Best practices question

Paul Theodoropoulos paul at
Wed Jun 26 02:07:45 UTC 2019

On 6/25/2019 17:08 PM, Hal Murray wrote:
>> Is this a reasonable setup ...
> Sure.  I expect it will mostly ignore the external servers - just use them for
> sanity checking.
> You are in a position to run some experiments.
> How good is your network connection?  Plot the round trip time.  Maybe try a
> server that is not nearby.  See if anything happens when you do a long
> download.
Network is average DSL, 55/in, 6/out, so hardly ideal. a-ntpsec has been 
in service for more than a year and is quite stable. Thing is - I'm not 
sure what I'd learn about this new setup with three servers on the LAN, 
from plotting RTT to remote NTP servers...? I do know that darned near 
*any* activity I do on the command line of any of them tends to cause the 
the graphs to become considerably more spikey. For that reason, I do 
compiling of ntpsec, gpsd, etc on the separate ntpviz server, then copy 
the 'built' environment across. Still gives a spike for the 'install' 
part, but only a blip.

> You can try using the NMEA driver direct rather than going through GPSD.  Use
> separate drivers for serial and PPS.  There is logic in there to give PPS
> special treatment.
So I can run NTPsec without GPSD? For some reason I thought it was 
required to interpret the PPS from the GPS. I'll do some further reading. 

> You can build a kernel with NoHz disabled so the in-kernel PPS processing will
> work.  (I forget what the option is called.  I'll fish it out if you get that
> far.)
Is this any different from declaring the 'nohz=off' in the cmd.txt boot 
file in raspbian?

Thanks for the tip re ntpsec directly without gpsd. I'll see what happens.

Paul Theodoropoulos

More information about the users mailing list