[dns-operations] GPS time glitch last night

[Mukund Sivaraman]

On Wed, Jan 01, 2020 at 10:07:20PM -0600, Chris Adams wrote:
> Once upon a time, Mukund Sivaraman <muks at mukund.org> said:
> > Some people are reporting a GPS time glitch list night, which seems to
> > have affected our ntpd instance too. From the logs, it appears to have
> > stopped trusting GPS clocks due to a greater than expected change in
> > $GPRMC time. A restart of the ntpd process fixed it.
> 
> Is this just with a particular clock type (or multiple types but all
> using NMEA sentences)?  I have an old Trimble Resolution T in TSIP
> binary mode rather than NMEA, and it doesn't appear to have had any
> issue.

We use the Garmin GPS 18x LVC model in NMEA mode, and they run current
firmware. Glitches have been reported with other devices on the nanog
list, and Bert's monitor seems to have picked up a spurious reading as
well. On the machine that I posted logs from in this thread, its logs
show it occurring about half hour away from another reported time. This
receiver is in Chennai, India and receives from a different set of
satellites at any time, vs. Americas, Europe, and South Africa. The
wiredata from the GPS receiver was limited to $GPRMC NMEA sentences and
PPS pulses, so there are no other logs.

Without being familiar with the firmware code, I guess that even if the
device were in binary mode, the time glitch would have still have
occurred as the mode is only a difference in encoding - the data in the
NMEA sentences, and the PPS signal, are generated by the GPS receiver.