Question about "host already uses address" on PE r610 with Broadcom NetXtreme II

Thomas_Chenault at Dell.com Thomas_Chenault at Dell.com
Wed Apr 28 19:01:52 CDT 2010


In this respect, Linux treats IP addresses in a host-based rather than interface-based fashion. A Linux host owning a given IP address will respond to an ARP request for that address on any and all interfaces that receive the request, supplying the MAC address of the responding interface. This behavior can be modified by enabling arp_filter for the desired interfaces.

      echo 1 > /proc/sys/net/ipv4/conf/<netdev>/arp_filter


Thomas

________________________________
From: linux-poweredge-bounces-Lists On Behalf Of sa
Sent: Wednesday, April 28, 2010 3:35 PM
To: linux-poweredge-Lists
Subject: Question about "host already uses address" on PE r610 with Broadcom NetXtreme II

I have a pair of Dell r610s, each of which has an on board quad port Broadcom network card.  Both are running a recently patched CentOS 5.4.  I set each server up to use DHCP to get an IP address for eth0.  On server 1, I set eth3 to 192.168.253.1.  Everything is working as expected.  I can access the server via either address.  Note that both ports are connected to the same switch.

Now I physically move the network cord from eth3 on server1 to eth3 on server2.  I turn on server 2 which gets its eth0 address from DHCP.  eth3 is set to use the same IP that eth3 on server1 was, 192.168.253.1.  However eth3 on server2 fails to come up, issuing this error:

Error, some other host already uses address 192.168.253.1.

So at this point, I have server1 that only has eth0 physically connected to the network.  I have server2 that has eth0 and eth3 connected to the network.  If I issue an arping 192.168.253.1, I get back the MAC address of eth0 on server1.  Then I do the following steps.

server2 # ifup eth3
Error, some other host already uses address 192.168.253.1.

server1 # ifdown eth3
server1 #

server2 # ifup eth3
server2 #
server2 # ifconfig eth3
eth3      Link encap:Ethernet  HWaddr 00:21:9B:99:1B:F2
          inet addr:192.168.253.1  Bcast:192.168.253.255  Mask:255.255.255.0
....
server2 # ifdown eth3

server1 # ifup eth3

server2 # ifup eth3
Error, some other host already uses address 192.168.253.1.

Physically disconnect eth0 on server1 while eth3 is configured and eth3 on server2 comes up with no error.

So my question comes down to how is an interface that is not physically connected to the network having this effect?  Why does the connection to eth0 seem to be influencing the eth3 settings?

Got questions or need more info, let me know.

Chris

---------------------
More info:



server2 # arping 192.168.253.1
Unicast reply from 192.168.253.1 [00:21:9B:99:11:E7]  0.700ms
Unicast reply from 192.168.253.1 [00:21:9B:99:11:E7]  0.682ms

server 1 # ifconfig
eth0      Link encap:Ethernet  HWaddr 00:21:9B:99:11:E7
          inet addr:10.1.0.189  Bcast:10.1.255.255  Mask:255.255.0.0
...
...
eth3      Link encap:Ethernet  HWaddr 00:21:9B:99:11:ED
          inet addr:192.168.253.1  Bcast:192.168.253.255  Mask:255.255.255.0


dmesg | grep Bro
Broadcom NetXtreme II Gigabit Ethernet Driver bnx2 v1.9.3 (March 17, 2009)
eth0: Broadcom NetXtreme II BCM5709 1000Base-T (C0) PCI Express found at mem d6000000, IRQ 90, node addr 00219b9911e7
eth1: Broadcom NetXtreme II BCM5709 1000Base-T (C0) PCI Express found at mem d8000000, IRQ 98, node addr 00219b9911e9
eth2: Broadcom NetXtreme II BCM5709 1000Base-T (C0) PCI Express found at mem da000000, IRQ 106, node addr 00219b9911eb
eth3: Broadcom NetXtreme II BCM5709 1000Base-T (C0) PCI Express found at mem dc000000, IRQ 114, node addr 00219b9911ed

-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.us.dell.com/pipermail/linux-poweredge/attachments/20100428/f193f7b1/attachment-0001.htm 


More information about the Linux-PowerEdge mailing list