Re: [chrony-users] Possible bug in PPS support

[ Thread Index | Date Index | More chrony.tuxfamily.org/chrony-users Archives ]


Bill Unruh wrote:
On Mon, 23 Oct 2017, Rob Janssen wrote:

Bill Unruh wrote:
If you really need 20usec, then relying on one gps is certainly a bad
decision. You should have two or three machines all with independent gps
sources so you could catch one of them going rogue, or quitting.
The GPSDOs we are using are 2-3 orders of magnitude better than that.
These are not your typical $50 modules, but professional GPSDO with OCXO
or better oscillator.

It is not the accuracy of the individual gps but the the fallback in case one
of them goes mad (as happened to you). You do not want them on the same
machine unless they have hardware timestamping, since the interrupt latency is
far larger than 1us for servicing each interrupt.
Again you are wandering away from the topic Bill!
The discussion is about detection of a possible problem, not about availability.
I did not specify availability of the system, it may well be down when there is a component
failure, but we only want to know about it.



Monitoring of their accuracy is done by their owners, we only get the signal
via distribution amplifiers.  That is why we would prefer to have some additional
validation, like the PPS signal completely missing.
(which could also be caused by a mistakenly unplugged or cut cable, which
would never be detected by the GPSDO monitoring)

As I said, you could do that with a cron job every 5 min cheching.
We already have a comprehensive monitoring system based on Nagios, that in case
of this service uses "chronyc -h host tracking" to regularly retrieve the status of chrony
and alerts responsible people when something is wrong.

The issue is that it monitors "stratum" and "last offset" and it failed to trigger when the
PPS signal went away, even after 13 hours.  It would have triggered when stratum
went above 1 or last offset above 20us, but it didn't.  Both of these values remain frozen
when there is no PPS.

That is the issue I want to rectify, but that won't happen when I discuss with you.
Fortunately there is Miroslav who gave me useful hints.

Rob

--
To unsubscribe email chrony-users-request@xxxxxxxxxxxxxxxxxxxx with "unsubscribe" in the subject. For help email chrony-users-request@xxxxxxxxxxxxxxxxxxxx with "help" in the subject.
Trouble?  Email listmaster@xxxxxxxxxxxxxxxxxxxx.


Mail converted by MHonArc 2.6.19+ http://listengine.tuxfamily.org/