|
Ticket ID: SIXXS #9534870 Ticket Status: User PoP: demuc02 - M-net Telekommunikations GmbH (Muenchen)
Routing between hosts on this PoP broken
![[de]](/s/countries/de.gif) Shadow Hawkins on Sunday, 09 June 2013 09:34:22
While hosts behind this PoP are reachable from the outside and can communicate with the v6-Internet just fine, two hosts behind demuc02 cannot talk to each other. This problem appeared sometime between 07 and 09 Jun 2013.
E.g.:
# ip r g 2001:a60:f069::1
2001:a60:f069::1 from :: via 2001:a60:902f:1::2 dev lan src 2001:a60:f0fb::1 metric 1024 mtu 1500 advmss 1440 hoplimit 0
# ping6 -nc1 2001:a60:902f:1::2
PING 2001:a60:902f:1::2(2001:a60:902f:1::2) 56 data bytes
64 bytes from 2001:a60:902f:1::2: icmp_seq=1 ttl=64 time=0.388 ms
--- 2001:a60:902f:1::2 ping statistics ---
1 packets transmitted, 1 received, 0% packet loss, time 0ms
rtt min/avg/max/mdev = 0.388/0.388/0.388/0.000 ms
# tracepath6 -n 2001:a60:f069::1
1?: [LOCALHOST] 0.111ms pmtu 1500
1: no reply
2: no reply
3: no reply
[...]
^C
But:
# ping6 -nc1 madduck.net
PING madduck.net(2001:1620:2018:2::4d6d:8b55) 56 data bytes
64 bytes from 2001:1620:2018:2::4d6d:8b55: icmp_seq=1 ttl=58 time=23.1 ms
--- madduck.net ping statistics ---
1 packets transmitted, 1 received, 0% packet loss, time 0ms
rtt min/avg/max/mdev = 23.156/23.156/23.156/0.000 ms
Routing between hosts on this PoP broken
This problem appeared sometime between 07 and 09 Jun 2013.
We have not done any code updates to sixxsd for quite a while, thus we can fortunately exclude that then.
I assume you mean that before the 7th everything worked fine?
Live tunnel status does have for T65681:
ICMPv4 Errors Received : 8, last: 188.174.198.149 2013-06-09 10:02:49 (1370772169; 0 days 05:48:40 ago)
And except for the non-standard 1432 MTU nothing non-standard, but that MTU should not hurt if the underlying network can manage.
# ping6 -nc1 2001:a60:902f:1::2
Unless they are filtered (or untimely suddenly unreachable), I cannot reach that host at this moment:
$ traceroute6 2001:a60:902f:1::2
traceroute to 2001:a60:902f:1::2 (2001:a60:902f:1::2), 30 hops max, 80 byte packets
1 ge-1-3-0.breda.ipv6.concepts-ict.net (2001:838:1:1::1) 0.592 ms 0.579 ms 0.571 ms
2 2001:838:5:11::2 (2001:838:5:11::2) 0.749 ms 0.753 ms 0.750 ms
3 gi2-11.nikhef.ipv6.concepts-ict.net (2001:838:5:a::1) 2.014 ms 2.034 ms 2.078 ms
4 amsix-v6.irt1.ams01.nl.as13237.net (2001:7f8:1::a501:3237:1) 2.236 ms 2.078 ms 2.044 ms
5 ae1.irt1.dus03.de.v6.as13237.net (2001:7f0:0:29::1) 5.070 ms 4.899 ms 5.051 ms
6 ae3.irt1.fra44.de.v6.as13237.net (2001:7f0:0:2b::2) 7.683 ms 7.861 ms 7.847 ms
7 ae2.irt2.nue22.de.v6.as13237.net (2001:7f0:0:f::1) 18.626 ms 18.377 ms 18.388 ms
8 M-net-NUE.v6.lambdanet.net (2001:7f0:1:6::2) 17.116 ms 17.060 ms 17.229 ms
9 * * *
$ ping6 2001:a60:902f:1::2
PING 2001:a60:902f:1::2(2001:a60:902f:1::2) 56 data bytes
^C
--- 2001:a60:902f:1::2 ping statistics ---
2 packets transmitted, 0 received, 100% packet loss, time 1008ms
# tracepath6 -n 2001:a60:f069::1
You should at minimum get a hop for a single gw-XXX.
Can you provide the full routing tables on both sides of the tunnel, maybe you got some RPF going on there or accidentally have a /48 where a /64 should be or other routing fuzzyness.
|
![[ch]](/s/countries/ch.gif)
on Sunday, 09 June 2013 15:57:35
Posting is only allowed when you are