Troubleshooting -- Bonding Problems on Red Hat Enterprise Linux

The following KCS Solutions address some of the more common questions related to bonding on Red Hat Enterprise Linux.

Contributors

• Kenji Fujii  (This content is not included.kfujii@redhat.com)
• Takao Omichi (This content is not included.tomichi@redhat.com)
• Marc Milgram (This content is not included.mmilgram@redhat.com)
• Flavio Leitner (This content is not included.fleitner@redhat.com)
• David Powles (This content is not included.dpowles@redhat.com)
• Red Hat Bonding Group (those who have contributed bonding-related information to the Knowledgebase)

Bonding configuration

To create a bonding configuration which is optimized for your environment and avoids known issues, Red Hat Customer Portal Labs provides a This content is not included.Network Bonding Configuration app

Bonding Troubleshooting

• A bonding's primary setting in /etc/sysconfig/network-scripts/ifcfg-bond0 is not effective
• After creating mode 5 or 6 bonds in Red Hat Enterprise Linux 4, the bonded network interfaces may refuse to shutdown properly causing system hangs when power cycling
• Bonding device on xen kernel: does not send arp request when cable is pulled out
• Bonding does not work after updating to Red Hat Enterprise Linux 5.4
• Bonding driver uses 100% of the CPU and causes a soft lockup to occur when using mode 4 (802.3ad)
• Clean up arp_ip_target with ifdown-eth network script for avoiding write error message
• Even though bonding has been configured for mode 1 [active backup], it still shows as Mode 0 [Round-Robin]
• Failed interface that is part of mode 1 bonding attached to xen bridge does not become active when it will works again
• Messages: "bond0:IPv6 duplicate address detected!" errors are seen in dmesg after a boot
• Messages: "Debug: sleeping function called from invalid context at mm/slab.c:206"
• Round-robin bonding (mode 0) with virtualization breaks networking for guests
• When a bonding device is down, a warning message (bonding: bond0: Warning: the permanent HWaddr of eth0 - xx:xx:xx:xx:xx:xx - is still in use by bond0 ...) is displayed
• This content is not included.When one slave of a mode 0 bonding device is removed, the server stops responding to pings from other hosts in VLANs working over the bond device
• Where can I find information about the status of my bonded NIC interfaces?
• Why do bonding device unload operations sometimes hang on IBM System p?
• Why does /proc/net/bond0/bonding show that the network interface card's link is up when the cable is unplugged within a bonding environment?
• Why is my bonding mode not being set correctly in Red Hat Enterprise Linux 4?
• Xen server crashes when used with network bonding modes 5(balance-tlb) or 6(balance-alb)