Bonded interface stops working

Solution In Progress - Updated -

Issue

  • Mode 4 LACP bonding goes down when physical interfaces on different 2-port NICs are enslaved to the one bond, and the system boots or network service is restarted.
  • Behaviour is correct on boot and when restarting network service, only if the we use eth0 and eth1 (both on same PCI card), or eth2 and eth3 (both on same PCI card).
  • Any combination of physical interfaces where the bond slaves are on different PCI NICs causes the bond0 not to be pingable on boot. The /proc/net/bonding/bond0 file shows bond0 up, and both physical interfaces up.
  • The ifconfig commands shows both physical interfaces sending and receiving traffic, but bond0 interface is not pingable. An ifdown, ifup on either physical interface fixes this (i.e. bond0 becomes pingable). But a reboot or restarting the network service breaks it again.

Environment

  • Red Hat Enterprise Linux 6.5
  • HP ProLiant DL560 Gen8
  • 2x Broadcom Corporation NetXtreme II BCM57810 10 Gigabit Ethernet (rev 10) network interfaces:
    • 1x HP Ethernet 10Gb 2-port 530SFP+ Adapter (PN: 647579-001) with firmware version bc 7.0.49
    • 1x HP Ethernet 10Gb 2-port 530FLR-SFP+ Adapter (PN: 652501-001) with firmware version bc 7.8.24

Subscriber exclusive content

A Red Hat subscription provides unlimited access to our knowledgebase, tools, and much more.

Current Customers and Partners

Log in for full access

Log In

New to Red Hat?

Learn more about Red Hat subscriptions

Using a Red Hat product through a public cloud?

How to access this content