Configuring the cluster

Types of interfaces used by takeover

Interface requirements for takeover

In a cluster, you must configure each filer so that if one filer fails, the partner node can assume its identity. To ensure a successful takeover, your cluster must meet the following requirements:

Clustered filers must be attached to the same networks and subnetworks.
Each local interface on a filer must have one local node address and one partner address, and for only one type of interface.

Interface types

There are two kinds of interface configurations you can have for your cluster. The System Administrator's Guide provides more information about the types of NICs that support dual-MAC addresses. The System Administrator's Guide also describes how to configure vif.

Interface type	Description
Single, stand-by NICs	Each filer has a pair of dedicated interfaces with a local IP address. One of the NICs is a standby interface with the partner's IP address. This configuration is used for adapters that support only a single MAC (media access control) address.
Dual MAC capability	Each filer has a single NIC that supports a pair of dual MAC interfaces, which duplicate each other. These NICs can be one of the following: Single-port 10Base-T/100Base-TX Ethernet Quad-port 10Base-T/100Base-TX Ethernet FDDI GbE

Interface type

Description

Single, stand-by NICs

Each filer has a pair of dedicated interfaces with a local IP address. One of the NICs is a standby interface with the partner's IP address.

This configuration is used for adapters that support only a single MAC (media access control) address.

Dual MAC capability

Each filer has a single NIC that supports a pair of dual MAC interfaces, which duplicate each other. These NICs can be one of the following:

Single-port 10Base-T/100Base-TX Ethernet
Quad-port 10Base-T/100Base-TX Ethernet
FDDI
GbE

Network interface modes

Each network interface in a cluster can act in one of two modes, as follows:

Normal mode, which is the state where each filer manages its own NICs.
Takeover mode, which is the state where the partner node takes over control of the failed filer's NICs.

Normal mode and takeover mode are described in the following table.

Interface type	Normal mode	Takeover mode
Dedicated	Supports the identity of the local node	Idle
Standby	Idle	Supports the identity of the partner node by assuming the partner's IP address
Shared	Supports the identity of the local node	Supports the identity of The local node by retaining the local IP address The failed filer by assuming the partner's IP address

Interface type

Normal mode

Takeover mode

Dedicated

Supports the identity of the local node

Idle

Standby

Idle

Supports the identity of the partner node by assuming the partner's IP address

Shared

Supports the identity of the local node

Supports the identity of

The local node by retaining the local IP address
The failed filer by assuming the partner's IP address