[ntp:questions] Re: Synchronisation lost with swisstime.ethz.ch
martin.burnicki at meinberg.de
Sat Aug 13 14:32:49 UTC 2005
Eugen COCA wrote:
> bash-2.05$ ntpq -p swisstime.ethz.ch
> remote refid st t when poll reach delay offset
> LOCAL(0) LOCAL(0) 12 l 39 64 177 0.000 0.000
> *GENERIC(0) .GPS. 0 l 36 64 177 0.000 -100.87
> PPS(0) .PPS. 0 l 43 64 7 0.000 -100.62
> +ntp1.ptb.de .PTB. 1 u 37 64 177 35.030 -86.828
> bash-2.05$ ntptrace swisstime.ethz.ch
> swisstime.ee.ethz.ch: stratum 1, offset 0.173229, synch distance
> 0.21143, refid 'GPS'
> As you can see, the server is 100ms off and the Stratum number is 1.
This means that the server's system time is off by 100 milliseconds from the
GPS/PPS, and 86 milliseconds off the PTB server. The difference between the
PTB server and the GPS is about 14 milliseconds, and the jitter of the PTC
server is 35 milliseconds.
Due to the fact that the reach values of all but the PSS source are 177, it
looks like they have just restarted the NTP server a few minutes before.
Since the NTP daemon accepts the PPS source only after it has synchronized
to the ref clock associated with the PPS (in this case the GPS clock) the
reach of the PPS source is below the reach values of the other ones.
So the NTP server's system time is has been off by about 100 milliseconds
when the NTP daemon was started. This is below the 128 millisecond limit
beyond which the system clock would have been stepped, so the system time
currently seems to be slewed.
More information about the questions