Yo Joshua!
On Wed, 2 Dec 2020 12:41:42 -0500
Joshua Quesenberry <
engnfrc@gmail.com> wrote:
> For the purposes of this discussion the startup script just launches
> gpsd as root, we have a sequential set of scripts that run under one
> systemd startup target.
OK. Your problem then.
> The distribution is Raspbian Buster and we have a customized kernel
> running optimized for low-latency.
So you are using the old Buster packages. With the old gpsd.
> I can add the "-n" if you think it will help, but this has been
> running for over a year without a need for it.
Without -n gpsd will stop serving time when no client is connected.
OTOH, no telling ow your package was built.
> The "-r" is so we use
> the GPS time even without a fix, this is because our GPS has a
> built-in real-time clock that is kept alive that we were try to
> leverage, but that hasn't been the best design choice, dedicated
> real-time clock next time.
"-r" is read only mode. That means gpsd will not reconfigure your
receiver. Often not a good thing.
gpsd will never send PPS to SHM without a valid fix. gpsd will not use
your RTC.
> I have GPSD connecting to a Python script that outputs the GPS feed
> over a TCP connection. GPSD is only acting as a listener and cannot
> send packets back to the GPS. The Python scripts is the single master
> setting up the device and doing some other useful tasks. Different
> use case than most GPSD users, but definitely working well for us for
> over a year.
OK. Thus the -r to be read-only. Your problem then.
> Note that the device is sitting on my desk at work right now, we have
> a GPS repeater setup for indoor testing. Yes it has 3D fix. Here's a
> sample output from gpsmon.
email totally manlged that. Best to send screen copies as text attachements.
Looks OK to me, but gpsmon is not really maintained.
> Regarding the version, as mentioned we used whatever was latest on
> apt. We avoid having to checkout and build 3rd party software as much
> as possible. If we find here that upgrading to a newer GPSD version
> is required then I can broach the topic with others, but for now if
> we can keep it at 3.17 that would be most ideal. I don't know who
> builds the packages for apt, I would hope it's someone in this group,
I do not know either.
> Not sure why the maintainer of the apt package deviated from
> the defaults,
Do we know they change the defaults?
> I don't know if it matters or not with regards to ntpshmmon, but we
> don't have any of the ntp stuff installed on our system and aren't
> planning to install it. I would think not having ntp would have no
> affect on the PPS field in gpsmon, but I could be mistaken.
The PPS field in gpsmon is not useful for much. It says nothing about
what ntpshmmon will see.
Try running gpsd this way for a minute:
/usr/sbin/gpsd -r -D 6 -N tcp://127.0.0.1:5760 /dev/pps0 > gpsd.log
And send along gpsd.log. That will explina what your PPS is doing.
RGDS
GARY
---------------------------------------------------------------------------
Gary E. Miller Rellim 109 NW Wilmington Ave., Suite E, Bend, OR 97703
gem@rellim.com Tel:+1 541 382 8588
Veritas liberabit vos. -- Quid est veritas?
"If you can't measure it, you can't improve it." - Lord Kelvin