[2079] in athena10
Re: Linerva can't resolve some unc.edu hosts
daemon@ATHENA.MIT.EDU (Nelson Elhage)
Sun Apr 5 23:29:36 2009
Date: Sun, 5 Apr 2009 23:29:04 -0400
From: Nelson Elhage <nelhage@MIT.EDU>
To: Xavid <Xavid@mit.edu>
Cc: linerva@mit.edu
Message-ID: <20090406032904.GM5764@mit.edu>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <1B446D4F-52E1-4913-A7B1-4BA141E3C050@MIT.EDU>
Yeah, I bet they're filtering us. I'll send them email tomorrow asking
for them to fix it unless someone beats me to it.
- Nelson
On Sun, Apr 05, 2009 at 11:02:19AM -0400, Xavid wrote:
> On linerva today, I was unable to resolve www.physics.unc.edu. Other
> machines seem to resolve it fine, so either it's something wacky about
> linerva or it got blacklisted somewhere. Zlogs below.
>
> ~Xavid
>
> sipb / linerva / xavid 10:18 ('I kinda want to be a religious
> zealot so I can
> On linerva:
> $ ping www.physics.unc.edu
> ping: unknown host www.physics.unc.edu
>
> I can resolve it just fine from a VM or locally...
> -> sipb / linerva / xavid 10:40 (`The Brazilians are gone, the vomit's
> cleaned up
> xavid@vinegar-pot:~$ dig www.physics.unc.edu
>
> ; <<>> DiG 9.5.1-P1 <<>> www.physics.unc.edu
> ;; global options: printcmd
> ;; connection timed out; no servers could be reached
>
> It doesn't seem to affect any other MIT machines I try.
> sipb / linerva / broder 10:43 (Evan Broder)
> I don't know what's going on. The resolver can talk to the NS
> servers
> for unc.edu, and if I ask any individual server where
> physics.unc.edu
> is, they can tell me, but bind can't seem to work it out
> sipb / linerva / keithw 10:49 (Keith Winstein)
> It's not a BIND problem; Linerva can't ping 152.2.5.15
> (ns1.physics.unc.edu) or 152.2.4.8, and a traceroute gets
> almost all the way but not to the last hop. Since
> lamp.mit.edu (just 6 away from linerva's 18.181 IP address)
> can ping both, I'm not sure why that is.
> -> sipb / linerva / broder 10:49 (Evan Broder)
> This wouldn't be the first time that linerva had gotten black-
> holed by
> some automated security monitor
> sipb / linerva.d / keithw 10:50 (Keith)
> (Confirmation would be making sure the problem is
> IP-address-specific instead of Linerva-configuration-specifi
> c, but I thought trying to have LAMP steal Linerva's IP
> address to try to nail that down would not be a good
> idea...)
>
