www.arin.net

[Alexander Gall]

On Mon, 21 Aug 2006 17:18:58 -0700, Doug Barton <dougb at dougbarton.us> said:

> Tristan Horn wrote:
>> Since reachability to 2001:440:2000:1::16/125 is shown to be
>> problematic, and if your resolver is preferring AAAA over A while doing
>> reverse DNS lookups, could this explain the long delay before the telnet
>> connection opens?

> Voila! Been so long since I used telnet that this one slipped my mind. I
> added that address to my hosts file, and the delay is gone.

And can you actually do a "GET /"?

> Jim, ignore everything I said about 2001:500:4:1::80, and focus on the fact
> that I can't reach 2001:440:2000:1::18 from here. Then if I'm still not
> making sense, well, have a nice day. :)

I have similar problems.  In fact, I can see three distinct issues
here.

1) Inverse mappings for 2001:500:4::/48 are badly broken.  All
   delegations for 4.0.0.0.0.0.5.0.1.0.0.2.ip6.arpa are lame.

2) Packets exceeding a certain size seem to be dropped on the path
   from 2001:500:4:1::80 to me, just like Larry has reported in
   another message in this thread.  For the record, my (incomplete)
   path towards 2001:500:4:1::80 is

   traceroute: Warning: Multiple interfaces found; using 2001:620:0:4:203:baff:fe4c:d99b @ bge0:1
   traceroute to 2001:500:4:1::80, 30 hops max, 60 byte packets
    1  swiLM1-V4.switch.ch (2001:620:0:4::1)  0.676 ms
    2  swiNM1-V610.switch.ch (2001:620:0:c047::1)  0.607 ms
    3  swiCS3-G3-3.switch.ch (2001:620:0:c046::1)  0.381 ms
    4  swiCS5-10GE-1-3.switch.ch (2001:620:0:c039::5)  0.430 ms
    5  swiZH2-G2-8.switch.ch (2001:620:0:c016::2)  0.549 ms
    6  swiCE3-10GE-1-1.switch.ch (2001:620:0:c027::1)  4.666 ms
    7  2001:450:2002:10::1  13.902 ms
    8  2001:504:0:2::3549:1  93.720 ms
    9  2001:504:0:2::3549:1  93.771 ms
   10  equi6ix-ash.ipv6.us.occaid.net (2001:504:0:2:0:3:71:1)  93.714 ms
   11  2001:4830:e6:a::2  94.074 ms
   12  *
   13  *
   14 ^C

   ping with large packets indicates that the MTU on the reverse path
   is at most 1496 bytes (1448 bytes payload, 40 bytes IPv6 header, 8
   bytes ICMP header) but PMTU is broken and 2001:500:4:1::80 is
   sending larger segments, which are blackholed

   $ ping -ns 2001:500:4:1::80 1448
   PING 2001:500:4:1::80 (2001:500:4:1::80): 1448 data bytes
   1456 bytes from 2001:500:4:1::80: icmp_seq=0. time=97.142 ms
   ^C
   ----2001:500:4:1::80 PING Statistics----
   1 packets transmitted, 1 packets received, 0% packet loss
   round-trip (ms)  min/avg/max/stddev = 97.142/97.142/97.142/NaN
   $ ping -ns 2001:500:4:1::80 1449
   PING 2001:500:4:1::80 (2001:500:4:1::80): 1449 data bytes
   ^C
   ----2001:500:4:1::80 PING Statistics----
   2 packets transmitted, 0 packets received, 100% packet loss


3) 2001:440:2000:1::18 is completely unreachable for me.  Here is as
   far as I can see with traceroute

   traceroute: Warning: Multiple interfaces found; using 2001:620:0:4:203:baff:fe4c:d99b @ bge0:1
   traceroute to 2001:440:2000:1::18, 30 hops max, 60 byte packets
    1  swiLM1-V4.switch.ch (2001:620:0:4::1)  0.630 ms
    2  swiNM1-V610.switch.ch (2001:620:0:c047::1)  0.575 ms
    3  swiCS3-G3-3.switch.ch (2001:620:0:c046::1)  0.411 ms
    4  swiCS5-10GE-1-3.switch.ch (2001:620:0:c039::5)  0.417 ms
    5  swiZH2-G2-8.switch.ch (2001:620:0:c016::2)  0.538 ms
    6  swiCE3-10GE-1-1.switch.ch (2001:620:0:c027::1)  4.379 ms
    7  2001:450:2002:10::1  13.915 ms
    8  2001:440:eeee:ffc8::2  164.711 ms
    9  2001:440:eeee:ffc8::2  164.725 ms
   10  2001:440:eeee:ffc8::1  242.038 ms
   11  2001:440:1239:7000::2  245.979 ms
   12  *
   13 ^C

--
Alex