On Dec 4, 2006, at 11:26 AM, Rob Janssen wrote: >> The server itself (10.0.1.1) is listed as the top-requesting >> client, producing about 25% of the total requests – I’m sure >> that’s not meant to be so, or am I wrong? > This is caused by your script. It is querying your server to get > the stats.
I don't believe that the NTPstats script does perform queries to the NTP server; it simply gathers data going by via tcpdump, rather than generating queries itself. Notice the exceptionally high rate: Clients with rapid updates (min requests of 100): Rank First Seen Client IP Requests Rate Usage Cumulative 1 12/02/06 18:53:47 10.0.1.1 87104 1.30 24.90% 24.90% * ! ...of about ~1 query per second; this almost always indicates a misconfiguration of the NTP access controls which prevents the server from trusting itself. By contrast, I'm running the same stats script, and localhost/the IP of my NTP servers, do not even appear in the top-100 query sources. My servers poll each other at a rate of 1024 seconds... -- -Chuck _______________________________________________ timekeepers mailing list [email protected] https://fortytwo.ch/mailman/cgi-bin/listinfo/timekeepers
