There are multiple eth interfaces on my box. I connected two of them directly back-to-back using a fiber cable. But I could not ping from one to the other successfully.

(1) The two interfaces are:

eth1 Link encap:Ethernet HWaddr 00:1B:21:A2:73:8C
inet addr:10.0.0.1 Bcast:10.0.0.255 Mask:255.255.255.0
inet6 addr: fe80::21b:21ff:fea2:738c/64 Scope:Link
UP BROADCAST RUNNING MULTICAST MTU:1500 Metric:1
RX packets:165 errors:0 dropped:0 overruns:0 frame:0
TX packets:670 errors:0 dropped:0 overruns:0 carrier:0
collisions:0 txqueuelen:1000
RX bytes:25574 (24.9 KiB) TX bytes:46748 (45.6 KiB)

eth7 Link encap:Ethernet HWaddr 00:1B:21:5C:FD:E8
inet addr:10.0.0.2 Bcast:10.0.0.255 Mask:255.255.255.0
inet6 addr: fe80::21b:21ff:fe5c:fde8/64 Scope:Link
UP BROADCAST RUNNING MULTICAST MTU:1500 Metric:1
RX packets:651 errors:0 dropped:0 overruns:0 frame:0
TX packets:179 errors:0 dropped:0 overruns:0 carrier:0
collisions:0 txqueuelen:1000
RX bytes:53068 (51.8 KiB) TX bytes:24290 (23.7 KiB)


(2) The ping output:
# ping 10.0.0.1 -I eth7
PING 10.0.0.1 (10.0.0.1) from 10.0.0.2 eth7: 56(84) bytes of data.
From 10.0.0.2 icmp_seq=2 Destination Host Unreachable
From 10.0.0.2 icmp_seq=3 Destination Host Unreachable
From 10.0.0.2 icmp_seq=4 Destination Host Unreachable
^C
--- 10.0.0.1 ping statistics ---
5 packets transmitted, 0 received, +3 errors, 100% packet loss, time 4000ms
pipe 3
# ping 10.0.0.2 -I eth1
PING 10.0.0.2 (10.0.0.2) from 10.0.0.1 eth1: 56(84) bytes of data.
From 10.0.0.1 icmp_seq=2 Destination Host Unreachable
From 10.0.0.1 icmp_seq=3 Destination Host Unreachable
From 10.0.0.1 icmp_seq=4 Destination Host Unreachable
^C
--- 10.0.0.2 ping statistics ---
4 packets transmitted, 0 received, +3 errors, 100% packet loss, time 2999ms
pipe 3

(3) The ARP output:
# arp
Address HWtype HWaddress Flags Mask Iface
10.0.0.1 (incomplete) eth7
10.0.0.2 ether 00:1b:21:5c:fd:e8 C eth1
10.126.135.254 ether 00:04:96:1e:4f:50 C eth0

(4) I added the ARP entry manually, but still failed to ping:
# arp -s 10.0.0.1 00:1B:21:A2:73:8C -i eth7
# arp
Address HWtype HWaddress Flags Mask Iface
10.0.0.1 ether 00:1b:21:a2:73:8c CM eth7
10.0.0.2 ether 00:1b:21:5c:fd:e8 C eth1
10.126.135.254 ether 00:04:96:1e:4f:50 C eth0

# ping 10.0.0.1 -I eth7
PING 10.0.0.1 (10.0.0.1) from 10.0.0.2 eth7: 56(84) bytes of data.
^C
--- 10.0.0.1 ping statistics ---
5 packets transmitted, 0 received, 100% packet loss, time 3999ms


# tcpdump -i eth7
tcpdump: verbose output suppressed, use -v or -vv for full protocol decode
listening on eth7, link-type EN10MB (Ethernet), capture size 65535 bytes
13:16:38.417302 IP 10.0.0.2 > 10.0.0.1: ICMP echo request, id 56000, seq 8, length 64
13:16:39.417289 IP 10.0.0.2 > 10.0.0.1: ICMP echo request, id 56000, seq 9, length 64
13:16:40.417367 IP 10.0.0.2 > 10.0.0.1: ICMP echo request, id 56000, seq 10, length 64


Just don't know why it did not work??!

Any help will be appreciated.

When connecting two interfaces directly together you will need a cross over cable. Are you sure you're using the right kind of cable?

Hi whizzit,

Thanks. I may not use the right cable. But the tcpdump on the receiver side showed the ICMP request was received.

For example, when "# ping 10.0.0.1 -I eth7",
On the sender interface:
On the receiver interface:

I guess the problem is, that you try to use two local IPs. Try tcpdump -i any, there you should see the response to the requests...The problem is, that you cannot (?) enforce the answer to go through the cable again. Instead it will be normally routed, which will end up in the local routing table and thus no packet will leave the box.