Author: Dude
Date: 01-Jun-2013
Version: B
Last updates:
23-Jul-2013, changes required due to forum upgrade
The following is provided in the hope that it might be useful. If you wish to use it for the purpose of reference, presentation or education, please include the link or information to this forum posting. Thank you!
What is Linux network interface bonding? Simply put, the Linux system provides a kernel driver, which can create a logical network device (bond) using 2 or more Network Interface Controllers (NIC) to accomplish link-aggregation and adapter fault-tolerance. For more information, please see the Deployment Guide at https://linux.oracle.com/documentation/
The channel bonding interface supports several modes of operation. In a typical active-backup set up for fault-tolerance, the system performs MII-based health monitoring of the network interface slaves (eth0, eth1) and assigns them according to bond options and interface availability. In active-backup mode (mode 1), only the current active network device is seen externally.
Linux bonding for fault-tolerance under Oracle VM VirtualBox does not seem to function as advertised. Searching the Internet for more information shows that other people have tried to set up Linux bonding using various virtualization products, virtual host network adapter and bonding options, but without success.
It appears the management of Media Access Control addresses (MAC) interferes with internal routing methods between the host and the virtual machine guest. Using the modinfo bonding command (fig. 1) shows a listing of available parameters for the bonding kernel module. The fail_over_mac parameter addresses the Mac address issue in VirtualBox.
Fig.1 |
# modinfo bonding filename: /lib/modules/2.6.39-400.17.1.el6uek.x86_64
/kernel/drivers/net/bonding/bonding.ko
parm: fail_over_mac:For active-backup,
do not set all slaves to the same MAC;
0 for none (default), 1 for active, 2 for follow (charp) |
Configuration Example
Virtualization Product: Oracle VM VirtualBox 4.2.10
VM Guest OS: Oracle Linux 6.4 (x64)
VM adapter settings: Bridged Adapter, Paravirtualized Network (virtio-net)
Adapter1 (eth0): 080027436617, en0, Ethernet 1
Adapter2 (eth1): 08002708F36A, en0, Ethernet 1
Note: If you wish to reset the network adapter names and MAC addresses used by the Oracle Linux guest system, erase the content of /etc/udev/rules.d/70-persistent-net.rules and restart the virtual machine.
1. Create or modify the following configuration files:
/etc/modprobe.d/bonding-eth0_eth1.conf |
alias bond0 bonding |
/etc/sysconfig/network-scripts/ifcfg-bond0 |
DEVICE=bond0 IPADDR=10.0.2.12 NETMASK=255.0.0.0 GATEWAY=10.0.0.138 DNS1=10.0.0.138 DNS2=8.8.8.8 ONBOOT=yes BOOTPROTO=none USERCTL=no BONDING_OPTS="mode=1 miimon=100 fail_over_mac=1" |
/etc/sysconfig/network-scripts/ifcfg-eth0 |
DEVICE=eth0 BOOTPROTO=none ONBOOT=yes MASTER=bond0 SLAVE=yes USERCTL=no |
/etc/sysconfig/network-scripts/ifcfg-eth1 |
DEVICE=eth1
BOOTPROTO=none ONBOOT=yes MASTER=bond0 SLAVE=yes USERCTL=no |
2. Restart the system or network services (root):
Fig. 2 |
# service network restart |
Configuration Test
Enter the following commands as user root to verify proper operation of NIC bonding:
Fig. 3 |
# tail /var/log/messages bond0: setting mode to active-backup (1).
bond0: Setting MII monitoring interval to 100.
bond0: Setting fail_over_mac to active (1).
bond0: Adding slave eth0.
bond0: making interface eth0 the new active one.
bond0: first active interface up!
bond0: enslaving eth0 as an active interface with an up link.
bond0: Adding slave eth1.
bond0: enslaving eth1 as a backup interface with an up link.
|
Fig. 4 |
# ifconfig bond0 Link encap:Ethernet HWaddr 08:00:27:43:66:17
inet addr:10.0.2.12 Bcast:10.255.255.255 Mask:255.0.0.0
inet6 addr: fe80::a00:27ff:fe43:6617/64 Scope:Link
UP BROADCAST RUNNING MASTER MULTICAST MTU:1500 Metric:1
eth0 Link encap:Ethernet HWaddr 08:00:27:43:66:17
UP BROADCAST RUNNING SLAVE MULTICAST MTU:1500 Metric:1
eth1 Link encap:Ethernet HWaddr 08:00:27:08:F3:6A
UP BROADCAST RUNNING SLAVE MULTICAST MTU:1500 Metric:1
|
Fig. 5 |
# netstat -rn Kernel IP routing table
Destination Gateway Genmask Flags MSS Window irtt Iface
0.0.0.0 10.0.0.138 0.0.0.0 UG 0 0 0 bond0
10.0.0.0 0.0.0.0 255.0.0.0 U 0 0 0 bond0
169.254.0.0 0.0.0.0 255.255.0.0 U 0 0 0 bond0
|
Fig. 6 |
# ifconfig eth0 down # tail /var/log/messages bond0: link status definitely down for interface eth0, disabling it
bond0: making interface eth1 the new active one. # ifconfig bond0 Link encap:Ethernet HWaddr 08:00:27:08:F3:6A
inet addr:10.0.2.12 Bcast:10.255.255.255 Mask:255.0.0.0
inet6 addr: fe80::a00:27ff:fe43:6617/64 Scope:Link
UP BROADCAST RUNNING MASTER MULTICAST MTU:1500 Metric:1
eth1 Link encap:Ethernet HWaddr 08:00:27:08:F3:6A
UP BROADCAST RUNNING SLAVE MULTICAST MTU:1500 Metric:1 |
Fig. 7 |
# ifconfig eth0 up # ifconfig eth1 down # tail /var/log/messages bond0: link status definitely up for interface eth0, 4294967295 Mbps full duplex. bond0: link status definitely down for interface eth1, disabling it bond0: making interface eth0 the new active one.
# ifconfig bond0 Link encap:Ethernet HWaddr 08:00:27:43:66:17
inet addr:10.0.2.12 Bcast:10.255.255.255 Mask:255.0.0.0
inet6 addr: fe80::a00:27ff:fe43:6617/64 Scope:Link
UP BROADCAST RUNNING MASTER MULTICAST MTU:1500 Metric:1
eth0 Link encap:Ethernet HWaddr 08:00:27:43:66:17
UP BROADCAST RUNNING SLAVE MULTICAST MTU:1500 Metric:1
|
During the above test, the client network connection from host 10.0.0.1 (ssh) was maintained without any noticeable interruption. As it can be seen from the syslog and ifconfig command output, the bond0 interface was automatically assigned either the MAC address of eth0 or eth1 and the fail-over worked as advertised.
Good luck!