RE: [chrony-dev] PPS reference clock rejected because of high dispersion

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


> > > I'm not sure exactly where you want to do this check. The
> dispersion
> > > calculated at the beginning of RCL_AddSample and RCL_AddPulse is
> > > actually based on the average dispersion from the filter. I don't
> > > see how rejecting samples when the filter result is too big will
> > > help making the average dispersion lower, as you will get in the
> > > same deadlock situation again. Rejected samples do not update the
> > > filter, so the average dispersion will never get lower anymore.
> >
> > If you do it in filter_get_sample() before the variance stat is
> > updated, it should never be larger and it shouldn't stuck.
> 
> Here is a patch doing that. Does it fix the problem for you?
> 

I've replaced my patch with your version and used the same tricks to
reproduce my problem. So far I haven't been able to recreate it so it
does fix my problem. I've set the maxdispersion of the SHM driver to
200ms, and the PPS to 1ms.

I think the solution is sound as long as the user knows what the
acceptable max dispersion should be. The 100ms default is still a
somewhat magic number, and users with very noisy local reference clocks
should be aware that this maxdispersion should be increased in their
case.

Thanks for helping us out with this issue, and I hope we can see these
changes soon in the next release!

Greetings,

Tjalling Hattink

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


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