>> For people who left the NTP pool, it might be usefull to run
>> http://www.vanheusden.com/time/jans/ for a while: it acts as a real NTP
>> server but then returing random or constant or backticking time. People
>> might then realise sooner that something is wrong (=they're using a no
>> longer existing server).
>
> This approach may not be as effective as you think. In a typical ntpd
> setup with multiple upstream NTP servers, ntpd doesn't really care if
> one of the upstream servers goes insane. It's simply marked as bad, and
> the other servers are used instead. However, this doesn't reduce the
> traffic to the bad server at all -- it might even slightly increase in
> certain situations.
> If you really must do something "active" instead of just not responding
> to the requests, please consider using Kiss-of-Death (KoD) packets. This
> is the official method of telling clients to stop sending requests to
> the server. Please read http://tools.ietf.org/html/rfc5905#section-7.4
> ntpd can also output these KoD packets with the proper configuration,
> there's no need for any external software.
When developing this software I was thinking of simplistic software
which only send SNTP messages querying time.
Folkert van Heusden
--
www.vanheusden.com/multitail - multitail is tail on steroids. multiple
windows, filtering, coloring, anything you can think of
----------------------------------------------------------------------
Phone: +31-6-41278122, PGP-key: 1F28D8AE, www.vanheusden.com
_______________________________________________
pool mailing list
[email protected]
http://lists.ntp.org/listinfo/pool
