2

I have a Linux NTP server with a ublox M8Q GPS module connected via GPIO (UART @ 9600bps and PPS).

What I observe is that the local GPS source — SHM(0) — gets marked as falseticker and I'd like to understand why.

clock@clock:~ $ ntpq -p
     remote           refid      st t when poll reach   delay   offset  jitter
==============================================================================
 uk.pool.ntp.org .POOL.          16 p    - 1024    0    0.000    0.000   0.000
*SHM(2)          .PPS.            0 l   50   64  377    0.000   -0.007   0.007
xSHM(0)          .GPS.            0 l   50   64  377    0.000  -112.12   1.233
-85.199.214.100  .GPS.            1 u  609 1024  377   14.412    1.151   0.932
-194.80.204.184  .GPS.            1 u  452 1024  377   22.394    1.048   1.570
-srv20.arnoldthe 193.11.166.36    2 u  413 1024  377   22.348    0.106   2.006
+183.ip-51-89-15 85.199.214.101   2 u  435 1024  377   11.956    0.722   0.661
-ns3145492.ip-51 194.80.204.184   2 u  188 1024  377   12.317    1.043   1.844
+85.199.214.101  .GPS.            1 u  104 1024  377   15.372    0.654   0.231
-195.195.221.100 .GPS.            1 u  948 1024  377   27.442    2.899   2.659

The relevant part of /etc/ntpd.conf is

server 127.127.28.2 prefer
fudge  127.127.28.2 refid PPS

server 127.127.28.0
fudge  127.127.28.0 refid GPS

What would be a good way to troubleshoot this?

NPE
  • 669
  • 3
  • 9
  • 19
  • `enable peerstats` and plot the offsets and dispersions of all your configured peers. A crowd of clocks that are systematically off can overrule a proper clock. 112ms offset seems a bit high. Did you play with `setserial`'s `low_latency`? – U. Windl Dec 30 '19 at 14:00

0 Answers0