We might have a general problem with IPv6...
We (AS34288) have native IPv6 connectivity, we're peering at several
exchanges and have about 500 direct and native IPv6 peerings.
Nevertheless the IPv6 monitoring sees constant problems with packet loss.
I cannot see any problem on my side. I monitored our servers (in
Switzerland) from he.net in Fremont, CA for the last 24h; sending about
10 NTP requests per second. I did not see a single ntp packet loss at all.
We are fully IPv6 enabled and see about 10% IPv6 traffic now. We see no
major difference between IPv4 and IPv6 traffic, the IPv6 paths are in
fact a bit shorter... 2.4 ASes for IPv6 vs. 2.9 ASes for IPv4. I have no
indication that this problem is actually related to poor IPv6
connectivity at large.
Our logs:
http://www.pool.ntp.org/scores/2001:4b20::beef:1:16
http://www.pool.ntp.org/scores/2001:4b20::beef:1:17
Interestingly, other (local) NTP servers show similar effects:
http://www.pool.ntp.org/scores/2001:4168:1::1
http://www.pool.ntp.org/scores/2001:620:1b0:2::9
http://www.pool.ntp.org/scores/2001:470:b69a:470::1
http://www.pool.ntp.org/scores/2001:1620:2018:2::123
http://www.pool.ntp.org/scores/2001:a60::123:2
http://www.pool.ntp.org/scores/2a01:4f8:63:12a1::6
http://www.pool.ntp.org/scores/2a02:b80:0:6:7b::1
[and so on]
We should track this issue down! Just blaming "poor IPv6 performance" is
not the way to go.
We should be able to rule out that the problems are related to the
monitor itself!
- poor local IPv6 connection of the monitor?
- too low timeouts on the monitor?
- software bugs?
Greetings from Switzerland
wiwi
_______________________________________________
pool mailing list
[email protected]
http://lists.ntp.org/listinfo/pool
