[76579] in North American Network Operators' Group
Re: Interesting DNS problem.
daemon@ATHENA.MIT.EDU (Eric Brunner-Williams in Portland )
Thu Dec 16 16:19:33 2004
To: "william(at)elan.net" <william@elan.net>
Cc: Bob Martin <bob@buckhorn.net>, NANOG list <nanog@merit.edu>,
brunner@nic-naa.net
In-Reply-To: Your message of "Thu, 16 Dec 2004 13:27:27 PST."
<Pine.LNX.4.44.0412161309450.31925-100000@sokol.elan.net>
Date: Thu, 16 Dec 2004 17:17:35 +0000
From: Eric Brunner-Williams in Portland Maine <brunner@nic-naa.net>
Errors-To: owner-nanog-outgoing@merit.edu
a related problem is having N ip addrs bound to M nics on a host, where N > M.
if an ssl connection fails and debug is needed between the M:N:host and some
other ssl-speaking box, then it makes a difference if the ssl connection is
associated with the primary, or some aliased (set N-1) ip addr. client failure
semantics are primary address specific, for some value of ssl clients.
in theory you could alias an ns box's ip addrs (just did that, renumbering),
and have multi-addrs on a server authoritative for multi-zones, and not have
a flag day.
have fun, jobs are scarce as hen's teeth.
