Re: [chrony-dev] [GIT] chrony/chrony.git branch, master, updated. 2.1.1-89-g3cd32ed

[Miroslav Lichvar <mlichvar@xxxxxxxxxx>]

On Fri, Sep 18, 2015 at 02:55:18PM +0000, Adri Koppes wrote:
> > > If the clock is too far behind or ahead, doesn't chrony already step the
> > clock?
> > 
> > Not by default. You need to specify the threshold and the number of
> > updates in which are allowed steps with the makestep directive in
> > chrony.conf. Or you can do that manually with the makestep command from
> > chronyc.
> 
> I use initstepslew instead of makestep, but I guess the result would be the same.

Yes, that should be the same. The difference with initstepslew is that
the servers are used only on start. If they are the same servers as
specified with the server or pool directive, the statistics are thrown
away needlessly.

> I just switched to the latest release and keep an eye on how it's performing.

The latest git now implements the combined slewing with adjtime() on
FreeBSD and NetBSD if you want to try it. Offsets larger than 1 second
should be slewed with 5000 ppm rate. I ran few tests and it seems to
be working as expected. In the tracking log the skew jumps up a bit
when the remaining offset is crossing the 1 second, but that's
expected I think. The fast slewing can be disabled by setting the
maxslewrate option to a value smaller than 5000 ppm.

BTW, anyone knows an easy way to step the clock from shell by a fixed
amount on FreeBSD and NetBSD? I had no luck with the date command.

-- 
Miroslav Lichvar

-- 
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.