[AusNOG] ICMP weirdness inside Internode's network

[Julien Goodwin]

As I don't have any great Internode contacts at the moment I thought I'd throw
this out to Ausnog where there'll be a few of them listening.

I was just debugging some IPv6 issues on a DSL session in MEL (all my side),
when as one of the final actions I traced from my DSL in SYD and got * steps,
but only on v4, not on v6.

My first guess is just broken ICMP handling inside MPLS tunnels, but without
knowledge of Internode's network internals I can't make an intelligent suggestion.

ex: (exact edge IP's removed, anyone at 'node will already know what they are)

traceroute to 203.x.y.z (203.x.y.z), 30 hops max, 60 byte packets
 1  fritz.box (203.x.y.z)  7.146 ms  7.113 ms  7.099 ms
 2  lns20.syd7.internode.on.net (150.101.199.219)  42.663 ms  42.675 ms  42.665 ms
 3  * * *
 4  * * *
 5  * * *
 6  * * *
 7  te0-0-0.lns20.mel4.internode.on.net (150.101.208.66)  54.865 ms  53.364 ms  55.901 ms
 8  X (203.x.y.z)  86.870 ms !H  82.892 ms !H  83.407 ms !H

traceroute to 2001:44b8:yyyy:zzzz::1 (2001:44b8:yyyy:zzzz::1), 30 hops max, 80 byte packets
 1  fritz.box (2001:44b8:yyyy:zzzz:a:b:c:d)  2.560 ms  4.610 ms  4.601 ms
 2  loop0.lns20.syd7.internode.on.net (2001:44b8:b070::6)  42.887 ms  42.899 ms  42.885 ms
 3  te3-3.cor2.syd7.internode.on.net (2001:44b8:b070:15::1)  45.797 ms  45.805 ms  48.745 ms
 4  gi6-0-0-143.bdr1.syd6.internode.on.net (2001:44b8:b060:143::1)  48.742 ms * *
 5  te0-0-0.bdr1.mel4.internode.on.net (2001:44b8:c040:d::1)  61.509 ms * *
 6  2001:44b8:c0d0:f::2 (2001:44b8:c0d0:f::2)  63.993 ms  50.571 ms  52.995 ms
 7  te0-0-0.lns20.mel4.internode.on.net (2001:44b8:c040:15::2)  55.147 ms  52.988 ms  52.961 ms
 8  2001-44b8-X.static.ipv6.internode.on.net (2001:44b8:yyyy:zzzz::1)  85.353 ms  82.677 ms  85.077 ms