| Re: [chrony-users] Possible bug in PPS support |
[ Thread Index |
Date Index
| More chrony.tuxfamily.org/chrony-users Archives
]
On Mon, 23 Oct 2017, Miroslav Lichvar wrote:
On Mon, Oct 23, 2017 at 10:24:54AM -0700, Bill Unruh wrote:
On Mon, 23 Oct 2017, Rob Janssen wrote:
You don't support my calculation that if the clock apparently wandered
away 3400us
Again, no evidence of that 3400 us.
If I understand it correctly, 3.4ms was the offset of the NTP source
My question is how he determined that the offset was 3.4 ms after 13 hours.
Simply looking at the offset from the one of the ntp servers does not cut it.
That is only 2 std dev from the mean.
13 hours after the PPS stopped working. The stddev of the NTP source
from sourcestats is ~50 microseconds, so if the offset was originally
better than 1.6ms (there are 3 different sources in the original
report with -0.1ms, 1.5ms, 1.5ms offsets), it drifted at least by
~1.8ms in that time.
If there was a significant change in the temperature, the error gained
in 13 hours could be much larger. On one of my servers with PPS I see
that the frequency offset can change by 0.5 ppm in just few seconds.
I agree. and chrony PPS does a bad job of measuring that. Perhaps chrony
should keep track of the drift over a much longer period than the measurement
period (max 64 samples are 16 sec per sample is only about 15 min. so, keeping
a list of the drift rate over say a day would give a much better feeling for
the drift wander due to temp differences, etc. It is certainly true that the
drift fluctuations are not guassian so an estimate derived from 15 min really
gives a very poor estimate of the fluctuations on the time scale of hours or
days.
So that part of his concern is certainly valid. On the other hand, being
worried about the loss of connectivity on the 15 min time scale probably is
not, unless he has evidence.
But the evidence is all there in the measurement logs when the pps is running.
He could use that to estimate what the skew is over a variety of time periods.
O
--
Miroslav Lichvar
--
To unsubscribe email chrony-users-request@xxxxxxxxxxxxxxxxxxxx
with "unsubscribe" in the subject.
For help email chrony-users-request@xxxxxxxxxxxxxxxxxxxx
with "help" in the subject.
Trouble? Email listmaster@xxxxxxxxxxxxxxxxxxxx.
--
To unsubscribe email chrony-users-request@xxxxxxxxxxxxxxxxxxxx
with "unsubscribe" in the subject.
For help email chrony-users-request@xxxxxxxxxxxxxxxxxxxx
with "help" in the subject.
Trouble? Email listmaster@xxxxxxxxxxxxxxxxxxxx.