ntpd wierdness.

Hi

ntpq/ntpd appear inconsistent since at least r284659, reporting my
refclocks as false tick and the other selected for PPS, but not
time source:

[brane] ~/graphing $ ntpq -c peers
     remote           refid      st t when poll reach   delay   offset  jitter
==============================================================================
xGPS_ONCORE(0)   .GPS.            0 l    8   16  377    0.000    0.088   0.003
oGPS_ONCORE(1)   .GPS.            0 l    7   16  377    0.000    0.087   0.003
+xxxx            .PPS.            1 u   21   64  377   31.655    2.514   9.528
*yyyy            .GPS.            1 u    5   64  377   31.913    2.705   2.684

[brane] ~/graphing $ ntpq -c assoc

ind assid status  conf reach auth condition  last_event cnt
===========================================================
  1 44551  913a   yes   yes  none falsetick    sys_peer  3
  2 44552  973a   yes   yes  none  pps.peer    sys_peer  3
  3 44553  963a   yes   yes  none  sys.peer    sys_peer  3
  4 44554  9424   yes   yes  none candidate   reachable  2

[brane] ~/graphing $ ntpq -crv
associd=0 status=0418 leap_none, sync_uhf_radio, 1 event, no_sys_peer,
version="ntpd 4.2.8p2-a (1)", processor="amd64",
system="FreeBSD/11.0-CURRENT", leap=00, stratum=1, precision=-20,
rootdelay=0.000, rootdisp=1.090, refid=GPS,
reftime=d945578a.3989f1fd  Mon, Jul  6 2015 21:37:46.224,
clock=d9455790.a5ebb2cf  Mon, Jul  6 2015 21:37:52.648, peer=44552, tc=4,
mintc=3, offset=0.089614, frequency=31.102, sys_jitter=0.002181,
clk_jitter=0.001, clk_wander=0.001

The Oncore GPS are slightly wierd in they almost always start 32000ms
ahead on the first time read from them after reset and then correct
themselves resulting in huge jitter at ntpd start, but then with
the previous version, the refclocks were always selected as pps.peer
and sys.peer.

Now, even though the refclocks have settled down, the peer selection
seems a little strange.  At least not what I'm used to seeing.

Ian

-- 
Ian Freislich