[ntp:questions] Same Pool Server in list

Dave Hart hart at ntp.org
Mon Jul 2 15:27:43 UTC 2012


On Mon, Jul 2, 2012 at 09:40 UTC, Marc-Andre Alpers <m-a.alpers at web.de> wrote:
> I have restarted my NTPd and the following happened.
>
> root at amd-x4:~# ntpq -p
>  evie.fra1.as411 78.47.148.174    3 u    1   64    1   36.242    8.622
>  4.338
>  evie.fra1.as411 78.47.148.174    3 u    2   64    1   27.690    9.442
>  2.231
>
> Twice the same NTP server in the list.
>
> root at amd-x4:~# ntpq -pn
>  2a01:4f8:d13:23 78.47.148.174    3 u    5   64    1   36.242    8.622
>  4.338
>  78.46.221.42    78.47.148.174    3 u    4   64    1   27.690    9.442
>  3.712
>
> I use the pool server and I have IPv6 connectivity. How can i
> prevent that the same NTP Server is in the list?

While you could add -4 to each of the server lines for the pool, that
would disable IPv6 use of the pool. not only the ones that happen to
be the same remote clock.  I don't recommend it.

Moreover, I argue you shouldn't be dismayed to be reaching the same
remote clock via two versions of IP, as the networks are in fact
different, though they share fate on at least your last mile (your
internet connection).  Even in a fully dual-stacked situation where
the physical path is identical and all the hops are the same routers,
IPv4 and IPv6 are going to be slightly different beasts.  Today, most
often the IPv4 and IPv6 paths are quite different.  Also, the
performance of a well-behaved remote clock probably varies little over
time, while your view of its performance from afar varies quite a bit
with network weather, and that view can be more stable over v4 or v6
at any given time, there's little lost in my opinion in failing to
treat the IPv4 and IPv6 addresses of a single server as duplicates of
each other.

Cheers,
Dave Hart


More information about the questions mailing list