[80675] in North American Network Operators' Group
Re: google.com outage?
daemon@ATHENA.MIT.EDU (Simon Waters)
Mon May 9 04:35:32 2005
To: nanog@merit.edu
Date: Mon, 9 May 2005 09:34:59 +0100
In-Reply-To: <g3is1tx87b.fsf@sa.vix.com>
From: Simon Waters <simonw@zynet.net>
Errors-To: owner-nanog@merit.edu
On Monday 09 May 2005 6:46 am, Paul Vixie wrote:
> > -- Chris Keladis <chris@cmc.optus.net.au> wrote:
> >
> > Just a guess, but perhaps with www.google.com returning NXDOMAIN the
> > gethostby* functions tried variations and ended up resolving sites like
> > www.google.com.net which inadvertantly sent people to the spoof site.
> >
> > Seems plausible to me, anyway.
>
> not to me. RFC 1535 is VERY widely deployed, perhaps even universally so.
I doubt it is universally so, as I think old versions of Windows did something
similar in the 4.8.3 resolver, and they are still out there. Software takes a
long time to die.
However a lot of people saw the search engine "sogo" because their browser
asked for www.google.com.net when it found www.Google.com didn't exist, and
this is caught by a wildcard (allegedly). Although my quick inspection
suggests that the DNS for these domains is very badly mangled, BIND 9 does
somehow manage to get an IP address out of the mess it finds.
