[106075] in North American Network Operators' Group
Re: SBCglobal routing loop.
daemon@ATHENA.MIT.EDU (Mike Lewinski)
Fri Jul 18 20:04:15 2008
Date: Fri, 18 Jul 2008 18:03:20 -0600
From: Mike Lewinski <mike@rockynet.com>
To: nanog@merit.edu
In-Reply-To: <18f601940807181608p3ba1c6e7m42feb61ce9baf189@mail.gmail.com>
Errors-To: nanog-bounces@nanog.org
Aaron Glenn wrote:
> I think it should be clear to those posting here as a last ditch
> effort that they should certainly outline the steps they've already
> taken -- basically justifying their post to NANOG: "I tried X, waited
> Y, got Z, and now I'm here"
To give an example:
http://mailman.nanog.org/pipermail/nanog/2008-June/001452.html
I started with the customer care support line (in a concall with our
client who was also their client). Then I moved onto various email
contacts. Since my problem was partially DNS related, I tried the SOA
mailto address. Since the other part of my problem was email related I
tried postmaster@. I also tried to mail the ARIN whois contacts (but a
bug in my client software prevented me from getting the record, and
since the error appeared to be on the RPs side it looked to be a dead end)
Then I went onto Jared's NOC list, as well as the INOC-DBA directory.
By the time I resorted to posting here, I had exhausted every feasible
contact method I could think of. And yes, that meant waiting days for
resolution BEFORE posting here, and some additional days AFTER posting
here before actually getting the help I needed.
INOC-DBA is still a great idea, BTW, and would be even more useful if it
were more widely used: http://www.pch.net/inoc-dba/
Also, to second another poster's point, most "routing loops" aren't
really routing problems per-se, as much as localized link failures and
the absence of naildown routes (we use the naildowns on all
point-to-point circuits for this reason, since we don't run routing
protocol on CPE).
