On 1/27/20 4:17 PM, Rick Moen wrote:
> Quoting Dr. Nikolaus Klepp (dr.klepp@???):
>
>> Just for your amusement: systemd breaks linuxcnc on RPi4 - who's not
>> into machining will most likely not see the fun part of it, but anyway
>> :)
>>
>> https://www.youtube.com/watch?v=RDKaFJmB254
> Nik --
>
> Thank you for the link. I think it should be of interest for people
> wanting to observe an inquisitive mind doing good diagnosis work,
> whether they are interested in linuxcnc or not, e.g., right around the
> 11:35 minute mark (out of 14:43) he arrives at the important conclusion,
> through logical testing, that systemd's timesyncd implements an
> un-gentle jerk-forward-or-back of time all at once -- as opposed to the
> gradual adjustment any of the real NTP daemons would do -- and that such
> rough time discontinuities are problematic for real-time operations.
>
> (It's distinctive that the entire RPi series lacks by default any RTC
> circuitry, so it's particularly vulnerable to such software problems.)
>
> Starting around 13:00, he tests his hypothesis by switching to ntpd, and
> indeed the observed symptom immediately goes away.
>
I have an OpenSUSE system that exhibits a similar thing... Periodically xymon tells me the time server isn't acceptable, and the shuttle motherboard HAS an RTC. It's not running ntpd.

I also have a PI with a GPS clock board and a special build of ntpd to enable support for it.  From time to time, even though checks of the GPS clock show it's running correctly,
IT does the same thing too.