Weird behavior in a router with more than 2 IPs in a interface

Linux Advanced Routing and Traffic Control

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

 



	Hi list,
	
	I'm asking here because it is last resource.  I already tried a lot of
configurations and I really don't know how to solve this problem.
	I installed a Linux box (Debian 3.0, kernel 2.4.19) with 4 network
interfaces:
 eth0 with 192.168/24 address (intranet);
 eth1 with 10/24 address (DMZ);
 eth2 with 200/24 address (link 1 - default);
 eth3 with 200/24 address too (link 2 - ADSL - just for non specific
traffic).

	Everything was working properly, including source routing rules and
SNAT/DNAT translations.  So I started the 2d phase, migrating servers
from link eth2 to DMZ on link eth1.  No problems during some days, but
now I seeing a lot of weird behaviors.  Interface eth2 has 4/5 address
for while.  Some of them are reachable, some not, i.e., I just can ping
some address while some others not.   Using tcpdump, I can see icmp
packets arriving, but no response from the server.  Resetting switch
works for some time, allowing all IP to be reachable, but after a time
(minutes, hours or, even, days) some IPs just stops.
	I already tried to get firewall down, but it didn't solve the problem,
so I'm believing the problem is with layer 2 and internal routing.  My
iproute rules follows below:

ip route add default via 200.1.7.1         table dmznet
ip rule  add from 192.168.0.0/24              table dmznet
ip route add 192.168.0.0/24 via 192.168.0.254 table dmznet
ip rule  add from 200.1.7.0/26             table dmznet

ip addr  add 200.2.8.3/26 dev eth2
ip addr  add 200.2.8.4/26 dev eth2
ip addr  add 200.2.8.5/26  dev eth2
ip addr  add 200.2.8.15/26 dev eth2
ip rule  add from 192.168.0.0/24 to 200.2.8.0/26 table dmznet
ip route add 200.2.8.0/26 via 200.2.8.20      table dmznet
ip route add 10.0.0.0/24 via 10.0.0.254             table dmznet
ip rule  add from 10.0.0.11/32                      table dmznet

Looking with tcpdump, I can see this:
cramulhao:~# tcpdump -i any not port 22 and host 200.2.2.68 -n
tcpdump: listening on any
17:18:51.305577 200.2.2.68 > 200.2.8.20: icmp: echo request (DF)
17:18:51.305735 200.2.8.20 > 200.2.2.68: icmp: echo reply
17:18:56.933199 200.2.2.68 > 200.2.8.3: icmp: echo request (DF)
17:19:08.595382 200.2.2.68 > 200.2.8.4: icmp: echo request (DF)
17:19:23.298808 200.2.2.68 > 200.2.8.5: icmp: echo request (DF)
17:19:23.298954 200.2.8.5 > 200.2.2.68: icmp: echo reply

	As anyone can see, IPs ending with 3 and 4 are not responding.  Have
anyone a clue about what is happening?

	Thanks and BR's

-- 

Hélio Alexandre Lopes Loureiro [helio.loureiro@edb.ericsson.se]
Regional Software Supply & Integration
South America
Tel.: + 55 11 6224-1795 
Public Key ID: FB5972D1@http://search.keyserver.net


_______________________________________________
LARTC mailing list / LARTC@mailman.ds9a.nl
http://mailman.ds9a.nl/mailman/listinfo/lartc HOWTO: http://lartc.org/


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