This looks like a tough routing issue

Applications IP Networking
1

Hi, The default gateway (172.17.220.1) cannot be pinged via the default interface bond0 (172.17.220.231), but can be pinged via interface eth1 (10.201.135.73).

# ping -I eth1 172.17.220.1
PING 172.17.220.1 (172.17.220.1) from 10.201.135.73 eth1: 56(84) bytes of data.
64 bytes from 172.17.220.1: icmp_seq=0 ttl=255 time=5.48 ms
64 bytes from 172.17.220.1: icmp_seq=1 ttl=255 time=0.523 ms

# ping -I bond0 172.17.220.1
PING 172.17.220.1 (172.17.220.1) from 172.17.220.231 bond0: 56(84) bytes of data.
From 172.17.220.231 icmp_seq=1 Destination Host Unreachable
From 172.17.220.231 icmp_seq=2 Destination Host Unreachable

Ping is supposed to work with the default interface bond0. Because the virtual interface bond0 can be pinged by systems in the same subnet, bond0 is working (right?). There must be a routing issue. Here is the interfaces info:

# cat /etc/sysconfig/network-scripts/ifcfg-bond0
DEVICE=bond0
IPADDR=172.17.220.231
NETMASK=255.255.252.0
GATEWAY=172.17.220.1
ONBOOT=yes
BOOTPROTO=static
USERCTL=no

# cat /etc/sysconfig/network-scripts/ifcfg-eth1
DEVICE=eth1
IPADDR=10.201.135.73
NETMASK=255.255.248.0
ONBOOT=yes
BOOTPROTO=static

OS is RHEL 4.9 (Linux). Please help. Thank you much in advance!

nz

2

Well, it is just the local net, bridging, switching, hubbing, broutering but not routing. Life there is just arp and go. See if the arp cache knows the mac of the gateway. Address Resolution Protocol (ARP) for Linux administrators Tournas Dimitrios

Can it ping itself by IP on both interfaces?

Can other hosts on either net ping it on both IP?

3 
ifconfig bond0

the interface needs to be configured. Is the bond up and running?

As you haven't shown any other interface configurations you should have this entry in one of them (i.e. /etc/sysconfig/network-scripts/ifcfg-eth0)

MASTER=bond0
SLAVE=yes
4

(1) It can ping itself by IP on both interfaces.
(2) Other hosts in the same subnet can ping bond0, but not eth1.

Thanks.

---------- Post updated at 09:07 AM ---------- Previous update was at 09:04 AM ----------

Here is the info:

# ifconfig bond0
bond0     Link encap:Ethernet  HWaddr 00:11:0A:59:B7:E8
          inet addr:172.17.220.231  Bcast:172.17.223.255  Mask:255.255.252.0
          inet6 addr: fe80::211:aff:fe59:b7e8/64 Scope:Link
          UP BROADCAST RUNNING MASTER MULTICAST  MTU:1500  Metric:1
          RX packets:35636 errors:0 dropped:0 overruns:0 frame:0
          TX packets:132845 errors:0 dropped:0 overruns:0 carrier:0
          collisions:0 txqueuelen:0
          RX bytes:3754935 (3.5 MiB)  TX bytes:7267244 (6.9 MiB)

# cat ifcfg-eth4
DEVICE=eth4
ONBOOT=yes
TYPE=Ethernet
USERCTL=no
MASTER=bond0
SLAVE=yes
BOOTPROTO=none

Thanks.

5

So, the 10. interface that can route is not pingable, and the 172. interface that cannot arp its gateway is pingable? Maybe we should get some traceroute's or something. The 10. might not ping because of routing setup, as 10. is only routable within an intranet. If the 172. pings from outside, it can obviously find a route, and we wonder what that is! Some tcpdump or whatever can tell us how far it gets on the lan. What does the arp cache look like for the 172. hosts? (ARP cache generally has a life of minutes.)