersion of 39ms.
>> remote refid st t when poll reach delay offset jitter
>>-18.104.22.168 .PTB. 1 u 400 1024 377 130.298 -1.174 0.244
>>-ntp1.twc.weathe .CDMA. 1 u 168 1024 377 129.244 -5.943 14.381
>> ip-207-145-113- .INIT. 16 u - 1024 0 0.000 0.000 4000.00
> I will agree with "excessive"! But one thing I've noted in other
> similar excessive configurations, is that ntpd starts really being
> really aggressive with declaring falsetickers. Of course it can afford
> to be, it only needs a few to survive the selection process.
Perhaps the documentation needs to emphasize that nptd can't use more
than ten servers at once and that this sort of thing borders on abusive!
I'm sure the server in question would benefit from a much more cluefull
selection. It's ridiculous for a server in Florida to use a server in
More information about the questions