HP A7503, A7510, A7506-V, A7506 Configuration Manual

Download

HP A7500 Switch Series

IRF Configuration Guide

Part number: 5998-1865

Software version: Release 6626 and later

Document version: 6W101-20111130

Legal and notice information

No part of this documentation may be reproduced or transmitted in any form or by any means without prior written consent of Hewlett-Packard Development Company, L.P.

The information contained herein is subject to change without notice.

HEWLETT-PACKARD COMPANY MAKES NO WARRANTY OF ANY KIND WITH REGARD TO THIS MATERIAL, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE. Hewlett-Packard shall not be liable for errors contained herein or for incidental or consequential damages in connection with the furnishing, performance, or use of this material.

The only warranties for HP products and services are set forth in the express warranty statements accompanying such products and services. Nothing herein should be construed as constituting an additional warranty. HP shall not be liable for technical or editorial errors or omissions contained herein.

Hardware compatibility ···················································································································································· 1

Switch model compatibility ····································································································································· 1 Card compatibility ···················································································································································· 1 Hardware compatibility for an IRF fabric with three or four members ······························································· 1

IRF overview ······································································································································································· 1

Benefits ······································································································································································ 1

Application scenario ················································································································································ 2 Basic concepts ··································································································································································· 3 Establishment, operation, and maintenance of an IRF fabric ······················································································· 6

Connecting the IRF member switches ····················································································································· 7

Topology collection ·················································································································································· 8

Master election ························································································································································· 9

IRF fabric management and maintenance ············································································································· 9 IRF fabric configuration task list ···································································································································· 10 Pre-configuring an IRF member switch in standalone mode ······················································································ 15

Setting a member ID for the switch······················································································································ 15

Specifying a priority for the switch ······················································································································ 15

Binding physical ports to IRF ports ······················································································································ 16 Saving configuration to the startup configuration file ································································································ 17 Switching operating mode ············································································································································ 17

IRF modes ······························································································································································· 17

Configuration file auto-conversion ······················································································································· 17

Switching the operating mode to IRF mode········································································································ 17 Accessing an IRF fabric ················································································································································· 18

Accessing the active MPU of an IRF fabric ········································································································· 18

Accessing a standby MPU of an IRF fabric ········································································································ 18 Configuring IRF member switches in IRF mode ··········································································································· 19

Specifying an IRF domain ID for an IRF fabric ··································································································· 19

Modifying a member ID for a switch ·················································································································· 20

Specifying a priority for a member switch ········································································································· 21

Configuring IRF ports ············································································································································ 21

Enabling auto reboot for IRF fabric merge ········································································································· 22

Configuring a description for a member switch ································································································ 23

Configuring load sharing criteria for IRF links ··································································································· 23

Specifying the preservation time of bridge MAC address ················································································ 24

Enabling automatic system software updating ··································································································· 25

Setting the IRF link down report delay ················································································································ 26

Configuring MAD detection ································································································································· 26 IRF configuration fast recovery ····································································································································· 32

Configuration prerequisites ·································································································································· 32

Configuration procedures ····································································································································· 33 Displaying and maintaining an IRF fabric ··················································································································· 33 IRF fabric configuration examples ································································································································ 34

LACP MAD detection-enabled IRF configuration example (with two member switches) ······························· 34

BFD MAD detection-enabled IRF configuration example (with two member switches) ·································· 38

LACP MAD detection-enabled IRF configuration example (with four member switches) ······························· 42

BFD MAD detection-enabled IRF configuration example (with four member switches) ································· 48

Switching the operating mode of IRF member switches from IRF to standalone ············································ 53

Index ··········································································································································································· 56

IRF configuration

NOTE:

Switches running Release 6635 or a later release support establishing an IRF fabric with up to four members. Switches running a release prior to Release 6635 support establishin two members.

Hardware compatibility

Switch model compatibility

• Among the A7500 Switch Series, these models support IRF: A7503, A7506, A7506-V, and A7510.

Only switches of the same models can form an IRF.

• The A7510 model supports establishing an IRF fabric with up to two members.

• Two A7510 switches can form an IRF fabric only when they use no LSQ1SRP1CB MPU.

an IRF fabric with up to

Card compatibility

The MPUs in the same IRF member switch must be of the same model. It is a good practice to use the same model of MPUs in all IRF member switches.

Hardware compatibility for an IRF fabric with three or four members

The following cards cannot be used in an IRF fabric with three or four members:

• SA cards

• LSQ1SRP1CB MPUs

IRF overview

The HP Intelligent Resilient Framework (IRF) technology creates a large IRF fabric from multiple switches to provide data center class availability and scalability. IRF virtualization technology offers processing power, interaction, unified management, and uninterrupted maintenance of multiple switches.

Benefits

IRF delivers the following benefits:

• Simplified topology and streamlined management. An IRF fabric appears as one node on the

network. You can log in at any member switch to manage all members in the IRF fabric.

• High availability and reliability. One member switch works as the master to manage and maintain

the entire IRF fabric, and all other member switches process services and provide backup. If the master fails, all other member switches elect a new master among them to prevent service

interruption. You can perform link aggregation not only for IRF links but also for physical links between the IRF fabric and its upper or lower layer devices for link redundancy.

• Network scalability and resiliency. You can increase ports, network bandwidth, and processing

capability of an IRF fabric simply by adding member switches. Each member switch has its own CPU and can independently process and forward protocol packets.

Application scenario

Figure 1 shows an IRF fabric that comprises two switches, which appear as a single node to the upper

and lower layer devices.

Figure 1 IRF application scenario

IP network

IRF

Master

IP network

IRF link

Slave

Equal to

Basic concepts

Figure 2 IRF implementation schematic diagram (with two member switches)

Device A and Device B in Figure 2 form an IRF fabric, which has four MPUs (one active and three standby) and two interface cards. The IRF fabric manages both the physical and software resources of Device A and Device B.

An IRF fabric with four member switches is larger in scale: it has eight MPUs (one active and seven standby) and four interface cards, as shown in Figure 3.

Figure 3 IRF implementation schematic diagram (with four member switches)

Device A

(MemberID=1)

Active MPU of the

member

Standby MPU of the

member

Service

interface

Physical IRF

Device C

(MemberID=3)

Active MPU of the

member

Standby MPU of the

member

Service

interface

Physical IRF

IRF

IRF-Port1

IRF-Port2

port

IRF-Port1

IRF-Port2 IRF-Port1

port

IRF link

After an IRF

is formed.

Suppose

Device A is

the master

Master

(MemberID=1)

Active MPU of the IRF

Standby MPU of the IRF

(MemberID=2)

Standby MPU of the IRF Standby MPU of the IRF

IRF-Port2

IRF-Port1

IRF-Port2

Slave

Device B

(MemberID=2)

Active MPU of the

member

Standby MPU of the

member

Physical IRF

port

Device D

(MemberID=4)

Active MPU of the

member

Standby MPU of the

member

Physical IRF

port

Service

interface

Service

interface

CAUTION:

In an IRF fabric, each slave switch must have at least one MPU to work normally.

This section uses Figure 2 to explain the concepts that you might encounter when working with IRF.

Operating mode

A switch can operate in one of the following two modes:

• Standalone mode—The switch cannot form an IRF fabric with other switches.

• IRF mode—The switch can connect with other switches to form an IRF fabric.

You can change the operating mode of a switch at the command line interface (CLI).

Slave

(MemberID=3)

Standby MPU of the IRF

Standby MPU of the IRF Standby MPU of the IRF

Slave

(MemberID=4)

Standby MPU of the IRF

IRF member switch roles

IRF uses two member switch roles: master and slave.

When switches form an IRF fabric, they elect a master to manage the IRF fabric, and all other switches back up the master. When the master switch fails, the other switches automatically elect a new master from among them to avoid service interruption. For more information about master election, see “Master

electi

on.”

Active MPU of a member switch

Each switch has one active MPU for device supervision.

After a switch joins an IRF fabric, its active MPU has the following responsibilities:

• Manages the local switch, for example, synchronizing configuration between the active MPU and

the standby MPU, processing protocol packets, and creating and maintaining route entries.

• Handles IRF related events, such as role election and topology collection.

Standby MPUs of a member switch

A switch can optionally have a standby MPU as the backup of its active MPU.

Active MPU of an IRF fabric

The active MPU of an IRF fabric is the active MPU of the master switch and manages the entire IRF fabric.

Standby MPUs of an IRF fabric

A standby MPU of an IRF fabric is a backup for the active MPU of the IRF fabric. An MPU of a member switch is a standby MPU of the IRF fabric unless it is the active MPU of the IRF fabric.

IRF port

An IRF port is a logical interface for the internal connection between IRF member switches. Each IRF member switch has two IRF ports: IRF-port 1 and IRF-port 2. An IRF port is activated when you bind a physical port to it.

NOTE:

In standalone mode, the IRF ports are named IRF-port 1 and IRF-port 2. In IRF mode, the IRF ports are named IRF-port and IRF-port 2 are used.

Physical IRF port

Physical IRF ports are physical ports bound to an IRF port. You can configure a 10 GE optical port as a physical IRF port.

By default, a 10 GE optical port functions as a common service port and forwards data traffic. When bound to an IRF port, it acts as an IRF physical port and forwards packets among member switches. Packets that can be forwarded include IRF-related negotiation packets, and data packets that need to be forwarded cross-switches.

IRF partition

/1 and IRF-port n/2, where n is the member ID of the switch. In this manual, IRF-port 1

IRF partition occurs when an IRF fabric splits into two or more IRF fabrics because of IRF link failures, as shown in Figure 4. T

he partitioned IRF fabrics operate with the same IP address and cause routing and

forwarding problems on the network.

Figure 4 IRF partition

IRF merge

IRF merge occurs when two partitioned IRF fabrics re-unite or when you configure and connect two independent IRF fabrics to be one IRF fabric, as shown in Figure 5.

Figure 5 IRF merge

Member priority

Member priority determines the role of a member switch duri ng the role election process. A member with a higher priority is more likely to be a master.

The priority of a switch defaults to 1. You can modify the priority at the CLI.

Establishment, operation, and maintenance of an IRF fabric

IRF fabric management involves these stages: Connecting the IRF member switches, IRF topology

An IRF fabric typically adopts daisy chain connection or ring connection, as shown in Figure 7.

• A daisy chain connection is mainly used in a network where member switches are distributedly

located.

• A rin g co nne ction is more re liab le t han the dai sy chain connection. In a daisy chained IRF fabric, the

failure of one link can cause the IRF fabric to partition into two independent IRF fabrics; where the failure of a link in a ring connection result in a daisy chain connection, not affecting IRF services.

Figure 6 IRF connections

NOTE:

The ring connection is supported only when the IRF fabric has three or four member switches.

Topology collection, Master election, and IRF fabric management and maintenance.

Connecting the IRF member switches

Connection medium

To establish an IRF fabric, physically connect the physical IRF ports of member switches.

The switches use 10 GE optical ports as physical IRF ports. You can connect physical IRF ports with fibers.

Connection requirements

As shown in Figure 7, connect the physical ports bound to IRF-Port1 on one switch to the physical ports bound to the IRF-Port2 on its neighbor switch.

Figure 7 IRF fabric physical connection

NOTE:

• An IRF port can be bound to a maximum of 8 physical ports to increase the bandwidth and reliability of

the IRF port.

• The links between two IRF members must be fibers, and no intermediate devices are allowed.

IRF topology

An IRF fabric typically adopts daisy chain connection or ring connection, as shown in Figure 7.

• A daisy chain connection is mainly used in a network where member switches are distributedly

• A rin g co nne ction is more re liab le t han the dai sy chain connection. In a daisy chained IRF fabric, the

Figure 8 IRF connections

located.

NOTE:

The ring connection is supported only when the IRF fabric has three or four member switches.

Topology collection

Each member exchanges IRF hello packets with neighbors to collect the topology data, including IRF port connection states, member IDs, priorities, and bridge MAC addresses.

Each member is managed by its active MPU, which records its known topology information locally. At the startup of a member switch, the active MPU of the member switch records topology information of the member switch. When an IRF port of the member switch is up, the active MPU of the switch performs the following operations:

1. Periodically sends its known topology information from this port.

2. When receiving the topology information from the directly connected neighbor, it updates the

local topology information.

3. If a standby MPU is available on the member switch, the active MPU synchronizes its recorded

topology information to the standby MPU to ensure consistent topology information on both boards.

After all member switches have obtained topology information (known as topology convergence), the IRF fabric enters the role election stage.

Master election

Master election is held each time the topology changes, for example, when the IRF fabric is established, a new member switch is plugged in, the master switch fails or is removed, or the partitioned IRF fabrics merge.

The master is elected based on the following rules in descending order:

• The current master, even if a new member has a higher priority. (When an IRF fabric is being formed,

all member switches consider themselves as the master, and this rule is skipped)

• The switch with higher priority.

• The switch with the longest system up-time. (The member switches exchange system up-time in the

IRF hello packets.)

• The switch with the lowest bridge MAC address.

The IRF fabric is formed on election of the master.

NOTE:

• During an IRF merge, the switches of the IRF fabric that fails the master election must reboot to re -join the

IRF fabric that wins the election. The reboot can be automatically performed or must be manually performed depends on the switch model.

• After a master election, all slave member switches initialize and reboot with the configuration on the

master, and their original configuration, even if has been saved, will be lost.

IRF fabric management and maintenance

After the IRF fabric is established, you can access the master from any member switch to manage all the resources of the member switches.

Member ID

An IRF fabric uses member IDs to uniquely identify and manage its members. Member IDs are used in interface numbering and file management:

• In interface numbering—Assume an interface on the switch that operates in standalone mode was

named GigabitEthernet 3/0/1. After the switch joined an IRF fabric, it receives a member ID of 2, and the name of the interface changes to GigabitEthernet 2/3/0/1.

• In file management—When the switch operates in standalone mode, the path of a file is, for

example, slot1#flash:/. After the switch joins an IRF fabric, the path changes to chassisA#slot1#flash:/, where A indicates the member ID of the switch in the fabric.

To guarantee the operation of the IRF fabric, you must assign each member switch a unique member ID.

NOTE:

If the active MPU and standby MPU of a member switch keep different member IDs of the switch, the member ID kept by the active MPU is applied when the switch starts up. If the switch with the member ID of 2 has only one active MPU, after you plug in a standby MPU that keeps a member ID of 3, the membe ID of the switch is still 2 and the member ID kept on the standby MPU is synchronized to 2.

IRF fabric topology maintenance

As soon as a member switch is down or an IRF link is down, its neighbor switches broadcast the leaving of the switch to other members. When a member switch receives the leave message, it looks up its IRF

topology database to determine whether the leaving switch is the master. If yes, the member switch starts

a master election and updates its IRF topology database. If the leaving switch is not a master, the member switch directly updates its IRF topology database.

NOTE:

n IRF port goes down only when all its physical IRF ports are down.

IRF multi-active detection

An IRF link failure causes an IRF fabric to split in two IRF fabrics operating with the same Layer 3 configurations, such as the same IP address. To avoid IP address collision and network problems, IRF uses multi-active detection (MAD) mechanisms to detect the presence of multiple identical IRF fabrics and handle collisions. MAD provides the following functions:

1. Detection

MAD detects active IRF fabrics with the same Layer 3 global configuration by extending the Link Aggregation Control Protocol (LACP) or the Bidirectional Forwarding Detection (BFD) protocol. For more information, see “Configuring MAD detection.”

2. Collision handling

If multiple identical active IRF fabrics are detected, only the IRF fabric that has the lowest master ID can operate in active state and forward traffic normally. MAD sets all other IRF fabrics in recovery (disabled) state and shuts down all physical ports but the console and physical IRF ports and other ports you have specified with the mad exclude interface command.

3. Failure recovery

An IRF link failure triggers IRF fabric partition and causes multi-active collision. In this case, repair the failed IRF link to make the collided IRF fabrics merge into one and recover the failure. If the IRF fabric in recovery state fails before the failure is recovered, repair both the failed IRF fabric and the failed IRF link, and then the collided IRF fabrics can merge into one and the failure is recovered. If the IRF fabric in the active state fails before the failure is recovered, enable the IRF fabric in recovery state at the CLI to make it take over the active IRF fabric and protect the services from being affected. Then, recover the MAD failure.

NOTE:

• For information about LACP, see

• For information about BFD, see

Layer 2 — LAN Switching Configuration Guide

High Availability Configuration Guide

IRF fabric configuration task list

IRF port, member ID, and member priority are three key elements for an IRF fabric. Configure them in one of the following methods:

• Pre-configure the three parameters when the switch operates in standalone mode. These

configurations do not affect the operation of the switch and take effect after the switch switches to IRF mode. Adopt this method to configure the IRF fabric before establishing it. To enable switches operating in standalone mode to form an IRF fabric, you must pre-configure their member IDs and switch their operating modes. If you configure the priority for a switch as the greatest value when the switch operates in standalone mode, this switch can win the role election and become the master after an IRF fabric is formed for the first time. If you configure IRF ports for member switches that operate in standalone mode, they can directly form an IRF fabric with other switches after their

operating mode is switched to IRF. Pre-configurations enable member switches to reboot only once to establish an IRF fabric.

• Re-configure the three parameters when the switch operates in IRF mode. Adopt this method when

you change the current configuration. For example, change the member ID of a switch to a specified value (note that the modification of the member ID may make some member ID-related configurations invalid); modify the priority of a member switch to make it being elected as the master in the next role election; modify the binding between an IRF port and physical IRF ports (such as deleting a binding or adding a new binding), and the configuration of IRF ports may affect the operation of the switch (for example, causing IRF fabric partition, or IRF fabric merge).

Although you can configure the three parameters in either of the two ways, they have different effects. HP recommends this IRF fabric establishment procedure:

1. Plan the network, and decide the master switch, IRF member IDs, and physical IRF ports.

2. Pre-configure the switches in standalone mode, including their IRF ports, member IDs, and member

priorities.

3. Save the configurations to the startup configuration file on each member switches.

4. Connect physical IRF ports and ensure that they are interoperable.

5. Switch the operating mode of the switches to IRF. The switches reboot to form an IRF fabric.

6. Log in to the IRF fabric.

7. Configure and manage the IRF fabric, for example, change an IRF port binding when the physical

IRF port in the binding fails).

Complete the following tasks to configure an IRF fabric:

Task

Pre-configuring an IRF member

Setting a member ID for the switch

Specifying a priority for the switch

Task

switch in standalone mode

Binding physical ports to IRF ports

Saving configuration to the startup configuration file

Connecting cables and make sure that the physical IRF ports are interconnected.

Switching operating mode

Accessing an IRF fabric

Configuring IRF member switches in IRF mode

Accessing the active MPU of an IRF fabric

Accessing a standby MPU of an IRF fabric

Specifying an IRF domain ID for an IRF fabric

Modifying a member ID for a switch

Specifying a priority for a member switch

Configuring IRF ports

Enabling auto reboot for IRF fabric merge

Task

Configuring a description for a member switch

Configuring load sharing criteria for IRF links

Task

Configuring load sharing criteria for IRF links

When an IRF port is bound with two or more physical IRF ports, there are multiple IRF links among IRF fabri

Determine how traffic is load-shared among members by configuring load sharing crite The criteria can be information carried in packets, such as IP addresses, MAC addr combination of them. The criteria can also be packet types, such as Layer 2 and Layer

Configure global (in system view) or port-specific (in IRF port view) load sharing criteria

• If you confi

ure the load sharing criteria in system view, the configuration is effectiv

all IRF ports.

• If you confi

ure load sharing criteria in IRF port view, the configuration is effective fo

IRF port.

• An IRF port preferentially uses the port-specific load sharing criteria. If no port-speci

criteria is available, it uses the global load sharing criteria.

NOTE:

The load sharing criterion or criteria you configured in the same view overwrite the old on

If you configure a load sharing criterion not supported by the switch, you will be prompted

does not support the criterion.

Before configuring the load sharing criteria, bind IRF ports to correspondin

load sharing criterion configuration fails.

Configuring global load sharing criteria

Follow these steps to configure the global IRF link load sharing criteria:

To do…

Enter system view system-view —

Configure the global IRF link load sharing criteria

physical IRF p

Use the command… Remarks

Optional

By default, the g

irf-port load-sharing mode { destination-ip | destination-mac | ingress-port | source-ip | source-mac } *

load sharing crit or destination M for Layer 2 pack destination IP ad Layer 3 packets, destination port Layer 4 packets.

Configuring port-specific load sharing criteria

Follow these steps to configure the port-specific load sharing criteria:

To do…

Enter system view system-view —

Enter IRF port view irf-port member-id/port-number —

Configure the port-specific load sharing criteria

Use the command… Remarks

irf-port load-sharing mode { destination-ip | destination-mac | ingress-port | source-ip | source-mac } *

Optional

By default, the p sharing criteria destination MA Layer 2 packets, destination IP ad Layer 3 packets, destination port Layer 4 packets.

Task

Enabling automatic system software updating

Setting the IRF link down report delay

Configuring MAD detection

IRF configuration fast recovery

Pre-configuring an IRF member switch in standalone mode

You can configure the IRF ports, member ID, and member priority for the switch when it is operating in standalone mode. These configurations take effect when the operating mode of the switch changes to IRF.

Setting a member ID for the switch

• A switch by default operates in standalone mode without an IRF member ID. You must assign it an

IRF member ID before you can set it in IRF mode. You can execute the display irf configuration command and check the MemberID field. If the switch has no IRF member ID, the field displays two hyphens (--).

• To avoid member ID collision with other members when the switch is added into an IRF fabric, plan

the member ID of this device when it operates in standalone mode.

Follow these steps to set a member ID for the switch:

To do… Use the command…

Enter system view system-view —

Set a member ID for the switch when it operates in standalone mode

NOTE:

• The member ID of an A7510 switch can only be 1 or 3.

• The member ID of a switch installed with an LSQ1SRP1CB card can only be 1 or 2.

• For an A7506 or A7506-V switch using the member ID 4, if it runs a release prior to Release 6635, the

last two slots cannot be used; if it runs Release 6635 or later, the last slot cannot be used.

irf member member-id

Specifying a priority for the switch

Follow these steps to specify a priority for the switch:

To do… Use the

command…

Remarks

Required

By default, no member ID is set for the switch.

Remarks

Enter system view system-view —

Specify a priority for the switch when it operates in standalone mode

irf priority priority

Optional

The priority of the switch defaults to 1.

Binding physical ports to IRF ports

NOTE:

The A7500 switch uses 10 GE optical ports, which are on the main processing unit (MPU), or on SC, SD, or EB card, as physical IRF ports to perform IRF connection. For more information about LPUs providin GE optical ports, see

To establish IRF connection, you must assign the physical ports that connect IRF member switches to IRF ports. An IRF port can be bound to either one or multiple physical IRF ports.

Follow these steps to bind physical ports to IRF ports:

To do… Use the command…

Enter system view system-view —

Create an IRF port and enter IRF port view when the switch is operating in standalone mode

Bind a physical IRF port to the IRF port

HP A7500 Switch Series Installation Guide

irf-port port-number

port group interface interface-type interface-number [ mode { enhanced | normal } ]

Remarks

Required

By default, no IRF port is created on the switch.

If the IRF port has been created, this command enters IRF port view.

Required

By default, no physical ports are bound to any IRF port.

CAUTION:

• A physical IRF port can be bound to an IRF port only when it operates in Layer 2 (bridge) mode. For

more information about port link mode, see

Layer 2 — LAN Switching Configuration Guide.

• To realize IRF link redundancy and load sharing and increase the bandwidth and reliability of IRF links,

bind one IRF port to multiple physical IRF ports by repeatedly executing the port group interface command. The maximum number of physical IRF ports that can be bound to an IRF port is eight. When the upper limit is reached, the execution of this command fails.

• The switch allows you to bind multiple physical IRF ports on different LPUs to one IRF port.

• The mode keyword in the port group command specifies the workin

mode of a physical IRF port. By default, the working mode of a physical IRF port is normal. An SC card does not support configuration of the working mode of physical IRF ports as enhanced.

• Physical IRF ports that connect two member switches of an IRF fabric must be confi

ured to work in the

same mode.

• To use the MPLS L2VPN or virtual private LAN service (VPLS) function in an IRF fabric, confi

ure the 10 GE optical ports on the SD or EB card as the physical IRF port, and configure the working mode of physical IRF ports as enhanced.

• In standalone mode, binding a physical port to an IRF port does not affect the current confi

uration of the port. However, when the operating mode changes to IRF mode, the default configuration of the physical IRF port restores, and you can only execute the shutdown, default, description, and flow-interval commands on the physical port. For more information about these commands, see

2 — LAN Switching Command Reference.

Laye

Saving configuration to the startup configuration file

Follow these steps to save the current configuration to the startup configuration file:

To do… Use the command…

Save the current configuration to the startup configuration file

save [ safely ] [ force ]

Switching operating mode

IRF modes

A switch works in either standalone or IRF mode.

• A standalone switch does not belong to any IRF fabric. To assign a standalone switch to an IRF

fabric, you must switch the operating mode to IRF mode before adding it to an IRF fabric.

• An IRF mode switch can itself form an IRF fabric or form an IRF fabric with other IRF mode switches.

To save management costs and system resources, set a switch in IRF mode only when you are creating a multi-member IRF fabric.

By default, a switch is operating in standalone mode. The switch reboots when its operating mode changes to IRF mode.

Configuration file auto-conversion

Remarks

Required

Available in any view.

When you change the operating mode of the switch from standalone to IRF, you can use the configuration file auto-conversion function to convert the startup configuration file to prevent some slot- or interface-related configurations from becoming invalid. For example, this function can convert the slot slot-number parameter set in standalone mode to the chassis chassis-number slot slot-number parameter in IRF mode, and add the chassis ID in an interface number.

Switching the operating mode to IRF mode

Follow these steps to switch the operating mode of a switch to IRF mode:

To do… Use the command…

Enter system view system-view —

Switch the operating mode to IRF mode

chassis convert mode irf

Remarks

Required

The default operating mode is standalone mode.

NOTE:

• When you switch the operating mode, the switch automatically reboots to make the change effective.

• If the member ID of the switch is not set, you must first set the member ID for the switch, and then execute

the chassis convert mode irf command. To display the member ID of the switch, use the display irf configuration command. If the MemberID field is displayed as —, the member ID of the switch is not se

yet.

• To switch the operating mode of the switch to standalone mode, use the undo chassis convert mode irf

command.

Accessing an IRF fabric

Accessing the active MPU of an IRF fabric

Access an IRF fabric in one of the following ways:

• Local login—Log in through the console port of a member switch.

• Remote login—Remotely log in at a Layer 3 interface on any member switch through Telnet or

SNMP.

When you log in to an IRF fabric, you are placed at the CLI of its active MPU, regardless of at which member switch you are logged in. The active MPU of the IRF fabric is the configuration and control center of the IRF fabric. You make configuration for the IRF fabric on the active MPU, and the IRF fabric synchronizes the configurations to all standby MPUs in the IRF fabric.

Accessing a standby MPU of an IRF fabric

You can log in to the CLI of a standby MPU of the IRF fabric to display its configurations and debug the standby MPU. When you switch from the active MPU’s CLI to the standby MPU’s CLI, you are placed in the user view of the standby MPU and the command prompt changes to <Sysname-Slave#member-ID/slot-number>, for example, <Sysname-Slave#1/0>. You can perform the following commands at the CLI of the standby MPU of an IRF fabric:

• display

• quit

• return

• system-view

• debugging

• terminal debugging

• terminal logging

• terminal monitor

• terminal trapping

To return to the CLI of the active MPU of the IRF fabric, use the quit command.

Follow these steps to log in to the CLI of a standby MPU in the IRF fabric:

To do… Use the command…

Remarks

Enter system view system-view —

To do… Use the command…

NOTE:

n IRF fabric allows at most 16 virtual type terminal (VTY) users to log in at the same time. The number of

supported console users equals the total number of MPUs in the fabric.

irf switch-to chassis chassis-number slot slot-number

Remarks

Required

By default, you are placed at the CLI of the active MPU of the IRF fabric when you log in to the IRF fabric.

Available in user view

Configuring IRF member switches in IRF mode

Specifying an IRF domain ID for an IRF fabric

Introduction to IRF domains

IRF uses IRF domain IDs to uniquely identify IRF fabrics. IRF domain IDs prevent IRF fabrics from interfering with one another.

See Figure 9, IRF fabrics have LACP MAD detection links between them. When a member switch in one IRF fabric receives an extended LACP packet for MAD detection, it looks at the domain ID in the packet to see whether the packet is from the local IRF fabric or from a different IRF fabric. Then, the switch can handle the packet correctly.

Switch A and Switch B form IRF fabric 1, and Switch C and Switch D form IRF fabric 2. The

Figure 9 A network that comprises two IRF domains

Core network

Switch A

Switch C

IRF 1 (domain 10)

IRF link

IRF 2 (domain 20)

IRF link

Access network

Assigning a domain ID to an IRF fabric

Follow these steps to assign a domain ID to an IRF fabric:

To do… Use the command…

Switch B

Switch D

Remarks

Enter system view system-view —

Assign a domain ID to the IRF fabric

irf domain domain-id

Modifying a member ID for a switch

An IRF fabric uses member IDs to uniquely identify its members. A lot of information and configurations relate to member IDs, such as port (physical or logical) numbers, configurations on ports, and member priorities.

After you change the member ID of a switch:

• If you do not reboot the switch, the original member ID still takes effect and all physical resources

are identified by the original member ID. In the configuration file, only the IRF port numbers, configurations on IRF ports, and priority of the switch change with the member ID, other configurations do not change.

• If you save the current configuration and reboot the switch, the new member ID takes effect and all

physical resources are identified by the new member ID. In the configuration file, only the IRF port numbers, configurations on IRF ports, and priority of the switch still take effect, other configurations (such as configuration for physical IRF ports, value of the chassis keyword which adopts the original member ID) no longer take effect and you need to configure them again.

By default, the domain ID of an IRF fabric is 0.

Follow these steps to modify a member ID for a member switch:

To do… Use the command…

Enter system view system-view —

Modify a member ID for a member switch

CAUTION:

irf member member-id renumber new-member-id

• Member ID changes take effect at the reboot of the switch.

• Change member IDs for the switches in an IRF fabric with caution. The change might cause

configuration change and even data loss.

• Different switch models and cards have different requirements for member IDs. See “Setting a member

for the switch.”

Specifying a priority for a member switch

Follow these steps to specify a priority for a member switch:

To do… Use the command…

Remarks

Optional

The switch has no default ID in standalone mode. After it switches to the IRF mode, it uses the member ID pre-configured in standalone mode by default.

Remarks

Enter system view system-view —

Specify a priority for a member of an IRF fabric

Configuring IRF ports

After an IRF fabric is formed, you can add physical ports used to connect member switches.

Follow these steps to configure IRF ports:

To do… Use the command…

Enter system view system-view —

Enter physical IRF port view

Shut down the port shutdown Required

Return to system view quit —

Create an IRF port and enter IRF port view

Optional

irf member member-id priority priority

The priority of a member defaults to 1.

Remarks

interface interface-type interface-number

irf-port member-id/port-number —

—

To do… Use the command…

Bind a physical port to the IRF port

Return to system view quit —

port group interface interface-type interface-number [ mode { enhanced |

normal } ]

Remarks

Required

By default, no physical ports are bound to any IRF port.

Repeat this step to assign multiple physical ports to the IRF port for link redundancy.

Enter physical IRF port view

Bring up the port undo shutdown Required

Connect the physical IRF ports on two IRF member switches

The system will prompts you that an IRF fabric merge will happen. You need to reboot one of switch that has failed in the election, and it will join the IRF fabric as a slave switch.

interface interface-type interface-number

— —

reboot Required

—

NOTE:

• The maximum number of physical IRF ports that can be bound to an IRF port is eight. When the upper

limit is reached, the execution of this command fails.

• The switch allows you to bind multiple physical IRF ports on different LPUs to one IRF port.

• The mode keyword in the port group command specifies the workin

mode of a physical IRF port. By default, the working mode of a physical IRF port is normal. An SC card does not support configuration of the working mode of physical IRF ports as enhanced.

• Physical IRF ports that connect two member switches of an IRF fabric must be confi

ured to work in the

same mode.

• To use the MPLS L2VPN or VPLS function in an IRF fabric, configure the 10 GE optical ports on the SD

or EB card as the physical IRF port, and configure the working mode of physical IRF ports as enhanced.

• Before you create or remove an IRF port binding, always shut down the physical IRF port. After you are

finished, perform the undo shutdown command to bring up the port. If the system prevents you from shutting down a port for avoiding anomalies, follow the system instructions to disable its peer port.

• On an Ethernet port functioning as a physical IRF port, you can execute only the shutdown, default,

description, and flow-interval commands. For more information about these commands, see

— LAN Switching Command Reference

Enabling auto reboot for IRF fabric merge

When merging two IRF fabrics, you must reboot the member switches in the IRF fabric that has been defeated in the master election. The auto reboot function enables the IRF fabric to automatically reboot all its member switches to complete the merge.

This function can work only when it is enabled on both IRF fabrics that are merging.

Follow these steps to enable auto reboot for IRF fabric merge:

Layer 2

To do… Use the command…

Enter system view system-view —

Remarks

Enable auto reboot for IRF fabric merge

irf auto-merge enable

Optional

Disabled by default.

NOTE:

The auto reboot function does not take effect on the IRF fabric merge caused by binding an IRF port with a physical IRF port in IRF mode. You must manually reboot the switches that have been defeated in the master selection to complete the merge.

Configuring a description for a member switch

You can configure a description for a member switch to identify its physical location, or for any other management purpose.

Follow these steps to configure a description for a member switch:

To do… Use the command…

Enter system view system-view —

Configure the description of a member

irf member member-id description text

Remarks

Optional

Not configured by default.

Configuring load sharing criteria for IRF links

When an IRF port is bound with two or more physical IRF ports, there are multiple IRF links among IRF fabrics.

Determine how traffic is load-shared among members by configuring load sharing criteria for IRF links. The criteria can be information carried in packets, such as IP addresses, MAC addresses, and any combination of them. The criteria can also be packet types, such as Layer 2 and Layer 3.

Configure global (in system view) or port-specific (in IRF port view) load sharing criteria for IRF links:

• If you configure the load sharing criteria in system view, the configuration is effective for IRF links on

all IRF ports.

• If you configure load sharing criteria in IRF port view, the configuration is effective for IRF links on this

IRF port.

• An IRF port preferentially uses the port-specific load sharing criteria. If no port-specific load sharing

criteria is available, it uses the global load sharing criteria.

NOTE:

The load sharing criterion or criteria you configured in the same view overwrite the old ones, if any.

If you configure a load sharing criterion not supported by the switch, you will be prompted that the switch

does not support the criterion.

Before configuring the load sharing criteria, bind IRF ports to correspondin

load sharing criterion configuration fails.

physical IRF ports. Otherwise,

Configuring global load sharing criteria

Follow these steps to configure the global IRF link load sharing criteria:

To do… Use the command…

Enter system view system-view —

irf-port load-sharing mode

Configure the global IRF link load sharing criteria

{ destination-ip | destination-mac |

ingress-port | source-ip | source-mac } *

Configuring port-specific load sharing criteria

Follow these steps to configure the port-specific load sharing criteria:

To do… Use the command…

Enter system view system-view —

Enter IRF port view irf-port member-id/port-number —

irf-port load-sharing mode

Configure the port-specific load sharing criteria

{ destination-ip | destination-mac |

ingress-port | source-ip | source-mac } *

Remarks

Optional

By default, the global IRF link load sharing criteria are source or destination MAC addresses for Layer 2 packets, source or destination IP addresses for Layer 3 packets, and source or destination port numbers for Layer 4 packets.

Remarks

Optional

By default, the port-specific load sharing criteria are source or destination MAC addresses for Layer 2 packets, source or destination IP addresses for Layer 3 packets, and source or destination port numbers for Layer 4 packets.

Specifying the preservation time of bridge MAC address

An IRF fabric uses the bridge MAC address of the master switch as its bridge MAC address. This bridge MAC address is used by Layer 2 protocols, MSTP and LACP, for example, to identify the IRF fabric. It is also used for generating an MAC address for a Layer 3 interface. This bridge MAC address must be unique on the switched LAN for proper communication.

To avoid duplicate bridge MAC addresses, an IRF fabric can automatically change its bridge MAC address after its master leaves, but the change can cause temporary service interruption. Depending on your network condition, you can enable the IRF fabric to preserve or change its bridge MAC address after the master leaves. Available options include:

• irf mac-address persistent timer—Preserves the bridge MAC address for 6 minutes after the master

leaves. If the master does not come back when the timer expires, the IRF fabric uses the bridge MAC address of the newly elected master as its bridge MAC address. This option avoids unnecessary switch of bridge MAC address due to a switch reboot or transient link failure.

• irf mac-address persistent always—Keeps the bridge MAC address even after the master leaves.

• undo irf mac-address persistent—Uses the bridge MAC address of the newly elected master to

replace the original one as soon as the master leaves.

Follow these steps to specify the preservation time of the bridge MAC address of an IRF fabric:

To do… Use the command…

Enter system view system-view —

Enable the IRF fabric to preserve its bridge MAC address permanently even after the master leaves

Enable the IRF fabric to preserve its bridge MAC address for six minutes after the master leaves

Enable the IRF fabric to change its bridge MAC address as soon as the master leaves

CAUTION:

irf mac-address persistent always

irf mac-address persistent timer

undo irf mac-address persistent

Remarks

Optional

By default, the IRF fabric preserves its bridge MAC address permanently even after the master leaves.

• Bridge MAC address change can cause transient traffic interruption.

• If two IRF fabrics have the same bridge MAC address, they cannot be merged into one IRF fabric.

Enabling automatic system software updating

• When you add a switch to the IRF fabric, the automatic system software updating function

compares the software versions of the switch and the IRF master. If the versions are different, the switch automatically downloads the system software image from the master, sets the downloaded file as the system software image for the next reboot, and automatically reboots with the new system software to re-join the IRF fabric.

• If this function is disabled, you must manually ensure that the joining switch uses the same system

software as the master switch. If not, the switch cannot join the IRF fabric.

Follow these steps to enable an IRF fabric to automatically synchronize system software of the master to the switch you are adding to the IRF fabric:

To do… Use the command…

Enter system view system-view —

Enable an IRF fabric to automatically synchronize system software of the master to the switch you are adding to the IRF fabric

CAUTION:

irf auto-update enable

Remarks

Optional

Disabled by default.

• Make sure that the switch you are adding to the IRF fabric is compatible with the software version

running on the master switch. If not, the automatic system software upgrading function cannot correctl work.

• Check that the switch has efficient space for the new system software.

• Changing the operating mode from IRF to standalone can cause the loss of the irf auto-update enable

command configuration, even if you have saved the configuration.

Setting the IRF link down report delay

You can avoid link flapping causing frequent IRF splits and merges during a short time by configuring the IRF ports to delay reporting link down events. An IRF port thus works as follows:

• If the IRF link changes from up to down, the port does not immediately report the change to the IRF

fabric. If the IRF link state is still down when the delay time is reached, the port reports the change to the IRF fabric.

• If the IRF link changes from down to up, the link layer immediately reports the event to the IRF fabric.

Follow these steps to set the IRF link down report delay:

To do… Use the command…

Enter system view system-view —

Set the IRF link down report delay irf link-delay interval

Configuring MAD detection

You have the following MAD mechanisms for detecting multi-active collisions in different network scenarios:

• LACP MAD

• BFD MAD

These MAD detection mechanisms operate independently, and you can configure all of them for an IRF fabric.

Configuring LACP MAD

1. LACP MAD detection mechanism

Remarks

Optional

The function is disabled by default.

The recommended value range (in milliseconds) is 200 to 500. The greater the interval, the slower the service recovery.

With LACP MAD, an IRF member switch sends extended LACP data units (LACPDUs) with a type length value (TLV) that conveys the domain ID and active ID of the IRF fabric for detecting an IRF split. The domain ID uniquely identifies an IRF fabric in the network, and the active ID is identical to the member ID of the master switch in the IRF fabric.

An IRF member switch compares the domain ID and the active ID in each received extended LACPDU with its domain ID and active ID:

• If the domain IDs are different, the extended LACPDU is from a different IRF fabric, and the switch

does not continue to process the extended LACPDU with the MAD mechanism.

• If the domain IDs are the same, the switch compares the active IDs:

{ If the active IDs are different, the IRF fabric has split.

{ If the active IDs are the same, the IRF fabric is operating normally.

2. Network requirements

Every IRF member switch has a link with an intermediate switch, and all these links form a dynamic link aggregation group, as shown in Figure 10.

The intermediate switch must be an HP switch capable of handling extended LACPDUs that carry the Active ID field. For more information about LACP and the support of the switch for extended LACPDUs, see Layer 2—LAN Switching Configuration Guide.

CAUTION:

If the intermediate switch is in an IRF fabric, you must assign this fabric a different domain ID than the LACP MAD-enabled IRF fabric to avoid false detection of IRF partition.

Figure 10 Network diagram

3. Configuring LACP MAD detection

Configure LACP MAD detection by following these steps:

• Create an aggregate interface (also required on the intermediate switch)

• Configure the aggregation group to work in dynamic aggregation mode (also required on the

intermediate switch)

• Enable LACP MAD detection on the dynamic aggregate interface

• Add member ports to the aggregation group.(also required on the intermediate switch)

Follow these steps to configure LACP MAD detection:

To do… Use the command…

Enter system view system-view —

Remarks

To do… Use the command…

Assign a domain ID to the IRF fabric

Irf domain domain-id

Remarks

Required if the intermediate device is an IRF fabric.

By default, the domain ID of an IRF fabric is 0.

Create an aggregate interface and enter aggregate interface view

Configure the aggregation group to work in dynamic aggregation mode

Enable LACP MAD detection mad enable

Return to system view quit —

Enter Ethernet interface view

Enter Layer 2 aggregate interface view

Enter Layer 3 aggregate interface view

interface bridge-aggregation interface-number

interface route-aggregation interface-number

link-aggregation mode dynamic

interface interface-type interface-number

Required

Use either command.

Required

By default, the aggregation group works in static aggregation mode.

Required

Disabled by default.

Even though this command can be configured on both static and dynamic aggregate interfaces, it takes effect only on dynamic aggregate interfaces. This is because this detection approach depends on LACP.

—

Assign the current Ethernet interface to the specified aggregation group

Configuring BFD MAD

1. BFD MAD detection mechanism

BFD MAD is implemented with the BFD protocol. To configure BFD MAD detection, configure a MAD IP address on a BFD-enabled Layer 3 interface for each member switch. This MAD address identifies the member during BFD MAD detection. The MAD IP addresses assigned to the member switches must belong to the same network segment.

• When the IRF fabric is operating normally, only the MAD IP address of the master is effective and

the BFD session is down.

• When the IRF fabric partitions, the MAD IP addresses of the masters in different IRF fabrics become

effective to activate the BFD sessions to detect for multi-active IRF fabric collision.

2. Network requirements

BFD MAD detection can work with or without intermediate switches. In a BFD MAD network without an intermediate switch, connect all IRF member switches with dedicated BFD MAD links in the full mesh topology, as shown in Figure 11. T

port link-aggregation group number Required

he interfaces connected by BFD MAD links must belong to the same

VLAN. In VLAN interface view, assign different IP addresses on the same network segment for different

member switches.

CAUTION:

Layer 3 interface used for BFD MAD must be dedicated. Do not configure any other services on them.

Otherwise, both the configured services and the BFD MAD detection function can be affected.

Figure 11 Network diagram

3. Configuring BFD MAD detection

Perform the following tasks to configure BFD MAD detection:

• Create a VLAN dedicated to BFD MAD detection. This is also required on the intermediate switch,

if it exists.

• Select the physical IRF ports for BFD MAD detection (at least one on each member switch) and add

them into the detection-dedicated VLAN. This is also required on the intermediate switch, if it exists.

• Create VLAN interfaces for the detection-dedicated VLAN, enable BFD MAD detection on these

interfaces, and then assign MAD IP addresses for them.

Follow these steps to configure BFD MAD:

To do… Use the command…

Enter system view system-view —

Create a new VLAN dedicated to BFD MAD detection

Return to system view quit —

vlan vlan-id

Remarks

Required

The default VLAN on the switch is VLAN 1.

To do… Use the command…

Remarks

Enter Ethernet interface view

Access port port access vlan vlan-id Required

Assign the port to the BFD MAD VLAN dedicated to BFD MAD detection

Return to system view quit —

Enter VLAN interface view

Enable BFD MAD mad bfd enable

Configure a MAD IP address for the VLAN interface on the specified member

Trunk port port trunk permit vlan vlan-id

Hybrid port

interface interface-type interface-number

port hybrid vlan vlan-id { tagged | untagged }

interface vlan-interface interface-number

mad ip address ip-address { mask | mask-length } member member-id

NOTE:

—

You can select one approach according to the port type.

BFD MAD detection has no requirement for the link type of the detection port.

The default link type of a port is access.

—

Required

Disabled by default.

Required

By default, no MAD IP address is configured for any interface.

• A VLAN interface enabled with BFD MAD detection and the interfaces in the BFD MAD VLAN do not

support any Layer 2 and Layer 3 protocol applications, including ARP and LACP.

• You cannot enable BFD MAD detection on VLAN-interface 1.

• Do not bind a BFD MAD enabled Layer 3 interface to any VPN instance. The MAD function is mutuall

exclusive with VPN.

• Do not enable the spanning tree function on any port in a BFD MAD VLAN. The MAD function is

mutually exclusive with the spanning tree function.

• To avoid anomalies, do not assign a BFD MAD enabled interface any other IP address, such as a

common IP address configured with the ip address command or a VRRP virtual IP address, than the MAD IP address.

• Do not assign a used IP address as the MAD IP address for the switch.

Excluding a port from the shut down action upon detection of multi-active collision

By default, all service ports of an IRF fabric except the console and physical IRF ports shut down when the IRF fabric transits to recovery state upon detection of a multi-active collision. You can exclude a service port from the shut down action for management or other special purposes. For example, you can exclude a port from the shut down action, so you can telnet to the port for managing the switch. For another example, you can exclude a VLAN interface and its Layer 2 Ethernet interface from the shut down action, so you can log in through the VLAN interface.

Follow these steps to configure a port not to shut down when the IRF fabric transits to recovery state:

To do… Use the command…

Enter system view system-view —

Remarks

To do… Use the command…

Remarks

Configure a service port not to shut down when the IRF fabric transits to recovery state

Manually recovering an IRF fabric

An IRF link failure causes an IRF fabric to divide into two IRF fabrics and multi-active collision occurs. When the system detects the collision, it holds a role election between the two collided IRF fabrics. The IRF fabric whose master’s member ID is smaller prevails and operates normally. The state of the other IRF fabric transitions to recovery state and temporarily cannot forward data packets, as shown in Figure 12. In this c

1. Repair the IRF link. Then, IRF fabric 1 prompts you that an IRF fabric merge will happen and you

2. Log in to the IRF fabric in recovery state and reboot it (execute the reboot command in user view).

After the switch reboots, the IRF fabric is recovered.

Figure 12 Recover IRF member switches (approach I)

ase, recover the IRF fabric by following these steps:

need to reboot the switch.

mad exclude interface interface-type interface-number

Required

When an IRF fabric transits to recovery state, all its service ports are shut down by default.

As shown in Figure 13, if the IRF fabric in active state fails due to exceptions (a member switch fails or link failure occurs, for example) before the IRF link is recovered, enable IRF fabric 2 (in recovery state) to change the state of IRF fabric 2 from recovery to active, and repair the IRF links. Then, the two IRF fabrics merge when the IRF link failure is recovered. The following are the operation steps:

3. Execute the mad restore command on IRF fabric 2 (in recovery state) to change its state from

recovery to active.

4. Repair IRF fabric 1 and IRF links, and then reboot IRF fabric 1.

After the reboot, IRF fabric 1 and IRF fabric 2 merge and the IRF fabric recovers.

Figure 13 Recover IRF member switches (approach II)

Follow these steps to manually recover an IRF fabric in recovery state:

To do… Use the command…

Enter system view system-view —

Recover an IRF fabric in recovery state

NOTE:

For a switch that transitioned to recovery state due to the multi-active collision, if you need to bring up one of its port that are already shut down, HP recommends that you execute the mad restore command to recover the switch and make it operate in the active state instead of executing the undo shutdown command to perform a manual recovery.

mad restore Required

IRF configuration fast recovery

If a member switch has only one MPU and that board is corrupted, use the following fast recover function to avoid complicated IRF configuration.

Remarks

Configuration prerequisites

• HP reco mmend s tha t you pe rform t he fast recover y ope ration wh en th e switch op erates i n IR F mod e.

• Save the IRF configuration file on all MPUs of the IRF fabric. Assume the name of the configuration

file is a.cfg.

Configuration procedures

• Scenario I: Member switch 1 has one MPU, and it is corrupted. Member switch 2 has two MPUs.

1. Modify the member ID of the standby MPU on member switch 2 to make it consistent with that of

member switch 1.

2. Unplug the corrupted MPU on member switch 1, and insert the standby MPU of member switch 2

into member switch 1.

3. Configure IRF ports on member switch 1, and connect member switch 1 to member switch 2

through physical IRF ports.

The fast recovery is completed.

• Scenario II: Both member switches have only one MPU, and the MPU on member switch 1 is

corrupted.

4. Insert another MPU into member switch 2 and use it as the standby MPU of member switch 2.

5. Copy the configuration file a.cfg on the active MPU of member switch 2 to the standby MPU, and

set the file as the startup configuration file.

6. Modify the member ID of the standby MPU on member switch 2 to make it consistent with that of

member switch 1.

7. Unplug the corrupted MPU on member switch 1, and insert the standby MPU of member switch 2

into member switch 1.

8. Configure IRF ports on member switch 1, and connect member switch 1 to member switch 2

through physical IRF ports.

The fast recovery is completed.

Follow the step below to modify the member ID of the standby MPU:

To do… Use the command…

Enter system view system-view —

In IRF mode:

set irf chassis chassis-number slot

Change the member ID of the standby MPU

slot-number member-id member-id

In standalone mode:

set irf slot slot-number member-id member-id

Remarks

This command is only applicable to the fast recovery of the IRF. Using this command in other cases may cause unknown errors. Use it with caution.

Displaying and maintaining an IRF fabric

To do… Use the command… Remarks

Display information about the IRF fabric

Display topology information about the IRF fabric

display irf [ | { begin | exclude | include } regular-expression ]

display irf topology [ | { begin | exclude | include }

regular-expression ]

Available in any view

Display all members’ configurations that take effect after device reboots

display irf configuration [ | { begin | exclude | include } regular-expression ]

Available in any view

To do… Use the command… Remarks

display irf-port load-sharing mode

Display the load sharing criteria for IRF links

[ irf-port [ member-id/port-number ] ] [ | { begin | exclude | include } regular-expression ]

Available in any view

Display the master/slave switchover states of a members of an IRF fabric

Display configuration information about MAD detection

display switchover state [ chassis chassis-number slot slot-number ] [ | { begin | exclude | include }

regular-expression ]

display mad [ verbose ] [ | { begin | exclude | include }

regular-expression ]

Available in any view

IRF fabric configuration examples

LACP MAD detection-enabled IRF configuration example (with two member switches)

Network requirements

The number of PCs on the enterprise network (see Figure 14) is outgrowing the number of ports available on the access switches. To address business growth, increase the number of ports at the access layer while protecting the present investments of the customer and ensuring ease of management and maintenance.

Figure 14 Network diagram

GE4/0/1 GE4/0/2

GE1/4/0/2

XGE1/3/0/1

Device A Device B

(IRF-port1/2)

Configuration considerations

• To increase the number of access ports, you must add extra devices. In this example, Device B is

added.

Device C

IP network

IRF

GE2/4/0/2

XGE2/3/0/1 (IRF-port2/1)

• To address the requirements for high availability, ease of management and maintenance, use IRF2

technology to create an IRF fabric with Device A and Device B at the access layer.

• To offset the risk of IRF fabric partition, configure MAD to detect multi-active collisions. In this

example, LACP MAD is adopted because the number of access devices tends to be large. In addition, for the purpose of LACP MAD, an intermediate switch that support extended LACPDUs must be used.

• To reduce system reboots involved to create the IRF fabric, pre-configure the IRF ports, member ID,

member priority on each device in standalone mode and change their operating mode to IRF after saving the configuration. The devices can thus form an IRF fabric.

Configuration procedure

1. Configure Device A

# Set the member ID of Device A to 1, create IRF port 2, and bind it to physical port Ten-GigabitEthernet 3/0/1.

<Sysname> system-view [Sysname] irf member 1 Info: Member ID change will take effect after the switch reboots and operates in IRF mode. [Sysname] interface ten-gigabitethernet 3/0/1 [Sysname-Ten-GigabitEthernet3/0/1] shutdown [Sysname-Ten-GigabitEthernet3/0/1] quit [Sysname] irf-port 2 [Sysname-irf-port2] port group interface ten-gigabitethernet 3/0/1 [Sysname-irf-port2] quit [Sysname] interface ten-gigabitethernet 3/0/1 [Sysname-Ten-GigabitEthernet3/0/1] undo shutdown [Sysname-Ten-GigabitEthernet3/0/1] quit

# Save the current configuration to the startup configuration file.

[Sysname] quit <Sysname> save

# Change the operating mode of the device to IRF.

<Sysname> system-view [Sysname] chassis convert mode irf The device will switch to IRF mode and reboot. You are recommended to save the current

running configuration and specify the configuration file for the next startup. Continue? [Y/N]:y

Do you want to convert the content of the next startup configuration file flash:/startup.cfg to make it available in IRF mode? [Y/N]:y

Please wait... Saving the converted configuration file to the main board succeeded. Slot 1: Saving the converted configuration file succeeded. Now rebooting, please wait...

The device operates in IRF mode after reboot.

2. Configure Device B

# Set the member ID of Device B to 2, create IRF port 1, and bind it to physical port Ten-GigabitEthernet 3/0/1.

<Sysname> system-view

[Sysname] irf member 2 Info: Member ID change will take effect after the switch reboots and operates in IRF mode. [Sysname] interface ten-gigabitethernet 3/0/1 [Sysname-Ten-GigabitEthernet3/0/1] shutdown [Sysname-Ten-GigabitEthernet3/0/1] quit [Sysname] irf-port 1 [Sysname-irf-port1] port group interface ten-gigabitethernet 3/0/1 [Sysname-irf-port1] quit [Sysname] interface ten-gigabitethernet 3/0/1 [Sysname-Ten-GigabitEthernet3/0/1] undo shutdown [Sysname-Ten-GigabitEthernet3/0/1] quit

# Save the current configuration to the startup configuration file.

[Sysname] quit <Sysname> save

# Connect the two devices as shown in Figure 14.

# Switch the operating mode of the device to IRF.

<Sysname> system-view [Sysname] chassis convert mode irf The device will switch to IRF mode and reboot. You are recommended to save the current

running configuration and specify the configuration file for the next startup. Continue? [Y/N]:y

Do you want to convert the content of the next startup configuration file flash:/startup.cfg to make it available in IRF mode? [Y/N]:y

Please wait... Saving the converted configuration file to the main board succeeded. Slot 1: Saving the converted configuration file succeeded. Now rebooting, please wait...

Device B and Device A form an IRF fabric after Device B reboots.

3. Configure LACP MAD

# Create a dynamic aggregate interface and enable LACP MAD.

<Sysname> system-view [Sysname] interface bridge-aggregation 2 [Sysname-Bridge-Aggregation2] link-aggregation mode dynamic [Sysname-Bridge-Aggregation2] mad enable [Sysname-Bridge-Aggregation2] quit

# Assign ports GigabitEthernet 1/4/0/2 and GigabitEthernet 2/4/0/2 to the aggregate interface for LACP MAD detection.

4. Configure Device C as the intermediate switch

Device C needs to act as the intermediate switch to forward and process extended LACPDUs exchanged between Device A and Device B for MAD detection.

CAUTION:

If the intermediate switch is in an IRF fabric, you must assign this fabric a different domain ID than the LACP MAD-enabled IRF fabric.

# Create a dynamic aggregate interface.

<Sysname> system-view [Sysname] interface bridge-aggregation 2 [Sysname-Bridge-Aggregation2] link-aggregation mode dynamic [Sysname-Bridge-Aggregation2] quit

# Assign ports GigabitEthernet 4/0/1 and GigabitEthernet 4/0/2 to the dynamic aggregate interface.

5. When the IRF links fail, the system outputs error messages for detecting an IRF link and card failure.

#May 7 09:13:42:388 2010 HP STM/4/LINK STATUS CHANGE: Trap 1.3.6.1.4.1.2011.10.2.91.6.0.1: Physical index of the member is 2, member ID is 1.

Link status of the IRF port with port index 1 and member ID 1 turned to 2. #May 7 09:13:42:720 2010 HP DEVM/1/BOARD STATE CHANGES TO FAILURE:

Because the member ID of Device B is 2, when the MAD mechanism detects errors, the state of Device B changes to recovery and all ports on Device B except for the reserved ones are shut down.

# If the master fails, log in to the console port of Device B and use the mad restore command to recover Device B to active state and bring up disabled ports on the device.

<Sysname> system-view [Sysname] mad restore This command will restore the device from multi-active conflict state. Continue? [Y/N]:y Restoring from multi-active conflict state, please wait... [Sysname] #May 7 09:23:16:050 2010 HP IFNET/4/INTERFACE UPDOWN: Trap 1.3.6.1.6.3.1.1.5.4: Interface 277872640 is Up, ifAdminStatus is 1, ifOperStatus

is 1 %May 7 09:23:16:069 2010 HP IFNET/3/LINK_UPDOWN: GigabitEthernet2/3/0/2 link status is

UP. #May 7 09:23:16:302 2010 HP LAGG/1/AggPortRecoverActive: Trap 1.3.6.1.4.1.2011.5.25.25.2.4: Aggregation Group 1: port member 277872640 becomes

ACTIVE! %May 7 09:23:16:322 2010 HP LAGG/5/LAGG_ACTIVE: Member port GigabitEthernet2/3/0/1 of

aggregation group BAGG1 becomes ACTIVE. %May 7 09:23:16:370 2010 HP IFNET/3/LINK_UPDOWN: Bridge-Aggregation2 link status is UP. %May 7 09:23:16:381 2010 HP IFNET/3/LINK_UPDOWN: Vlan-interface1 link status is UP. %May 7 09:23:16:391 2010 HP IFNET/5/LINEPROTO_UPDOWN: Line protocol on the interface

Vlan-interface1 is UP.

The output information shows that Device B is restored. You can repair Device A and the IRF links.

# When Device A and IRF links are recovered, reboot Device A. Device B prompts that the IRF port is up and a new card is inserted.

%May 7 09:30:12:122 2010 HP STM/6/STM_LINK_STATUS_UP:

IRF port 2 is up. #May 7 09:30:36:566 2010 HP DEVM/1/BOARD INSERTED:

# Execute the display irf command. You can see that the IRF fabric is recovered, Device B is the master and Device A is the slave.

<Sysname> display irf Switch Slot Role Priority CPU-Mac Description 1 1 Slave 1 00e0-fc0f-8c02 ---- *+2 1 Master 1 00e0-fc0f-8c20 -----

--------------------------------------------------

* indicates the device is the master. + indicates the device through which the user logs in.

The Bridge MAC of the IRF is: 0023-895f-954f Auto upgrade : yes Mac persistent : no Domain ID : 30 Auto merge : no

BFD MAD detection-enabled IRF configuration example (with two member switches)

Network requirements

The network as shown in Figure 15 is outgrowing the forwarding capability of the existing core switch, Device A. To address business growth, scale up the network to extend its forwarding capability while protecting the present investments of the customer and ensuring ease of management and maintenance.

Figure 15 Network diagram

Device A

Configuration considerations

• Device A is located at the distribution layer of the network. To improve the forwarding capability at

this layer, you must add extra devices. In this example, Device B is added.

• To address the requirements for high availability, scalability, and ease of management and

maintenance, use IRF2 technology to create an IRF fabric with Device A and Device B at the distribution layer. The access devices are each connected to the distribution layer with dual links.

XGE1/3/0/1 (IRF-port1/2)

GE1/4/0/1

IRF

BFD MAD link

……

XGE2/3/0/1 (IRF-port2/1)

GE2/4/0/1

Device B

• To offset the risk of IRF fabric partition, configure MAD to detect multi-active collisions. In this

example, BFD MAD is adopted because the number of member switches in the IRF fabric is small.

• Specify priorities for member switches to make sure that Device A is elected as the master of the IRF

fabric.

• When IRF links fail, first repair IRF links and then reboot the device in recovery state to make it

re-join the IRF fabric.

Configuration procedure

1. Configure Device A

# Set the member ID of Device A to 1, create IRF port 2, and bind it to physical port Ten-GigabitEthernet 3/0/1.

[Sysname-irf-port2] quit [Sysname] interface ten-gigabitethernet 3/0/1 [Sysname-Ten-GigabitEthernet3/0/1] undo shutdown [Sysname-Ten-GigabitEthernet3/0/1] quit

# Specify the priority of Device A as 10 to make sure that it is elected as the master when the IRF fabric is established, and save the configuration.

[DeviceA] irf priority 10

# Save the current configuration to the startup configuration file.

[Sysname] quit <Sysname> save

# Change the operating mode of the device to IRF.

<Sysname> system-view [Sysname] chassis convert mode irf The device will switch to IRF mode and reboot. You are recommended to save the current

running configuration and specify the configuration file for the next startup. Continue? [Y/N]:y

Do you want to convert the content of the next startup configuration file flash:/startup.cfg to make it available in IRF mode? [Y/N]:y

Please wait... Saving the converted configuration file to the main board succeeded. Slot 1: Saving the converted configuration file succeeded. Now rebooting, please wait...

The device operates in IRF mode after reboot.

2. Configure Device B

# Set the member ID of Device B to 2, create IRF port 1, and bind it to physical port Ten-GigabitEthernet 3/0/1.

<Sysname> system-view [Sysname] irf member 2 Info: Member ID change will take effect after the switch reboots and operates in IRF

mode. [Sysname] interface ten-gigabitethernet 3/0/1 [Sysname-Ten-GigabitEthernet3/0/1] shutdown [Sysname-Ten-GigabitEthernet3/0/1] quit [Sysname] irf-port 1 [Sysname-irf-port1] port group interface ten-gigabitethernet 3/0/1 [Sysname-irf-port1] quit [Sysname] interface ten-gigabitethernet 3/0/1 [Sysname-Ten-GigabitEthernet3/0/1] undo shutdown [Sysname-Ten-GigabitEthernet3/0/1] quit

# Save the current configuration to the startup configuration file.

[Sysname] quit <Sysname> save

# Connect the two devices as shown in Figure 15.

# Switch the operating mode of the device to IRF.

<Sysname> system-view

[Sysname] chassis convert mode irf The device will switch to IRF mode and reboot. You are recommended to save the current

running configuration and specify the configuration file for the next startup. Continue? [Y/N]:y

Do you want to convert the content of the next startup configuration file flash:/startup.cfg to make it available in IRF mode? [Y/N]:y

Please wait... Saving the converted configuration file to the main board succeeded. Slot 1: Saving the converted configuration file succeeded. Now rebooting, please wait...

Device B and Device A form an IRF after Device B reboots.

3. Configure BFD MAD

# Create VLAN 3, and add port GigabitEthernet 1/4/0/1 (located on Device A) and port GigabitEthernet 2/4/0/1 (located on Device B) to VLAN 3.

<Sysname> system-view [Sysname] vlan 3 [Sysname-vlan3] port gigabitethernet 1/4/0/1 gigabitethernet 2/4/0/1 [Sysname-vlan3] quit

# Create VLAN-interface 3 and configure a MAD IP address for the interface.

[Sysname] interface vlan-interface 3 [Sysname-Vlan-interface3] mad bfd enable [Sysname-Vlan-interface3] mad ip address 192.168.2.1 24 member 1 [Sysname-Vlan-interface3] mad ip address 192.168.2.2 24 member 2 [Sysname-Vlan-interface3] quit

# Disable the spanning tree function on ports GigabitEthernet 1/4/0/1 and GigabitEthernet 2/4/0/1.

[Sysname] interface gigabitethernet 1/4/0/1 [Sysname-gigabitethernet-1/4/0/1] undo stp enable [Sysname-gigabitethernet-1/4/0/1] quit [Sysname] interface gigabitethernet 2/4/0/1 [Sysname-gigabitethernet-2/4/0/1] undo stp enable

4. When the IRF links fail, the system outputs error messages for detecting multi-active devices and

prompts you to repair the links.

%May 6 15:10:05:477 2010 HP MAD/1/MAD_COLLISION_DETECTED: Multi-active devices detected, please fix it.

Because the member ID of Device B is 2, when the MAD mechanism detects errors, the state of Device B changes to recovery and all ports on Device B except for the reserved ones are shut down.

# Repair IRF links. Then, you are prompted that an IRF fabric merge will happen and you need to reboot the IRF fabric.

%May 6 15:12:52:935 2010 HP STM/6/STM_LINK_STATUS_UP: IRF port 1 is up. %May 6 15:13:02:828 2010 HP STM/4/STM_MERGE_NEED_REBOOT: IRF merge occurs and the IRF system needs a reboot.

# Log in to the console port of Device B and reboot the device.

<Sysname> reboot

Start to check configuration with next startup configuration file, please wait.

.........DONE!

This command will reboot the device. Continue? [Y/N]:y #May 6 15:31:09:724 2010 HP DEVM/1/REBOOT: Reboot device by command. %May 6 15:31:09:734 2010 HP DEVM/5/SYSTEM_REBOOT: System is rebooting now.

# After device reboot, Device B re-joins the IRF fabric. You can use the display irf command to view the topology information about the IRF fabric.

<Sysname> display irf topology Topology Info

------------------------------------------------------------------------ IRF-Port1 IRF-Port2 Switch Link neighbor Link neighbor Belong To 2 DOWN -- UP 1 00e0-fc0f-8c02 1 UP 2 DIS -- 00e0-fc0f-8c02

LACP MAD detection-enabled IRF configuration example (with four member switches)

Network requirements

The number of PCs on the enterprise network (see Figure 16) is outgrowing the number of ports available on the access switches. To address business growth, increase the number of ports at the access layer while protecting the present investments of the customer and ensuring ease of management and maintenance.

Figure 16 Network diagram

Device E

GE4/0/1

GE1/4/0/2

Device A

XGE1/3/0/2 (IRF-port1/1)

XGE3/2/0/2 (IRF-port3/2)

Device C Device D

XGE1/3/0/1 (IRF-port1/2)

GE3/3/0/1

XGE3/2/0/1 (IRF-port3/1)

GE4/0/2

GE4/0/3

IRF

GE4/0/4

XGE2/3/0/2 (IRF-port2/1)

GE4/3/0/1

XGE4/2/0/2 (IRF-port4/2)

IP network

GE2/4/0/1

Device B

XGE2/3/0/1 (IRF-port2/2)

XGE4/2/0/1 (IRF-port4/1)

Configuration considerations

• To increase the number of access ports, you must add extra devices. In this example, Device B,

Device C, and Device D are added.

• To address the requirements for high availability, scalability, and ease of management and

maintenance, use IRF2 technology to create an IRF fabric with Device A, Device B, Device C, and Device D at the access layer.

• To offset the risk of IRF fabric partition, configure MAD to detect multi-active collisions. In this

example, LACP MAD is adopted because the number of access devices tends to be large. In addition, for the purpose of LACP MAD, an intermediate switch that support extended LACPDUs must be used.

• To reduce system reboots involved to create the IRF fabric, pre-configure the IRF ports, member ID,

and member priority on each device in standalone mode, and change their operating mode to IRF after saving the configuration. The devices can thus form an IRF fabric.

Configuration procedure

1. Configure Device A

# Set the member ID of Device A to 1, create IRF ports 1 and 2, and bind them to physical ports Ten-GigabitEthernet 3/0/2 and Ten-GigabitEthernet 3/0/1 respectively.

<Sysname> system-view [Sysname] irf member 1 Info: Member ID change will take effect after the switch reboots and operates in IRF mode. [Sysname] irf-port 1 [Sysname-irf-port1] port group interface ten-gigabitethernet 3/0/2 [Sysname-irf-port1] quit [Sysname] irf-port 2 [Sysname-irf-port2] port group interface ten-gigabitethernet 3/0/1 [Sysname-irf-port2] quit

# Save the current configuration to the startup configuration file.

[Sysname] quit <Sysname> save

# Change the operating mode of the device to IRF.

<Sysname> system-view [Sysname] chassis convert mode irf The device will switch to IRF mode and reboot. You are recommended to save the current

running configuration and specify the configuration file for the next startup. Continue? [Y/N]:y

Do you want to convert the content of the next startup configuration file flash:/startup.cfg to make it available in IRF mode? [Y/N]:y

Please wait... Saving the converted configuration file to the main board succeeded. Slot 1: Saving the converted configuration file succeeded. Now rebooting, please wait...

The device operates in IRF mode after reboot.

2. Configure Device B

# Set the member ID of Device B to 2, create IRF ports 1 and 2, and bind them to physical ports Ten-GigabitEthernet 3/0/2 and Ten-GigabitEthernet 3/0/1 respectively.

<Sysname> system-view [Sysname] irf member 2 Info: Member ID change will take effect after the switch reboots and operates in IRF mode. [Sysname] irf-port 1 [Sysname-irf-port1] port group interface ten-gigabitethernet 3/0/2 [Sysname-irf-port1] quit [Sysname] irf-port 2 [Sysname-irf-port2] port group interface ten-gigabitethernet 3/0/1 [Sysname-irf-port2] quit

# Save the current configuration to the startup configuration file.

[Sysname] quit <Sysname> save

# Connect Device B to Device A as shown in Figure 16.

# Switch the operating mode of Device B to IRF.

<Sysname> system-view [Sysname] chassis convert mode irf The device will switch to IRF mode and reboot. You are recommended to save the current

running configuration and specify the configuration file for the next startup. Continue? [Y/N]:y

Do you want to convert the content of the next startup configuration file flash:/startup.cfg to make it available in IRF mode? [Y/N]:y

Please wait... Saving the converted configuration file to the main board succeeded. Slot 1: Saving the converted configuration file succeeded. Now rebooting, please wait...

Device B and Device A form an IRF fabric after Device B reboots.

3. Configure Device C

# Set the member ID of Device C to 3, create IRF ports 1 and 2, and bind them to physical ports Ten-GigabitEthernet 2/0/1 and Ten-GigabitEthernet 2/0/2 respectively.

<Sysname> system-view [Sysname] irf member 3 Info: Member ID change will take effect after the switch reboots and operates in IRF mode. [Sysname] irf-port 1 [Sysname-irf-port1] port group interface ten-gigabitethernet 2/0/1 [Sysname-irf-port1] quit [Sysname] irf-port 2 [Sysname-irf-port2] port group interface ten-gigabitethernet 2/0/2 [Sysname-irf-port2] quit

# Save the current configuration to the startup configuration file.

[Sysname] quit <Sysname> save

# Connect Device C to Device A as shown in Figure 16.

# Switch the operating mode of Device C to IRF.

<Sysname> system-view [Sysname] chassis convert mode irf The device will switch to IRF mode and reboot. You are recommended to save the current

running configuration and specify the configuration file for the next startup. Continue? [Y/N]:y

Do you want to convert the content of the next startup configuration file flash:/startup.cfg to make it available in IRF mode? [Y/N]:y

Please wait... Saving the converted configuration file to the main board succeeded. Slot 1: Saving the converted configuration file succeeded. Now rebooting, please wait...

Device C joins the IRF fabric after reboot.

4. Configure Device D

# Set the member ID of Device D to 4, create IRF ports 1 and 2, and bind them to physical ports Ten-GigabitEthernet 2/0/1 and Ten-GigabitEthernet 2/0/2 respectively.

<Sysname> system-view [Sysname] irf member 4 Info: Member ID change will take effect after the switch reboots and operates in IRF mode. [Sysname] irf-port 1 [Sysname-irf-port1] port group interface ten-gigabitethernet 2/0/1 [Sysname-irf-port1] quit [Sysname] irf-port 2 [Sysname-irf-port2] port group interface ten-gigabitethernet 2/0/2 [Sysname-irf-port2] quit

# Save the current configuration to the startup configuration file.

[Sysname] quit <Sysname> save

# Connect Device D to Device B and Device C as shown in Figure 16.

# Switch the operating mode of Device D to IRF.

<Sysname> system-view [Sysname] chassis convert mode irf The device will switch to IRF mode and reboot. You are recommended to save the current

running configuration and specify the configuration file for the next startup. Continue? [Y/N]:y

Do you want to convert the content of the next startup configuration file flash:/startup.cfg to make it available in IRF mode? [Y/N]:y

Please wait... Saving the converted configuration file to the main board succeeded. Slot 1: Saving the converted configuration file succeeded. Now rebooting, please wait...

Device D joins the IRF fabric after reboot. Now, the IRF fabric has four member switches.

5. Configure LACP MAD

# Create a dynamic aggregate interface and enable LACP MAD.

<Sysname> system-view

[Sysname] interface bridge-aggregation 2 [Sysname-Bridge-Aggregation2] link-aggregation mode dynamic [Sysname-Bridge-Aggregation2] mad enable [Sysname-Bridge-Aggregation2] quit

# Assign ports GigabitEthernet 1/4/0/2, GigabitEthernet 2/4/0/2, GigabitEthernet 3/3/0/1, and GigabitEthernet 4/3/0/1 to the aggregate interface for LACP MAD detection.

[Sysname] interface gigabitethernet 1/4/0/2 [Sysname-GigabitEthernet1/4/0/2] port link-aggregation group 2 [Sysname-GigabitEthernet1/4/0/2] quit [Sysname] interface gigabitethernet 2/4/0/1 [Sysname-GigabitEthernet2/4/0/1] port link-aggregation group 2 [Sysname-GigabitEthernet2/4/0/1] quit [Sysname] interface gigabitethernet 3/3/0/1 [Sysname-GigabitEthernet3/3/0/1] port link-aggregation group 2 [Sysname-GigabitEthernet3/3/0/1] quit [Sysname] interface gigabitethernet 4/3/0/1 [Sysname-GigabitEthernet4/3/0/1] port link-aggregation group 2 [Sysname-GigabitEthernet4/3/0/1] quit

6. Configure Device E as the intermediate switch

Device E needs to act as the intermediate switch to forward and process extended LACPDUs exchanged between the member switches for MAD detection.

CAUTION:

If the intermediate switch is in an IRF fabric, you must assign this fabric a different domain ID than the LACP MAD-enabled IRF fabric.

# Create a dynamic aggregate interface.

<Sysname> system-view [Sysname] interface bridge-aggregation 2 [Sysname-Bridge-Aggregation2] link-aggregation mode dynamic [Sysname-Bridge-Aggregation2] quit

# Assign ports GigabitEthernet 4/0/1, GigabitEthernet 4/0/2, GigabitEthernet 4/0/3, and GigabitEthernet 4/0/4 to the aggregate interface for LACP MAD detection.

[Sysname] interface gigabitethernet 4/0/1 [Sysname-GigabitEthernet4/0/1] port link-aggregation group 2 [Sysname-GigabitEthernet4/0/1] quit [Sysname] interface gigabitethernet 4/0/2 [Sysname-GigabitEthernet4/0/2] port link-aggregation group 2 [Sysname-GigabitEthernet4/0/2] quit [Sysname] interface gigabitethernet 4/0/3 [Sysname-GigabitEthernet4/0/3] port link-aggregation group 2 [Sysname-GigabitEthernet4/0/3] quit [Sysname] interface gigabitethernet 4/0/4 [Sysname-GigabitEthernet4/0/4] port link-aggregation group 2 [Sysname-GigabitEthernet4/0/4] quit

7. Troubleshoot the IRF when the MAD mechanism detects errors

Because the member switches are connected with a ring connection, when a single IRF link fails, the connection turns into a daisy chain connection and the IRF fabric is not partitioned. When both the IRF link between Device A and Device C and the IRF link between Device B and Device D go down, however, the IRF fabric will break into two IRF fabrics: one (IRF 1) with Device A and Device C, and one (IRF 2) with Device B and Device D. In addition, IRF link status error and card failure messages will be displayed. The following takes IRF 1 for example:

#May 7 09:13:42:388 2010 HP STM/4/LINK STATUS CHANGE: #May 7 09:13:42:720 2010 HP DEVM/1/BOARD STATE CHANGES TO FAILURE:

Because the master member ID of IRF 2 is bigger, when the MAD mechanism detects errors, the state of IRF 2 changes to recovery and all ports on the member devices except for the reserved ones are shut down.

# If IRF 1 also fails, log in to the console port of Device B or Device D and use the mad restore command to restore IRF 2 to active state and bring up the disabled ports on the member devices.

<Sysname> system-view [Sysname] mad restore This command will restore the device from multi-active conflict state. Continue? [Y/N]:y Restoring from multi-active conflict state, please wait... [Sysname] #May 7 09:23:16:050 2010 HP IFNET/4/INTERFACE UPDOWN: Trap 1.3.6.1.6.3.1.1.5.4<linkUp>: Interface 277872640 is Up, ifAdminStatus is 1,

ifOperStatus is 1 %May 7 09:23:16:069 2010 HP IFNET/3/LINK_UPDOWN: GigabitEthernet2/3/0/2 link status is

UP. #May 7 09:23:16:302 2010 HP LAGG/1/AggPortRecoverActive: Trap 1.3.6.1.4.1.2011.5.25.25.2.4<hwAggPortActiveNotification>: Aggregation Group 2:

port member 277872640 becomes ACTIVE! %May 7 09:23:16:322 2010 HP LAGG/5/LAGG_ACTIVE: Member port GigabitEthernet2/4/0/1 of

aggregation group BAGG2 becomes ACTIVE. %May 7 09:23:16:370 2010 HP IFNET/3/LINK_UPDOWN: Bridge-Aggregation2 link status is UP. %May 7 09:23:16:381 2010 HP IFNET/3/LINK_UPDOWN: Vlan-interface1 link status is UP. %May 7 09:23:16:391 2010 HP IFNET/5/LINEPROTO_UPDOWN: Line protocol on the interface

Vlan-interface1 is UP.

The output information shows that IRF 2 is recovered. You can repair IRF 1 and the IRF links.

# When IRF 1 and the IRF links recover, reboot IRF 1. IRF 2 prompts that the IRF port is up and a new card is inserted, which means that the two IRF fabrics has emerged.

%May 7 09:30:12:122 2010 HP STM/6/STM_LINK_STATUS_UP: #May 7 09:30:36:566 2010 HP DEVM/1/BOARD INSERTED:

# Execute the display irf command. You can see that the IRF fabric is recovered, and Device B is the master.

<Sysname> display irf Switch Slot Role Priority CPU-Mac Description 1 1 Slave 1 00e0-fc0f-8c02 ---- *+2 1 Master 1 00e0-fc0f-8c20 ---- 3 1 Slave 1 00e0-fc00-5801 ---- 4 1 Slave 1 00e0-fc00-3583 -----

------------------------------------------------- * indicates the device is the master. + indicates the device through which the user logs in.

The Bridge MAC of the IRF is: 0023-895f-954f Auto upgrade : yes Mac persistent : no Domain ID : 0 Auto merge : no

BFD MAD detection-enabled IRF configuration example (with four member switches)

Network requirements

The network as shown in Figure 17 is outgrowing the forwarding capability of the existing core switch, Device A. To address business growth, scale up the network to extend its forwarding capability while protecting the present investments of the customer and ensuring ease of management and maintenance.

Figure 17 Network diagram

Device A

XGE1/3/0/2

(IRF-port1/2)

XGE3/2/0/1

(IRF-port3/1)

Device C Device D

IRF link BFD link Data link

Configuration considerations

GE1/0/1

GE1/2/0/1 GE2/3/0/1 XGE1/3/0/1

(IRF-port1/1)

GE3/3/0/1 GE4/3/0/1

XGE3/2/0/2 (IRF-port3/2)

GE1/0/2

GE1/0/3 GE1/0/4

XGE2/2/0/2

(IRF-port2/2)

XGE4/2/0/1

(IRF-port4/1)

IRF

Device B

XGE2/2/0/1 (IRF-port2/1)

XGE4/2/0/2 (IRF-port4/2)

• Device A is located at the distribution layer of the network. To improve the forwarding capability at

this layer, you must add three devices. In this example, Device B, Device C, and Device D are added.

• To address the requirements for high availability, scalability, and ease of management and

maintenance, use IRF2 technology to create an IRF fabric with four member switches at the distribution layer.

• To offset the risk of IRF fabric partition, configure MAD to detect multi-active collisions. In this

example, BFD MAD is adopted.

• Specify priorities for member switches to make sure that Device A is elected as the master of the IRF

fabric.

• When IRF links fail, first repair IRF links and then reboot the device in recovery state to make it re-join

the IRF fabric.

Configuration procedure

1. Configure Device A

# Set the member ID of Device A to 1, create IRF ports 1 and 2, and bind them to physical ports Ten-GigabitEthernet 3/0/1 and Ten-GigabitEthernet 3/0/2 respectively.

<Sysname> system-view [Sysname] irf member 1 Info: Member ID change will take effect after the switch reboots and operates in IRF mode. [Sysname] irf-port 1 [Sysname-irf-port1] port group interface ten-gigabitethernet 3/0/1 [Sysname-irf-port1] quit [Sysname] irf-port 2 [Sysname-irf-port2] port group interface ten-gigabitethernet 3/0/2 [Sysname-irf-port2] quit

# Specify the priority of Device A as 10 to make sure that it is elected as the master when the IRF fabric is established.

[DeviceA] irf priority 10

# Save the current configuration to the startup configuration file.

[Sysname] quit <Sysname> save

# Change the operating mode of Device A to IRF.

<Sysname> system-view [Sysname] chassis convert mode irf The device will switch to IRF mode and reboot. You are recommended to save the current

running configuration and specify the configuration file for the next startup. Continue? [Y/N]:y

Do you want to convert the content of the next startup configuration file flash:/startup.cfg to make it available in IRF mode? [Y/N]:y

Please wait... Saving the converted configuration file to the main board succeeded. Slot 1: Saving the converted configuration file succeeded. Now rebooting, please wait...

Device A operates in IRF mode after reboot.

2. Configure Device B

# Set the member ID of Device B to 2, create IRF ports 1 and 2, and bind them to physical ports Ten-GigabitEthernet 2/0/1 and Ten-GigabitEthernet 2/0/2.

<Sysname> system-view [Sysname] irf member 2 Info: Member ID change will take effect after the switch reboots and operates in IRF mode.

[Sysname] irf-port 1 [Sysname-irf-port1] port group interface ten-gigabitethernet 2/0/1 [Sysname-irf-port1] quit [Sysname] irf-port 2 [Sysname-irf-port2] port group interface ten-gigabitethernet 2/0/2

# Save the current configuration to the startup configuration file.

[Sysname] quit <Sysname> save

# Connect Device B to Device A as shown in Figure 17.

# Switch the operating mode of Device B to IRF.

<Sysname> system-view [Sysname] chassis convert mode irf The device will switch to IRF mode and reboot. You are recommended to save the current

running configuration and specify the configuration file for the next startup. Continue? [Y/N]:y

Do you want to convert the content of the next startup configuration file flash:/startup.cfg to make it available in IRF mode? [Y/N]:y

Please wait... Saving the converted configuration file to the main board succeeded. Slot 1: Saving the converted configuration file succeeded. Now rebooting, please wait...

Device B and Device A form an IRF fabric after Device B reboots.

3. Configure Device C

# Set the member ID of Device C to 3, create IRF ports 1 and 2, and bind them to physical ports Ten-GigabitEthernet 2/0/1 and Ten-GigabitEthernet 2/0/2 respectively.

# Save the current configuration to the startup configuration file.

[Sysname] quit <Sysname> save

# Connect Device C to Device A as shown in Figure 17.

# Switch the operating mode of Device C to IRF.

<Sysname> system-view [Sysname] chassis convert mode irf The device will switch to IRF mode and reboot. You are recommended to save the current

running configuration and specify the configuration file for the next startup. Continue? [Y/N]:y

Do you want to convert the content of the next startup configuration file flash:/startup.cfg to make it available in IRF mode? [Y/N]:y

Please wait... Saving the converted configuration file to the main board succeeded. Slot 1: Saving the converted configuration file succeeded. Now rebooting, please wait...

Device C joins the IRF fabric after reboot.

4. Configure Device D

# Set the member ID of Device D to 4, create IRF ports 1 and 2, and bind them to physical ports Ten-GigabitEthernet 2/0/1 and Ten-GigabitEthernet 2/0/2 respectively.

# Save the current configuration to the startup configuration file.

[Sysname] quit <Sysname> save

# Connect Device D to Device B and Device C as shown in Figure 17.

# Switch the operating mode of Device D to IRF.

<Sysname> system-view [Sysname] chassis convert mode irf The device will switch to IRF mode and reboot. You are recommended to save the current

running configuration and specify the configuration file for the next startup. Continue? [Y/N]:y

Do you want to convert the content of the next startup configuration file flash:/startup.cfg to make it available in IRF mode? [Y/N]:y

Please wait... Saving the converted configuration file to the main board succeeded. Slot 1: Saving the converted configuration file succeeded. Now rebooting, please wait...

Device D joins the IRF fabric after reboot. Now, the IRF fabric has four member switches.

5. Configure BFD MAD

# Create VLAN 3, and add ports GigabitEthernet 1/2/0/1, GigabitEthernet 2/3/0/1, GigabitEthernet 3/3/0/1, and GigabitEthernet 4/3/0/1 to VLAN 3.

<Sysname> system-view [Sysname] vlan 3 [Sysname-vlan3] port gigabitethernet 1/2/0/1 gigabitethernet 2/3/0/1 gigabitethernet

3/3/0/1 gigabitethernet 4/3/0/1 [Sysname-vlan3] quit

# Create VLAN-interface 3 and configure a MAD IP address for the interface.

[Sysname] interface vlan-interface 3 [Sysname-Vlan-interface3] mad bfd enable

[Sysname-Vlan-interface3] mad ip address 192.168.2.1 24 member 1 [Sysname-Vlan-interface3] mad ip address 192.168.2.2 24 member 2 [Sysname-Vlan-interface3] mad ip address 192.168.2.3 24 member 3 [Sysname-Vlan-interface3] mad ip address 192.168.2.4 24 member 4 [Sysname-Vlan-interface3] quit

# Disable the spanning tree function on ports GigabitEthernet 1/2/0/1, GigabitEthernet 2/3/0/1, GigabitEthernet 3/3/0/1, and GigabitEthernet 4/3/0/1.

[Sysname] interface gigabitethernet 1/2/0/1 [Sysname-gigabitethernet-1/2/0/1] undo stp enable [Sysname-gigabitethernet-1/2/0/1] quit [Sysname] interface gigabitethernet 2/3/0/1 [Sysname-gigabitethernet-2/3/0/1] undo stp enable [Sysname-gigabitethernet-2/3/0/1] quit [Sysname] interface gigabitethernet 3/3/0/1 [Sysname-gigabitethernet-3/3/0/1] undo stp enable [Sysname-gigabitethernet-3/3/0/1] quit [Sysname] interface gigabitethernet 4/3/0/1 [Sysname-gigabitethernet-4/3/0/1] undo stp enable [Sysname-gigabitethernet-4/3/0/1] quit

6. Configure Device E as the intermediate switch

To make Device E act as the intermediate switch to transparently forwards BFD protocol packets exchanged between the member switches for MAD detection, add all ports for BFD MAD on Device E to one VLAN dedicated for BFD MAD. (Details not shown)

7. Troubleshoot the IRF when the MAD mechanism detects errors

Because the member switches are connected with a ring connection, when a single IRF link fails, the connection turns into a daisy chain connection and the IRF is not partitioned. When both the IRF link between Device A and Device C and the IRF link between Device B and Device D go down, however, the IRF fabric will break into two IRF fabrics: one (IRF 1) with Device A and Device C, and one (IRF 2) with Device B and Device D. In addition, IRF link status error and card failure messages will be displayed. The following takes IRF 1 for example:

%May 6 15:10:05:477 2010 HP MAD/1/MAD_COLLISION_DETECTED: Multi-active devices detected, please fix it.

# Fix the IRF link. When the IRF link recovers, an IRF merge occurs, and you must reboot the IRF fabric.

%May 6 15:12:52:935 2010 HP STM/6/STM_LINK_STATUS_UP: IRF port 1 is up. %May 6 15:13:02:828 2010 HP STM/4/STM_MERGE_NEED_REBOOT: IRF merge occurs and the IRF system needs a reboot.

# Log in to the console port of Device B or Device D and reboot IRF 2.

<Sysname> reboot Start to check configuration with next startup configuration file, please wait.

.........DONE!

This command will reboot the device. Continue? [Y/N]:y #May 6 15:31:09:724 2010 HP DEVM/1/REBOOT:

Reboot device by command. %May 6 15:31:09:734 2010 HP DEVM/5/SYSTEM_REBOOT: System is rebooting now.

# After IRF 2 starts up, it re-joins IRF 1. You can use the display irf topology command to view the topology information of the IRF fabric.

<Sysname> display irf topology Topology Info

------------------------------------------------------------------------ IRF-Port1 IRF-Port2 Switch Link neighbor Link neighbor Belong To 1 UP 3 UP 2 00e0-fc0f-8c0f 2 UP 1 UP 4 00e0-fc0f-8c0f 3 UP 4 UP 1 00e0-fc0f-8c0f 4 UP 2 UP 3 00e0-fc0f-8c0f

Switching the operating mode of IRF member switches from IRF to standalone

Network requirements

As shown in Figure 18, Device A and Device B are member switches of an IRF fabric.

Switch the operating mode of Device A and Device B from IRF to standalone.

Figure 18 An IRF fabric

GE4/0/1 GE4/0/2

GE1/4/0/2

XGE1/3/0/1

Device A Device B

(IRF-port1/2)

Configuration considerations

• Disconnect IRF connections. Unplug cables for the IRF connections or shut down all physical IRF

ports on the master at the CLI. This example uses the latter method.

Device C

IP network

IRF

GE2/4/0/2

XGE2/3/0/1 (IRF-port2/1)

• After the IRF fabric is partitioned, switch the operating mode of the two member switches from IRF

to standalone.

Configuration procedure

1. Identify the master.

<IRF> display irf

Switch Slot Role Priority CPU-Mac Description *+1 0 Master 1 00e0-fc0a-15e0 DeviceA 1 1 Slave 1 00e0-fc0f-8c02 DeviceA 2 0 Slave 1 00e0-fc0f-15e1 DeviceB 2 1 Slave 1 00e0-fc0f-15e2 DeviceB

--------------------------------------------------

* indicates the device is the master. + indicates the device through which the user logs in.

The Bridge MAC of the IRF is: 000f-e26a-58ed Auto upgrade : no Mac persistent : always Domain ID : 0

The output shows that Device A is the master.

2. Disconnect IRF connections: disable the physical IRF port Ten-Gigabitethernet 1/3/0/1 on the

master at the CLI. (In this example, there is only one IRF physical link.)

<IRF> system-view [IRF] interface ten-gigabitethernet 1/3/0/1 [IRF-Ten-Gigabitethernet1/3/0/1] shutdown [IRF-Ten-Gigabitethernet1/3/0/1] quit

3. Save the current configuration to the startup configuration file.

[IRF] save The current configuration will be written to the device. Are you sure? [Y/N]:y Please input the file name(*.cfg)[flash:/startup.cfg] (To leave the existing filename unchanged, press the enter key): flash:/startup.cfg exists, overwrite? [Y/N]:y

Validating file. Please wait.....................................

The current configuration is saved to the active main board successfully. Configuration is saved to device successfully.

4. Switch the operating mode of Device A from IRF to standalone.

[IRF] undo chassis convert mode The device will switch to stand-alone mode and reboot. You are recommended to save the

current running configuration and specify the configuration file for the next startup. Continue? [Y/N]:y

Do you want to convert the content of the next startup configuration file flash:/vrpcfg.cfg to make it available in stand-alone mode? [Y/N]:y

Please wait.............

Saving the converted configuration file to main board succeeded. Chassis 1 Slot 1: Saving the converted configuration file succeeded. Now rebooting, please wait...

Device A automatically reboots to complete the operating mode switch.

5. Log in to Device B and switch the operating mode of Device B from IRF to standalone.

<IRF> system-view [IRF] undo chassis convert mode

The device will switch to stand-alone mode and reboot. You are recommended to save the

current running configuration and specify the configuration file for the next startup. Continue? [Y/N]:y

Do you want to convert the content of the next startup configuration file flash:/vrpcfg.cfg to make it available in stand-alone mode? [Y/N]:y

Please wait.............

Saving the converted configuration file to main board succeeded. Chassis 2 Slot 1: Saving the converted configuration file succeeded. Now rebooting, please wait...

Device B automatically reboots to complete the operating mode switch.

NOTE:

If you have created a VLAN interface and configured an IP address for the interface on the IRF fabric, and both Device A and Device B have member ports of this VLAN, IP address collision occurs after the operating mode of Device A and Device B switches to standalone. To remove the collision, lo device and modify the IP address of this VLAN interface.

in to one

Index

A B C D E I P S

Accessing an IRF fabric,18

sic concepts,3

onfiguring IRF member switches in IRF mode,19

Displa

ying and maintaining an IRF fabric,33

stablishment, operation, and maintenance of an IRF

fabric,6

RF configuration fast recovery,32

RF fabric configuration examples,34 RF fabric configuration task list,10

overview,1

IRF

e-configuring an IRF member switch in standalone

Pr mode,15

aving configuration to the startup configuration

file,17 S

witching operating mode,17

Support and other resources

Contacting HP

For worldwide technical support information, see the HP support website:

http://www.hp.com/support

Before contacting HP, collect the following information:

• Product model names and numbers

• Technical support registration number (if applicable)

• Product serial numbers

• Error messages

• Operating system type and revision level

• Detailed questions

Subscription service

HP recommends that you register your product at the Subscriber's Choice for Business website:

http://www.hp.com/go/wwalerts

After registering, you will receive email notification of product enhancements, new driver versions, firmware updates, and other product resources.

Related information

Documents

To find related documents, browse to the Manuals page of the HP Business Support Center website:

http://www.hp.com/support/manuals

• For related documentation, navigate to the Networking section, and select a networking category.

• For a complete list of acronyms and their definitions, see HP A-Series Acronyms.

Websites

• HP.com http://www.hp.com

• HP Networking http://www.hp.com/go/networking

• HP manuals http://www.hp.com/support/manuals

• HP download drivers and software http://www.hp.com/support/downloads

• HP software depot http://www.software.hp.com

• HP Education http://www.hp.com/learn

HP A7503, A7510, A7506-V, A7506 Configuration Manual

Specifications and Main Features

Frequently Asked Questions

User Manual

Contents

IRF configuration

Hardware compatibility

Switch model compatibility

Card compatibility

Hardware compatibility for an IRF fabric with three or four members

IRF overview

Benefits

Application scenario

Basic concepts

Operating mode

IRF member switch roles

Active MPU of a member switch

Standby MPUs of a member switch

Active MPU of an IRF fabric

Standby MPUs of an IRF fabric

IRF port

Physical IRF port

IRF partition

IRF merge

Member priority

Establishment, operation, and maintenance of an IRF fabric

Connecting the IRF member switches

Connection medium

Connection requirements

IRF topology

Topology collection

Master election

IRF fabric management and maintenance

Member ID

IRF fabric topology maintenance

IRF multi-active detection

IRF fabric configuration task list

Configuring global load sharing criteria

Configuring port-specific load sharing criteria

Pre-configuring an IRF member switch in standalone mode

Setting a member ID for the switch

Specifying a priority for the switch

Binding physical ports to IRF ports

Saving configuration to the startup configuration file

Switching operating mode

IRF modes

Configuration file auto-conversion

Switching the operating mode to IRF mode

Accessing an IRF fabric

Accessing the active MPU of an IRF fabric

Accessing a standby MPU of an IRF fabric

Configuring IRF member switches in IRF mode

Specifying an IRF domain ID for an IRF fabric

Introduction to IRF domains

Assigning a domain ID to an IRF fabric

Modifying a member ID for a switch

Specifying a priority for a member switch

Configuring IRF ports

Enabling auto reboot for IRF fabric merge

Configuring a description for a member switch

Configuring load sharing criteria for IRF links

Configuring global load sharing criteria

Configuring port-specific load sharing criteria

Specifying the preservation time of bridge MAC address

Enabling automatic system software updating

Setting the IRF link down report delay

Configuring MAD detection

Configuring LACP MAD

Configuring BFD MAD

Excluding a port from the shut down action upon detection of multi-active collision

Manually recovering an IRF fabric

IRF configuration fast recovery

Configuration prerequisites

Configuration procedures

Displaying and maintaining an IRF fabric

IRF fabric configuration examples

LACP MAD detection-enabled IRF configuration example (with two member switches)

Network requirements

Configuration considerations

Configuration procedure