Strange Reachability Issue

virendra rode virendra.rode at
Tue Sep 4 15:12:19 UTC 2012

Hash: SHA256


On 09/04/2012 07:20 AM, Bryn Sadler wrote:
> Many thanks to Jared for jumping on this so quickly off-list, it's
> much appreciated and hopefully we're getting towards a solution
> now.
> Bryn
- ------------------
yup you are in good hands, sounds more like filtering related among
peers. Use can use communities to test this theory.

> On 04/09/2012 15:12, "Jared Mauch" <jared at> wrote:
>> I know a few folks from NTT have looked into this.  If someone 
>> from KPN would get in touch with Bryn I'm sure the issue could
>> be quickly resolved.
>> - Jared
>> On Sep 4, 2012, at 9:18 AM, Brandt, Ralph wrote:
>>> I will bet that will bet that within 48 hours of you checking
>>> and posting this the problem will mysteriously go away.
>>> Ralph Brandt Mechanicsburg PA 17055
>>> -----Original Message----- From: Bryn Sadler
>>> [mailto:bryn.sadler at] Sent: Tuesday, September
>>> 04, 2012 9:02 AM To: nanog at Subject: Strange
>>> Reachability Issue
>>> Hello all,
>>> I was wondering if anyone might be able to share their thoughts
>>> on a strange issue we're experiencing with NTT at the moment.
>>> We're AS48273 and are advertising a prefix
>>> through AS 8190 (single upstream provider at the moment). We've
>>> been doing this for some years now, and all has been fine. The
>>> last few days we seem to have disappeared from NTT's worldwide
>>> routing tables, with the exception of Europe. If we use their
>>> looking glass to query any NTT european router we can see the
>>> prefix, being learned via AS 286 then AS 8190, as expected. If
>>> we look at any other router outside of europe, there is simply
>>> no entry for that prefix. Elsewhere in the world we seem to be 
>>> fine, we're in every other network I've looked at and general 
>>> reachability is fine.
>>> First step was to contact the NTT NOC, and they've confirmed
>>> that there's no Internal routing/config issue on their network,
>>> but that they cannot discuss their peering arrangements due to
>>> NDAs. We've also picked it up with KPN (AS 286), who have
>>> verified that the prefix is visible throughout their network
>>> (true), and that they are advertising it to NTT. One thing of
>>> note is that Global Crossing are learning our prefix from KPN
>>> as well, and the prefix is showing fine in their global
>>> tables.
>>> We seem to be caught in the classic finger pointing scenario,
>>> but I can't get my head around why the prefix is visible in NTT
>>> Europe, but not anywhere else, unless they have a config error
>>> somewhere on their network. We've checked a few of the routes
>>> from AS 8190 and they are in the NTT global tables just fine,
>>> and at least in Europe they seem to have the same community
>>> values etc. We're still following on with NTT, but can anyone
>>> offer some wisdom for new avenues to pursue?
>>> Thanks for your help,
>>> Bryn Sadler
>>> This message has been scanned for viruses by essensys
>>> mailcontrol
Version: GnuPG v1.4.11 (GNU/Linux)
Comment: Using GnuPG with Mozilla -


More information about the NANOG mailing list