[ntp:questions] Re: Estimate of the number of people using the pool system
Adrian 'Dagurashibanipal' von Bidder
grazdan at fortytwo.ch
Sun Oct 19 19:24:21 UTC 2003
Clinging to sanity, David L. Mills mumbled in his beard:
> Frederick,
>
> The default minpoll and maxpoll were not chosen lightly, but only after
> a good deal of experience in the trenches. I question the wisdom of the
> NetBSD folks when casually overriding those defaults in a model
> configuration.[...]
>> server pool.ntp.org maxpoll 12
>> server pool.ntp.org maxpoll 12
>> server pool.ntp.org maxpoll 12
[...]
There's me to blame for this one, not the NetBSD folks - see http:/
www.pool.ntp.org
There's three things:
- I was trying to minimize the server load as far as possible - I don't
want the pool project shut down because servers (donated by volounteers, as
you know) get loaded too much and stop participating.
- if you use pool.ntp.org as your timeserver, you'll have some server of
unknown quality, and possibly with big network delays. It should get your
time to within a few 10 ms of the True Time(tm), and that's enough for
everyday work - but we're not speaking about keeping the clock within 1ms,
we're not even speaking of keepeing it synced enough to not fool make when
using NFS-mounted filesystems. It should be able to do this with maxpoll
12.
- I was under the impression that the ntp algorithm would only really
increase to the maxpoll if this is possible without a big degradation in
quality.
thanks & greets
-- vbi
--
Let's all show human CONCERN for REVERAND MOON's legal difficulties!!
More information about the questions
mailing list