[ntp:questions] Bug 2341 - ntpd fails to keep up with clock drift at poll>7
david-taylor at blueyonder.co.uk.invalid
Tue Dec 3 08:15:41 UTC 2013
On 03/12/2013 04:08, Mischanko, Edward T wrote:
> It appears that you may misunderstand my observation? Raspberry-Pi servers seem to be economical and functional as NTP servers, but how well do they work in the network configuration that I am using. Yes, there are things that I can do to have a more accurate server; but how well does NTP work in default usage?
I did try the Raspberry Pi, LAN-connected, using purely Internet
servers, and the results are plotted here:
Offsets were reported in the +/- 5 millisecond range. This on my Cable
Modem connection which is rather asymmetric with 60 Mb/s downstream and
3 Mb/s upstream (I appreciate that the delays may not reflect the
allowed bandwidths). There was no attempt to limit the poll to less
than the 1024 second maximum. It's likely that the version in use was
I did not detect any failure to keep up with clock drift, but the RPi
was not subject to large temperature extremes, just room central heating
which was switched off between about 22:00 and 06:00 UTC.
I hope that serves as one extra data point.
More information about the questions