Linux Netfilter discussions
 help / color / mirror / Atom feed
From: Phil Stoneman <pws@corefiling.com>
To: netfilter@vger.kernel.org
Subject: TTL Exceeded packets coming from the original packet destination
Date: Tue, 24 May 2011 16:44:51 +0100	[thread overview]
Message-ID: <4DDBD273.3070405@corefiling.com> (raw)

Hi all

I've been searching google and these archives, and not found a useful 
answer to this, so I wonder here if someone can help me?

We have a range of public IPs allocated to us by our ISP. Our linux 
router sits on one of those, and another of those is NATed through to a 
host in our DMZ. Traceroutes from the outside world which get dropped by 
the linux router have the TTL exceeded packet generated with the source 
set to the original UDP packet destination - not set to an interface 
address of the linux router.

This gives us something like this:

pws@externalhost:~$ traceroute -n 90.155.100.39
traceroute to 90.155.100.39 (90.155.100.39), 30 hops max, 40 byte packets
<snip>
  4  195.66.224.223  1.236 ms  1.249 ms  1.275 ms
  5  213.229.97.181  2.328 ms  2.327 ms  2.304 ms
  6  90.155.100.39  5.476 ms  5.050 ms  5.087 ms
  7  90.155.100.39  5.266 ms  5.381 ms  5.406 ms

Which isn't as I'd expect (I'd expect hop 6 to be the interface address 
of the linux router).


Network details (monospace font):

InsideHost<----->Linux Router<----->Outside world
          |       |          |       |
10.0.250.2       |          |       |
         10.0.250.1          |       |
                90.155.100.198       |
                        90.155.100.193

Relevant iptables rules:
iptables.eth-ext:-A PREROUTING -i vlan90 -j preroutingaa
iptables.eth-ext:-A preroutingaa -d 90.155.100.39    -j DNAT 
--to-destination 10.0.250.2
iptables.eth-ext:-A POSTROUTING -j sourcenat
iptables.eth-ext:-A sourcenat -s 10.0.250.2 -o vlan90 -j SNAT 
--to-source 90.155.100.39

I'm on a Debian 6.0.1 system, with kernel 2.6.32-5-686 in case that's 
relevant.


I don't know if this is being caused by the conntrack stuff doing some 
undesired NAT on the ICMP responses, or if it's related to this posting 
I found on google: http://www.spinics.net/lists/lartc/msg18454.html

Anyway, could someone help me work out what I need to do to make 
traceroutes happen properly please?


Thanks!

Phil

                 reply	other threads:[~2011-05-24 15:44 UTC|newest]

Thread overview: [no followups] expand[flat|nested]  mbox.gz  Atom feed

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=4DDBD273.3070405@corefiling.com \
    --to=pws@corefiling.com \
    --cc=netfilter@vger.kernel.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox