On Mon, Jun 05, 2000 at 11:38:20PM -0500, Frederick Bruckman wrote:
> On Mon, 5 Jun 2000, Manuel Bouyer wrote:
> > On Mon, Jun 05, 2000 at 02:45:31AM -0700, Hal Murray wrote:
> > > I see things like "time reset 15.640742 s" in the log file when running 
> > > heavy network tests.
> > This is a message from ntpd, rigth ? Maybe it's just that ntp lost sync
> > messages because of the heavy network tests (I didn't look at the sources
> > so I'm not sure what this really means :)
> You know, it may not be lost interrupts at all, but rather, an asymetric
> latency in the network or in the network driver that destablized ntp
> during the network tests. If that's the case, it would be obvious from
Perhaps it's not related to network load etc. at all:

1.  If you see the clock stepping forth and back nearly the same amount it
may be just a bogus timeserver. IMO 15.6s is quite a lot. Even if ntp looses
connection completely for a day or two and you have a very inaccurate
clock(*) you should not see steps like this.

(inaccurate in t dot dot+ since [x]ntpd handles t dot (drift) very well
once it has been up a day or two [AFAIK]).

2. timed?
Are you running another time daemon, like timed? If not configured correctly
it may "adjust" the system clock as well and *bad* things may happen.


I had #2 on one of my systems once wich caused #1 on another box :-)

-- 
Dies ist Thilos Unix Signature! Viel Spass damit.