Re: first of the two IPs reachable only if second interface is down??

Linux Advanced Routing and Traffic Control

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



Hi Edward

but now, if I do - ip route show
I see: (change IPs for our purposes here)

10.0.0.0/24 dev eth1  proto kernel  scope link  src 10.0.0.2
public_net/24 dev eth0 proto kernel scope link src "public_IP_of_the_box"

169.254.0.0/16 dev eth0  scope link  metric 1002
169.254.0.0/16 dev eth1  scope link  metric 1003

default via "public_IP_of_the_box" dev eth0
default via "private_IP_of_the_box" dev eth1  metric 1

isn't this already ok?
both IP get set up by dhcpd by another admin, only setting by me is the metric, 0 for public inteface and 1 for private.

now I'll apply your instructions and let you know if there are any changes

I believe that if "a_client(s)" on private network (the same network the box operates on with its 'private interface') that has set the same gateway as "the_box", that this a_client(other computers on private network) if needs to reach 'public interface' of 'the_box' then the local gateway and other routers outside should take care of routing the traffic, and 'the_box' should not need any special configs. And if remember correctly it all used to work some long time ago, like:


a_client (10.0.0.100) <--> (10.0.0.1) router <---> (public IP) the_BOX (10.0.0.101) .. the rest of the stack

many! thanks for your help Edward

On 21/06/12 12:34, Edward Murphy wrote:
Hi,

What you'll find if I'm reading this right, is that packets are coming in via your public interface and out the private one. The requesting host simply discards The packet.

Solutio: use Iproute2 to have 2 network tables for the system independent of eachother.

Firstly. Please forgive syntax errors here as I'm doing this from memory.

Edit /etc/iproute/rt_tables and add 2 lines:

200 external
201 internal

Note the names are irrelevant.

Reboot the system.

Now in your init scripts for network setup the interface.

Let's say your private interface is 10.0.0.2/24

ip route add 10.0.0.0/24 dev eth0 source 10.0.0.2 table private ip route add default via 10.0.0.1 source 10.0.0.2 table internal
ip rule add from 10.0.0.2 dev eth0 table internal.


By doing this you are firstly setting up the routing table, and then pushing any data that has a source of your internal interface (replies to packets mostly) to use this routing table.

Do the same for your public interface but specify the public network card, and the routing table external.

Kind regards
Edward Murphy


----- Reply message -----
From: "lejeczek" <peljasz@xxxxxxxxxxx>
Date: Thu, Jun 21, 2012 01:08
Subject: first of the two IPs reachable only if second interface is down??
To: <lartc@xxxxxxxxxxxxxxx>

hello everybody

apologies if this may feel off the topic, I was hoping some
net experts could shed some lights on some peculiar symptoms
I experience
I would very! much appreciate any help

a BOX that has two net interfaces, a public and private one,
both configured via dhcp by net admin

public IP is reachable from/via the Internet/public network
just fine, but at the same time
same public IP is NOT reachable from within PRIV network,
the same network the BOX's second interface is on

default gateway for the PRIV subnet is a separate another
system, BOX uses PRIV gateway as the default one (but have
tried the public gateway too)

arpinging both IPs from PRIV gives me the same one MAC of
BOX private net interface

public IP becomes reachable to PRIV subnet immediately after
second(private) interface was turned down

we have no control over the whole network stack but know
that PUBLIC IP connection goes via unmanaged switches to the
gateway

is it the BOX or the network stack somewhere?
unmanaged switches should not cache any ARP, also, how come
that it gets corrected (public becomes available to PRIV)
only if I turn PRIV interface down??

box is SL 6.1, firewall is off, ip_forward=0
this is the most peculiar problem of this nature I've ever
experienced

the goal is simple, have other systems on the same private
subnet as the BOX's second interface to be able to talk to
the BOX's public IP
obviously not routed via BOX's private address, this
workaround solves the problem, but routing via PRIV's gataway

many thanks
--
To unsubscribe from this list: send the line "unsubscribe lartc" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html


--
To unsubscribe from this list: send the line "unsubscribe lartc" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html


[Index of Archives]     [LARTC Home Page]     [Netfilter]     [Netfilter Development]     [Network Development]     [Bugtraq]     [GCC Help]     [Yosemite News]     [Linux Kernel]     [Fedora Users]
  Powered by Linux