Re: [chrony-users] Timeout for initstepslew?

[Zack Shivers <zack@xxxxxxxxx>]

> Were you getting significantly worse RMS with longer filters?
I didn't try longer filters, because I was meeting your criteria using
this number. Does a filter of 5 seem too short?


On Wed, Oct 2, 2019 at 12:31 AM Miroslav Lichvar <mlichvar@xxxxxxxxxx> wrote:
>
> On Tue, Oct 01, 2019 at 12:48:26PM -0700, Zack Shivers wrote:
> > > Yes, you could, at least in theory. Is the server free running (using
> > > local directive), or synchronized to another server? If it is not very
> > > stable, there is no point in trying to tightly synchronize its
> > > clients.
> >
> > Yes, the server is free running with the local directive. I'm curious
> > why you say there's no point.
>
> I meant if it was synchronized over Internet, its clock would be too
> unstable relative to what is possible between two local hosts. You
> would need to enable the smoothtime directive to stabilize the server
> time. If the server is free running, it's ok. The smoothtime directive
> could still be useful if you needed to prevent the server from
> drifting too far from true time with manual input.
>
> > > Best values of minpoll/maxpoll and filter will depend on your network,
> > > clock, etc. Set the polling interval to the shortest one you are
> > > comfortable with (not overloading the network or server) and try to
> > > increase the filter length so that the number of samples reported in
> > > sourcestats is close to 64, but not 64, most of the time. If it's
> > > stuck at 64 most of the time, the filter and the update interval is
> > > probably too short.
> >
> > Thank you for the practical advice. If RMS offset is a good indicator,
> > I seem to be getting an excellent synchronization:
> > RMS offset      : 0.000000093 seconds
>
> That looks good.
>
> > Relevant client configuration:
> > server 192.168.0.2 minpoll -6 maxpoll -4 xleave filter 5
>
> Were you getting significantly worse RMS with longer filters?
>
> --
> Miroslav Lichvar
>
> --
> 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.
>

-- 
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.