NTPsec flags my GPS as a falseticker

[Nick Burkitt]

Hi Hal.
>
>users at ntpsec.org said:
>>  2020-03-19T02:39:11 ntpd[19607]: CONFIG: configpeers: Ignoring duplicate '127.127.28.2'
>>  2020-03-19T02:39:11 ntpd[19607]: CONFIG: configpeers: Ignoring duplicate '127.127.28.0'
>
>Something fishy there.  I'd fix that just to make sure you are using the
>correct lines from the config file.
Yeah - I have no idea where that's coming from. The PPS and GPS devices 
are each mentioned only once in my ntpd.conf file.
>
>The combination of 0 and 2 is a bit strange.  The usual pairing is 0 and 1 for
>the first GPS device and 2 and 3 for the second GPS device.
That one cost me a lot of time. :-) gpsd assigns two shared memory 
segments to each device listed on its command line. Since my PPS device 
is separate from my GPS device, I have to pass both devices on the gpsd 
command line, GPS (/dev/ttyACMx) followed by PPS (/dev/pps0). gpsd 
assigns shm unit 0 to the GPS and shm unit 1 to the (non-existent) 
associated PPS, shm unit 2 to the PPS, and it's not clear what it does 
with the shm unit 3.

ntpshmmon does a good job of detecting active shm units, though.

>I'm not a gpsd wizard.  If you don't need it for some other reason, try using
>the NMEA and PPS drivers.  This works for me on a Raspberry Pi.
I do need the capabilities of gpsd, and it allows me to support 
different GPS devices. At the moment that's only two, but you never 
know. But both the GPS and PPS times are recognized by ntpd. The only 
issue (I think) is that ntpd declares my GPS as a falseticker, at least 
most of the time. I'd like to understand why that is. So far, I've 
managed to improve things a bit by adjust the GPS offset

># PPS via GPIO, Adafruit GPS card with ???? NMEA driver: GPRMC, 9600, stats
>server 127.127.20.0 prefer path /dev/ttyAMA0  mode 0x010011
>fudge  127.127.20.0 flag1 0            # disable PPS
>fudge  127.127.20.0 time2 0.600        # Fixup offset
>server 127.127.22.0                    # PPS signal, needs prefer
>fudge  127.127.22.0 flag2 0            # Rising edge
>
>      remote           refid      st t when poll reach   delay   offset   jitter
>===============================================================================
>*NMEA(0)         .GPS.            0 l   59   64  377   0.0000 -31.9530  35.5048
>oPPS(0)          .PPS.            0 l   58   64  377   0.0000  -0.1953   1.6976
>
So you define your GPS and PPS as peer servers, rather than refclocks. 
How does that affect the way ntpd handles them?

Thanks,

-Nick