[48469] in North American Network Operators' Group
Re: Bogon list
daemon@ATHENA.MIT.EDU (Miquel van Smoorenburg)
Wed Jun 5 11:32:39 2002
To: nanog@merit.edu
From: "Miquel van Smoorenburg" <miquels@cistron.nl>
Date: Wed, 5 Jun 2002 15:29:52 +0000 (UTC)
X-Complaints-To: abuse@cistron.nl
Errors-To: owner-nanog-outgoing@merit.edu
In article <cistron.20020605144803.GB12164@overlord.e-gerbil.net>,
Richard A Steenbergen <ras@e-gerbil.net> wrote:
>On Wed, Jun 05, 2002 at 08:34:58AM +0000, Miquel van Smoorenburg wrote:
>>
>> I haven't seen a 'icmp source lo0' interface command yet. Hopefully
>> it will be added for ipv6 so exchanges can use link-local addressing
>> (ipv6 has no fragmentation, PMTUd is mandatory).
>
>I'm not terribly sure why you would want to make traceroutes lose all
>information about the circuits you're traveling through. It would make
>diagnostics an everloving nightmare, IMHO.
With link-local addressing and ip verify unicast reverse you're
not going to see any TTL exceeded traceroute packets from link-local
addresses anyway. And no ICMP size exceeded packets either, which
was my point.
But indeed it should be 'ip icmp type <TYPE> source <IP|INT>',
with at least a default 'ip icmp type 4 source lo0' or similar for
interfaces with link-local addressing.
Mike.
