On 12/13/2011 10:16 PM, unruh wrote:
On 2011-12-13, Harlan Stenn<[email protected]> wrote:
Richard wrote:
[much elided]
Like it or not, NTPD, when started, will need up to TEN HOURS to
settle down with the best time that you are going to get! This is not
a hardship if you run NTPD 24x365 (366 in leap years). If you have to
shutdown frequently and can't wait for NTPD to reach steady state then
NTP is the wrong tool for you!
You say this as if it is necessarily a bad thing.
If it were the only way of ensuring good timing, it would not be. It is
not. It is an extremely primative feedback loop, which IMHO wastes a
huge amount of information in the data. And the slowness of the settling
down initially is just the other side of ntpd's poor handling of changes
in the clock frequency ( due for example to heating because of the
computer suddenly having to do some heavy computing).
You don't have to convince me. Dave Mills is the one to be convinced
and I'm inclined to believe that he knows what he is doing.
NTP is *careful* about choosing reference time. It acts conservatively.
In some ways, yes, in some ways it is extremely profligate.
There are a bunch of folks who consider this to be a feature.
Are you implying that before ntpd gets "the best time that you are going
to get" that the actual time on the box is inadequate?
It has a much larger displacement from the true time than it needs to
have.
I suppose it depends on what you want.
In that interim, is the time "bad" or are the error numbers just
shrinking?
_______________________________________________
questions mailing list
[email protected]
http://lists.ntp.org/listinfo/questions
