Re: [chrony-dev] Bug -- interaction between ntpdate and chronyd at bootup -- will never sync up

[Miroslav Lichvar <mlichvar@xxxxxxxxxx>]

On Mon, Feb 04, 2013 at 10:49:45PM -0600, ray vantassle wrote:
> > This reading from chrony.drift is absurd. Where did it come from?
> 
> Um, no RTC.  Also, ntpdate and chronyd are stepping on one another at
> startup.  I think that the timing is such that chronyd is in the midst
> of conditioning the clock when ntpdate sets the time, so chronyd
> thinks the clock is fubar.

I'm wondering how it's possible the two were running at the same time
as both normally bind to the port 123. Or was ntpdate using the -u
option?

> In the case here, chronyd NEVER gets straightened out.  I had to stop
> and restart it, and then it got sane.

I agree chronyd should be able to recover from this situation and I'll
see if I can reproduce it.

> Since the clock starts out at Dec 31 1969 18:00:00 CST, it is
> important to set the date/time to approximately the actual date/time
> as soon as possible in the bootup sequence.  Ntpdate with 1 or 2
> servers and only 1 sample does that.  Chronyd takes quite a bit
> longer.

With the initstepslew directive it should be much faster. It's
basically an ntpdate client inside chrony.

Thanks for the report.

-- 
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.