[82022] in North American Network Operators' Group
Re: OT? /dev/null 5.1.1 email
daemon@ATHENA.MIT.EDU (Valdis.Kletnieks@vt.edu)
Fri Jul 8 15:34:07 2005
To: jm@jmason.org (Justin Mason)
Cc: Patrick Muldoon <doon@inoc.net>,
Brad Knowles <brad@stop.mail-abuse.org>, nanog@nanog.org
In-Reply-To: Your message of "Tue, 05 Jul 2005 21:27:23 PDT."
<20050706042723.CE4272F0439@radish.jmason.org>
From: Valdis.Kletnieks@vt.edu
Date: Wed, 06 Jul 2005 01:24:31 -0400
Errors-To: owner-nanog@merit.edu
--==_Exmh_1120627471_3612P
Content-Type: text/plain; charset=us-ascii
On Tue, 05 Jul 2005 21:27:23 PDT, Justin Mason said:
> BTW, someone (possibly Randal L. Schwartz) came up with a neat related
> trick to the above -- set up an interface alias on *the same machine* as
> the primary MX, list that as the last MX in the list, and (assuming that
> the software side of the primary MX is reliable) you're then assured that
> any SMTP traffic that arrives on that IP's port 25 is spam, since when
> the primary MX's hardware goes down, this MX will, too.
That's got the same failure mode - if I take a 30-second hit and can't reach
the first MX, then the link comes up before I try the last MX, I hit the "bad" one.
And since the link burp is at *my* end, you don't even know about it, unless
you hook it into a full BGP feed with Zebra or something and see AS1312's routes
flap (and even then, it's dicy - a very short burp may not cause the routes to
be withdrawn)....
--==_Exmh_1120627471_3612P
Content-Type: application/pgp-signature
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.1 (GNU/Linux)
Comment: Exmh version 2.5 07/13/2001
iD8DBQFCy2sPcC3lWbTT17ARApdWAJoCgzrk6sH/7dEMoSa9YILZI8wtoACfbvpe
af+UKgDCUm44HfLoBbNbH0g=
=tv6J
-----END PGP SIGNATURE-----
--==_Exmh_1120627471_3612P--
