bonding problem

freeindy · 05-22-2008, 09:38 AM

Hi,

I am running WindRiver Linux on a pentium4 system. It seems to me that it's has the same basic architecture as red hat so the settings can be set as the same (although I might be wrong).

Now, I've run into problems bonding two NIC's. This is what I've done:

In /etc/sysconfig/networks-scripts/ifcfg-bond0:

Code:

DEVICE=bond0
IPADDR=10.90.90.6
NETWORK=10.90.90.0
NETMASK=255.255.255.0
BROADCAST=10.90.90.255
ONBOOT=yes
BOOTPROTO=none
USERCTL=no

In /etc/sysconfig/networks-scripts/ifcfg-eth2:

Code:

DEVICE=eth2
USERCTL=no
ONBOOT=yes
MASTER=bond0
SLAVE=yes
BOOTPROTO=none

In /etc/sysconfig/networks-scripts/ifcfg-eth3:

Code:

DEVICE=eth3
USERCTL=no
ONBOOT=yes
MASTER=bond0
SLAVE=yes
BOOTPROTO=none

In /etc/modeprobe.conf AND /etc/modules.conf (None of them were existing so I created them both):

Code:

alias bond0 bonding
options bond0 miimon=100 mode=1

With all this settings, I can't ping 10.90.90.9 (I have the same setting on the this machine as well) and 10.90.90.20 which is in the same network. I'm suspecting the bonding is not working.

ifconfig says:

Code:

bond0     Link encap:Ethernet  HWaddr 00:00:00:00:00:00  
          inet addr:10.90.90.6  Bcast:10.90.90.255  Mask:255.255.255.0
          inet6 addr: fe80::200:ff:fe00:0/64 Scope:Link
          UP BROADCAST RUNNING MASTER MULTICAST  MTU:1500  Metric:1
          RX packets:0 errors:0 dropped:0 overruns:0 frame:0
          TX packets:0 errors:0 dropped:0 overruns:0 carrier:0
          collisions:0 txqueuelen:0 
          RX bytes:0 (0.0 b)  TX bytes:0 (0.0 b)

eth2      Link encap:Ethernet  HWaddr 00:A0:A5:5D:AD:82  
          inet6 addr: fe80::2a0:a5ff:fe5d:ad82/64 Scope:Link
          UP BROADCAST RUNNING MULTICAST  MTU:1500  Metric:1
          RX packets:0 errors:0 dropped:0 overruns:0 frame:0
          TX packets:3 errors:0 dropped:0 overruns:0 carrier:0
          collisions:0 txqueuelen:1000 
          RX bytes:0 (0.0 b)  TX bytes:742 (742.0 b)
          Base address:0x5c00 Memory:f43e0000-f4400000 

eth3      Link encap:Ethernet  HWaddr 00:A0:A5:5D:AD:83  
          inet6 addr: fe80::2a0:a5ff:fe5d:ad83/64 Scope:Link
          UP BROADCAST RUNNING MULTICAST  MTU:1500  Metric:1
          RX packets:0 errors:0 dropped:0 overruns:0 frame:0
          TX packets:3 errors:0 dropped:0 overruns:0 carrier:0
          collisions:0 txqueuelen:1000 
          RX bytes:0 (0.0 b)  TX bytes:742 (742.0 b)
          Base address:0x5800 Memory:f43c0000-f43e0000

route says:

Code:

Destination     Gateway         Genmask         Flags Metric Ref    Use Iface
192.168.0.0     *               255.255.255.0   U     0      0        0 eth6
10.90.90.0      *               255.255.255.0   U     0      0        0 bond0
default         10.90.90.101    0.0.0.0         UG    0      0        0 bond0

I had a setting in a diffent system and it works:

Code:

auto bond0
iface bond0 inet static
                address 10.90.90.6
                network 10.90.90.0
                netmask 255.255.255.0
                broadcast 10.90.90.255
                up ifenslave bond0 eth2 eth3
                down ifenslave -d bond0 eth0 eth3
                up route add default gw 10.90.90.101 bond0
                down route del default gw 10.90.90.101 bond

Any ideas ?

Thanks,
Indy

born4linux · 05-23-2008, 04:09 AM

something definitely wrong there - the HW address of bond0 should
be one of the physical NICs HW address (in failover mode, bond0 and
the rest of the slave NICs will show the same HW address).

try reloading the bonding module with the primary device included.

also, run ethtool on your network devices (ethX).

freeindy · 05-23-2008, 08:38 AM

born4linux,

How do you mean 'reloading the bonding module with the primary device included.' ?

Can you give an example?

Indy

born4linux · 05-24-2008, 11:29 AM

1. stop your network
2. unload the bonding module:

rmmod bonding

3. load it:

modprobe bonding miimon=100 mode=1 primary=ethX

or just edit /etc/modprobe.conf and add

primary=eth3

and reboot.

freeindy · 05-26-2008, 04:32 AM

ok. I found the problem. Could be specific to WindRiver, I don't know.

The NIC's are not slaved. I had to use ifenslave specifically to make it work. I presume that it should make the NIC's enslaved by the cfg files but it didn't. And the reason comes down to that the the utility ifenslave is not located in /sbin/ where the network looks for. so simply by putting an soft link in the /sbin/ directory, it works.

Thanks for your help born4linux. Your HW address problem helped. It didn't even look at it from the beginning.

Indy