Hi
I have a Ubuntu 10.04 server with the the HE tunnel correctly configured on br0 acting as a gateway. From the server, at all times, ipv6 connectivity works. radvd is correctly advertising the link-local of the interface. However, I am having problems from Ubuntu 10.10 desktops. I haven't provided routes and interface details because it does work - just not all the time - but I can do it it would help.
From the desktop I can ping the server, and ssh to it by its ipv6 address. But I cannot access ::/0 until I have run tracepath6 first - and then it will work for a maximum of 10 minutes.
e.g.
simon@hadrian:~$ ping6 ipv6.google.com
PING ipv6.google.com(2a00:1450:8006::93) 56 data bytes
^C
--- ipv6.google.com ping statistics ---
5 packets transmitted, 0 received, 100% packet loss, time 4033ms
simon@hadrian:~$ tracepath6 ipv6.google.com
1?: [LOCALHOST] 0.033ms pmtu 1500
1: 2001:470:1f08:xxxx::3 1.207ms
1: 2001:470:1f08:xxxx::3 0.770ms
2: 2001:470:1f08:xxxx::3 0.690ms pmtu 1480
2: smitten-1.tunnel.tserv5.lon1.ipv6.he.net 41.147ms
3: gige-g4-6.core1.lon1.he.net 34.579ms
4: 2001:7f8:4::3b41:1 34.426ms !A
Resume: pmtu 1480
simon@hadrian:~$ ping6 ipv6.google.com
PING ipv6.google.com(2a00:1450:8006::93) 56 data bytes
64 bytes from 2a00:1450:8006::93: icmp_seq=1 ttl=56 time=27.0 ms
64 bytes from 2a00:1450:8006::93: icmp_seq=2 ttl=56 time=27.0 ms
64 bytes from 2a00:1450:8006::93: icmp_seq=3 ttl=56 time=27.3 ms
64 bytes from 2a00:1450:8006::93: icmp_seq=4 ttl=56 time=27.5 ms
64 bytes from 2a00:1450:8006::93: icmp_seq=5 ttl=56 time=27.7 ms
^C
--- ipv6.google.com ping statistics ---
5 packets transmitted, 5 received, 0% packet loss, time 4004ms
rtt min/avg/max/mdev = 27.065/27.366/27.745/0.286 ms
And then a few minutes later, I would have to repeat the process. If I ssh to the gateway by ipv6 then connectivity remains for as long as I remain connected, but not if I connect by ipv4. I don't believe it is a firewall issue, as it happens with ufw enabled or disabled on both client and server.
Any ideas why this may be happening?