Hello everyone, this time my graphs look much better, score permanently at 20.0, offset used to dance in (-1m,1m) interval, what made me really happy, but now I can see some "random" +/- 10ms jumps:
http://www.pool.ntp.org/scores/193.87.160.18/graph/offset.png I know it's still "fine", but I am curious if there is something I can do with that (I want to have a REALLY stable time server). - There was no special activity, higher cpu or net usage, nothing - Temperatures are just fine - I made no config changes - So I guess it is either very special network problem, or (more realistic guess) problem with upstream ntp servers? This is what I have: :~# ntptrace localhost: stratum 2, offset -0.000255, synch distance 0.002466 8*****************.at: stratum 1, offset 0.000003, synch distance 0.000407, refid 'GPS' :~# ntpq -p remote refid st t when poll reach delay offset jitter ============================================================================== +ntp.*******.cz .PPS. 1 u 78 1024 377 7.719 0.381 0.095 +time.******.cz .ATOM. 1 u 453 1024 377 5.695 -0.386 0.095 *8********51 .GPS. 1 u 613 1024 377 4.933 -0.351 0.320 I am now syncing five servers with pool zone sk, which contains this server too and I get spammed with "kernel time sync status change 4001" and "kernel time sync status change 0001". Maybe it's related or important? I can also post my ntp conf if necessary, but it's pretty much default deb config. Thanks for any ideas. Cheers. _______________________________________________ timekeepers mailing list [email protected] https://fortytwo.ch/mailman/cgi-bin/listinfo/timekeepers
