Shared Ethernet Adapter failover

Shared Ethernet Adapter failover provides redundancy by configuring a backup Shared Ethernet Adapter on a different Virtual I/O Server logical partition that can be used if the primary Shared Ethernet Adapter fails. The network connectivity in the client logical partitions continues without disruption.

A Shared Ethernet Adapter is comprised of a physical adapter (or several physical adapters grouped under a Link Aggregation device) and one or more virtual Ethernet adapters. It can provide layer 2 connectivity to multiple client logical partitions through the virtual Ethernet adapters.

The Shared Ethernet Adapter failover configuration uses the priority value that is given to the virtual Ethernet adapters during their creation to determine which Shared Ethernet Adapter will serve as the primary and which will serve as the backup. The Shared Ethernet Adapter that has the virtual Ethernet configured with the numerically lesser priority value will be used preferentially as the primary adapter. For the purpose of communicating between themselves to determine when a failover should take place, Shared Ethernet Adapters in failover mode use a VLAN dedicated for such traffic, called the control channel. For this reason, a virtual Ethernet (created with a PVID that is unique on the system) must be specified as the control channel virtual Ethernet when each Shared Ethernet Adapter is created in failover mode. Using the control channel, the backup Shared Ethernet Adapter is notified when the primary adapter fails, and network traffic from the client logical partitions is sent over the backup adapter. If and when the primary Shared Ethernet Adapter recovers from its failure, it again begins actively bridging all network traffic.

A Shared Ethernet Adapter in failover mode might optionally have more than one trunk virtual Ethernet. In this case, all the virtual Ethernet adapters in a Shared Ethernet Adapter must have the same priority value. Also, the virtual Ethernet adapter used specifically for the control channel does not need to have the trunk adapter setting enabled. The virtual Ethernet adapters used for the control channel on each Shared Ethernet Adapter in failover mode must have an identical PVID value, and that PVID value must be unique in the system, so that no other virtual Ethernet adapters on the same system are using that PVID.

To ensure prompt recovery times, when you enable the Spanning Tree Protocol on the switch ports connected to the physical adapters of the Shared Ethernet Adapter, you can also enable the portfast option on those ports. The portfast option allows the switch to immediately forward packets on the port without first completing the Spanning Tree Protocol. (Spanning Tree Protocol blocks the port completely until it is finished.)

The Shared Ethernet Adapter is designed to prevent network loops. However, as an additional precaution, you can enable Bridge Protocol Data Unit (BPDU) Guard on the switch ports connected to the physical adapters of the Shared Ethernet Adapter. BPDU Guard detects looped Spanning Tree Protocol BPDU packets and shuts down the port. This helps prevent broadcast storms on the network. A broadcast storm is a situation where one message that is broadcast across a network results in multiple responses. Each response generates more responses, causing excessive transmission of broadcast messages. Severe broadcast storms can block all other network traffic, but they can usually be prevented by carefully configuring a network to block illegal broadcast messages.

Note: When the Shared Ethernet Adapter is using GARP VLAN Registration Protocol (GVRP), it generates BPDU packets, which causes BPDU Guard to shut down the port unnecessarily. Therefore, when the Shared Ethernet Adapter is using GVRP, do not enable BPDU Guard.

For information about how to enable the Spanning Tree Protocol, the portfast option, and BPDU Guard on the ports, see the documentation provided with the switch.

Last updated: Thu, October 15, 2020