I have a PC directly connected to a GPS clock with PPS signal that sets the clock to an accuracy of 5 µs (!) or less (shows in NTP as a normal server).
Now I want to measure how bad the internal software clock is. Therefore I am searching for an option to actually poll a server and don’t set the time but actually let it go off.
I could mark the server as noselect in the ntp.conf file, but then ntpd won’t write the loopstats file.
How to let the ntpd write the loopstats file with the offset to the server, but don’t acutally set the time?
ntpdate command line tool provides the
-q (query) option for querying a remote NTP server and displaying the offset from your local clock. For example:
$ ntpdate -q us.pool.ntp.org server 188.8.131.52, stratum 2, offset 0.004806, delay 0.04530 server 184.108.40.206, stratum 2, offset 0.003129, delay 0.13052 server 220.127.116.11, stratum 2, offset 0.006828, delay 0.08769 2 Feb 06:32:54 ntpdate: adjust time server 18.104.22.168 offset 0.004806 sec
The interesting data here is that the very end:
offset 0.004806 sec. If this is the metric you’re most interested in then you should be all set. It doesn’t provide the level of details of the
loopstats file, but it should help out.
Leave a comment
- Understanding redundant power supplies
- Is there a way for administrators to disable users from installing Firefox extensions?
- Is there research material on NTP accuracy available?
- How to create a limited “domain admin” that does not have access to domain controllers?
- Can Windows RDC admin users be immune from being kicked?