On 2/4/2012 15:28, David Lord wrote:
A C wrote:
For reference, this is what the system looks like normally until the
oscillations go out of control and wipe out clock sync:
remote refid st t when poll reach delay offset jitter
==============================================================================
o127.127.22.0 .PPS. 0 l 2 32 377 0.000 0.005 0.061
127.127.28.0 .GPSD. 4 l 115 128 377 0.000 -23.269 4.368
-74.118.152.85 69.36.224.15 2 u 456 512 377 42.365 0.631 0.334
+68.0.14.76 .GPS. 1 u 69 512 377 79.443 2.383 0.249
+173.203.122.111 209.51.161.238 2 u 67 512 377 74.682 1.053 0.361
*130.207.165.28 130.207.244.240 2 u 232 512 377 94.749 -4.766 0.619
-131.144.4.10 130.207.244.240 2 u 404 512 377 84.795 2.599 0.804
That all looks ok to me.
Can you disable running gpsd, as you're not using it, in
case it's having an effect on ntpd?
Have you tried using the stable version of ntpd or if not,
what is reason for running development version?
I am using gpsd elsewhere to export the GPS data to another machine but
I'll turn it off for testing, too, if the most recent replacement
snprintf doesn't work.
I'm using the development version because I needed an snprintf
workaround for a broken dtoa() in libc.
_______________________________________________
questions mailing list
questions@lists.ntp.org
http://lists.ntp.org/listinfo/questions
