| Re: [chrony-users] understanding tracking.log |
[ Thread Index |
Date Index
| More chrony.tuxfamily.org/chrony-users Archives
]
On Tue, Jul 21, 2015 at 08:23:10AM +1200, Bryan Christianson wrote:
> In looking at /var/log/tracking.log I see
>
> Date (UTC) Time IP Address St Freq ppm Skew ppm Offset L Co Offset sd Rem. corr.
> ==================================================================================================
> 2015-07-20 20:13:34 10.64.1.25 2 30.665 0.177 -3.530e-06 N 2 6.553e-05 1.151e-05
>
> I am having trouble understanding the field 'Offset sd' and what i represents. It is consistently an order of magnitude greater in value than the 'Offset' field so I presume it is not the standard deviation of offset.
Yeah, I agree this may be confusing. The offset sd value is meant to
be an error estimate of the offset, it's not actually calculated from
the values printed in the offset column.
It's the estimated deviation of the intercept from the linear
regression. Depending on how well the clock tracks the source, it may
be consistently larger or consistently smaller than the offset value
(or its RMS value in the chronyc tracking output).
The other reason why it may be significantly larger is that when
combining sources (Co > 1 in the tracking log) it also includes the
offset between the combined sources. So even if each source is
perfectly stable, their offsets may not be very close and the offset
sd value will still be large.
> "chronyc tracking" displays a result for RMS Offset consistent with 'Last Offset' as seen here:
>
> Reference ID : 10.64.1.25 (fleur.lan.seaviewsound.co.nz)
> Stratum : 2
> Ref time (UTC) : Mon Jul 20 20:20:34 2015
> System time : 0.000007302 seconds slow of NTP time
> Last offset : +0.000004646 seconds
> RMS offset : 0.000005560 seconds
--
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.