[ntp:questions] Re: How often should server hop happen?
brad at stop.mail-abuse.org
Wed Feb 2 17:08:14 UTC 2005
At 8:32 AM -0800 2005-02-02, David Schwartz wrote:
> It's not a good thing because there will inevitably be a jump when you
> change servers. The jump will be smoothed over, but medium-term stability
> goes down quite a bit when you change servers.
No, there won't be a jump. You will get a shift, as ntpd slowly
adjusts the clock from the previous syspeer to the new one, but
that's part of the design of the NTP protocol and the ntpd program.
That said, an excessive amount of clock hopping is bad, and will
keep your jitter and stability from getting as low as they should be.
> It's not normal because the
> server you are synched to should look a bit better than other servers that
> you are not synched to.
Situations change over time. Network configurations change.
Servers that used to be good, and the best you could see, may now be
worse than others you've been monitoring.
This is the normal course of events. What is undesirable is when
these sorts of things happen too often.
> He's on a cable modem connection. It's hard to imagine that there are
> two servers that are so closely matched in quality and also both so high
> that they're his first choice.
It all depends on where those clocks are and who runs them, and
what's between those two end-points.
If they're something like tick.usno.navy.mil and
tock.usno.navy.mil, then the odds are that all the network issues
between the two end-points are the same, at which point very minor
operational differences between the two upstream servers might make
all the difference when looked at using the exceptionally sensitive
statistical algorithms found in NTPv4 and ntpd.
Brad Knowles, <brad at stop.mail-abuse.org>
"Those who would give up essential Liberty, to purchase a little
temporary Safety, deserve neither Liberty nor Safety."
-- Benjamin Franklin (1706-1790), reply of the Pennsylvania
Assembly to the Governor, November 11, 1755
SAGE member since 1995. See <http://www.sage.org/> for more info.
More information about the questions