Udo van den Heuvel wrote:
> Hello,
> 
> Perhaps someone can verify the ntp functionality at 2001:888:10:abc::2?
> This is a local IPv6 over IPv4 tunnel end which is also serving at
> 80.101.128.228. Not so great GPS reception but for the rest it sorta works.

mnordh...@cheezum:~$ ping6 2001:888:10:abc::2
PING 2001:888:10:abc::2(2001:888:10:abc::2) 56 data bytes
64 bytes from 2001:888:10:abc::2: icmp_seq=1 ttl=55 time=150 ms
64 bytes from 2001:888:10:abc::2: icmp_seq=2 ttl=55 time=138 ms
^C
--- 2001:888:10:abc::2 ping statistics ---
2 packets transmitted, 2 received, 0% packet loss, time 1001ms
rtt min/avg/max/mdev = 138.221/144.423/150.625/6.202 ms
mnordh...@cheezum:~$ ntpdate -q 2001:888:10:abc::2
server 2001:888:10:abc::2, stratum 1, offset 0.035882, delay 0.16173
27 Dec 13:47:28 ntpdate[3250]: adjust time server 2001:888:10:abc::2
offset 0.035882 sec
mnordh...@cheezum:~$ ping 80.101.128.228
PING 80.101.128.228 (80.101.128.228) 56(84) bytes of data.
64 bytes from 80.101.128.228: icmp_seq=1 ttl=48 time=154 ms
64 bytes from 80.101.128.228: icmp_seq=2 ttl=48 time=137 ms
^C
--- 80.101.128.228 ping statistics ---
2 packets transmitted, 2 received, 0% packet loss, time 1001ms
rtt min/avg/max/mdev = 137.484/145.956/154.428/8.472 ms
mnordh...@cheezum:~$ ntpdate -q 80.101.128.228
server 80.101.128.228, stratum 1, offset 0.034292, delay 0.16205
27 Dec 13:48:36 ntpdate[3254]: adjust time server 80.101.128.228 offset
0.034292 sec

Seems good, though the clock is a couple ms off. (Only about 2.4 from
two other machines, one of them only ~15 ms away, though.)

It's very nice that latency is about the same over both IPv4 and IPv6.

(Those tests were from tick.mattnordhoff.com, which uses an HE tunnel,
FWIW.)

> If IPv6 works well enough: how to make this known to the pool?

Just add the IP, same as anything else. (Last I tried, adding a hostname
didn't work; apparently it doesn't look up AAAA records.)

Currently, the pool doesn't actually support IPv6 NTP servers, so,
though you can add it, it won't be distributed, and the monitoring and
statistics won't work either.

> Udo
-- 
Matt Nordhoff
_______________________________________________
timekeepers mailing list
[email protected]
https://fortytwo.ch/mailman/cgi-bin/listinfo/timekeepers

Reply via email to