Traffic going through single interface.

Hello,

I have a server with two Tigon3 network cards.

I want them to have distinct IP addresses, in the same subnet.

But, the traffic is incoming/outgoing only through one interface card.

See:

FO[/etc]#netstat -i
Kernel Interface table
Iface MTU Met RX-OK RX-ERR RX-DRP RX-OVR TX-OK TX-ERR TX-DRP
TX-OVR Flg
eth1 1500 0 104689 0 0 0 0 0 0
0 BMRU
eth2 1500 0 105319 0 0 0 409 0 0
0 BMRU
lo 16436 0 39 0 0 0 39 0 0
0 LRU

FO[/etc]#netstat -rn
Kernel IP routing table
Destination Gateway Genmask Flags MSS Window
irtt Iface
161.148.173.0 10.0.0.3 255.255.255.0 UG 0 0
0 eth2
10.0.0.0 0.0.0.0 255.255.240.0 U 0 0
0 eth2
10.0.0.0 0.0.0.0 255.255.240.0 U 0 0
0 eth1
0.0.0.0 10.0.0.2 0.0.0.0 UG 0 0
0 eth1
0.0.0.0 10.0.0.2 0.0.0.0 UG 0 0
0 eth2


FO[/etc]#ifconfig
eth1 Link encap:Ethernet HWaddr 00:0F:1F:6EC:5E
inet addr:10.0.3.245 Bcast:10.0.15.255 Mask:255.255.240.0
UP BROADCAST RUNNING MULTICAST MTU:1500 Metric:1
RX packets:105399 errors:0 dropped:0 overruns:0 frame:0
TX packets:0 errors:0 dropped:0 overruns:0 carrier:0
collisions:0 txqueuelen:1000
RX bytes:11616629 (11.0 MiB) TX bytes:0 (0.0 b)
Interrupt:233

eth2 Link encap:Ethernet HWaddr 00:0F:1F:6EC:5F
inet addr:10.0.3.250 Bcast:10.0.15.255 Mask:255.255.240.0
UP BROADCAST RUNNING MULTICAST MTU:1500 Metric:1
RX packets:106073 errors:0 dropped:0 overruns:0 frame:0
TX packets:445 errors:0 dropped:0 overruns:0 carrier:0
collisions:0 txqueuelen:1000
RX bytes:11674467 (11.1 MiB) TX bytes:53389 (52.1 KiB)
Interrupt:50

Any suggestions? Debian 3.0, Kernel 2.6.11.10, SMP.

Rod.

"Rodrigo A B Freire" <(E-Mail Removed)> writes:

>Hello,

>I have a server with two Tigon3 network cards.

>I want them to have distinct IP addresses, in the same subnet.

>But, the traffic is incoming/outgoing only through one interface card.

Yes. Traffic goes through the interface that the routing table tells the
system to send traffic through.
What are you trying to do? If you want load sharing, then the system is not
set up to do that.

>See:

>FO[/etc]#netstat -i
>Kernel Interface table
>Iface MTU Met RX-OK RX-ERR RX-DRP RX-OVR TX-OK TX-ERR TX-DRP
>TX-OVR Flg
>eth1 1500 0 104689 0 0 0 0 0 0
>0 BMRU
>eth2 1500 0 105319 0 0 0 409 0 0
>0 BMRU
>lo 16436 0 39 0 0 0 39 0 0
>0 LRU

>FO[/etc]#netstat -rn
>Kernel IP routing table
>Destination Gateway Genmask Flags MSS Window
>irtt Iface
>161.148.173.0 10.0.0.3 255.255.255.0 UG 0 0
>0 eth2
>10.0.0.0 0.0.0.0 255.255.240.0 U 0 0
>0 eth2
>10.0.0.0 0.0.0.0 255.255.240.0 U 0 0
>0 eth1
>0.0.0.0 10.0.0.2 0.0.0.0 UG 0 0
>0 eth1
>0.0.0.0 10.0.0.2 0.0.0.0 UG 0 0
>0 eth2

Only the last default route is operative. Remove one or the other of these.
Similarly only one of the specific routes is operative.

Again what did you think you wanted to do?




>FO[/etc]#ifconfig
>eth1 Link encap:Ethernet HWaddr 00:0F:1F:6EC:5E
> inet addr:10.0.3.245 Bcast:10.0.15.255 Mask:255.255.240.0
> UP BROADCAST RUNNING MULTICAST MTU:1500 Metric:1
> RX packets:105399 errors:0 dropped:0 overruns:0 frame:0
> TX packets:0 errors:0 dropped:0 overruns:0 carrier:0
> collisions:0 txqueuelen:1000
> RX bytes:11616629 (11.0 MiB) TX bytes:0 (0.0 b)
> Interrupt:233

>eth2 Link encap:Ethernet HWaddr 00:0F:1F:6EC:5F
> inet addr:10.0.3.250 Bcast:10.0.15.255 Mask:255.255.240.0
> UP BROADCAST RUNNING MULTICAST MTU:1500 Metric:1
> RX packets:106073 errors:0 dropped:0 overruns:0 frame:0
> TX packets:445 errors:0 dropped:0 overruns:0 carrier:0
> collisions:0 txqueuelen:1000
> RX bytes:11674467 (11.1 MiB) TX bytes:53389 (52.1 KiB)
> Interrupt:50

>Any suggestions? Debian 3.0, Kernel 2.6.11.10, SMP.

>Rod.

Unruh,

> What are you trying to do? If you want load sharing, then the system is not set up to do that

Yes, I wanted kind of load balance the traffic between the interfaces.

I've set a round-robin in the DNS table, the machine has two squid
instances running and, of course, each instance is bonded to a IP
address.

What are using for this load balancing?



--
Raqueeb Hassan
Bangladesh

Raqueeb,

>What are using for this load balancing?

A few background.

Squid isn't a multi-threaded process, so it can't take advantage of a
SMP system.
The solution that I've found: Run two squid processes in the machine,
each one bond to a different IP address.

The 'load balancing' is provided by the round-robin in the DNS (each
resolution points to a different IP address).

I also wanted to distribute the load between the two NICs of the
machine. But, the problem is, ALL the traffic is going through a single
interface. And I ask the group members: what can I do to each interface
handle its IP address?

Rod.