|
Ticket ID: SIXXS #583663 Ticket Status: Remote Problem PoP: usqas01 - OCCAID Inc. (Ashburn, Virginia)
IPv6 routing problem from 2001:4830:164c::/48 to argo.pyxos.net and plasma.querx.com
![[us]](/s/countries/us.gif) Shadow Hawkins on Thursday, 27 September 2007 03:24:15
My tunnel appears to be up and working, but I am experiencing a problem routing to some IPv6 space; specifically argo.pyxos.net (2002:423b:6d88::1), and plasma.querx.com (2002:423b:6dbf::1) - both are tun6to4 IPs.
My handle is DEW1-SIXXS
Tunnel T12597
Subnet R5386
$ ip -6 addr show
1: lo: <LOOPBACK,UP,LOWER_UP> mtu 16436
inet6 ::1/128 scope host
valid_lft forever preferred_lft forever
4: wlan0: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qlen 1000
inet6 2001:4830:164c:30:290:4bff:fe4b:d76f/64 scope global dynamic
valid_lft 2591905sec preferred_lft 604705sec
inet6 fe80::290:4bff:fe4b:d76f/64 scope link
valid_lft forever preferred_lft forever
$ traceroute6 -n argo.pyxos.net
traceroute to argo.pyxos.net (2002:423b:6d88::1), 30 hops max, 40 byte packets
1 2001:4830:164c:30::1 1.392 ms 1.947 ms 3.633 ms
2 2001:4830:1600:bc::1 25.851 ms 32.401 ms 39.501 ms
3 2001:4830:e6:7::1 39.730 ms 39.929 ms *
4 2001:4810:0:100::3 40.110 ms 40.193 ms 40.324 ms
5 * * *
[snip]
$ traceroute6 -n plasma.querx.com
traceroute to plasma.querx.com (2002:423b:6dbf::1), 30 hops max, 40 byte packets
1 2001:4830:164c:30::1 2.492 ms 2.675 ms 2.837 ms
2 2001:4830:1600:bc::1 23.639 ms 28.456 ms 38.979 ms
3 2001:4830:e6:7::1 39.690 ms 39.910 ms 40.685 ms
4 2001:4810:0:100::3 40.882 ms 41.136 ms 41.318 ms
5 * * *
[snip]
$ traceroute -n argo.pyxos.net
traceroute to argo.pyxos.net (66.59.109.136), 30 hops max, 40 byte packets
1 192.168.30.1 2.418 ms 2.643 ms 2.874 ms
2 73.149.92.1 9.861 ms 19.973 ms 20.800 ms
3 68.86.213.113 21.052 ms 21.223 ms 21.425 ms
4 68.86.209.33 21.616 ms 21.875 ms 22.078 ms
5 68.86.209.29 22.312 ms 22.732 ms 26.206 ms
6 68.86.209.25 26.431 ms 25.001 ms 24.899 ms
7 68.86.209.21 25.119 ms 18.383 ms 12.427 ms
8 68.86.208.14 17.879 ms 20.824 ms 24.759 ms
9 68.86.208.13 32.089 ms 31.749 ms 32.004 ms
10 68.86.90.26 33.420 ms 33.654 ms 42.371 ms
11 4.78.169.50 39.815 ms 48.184 ms 47.929 ms
12 4.78.169.49 27.970 ms 27.981 ms 47.496 ms
13 4.68.16.69 43.119 ms 4.68.16.197 43.323 ms 4.68.16.69 46.451 ms
14 192.205.33.93 49.115 ms 46.691 ms 48.736 ms
15 12.123.3.61 51.149 ms 50.343 ms 54.804 ms
16 12.122.105.65 49.127 ms 49.427 ms 39.356 ms
17 12.86.0.214 52.956 ms 47.237 ms 48.964 ms
18 66.59.109.136 51.587 ms 31.338 ms 37.409 ms
$ traceroute -n plasma.querx.com
traceroute to plasma.querx.com (66.59.109.191), 30 hops max, 40 byte packets
1 192.168.30.1 2.227 ms 2.439 ms 2.641 ms
2 73.149.92.1 19.446 ms 22.492 ms 29.249 ms
3 68.86.213.113 29.631 ms 29.841 ms 30.031 ms
4 68.86.209.33 30.265 ms 30.879 ms 35.358 ms
5 68.86.209.29 35.601 ms 35.837 ms 35.985 ms
6 68.86.209.25 36.215 ms 35.137 ms 34.991 ms
7 68.86.209.21 74.520 ms 44.083 ms 44.297 ms
8 68.86.208.14 21.912 ms 22.225 ms 22.455 ms
9 68.86.208.13 24.214 ms 24.769 ms 24.967 ms
10 68.86.90.38 26.674 ms 26.906 ms 27.139 ms
11 4.78.169.46 27.370 ms 29.816 ms 30.021 ms
12 4.78.169.45 29.016 ms 23.078 ms 30.190 ms
13 4.68.16.197 32.697 ms 4.68.16.133 29.777 ms 4.68.16.5 36.834 ms
14 192.205.33.93 37.683 ms 36.812 ms 36.224 ms
15 12.123.3.61 35.004 ms 35.326 ms 34.958 ms
16 12.122.105.65 33.727 ms 33.728 ms 31.631 ms
17 12.86.0.214 65.144 ms 71.930 ms 70.861 ms
18 * * *
[snip]
So IPv6 to argo.pyxos.net and plasma.querx.com fails, while IPv4 to both works (despite the traceroute to plasma.querx.com seeming like it fails, I can ping it fine - I can't ping6 either host).
Both of these hosts are on the same IPv4 subnet with tun6to4 addresses.
-Doug
IPv6 routing problem from 2001:4830:164c::/48 to argo.pyxos.net and plasma.querx.com
Shadow Hawkins on Thursday, 27 September 2007 03:26:11
I wasn't able to find this ticket before, but my ticket might be a duplicate of this one:
Ticket#545976 (Several destinations outside OCCAID unreachable)
State change: remoteproblem
The state of this ticket has been changed to remoteproblem
IPv6 routing problem from 2001:4830:164c::/48 to argo.pyxos.net and plasma.querx.com
The big problem here is that it is 6to4, which means that there can be a lot of issues in the configuration of the various 6to4 hops. You will need to be able to diagnose the forward and reverse path in both IPv4 and IPv6 to find out where it goes wrong.
For instance what is the IPv6 traceroute back from those hosts, are they actually able to reach 2001:4830::/32 and if there is a route, over which IPv4 and IPv6 hops does it lead and are these correctly configured.
I suggest contacting the people who run those 6to4 nodes and letting them do a lot of debugging, or more easily just get a working tunnel.
6to4 is a great idea and it can work, but with the asymmetry in the Internet it is very hard if not impossible to debug.
IPv6 routing problem from 2001:4830:164c::/48 to argo.pyxos.net and plasma.querx.com
Shadow Hawkins on Thursday, 27 September 2007 13:53:07
Thanks for the information about potential problems with 6to4. I own one of the servers that has the 6to4 address (argo.pyxos.net), so I should be able to do debugging.
The problem seems to have resolved itself overnight, so unfortunately that will have to wait until the problem arises again.
Thanks again for your time.
-Doug
|
![[ch]](/s/countries/ch.gif)
on Thursday, 27 September 2007 04:32:06
Posting is only allowed when you are