HP A5820X Series User manual

HP A5820X & A5800 Switch Series
IRF
Configuration Guide
Abstract
This document describes the software features for the HP A Series products and guides you through the
software configuration procedures. These configuration guides also provide configuration examples to
help you apply software features to different network scenarios.
This documentation is intended for network planners, field technical support and servicing engineers,
and network administrators working with the HP A Series products.
Part number: 5998-1626
Software version: Release 1211
Document version: 5W100-20110430

Legal and notice information
© Copyright 2011 Hewlett-Packard Development Company, L.P.
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.

iii
Contents
IRF configuration ·························································································································································· 1
IRF overview ······································································································································································1
Benefits ······································································································································································1
Application scenario················································································································································2
IRF topologies ···························································································································································2
Basic concepts···································································································································································3
Establishment, operation, and maintenance of an IRF virtual device··········································································4
Connecting the IRF member switches·····················································································································4
Topology collection··················································································································································7
Master election ·························································································································································7
IRF virtual device management and maintenance ································································································7
IRF multi-active detection··········································································································································9
IRF virtual device configuration task list······················································································································· 10
Configuring an IRF virtual device ································································································································· 11
Specifying a domain ID for an IRF virtual device ······························································································ 11
Changing the IRF member ID of a switch ··········································································································· 13
Configuring IRF ports ············································································································································ 13
Specifying a priority for a member switch ········································································································· 15
Configuring a description for a member switch ································································································ 15
Configuring load sharing criteria for IRF links ··································································································· 15
Specifying the preservation time of bridge MAC address ··············································································· 17
Enabling automatic boot file updating················································································································ 18
Setting the IRF link down report delay ················································································································ 19
Configuring MAD detection·········································································································································· 19
Configuring LACP MAD ······································································································································· 20
Configuring BFD MAD·········································································································································· 22
Accessing an IRF virtual device ···································································································································· 28
Accessing the master ············································································································································ 28
Accessing a slave switch······································································································································ 29
Displaying and maintaining an IRF virtual device······································································································ 29
IRF virtual device configuration examples ··················································································································· 30
LACP MAD detection-enabled IRF configuration example················································································ 30
BFD MAD detection-enabled IRF configuration example·················································································· 32
ARP MAD detection-enabled IRF configuration example·················································································· 35
Support and other resources ·····································································································································38
Contacting HP ································································································································································ 38
Subscription service ·············································································································································· 38
Related information························································································································································ 38
Documents······························································································································································ 38
Websites ································································································································································ 38
Conventions ···································································································································································· 39
Index ···········································································································································································91H41

1
IRF configuration
Establish an IRF virtual device that comprises switches of the HP A5820X series or A5800 series, or
establish a mixed IRF virtual device that comprises both the HP A5820X and A5800 switches.
IRF overview
The HP proprietary IRF technology creates a large IRF virtual device from multiple switches to provide
data center class availability and scalability. IRF virtualization technology takes advantage of the
augmented processing power, interaction, unified management, and uninterrupted maintenance of
multiple switches.
Benefits
IRF delivers the following benefits:
•Simplified topology and streamlined management. An IRF virtual device appears as one node on
the network. Log in at any member switch to manage all members of the IRF virtual device.
•High availability and reliability. The member switches in an IRF virtual device work in 1:N
redundancy. One member switch works as the master to manage and maintain the entire IRF virtual
device. All other member switches process services as well as back up the master. As soon as the
master fails, all other member switches elect a new master among them to prevent service
interruption. In addition, you can perform link aggregation not only for IRF links, but also for
physical links between the IRF virtual device and its upper or lower layer devices for link
redundancy.
•Network scalability and resiliency. Increase ports, bandwidth, and processing capability of an IRF
virtual device simply by adding member switches.

2
Application scenario
Figure 1 shows an IRF virtual device 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 virtual device
IP network
IRF link Equal to
Master Slave
IRF topologies
Create an IRF virtual device in daisy chain topology, or more reliably, ring topology, as shown in Figure
2.
In ring chain topology, the failure of one IRF link does not cause the IRF virtual device to split as in daisy
chain topology. Rather, the IRF virtual device changes to a daisy chain topology without affecting
network services.
Figure 2 IRF connections

3
Basic concepts
IRF member switch roles
IRF uses two member switch roles: master and subordinate (slave).
When switches form an IRF virtual device, they elect a master to manage the IRF virtual device, 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 election.”
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.
Physical IRF port
Physical IRF ports are physical (copper or fiber) ports bound to an IRF port. They connect IRF member
switches and forward IRF protocol packets and data traffic between IRF member switches.
For more information about ports that can be used as IRF physical ports on the A5800 and A5820X
switches, see “Connecting the IRF member switches.”
IRF partition
IRF partition occurs when an IRF virtual device splits into two or more IRF virtual devices because of IRF
link failures, as shown in Figure 3. The partitioned IRF virtual devices operate with the same IP address
and cause routing and forwarding problems on the network.
Figure 3 IRF partition
IRF merge
IRF merge occurs when two partitioned IRF virtual devices re-unite or when you configure and connect
two independent IRF virtual devices to be one IRF virtual device, as shown in Figure 4.
Figure 4 IRF merge

4
Member priority
Member priority determines the role that a member switch can play in an IRF virtual device. A member
with a higher priority is more likely to be a master. The member priority of a switch is user configurable,
and defaults to 1. Modify the priority at the CLI.
Establishment, operation, and maintenance of an
IRF virtual device
IRF virtual device management involves these stages: Connecting the IRF member switches, Topology
collection, Master election, IRF virtual device management and maintenance, and IRF multi-active
detection.
Connecting the IRF member switches
Connecting the IRF member switches includes two tasks:
•Binding physical ports to IRF ports at the CLI
•Connecting the neighbor switches
Binding physical ports to IRF ports at the CLI
Bind one physical port, or for link redundancy, multiple physical ports, to an IRF port (see “Configuring
IRF ports“). Table 1 shows the physical ports that can be used as IRF ports and the port use restrictions.
Table 1 Physical IRF port requirements
Switch model Candidate
p
h
y
sical IRF
p
orts Re
q
uirements
•A5800-48G-PoE+
Switch with 2 Interface
Slots (JC101A)
•A5800-48G-PoE+
TAA Switch with 2
Interface Slots
(JG242A)
Ports on the two expansion cards (if
installed) on the front panel
All physical ports of an IRF port must be
located on the same interface card.
A5800AF-48G Switch
(JG225A)
The six fixed SFP+ ports (in two
groups) on the front panel:
•The rightmost two SFP+ ports in
one group
•The rest four SFP+ ports in the
other group
All physical ports of an IRF port must be
in the same group.

5
Switch model Candidate
p
h
y
sical IRF
p
orts Re
q
uirements
•A5800-48G Switch
with 1 Interface Slot
(JC105A)
•A5800-48G TAA
Switch with 1 Interface
Slot (JG258A)
•A5800-48G-PoE+
Switch with 1 Interface
Slot (JC104A)
•A5800-48G-PoE+
TAA Switch with 1
Interface Slot
(JG257A)
•The four fixed SFP+ ports on the
front panel
•Ports on the expansion card (if
installed) on the rear panel
All physical ports of an IRF port must be
located on the front panel or the
expansion interface card on the rear
panel.
•A5800-24G Switch
(JC100A)
•A5800-24G TAA
Switch (JG255A)
•A5800-24G-PoE+
Switch (JC099A)
•A5800-24G-PoE+TAA
Switch (JG254A)
•The four fixed SFP+ ports on the
front panel
•Ports on the expansion interface
card (if installed) on the rear
panel
An IRF port can use physical ports
distributed on different cards.
•A5800-24G-SFP
Switch with 1 Interface
Slot (JC103A)
•A5800-24G-SFP TAA
Switch with 1 Interface
Slot (JG256A)
The four fixed SFP+ ports and the
ports on the expansion interface
card (if installed) on the front panel
•A5820X-14XG-SFP+
Switch with 2 Interface
Slots (JC106A)
•A5820X-14XG-SFP+
TAA Switch with 2
Interface Slots
(JG259A)
The 14 fixed SFP+ ports and the
ports on the expansion interface
cards (if installed) on the front panel
•A5820X-24XG-SFP+
Switch (JC102A)
•A5820X-24XG-SFP+
TAA-compliant Switch
(JG243A)
The 24 SFP+ ports on the front
panel
No location limitation to the physical IRF
ports.
HP A5820AF-24XG
Switch
For a complete list of supported SFP+ ports and interface cards, see HP A5800 Switch Series Installation
Guide and HP A5820X Switch Series Installation Guide.

6
Connecting the neighbor switches
Connect the physical ports of IRF-Port1 on one switch to the physical ports of IRF-Port2 on its neighbor
switch, as shown in Figure 5.
Figure 5 IRF virtual device physical connection
Table 2 lists the SFP+ transceiver modules, optical fibers, and SFP+ cables available for connecting the
physical IRF ports on the HP A5800 and A5820X switches.
Table 2 SFP+ transceivers, optical fibers, and SFP+ cables available for physical IRF ports
Product code Module
description
Central
wavelengt
h (in nm)
Connector Fiber
specifications
Maximum
transmission distance
JD092B
HP X130 10G
SFP+ LC SR
Transceiver
850 LC
50/125 μm
multimode fiber
300 m (984.25 ft)
82 m (269.03 ft)
66 m (216.54 ft)
62.5/125 μm
multimode fiber
33 m (108.27 ft)
26 m (85.30 ft)
JD093B
HP X130 10G
SFP+ LC LRM
Transceiver
1310 LC
62.5/125 μm
multimode fiber 220 m (721.78 ft)
50/125 μm
multimode fiber
220 m (721.78 ft)
100 m (328.08 ft)
JD094B
HP X130 10G
SFP+ LC LR
Transceiver
1310
LC 9/125 μm
single-mode fiber
10 km (6.21 miles)
1550 40 km (24.86 miles)
JD095B
HP X240 10G
SFP+ SFP+
0.65m DA
Cable
N/A N/A SFP+ cable
0.65 m (2.1 ft)
JD096B
HP X240 10G
SFP+ SFP+
1.2m DA
Cable
1.2 m (3.9 ft)
JD097B
HP X240 10G
SFP+ SFP+ 3m
DA Cable
3 m (9.8 ft)
JG081B
HP X240 10G
SFP+ SFP+ 5m
DA Cable
5 m (16.4 ft)

7
If the IRF member switches are located far from each other, use the SFP+ transceivers with optical fibers.
If the IRF member switches are all in one equipment room, use the SFP+ cable.
For more information about the interface modules, see HP A-Series Switches Transceiver Modules User
Guide.
The SFP+ modules and SFP+ cables available for this switch series are subject to change over time. For
the most up-to-date list of SFP+ modules and cables, contact HP technical support or marketing staff.
Topology collection
Each member switch exchanges IRF hello packets with its directly connected neighbors to collect
topology data, including IRF port connection states, member IDs, priorities, and bridge MAC addresses.
Each member switch has a local topology database. At startup, an IRF member switch has only local
topology data. When an IRF port goes up, the member switch sends its topology data out of the port
periodically. The neighbor switch then updates its topology database with the received topology data.
The topology collection lasts for a period of time. After all members eventually get complete topology
information (topology convergence), the IRF virtual device enters the next stage: master election.
Master election
Master election is held each time the topology changes, for example, when the IRF virtual device is
established, a new member switch is plugged in, the master switch fails or is removed, or the partitioned
IRF virtual devices merge.
The master is elected based on the following rules in descending order:
1. The current master, even if a new member has a higher priority. When an IRF virtual device is
being formed, all member switches consider themselves as the master, this rule is skipped.
2. The member with a higher priority.
3. The member with the longest system up-time. The member switches exchange system up-time in the
IRF hello packets.
4. The member with the lowest bridge MAC address
The IRF virtual device is formed on election of the master.
During an IRF merge, the switches of the IRF virtual device that fails the master election automatically
reboot to join the IRF virtual device that wins the election.
After a master election, all subordinate member switches initialize and reboot with the configuration on
the master, and their original configuration, even if has been saved, is lost.
IRF virtual device management and maintenance
After the IRF virtual device is established, access the master from any member switch to manage all the
resources of the member switches.
Member ID
An IRF virtual device uses member IDs to uniquely identify its members. Member IDs are also included in
interface names and file system names for interface and file system identification. To guarantee the
operation of the IRF virtual device, you must assign each member switch a unique member ID.

8
Interface naming conventions
The interfaces are named in the format of member ID/subslot number/interface serial number, where
•The member ID identifies the IRF member switch on which the interface resides. If the switch is
standalone, the member ID defaults to 1. If the standalone switch was once an IRF member switch,
it uses the same member ID as it was in the IRF virtual device.
•The subslot number is the number of the slot in which the interface card resides. On the A5800
series or A5820X series, the subslot for the fixed ports on the front panel is numbered 0. If the
switch has one expansion slot, the number of the slot is 1. If the switch has two expansion slots,
their numbers are 1 and 2, from left to right.
•The interface serial number depends on the number of interfaces provided by the switch. Look at
the number on the silkscreen on the interface card for the number of supported interfaces.
For example, on the standalone switch Sysname, GigabitEthernet 1/0/1 represents the first fixed port
on the front panel. Set its link type to trunk:
<Sysname> system-view
[Sysname] interface gigabitethernet 1/0/1
[Sysname-GigabitEthernet1/0/1] port link-type trunk
For another example, on the IRF virtual device Master, GigabitEthernet 3/0/1 represents the first fixed
port on the front panel of member switch 3. Set its link type to trunk:
<Master> system-view
[Master] interface gigabitethernet 3/0/1
[Master-GigabitEthernet3/0/1] port link-type trunk
File system naming conventions
On a standalone switch, use the name of storage device to access its file system. For more information
about storage device naming conventions, see the chapter “File management” in Fundamentals
Configuration Guide.
On an IRF virtual device, you can also use the name of storage device to access the file system of the
master. To access the file system of any other member switch, use the name in the following format:
Member-ID#Storage-device-name. For example:
1. To access the test folder under the root directory of the Flash on the master switch, perform the
following steps:
<Master> mkdir test
...
%Created dir flash:/test.
<Master> dir
Directory of flash:/
0 -rw- 10105088 Apr 26 2000 13:44:57 test.app
1 -rw- 2445 Apr 26 2000 15:18:19 config.cfg
2 drw- - Jul 14 2008 15:20:35 test
30861 KB total (20961 KB free)
2. To create and access the test folder under the root directory of the Flash on member switch 3,
perform the following steps:
<Master> mkdir slot3#flash:/test
%Created dir slot3#flash:/test.
<Master> cd slot3#flash:/test
<Master> pwd
slot3#flash:/test

9
Or:
<Master> cd slot3#flash:/
<Master> mkdir test
%Created dir slot3#flash:/test.
3. To copy the test.app file on the master to the root directory of the Flash on member switch 3,
perform the following steps:
<Master> pwd
slot3#flash:
//The current working path is the root directory of the Flash on slave 3.
<Master> cd flash:/
<Master> pwd
flash:
//The current working path is the root directory of the Flash on the master.
<Master> copy test.app slot3#flash:/
Copy flash:/test.app to slot3#flash:/test.app?[Y/N]:y
%Copy file flash:/test.app to slot3#flash:/test.app...Done.
Configuration file synchronization
IRF uses a strict configuration file synchronization mechanism to ensure that all switches in an IRF virtual
device can work as a single node on the network, and to ensure that after the master fails, the other
switches can operate normally.
•When a subordinate switch starts up, it automatically gets and runs the master's configuration file.
If all switches in an IRF virtual device start up simultaneously, the subordinate switches get and run
the master's startup configuration file.
•Any configuration you made on the IRF virtual device is stored on the master and synchronized in
real time to each member switch. When you save the current configuration to the startup
configuration file of the master by using the save command, all subordinate switches execute the
same saving operation.
This real-time configuration synchronization ensures that all the IRF member switches keep the same
configuration file. If the master fails, all the other switches can still operate with the same configuration
file.
IRF virtual device 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.
An IRF port goes down only when all its physical IRF ports are down.
IRF multi-active detection
An IRF link failure causes an IRF virtual device to split in two IRF virtual devices operating with the same
Layer 3 configurations, such as the same IP address. To avoid IP address collision and network
problems, IRF uses the MAD mechanism to detect the presence of multiple identical IRF virtual devices
and handle collisions. MAD provides the following functions:
1. Detection

10
MAD detects active IRF devices with the same Layer 3 global configuration by extending the LACP, the
BFD protocol, or the gratuitous ARP. For more information, see “Configuring MAD detection.”
2. Collision handling
If multiple identical active IRF virtual devices are detected, only the one that has the lowest master ID can
operate in active state and forward traffic normally. MAD sets all other IRF virtual devices in the recovery
state (disabled) and shuts down all physical ports but the IRF ports and any other ports you have
specified with the mad exclude interface command.
3. Failure recovery
An IRF link failure triggers IRF virtual device partition and causes multi-active collision. In this case, repair
the failed IRF link to make the collided IRF virtual devices merge into one and recover the failure. If the
IRF virtual device in the recovery state fails before the failure is recovered, repair both the failed IRF
virtual device and the failed IRF link, and then the collided IRF virtual devices can merge into one and
the failure is recovered. If the IRF virtual device in the active state fails before the failure is recovered,
enable the IRF virtual device in the recovery state at the CLI to make it take over the active IRF virtual
device and protect the services from being affected. Then, recover the MAD failure.
For more information about LACP, see Layer 2—LAN Switching Configuration Guide. For information
about BFD, see High Availability Configuration Guide. For information about gratuitous ARP, see Layer
3—IP Services Configuration Guide.
IRF virtual device configuration task list
Before configuring an IRF virtual device, plan the roles and functions of all member switches. HP
recommends the configuration procedure in Figure 6.
Figure 6 IRF configuration flow chart
Connect physical IRF ports with SFP+ cables or fibers after activating IRF port configurations. After the
device detects that the IRF ports are connected normally, master election starts immediately, and then the
elected subordinate switches reboot automatically.
After an IRF virtual device is formed, configure and manage the IRF virtual device by logging in to any
device in the IRF.
Complete the following tasks to configure an IRF virtual device:
Task Remarks
Specifying a domain ID for an IRF virtual device Optional.
Changing the IRF member ID of a switch Required.
Configuring IRF ports Required.
Specifying a priority for a member switch Optional.
Configuring a description for a member switch Optional.

11
Task Remarks
Configuring load sharing criteria for IRF links Optional.
Specifying the preservation time of bridge MAC address Optional.
Enabling automatic boot file updating Optional.
Setting the IRF link down report delay Optional.
Connect the physical IRF ports of devices and make sure that the physical IRF ports are interconnected (a
ring connection is recommended).
Configuring MAD
detection
Configuring LACP MAD Optional.
Use one of the approaches
Configure the MAD
detection after an IRF
virtual device is
established.
Configuring BFD MAD
Configuring ARP MAD
Excluding a port from the shutdown action on
detection of multi-active collision Optional.
Manually recovering an IRF virtual device Optional.
Accessing an IRF
virtual device
Accessing the master Required.
Accessing a slave switch Optional.
Configuring an IRF virtual device
Specifying a domain ID for an IRF virtual device
Introduction to IRF domain
To differentiate IRF virtual devices, each IRF virtual device is assigned a domain ID.
As shown in Figure 7, Switch A and Switch B form IRF virtual device 1, and Switch C and Switch D form
IRF virtual device 2. If there is a MAD detection link between the two IRF virtual devices, they send MAD
detection packets to each other through the detection link. The system statuses and operations of both IRF
virtual devices are affected. To solve this problem, specify different domain IDs for the two IRF virtual
devices.
After assigning a domain ID to an IRF virtual device, the extended LACPDUs sent by the member
switches carry the IRF domain information to distinguish the LACP detection packets from different IRF
virtual devices.

12
Figure 7 Network diagram for multiple domains
Switch A Switch B
IRF virtual device 1
(domain 10)
IRF link
Core network
IRF virtual device 2
(domain 20)
IRF link
Switch C Switch D
Access network
Assigning a domain ID to an IRF virtual device
If LACP MAD detection is enabled on multiple IRF virtual devices, and LACP MAD detection links exist
among the IRF virtual devices, assign different domain IDs to the IRF virtual devices.
If there is no LACP MAD detection link among IRF virtual devices, or BFD MAD or ARP MAD is used, you
do not need to assign domain IDs to them.
To assign a domain ID to an IRF virtual device:
To do… Use the command… Remarks
1. Enter system view system-view —
2. Assign a domain ID to the
IRF virtual device irf domain domain-id
Required if LACP MAD is adopted.
Optional if ARP MAD is adopted.
By default, the domain ID of an IRF
virtual device is 0.
You must assign a domain ID for an IRF virtual device before enabling LACP MAD detection.
Although switches with different domain IDs can form an IRF virtual device, HP recommends that you
assign the same domain ID to the members of the same IRF virtual device. Otherwise, the LACP MAD
detection function cannot function properly.
To display the domain IDs and verify your configuration, execute the display irf command in any view.

13
Changing the IRF member ID of a switch
CAUTION:
Change member ID for the switches in an IRF virtual device with caution. The change might cause
configuration change and even data loss. Consider an IRF virtual device that comprises three member
switches of the same model with member IDs 1, 2, and 3. If you change the member ID of switch 2 to 3
and that of switch 3 to 2, then switch 2 uses the original port configurations of switch 3, and switch 3
uses those of switch 2 after they are rebooted.
An IRF virtual device uses member IDs to uniquely identify its members. After you change the member ID
of a switch, you must reboot the switch to validate the setting.
•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 device 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 device still take effect, other configurations
(such as configuration for physical IRF ports) no longer take effect and you will need to configure
them again.
Change the IRF member ID of a switch when it is standalone or after it joins an IRF virtual device. If the
switch is standalone, make sure that the member ID of the switch does not conflict with the member ID of
any other switch, so the change does not affect the operation of the IRF virtual device. After changing
the member ID, save the current configuration, power off the switch, connect the switch to its neighbor
switch, power it on, and configure the IRF port to enable IRF on the switch.
To change the IRF member ID of a switch:
To do… Use the command… Remarks
1. Enter system view system-view —
2. Change the IRF member ID
of the switch
irf member member-id renumber new-
member-id
Optional.
The member ID of a switch
defaults to 1.
Verify the IRF member ID of a switch with the display irf configuration command.
Configuring IRF ports
To bring the IRF function into work, you must connect the IRF member switches, assign the connected
physical ports to the appropriate IRF port on each member switch, and activate the IRF port
configuration. After the IRF port configuration is activated, the IRF ports go up, a master election is held,
and the switches that failed in the election automatically reboot to join the IRF virtual device as
subordinate switches.
When binding a physical port, use the display irf topology command to check that its link state is DIS or
DOWN.

14
To configure an IRF port:
To do… Use the command… Remarks
1. Enter system view system-view —
2. Enter the view of the port
you are binding to the IRF
port
interface interface-type interface-
number —
3. Shut down the port shutdown Required.
4. Return to system view quit —
5. Create the IRF port and enter
IRF port view irf-port member-id/port-number —
6. Bind the physical port to the
IRF port
port group interface interface-type
interface-number [ mode { enhanced
| normal } ]
Required.
By default, no physical port is
bound to any IRF port.
If you specify a binding mode
with the mode keyword, ensure
that the two ends of the IRF link
are using the same mode.
7. Return to system view quit —
8. Enter physical IRF port view interface interface-type interface-
number —
9. Bring up the physical port undo shutdown Required.
10. Return to system view quit —
11. Save the current
configuration save Required.
12. Activate the IRF port
configuration irf-port-configuration active Required.
Bind up to four physical ports to an IRF port for link redundancy and load sharing. The physical ports
must meet the requirements in Table 1.
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.
Before unplugging an interface card that contains any IRF physical port, unplug the cable of the port or
shut down the port by using the shutdown command in IRF physical port view.
Perform only the shutdown, description and flow-interval commands on the physical port bound to an
IRF port. For more information about the shutdown, description, and flow-interval commands, see Layer
2—LAN Switching Command Reference.

15
Specifying a priority for a member switch
The greater the priority value, the higher the priority. A member with a higher priority is more likely to
be a master.
To specify a priority for a member switch:
To do… Use the command… Remarks
1. Enter system view system-view —
2. Specify a priority for a
member of an IRF virtual
device
irf member member-id priority
priority
Optional.
The priority of a member
defaults to 1.
The priority setting takes effect immediately after configuration without the need to reboot the switch.
Configuring a description for a member switch
Configure a description for a member switch to identify its physical location, or for any other
management purpose.
To configure a description for a member switch:
To do… Use the command… Remarks
1. Enter system view system-view —
2. Configure a description for a
member switch
irf member member-id description
text
Optional
Not configured by default
Configuring load sharing criteria for IRF links
Bind multiple physical ports to an IRF port for link redundancy and load sharing. You can also configure
the switch to distribute traffic across the physical ports of an IRF port based on one of the following
criteria:
•Source IP address
•Destination IP address
•Source MAC address
•Destination MAC address
•The combination of source and destination IP addresses
•The combination of source and destination MAC addresses
Configure global or IRF port specific load sharing criteria. The switch preferentially uses the port-specific
load sharing criteria. If no port-specific load sharing criteria is available, it uses the global load sharing
criteria.

16
Configuring global load sharing criteria
To configure the global IRF link load sharing criteria:
To do… Use the command… Remarks
1. Enter system view system-view —
2. Configure the global IRF
link load sharing criteria
irf-port load-sharing mode {
destination-ip | destination-mac |
source-ip | source-mac } *
Required.
By default:
•The A5800-48G Switch with 1
Interface Slot (JC105A), A5800-
48G TAA Switch with 1 Interface
Slot (JG258A), A5800-48G-PoE+
Switch with 1 Interface Slot
(JC104A), and A5800-48G-PoE+
TAA Switch with 1 Interface Slot
(JG257A) base load sharing
calculation on the incoming port
number.
•Other models in the A5800 and
A5820X series use the
combination of the source and
destination MAC addresses as the
load sharing criteria for Layer 2
packets, and the combination of
the source and destination IP
addresses for Layer 3 packets.
Configuring port-specific load sharing criteria
To configure the port-specific load sharing criteria:
To do… Use the command… Remarks
1. Enter system view system-view —
2. Enter IRF port view irf-port member-id/port-number —

17
To do… Use the command… Remarks
3. Configure the port-
specific load sharing
criteria
irf-port load-sharing mode {
destination-ip | destination-mac |
source-ip | source-mac } *
Required.
By default:
•The A5800-48G Switch with 1
Interface Slot (JC105A), A5800-
48G TAA Switch with 1 Interface
Slot (JG258A), A5800-48G-PoE+
Switch with 1 Interface Slot
(JC104A), and A5800-48G-PoE+
TAA Switch with 1 Interface Slot
(JG257A) base load sharing
calculation on the incoming port
number.
•Other models in the A5800 and
A5820X series use the
combination of source and
destination MAC addresses as the
load sharing criteria for Layer 2
packets, and the combination of
source and destination IP
addresses for Layer 3 packets.
Specifying the preservation time of bridge MAC address
CAUTION:
•Bridge MAC address change can cause transient traffic interruption.
•When deploying the ARP MAD with MSTP solution, you must enable the IRF virtual device to change
its bridge MAC address as soon as the master leaves.
An IRF virtual device uses the bridge MAC address of the master switch as its bridge MAC address. The
IRF virtual device uses bridge MAC address to identify the IRF virtual device by Layer 2 protocols such as
MSTP and LACP. It uses bridge MAC address for generating an MAC address for a Layer 3 interface.
As with any other node on a switched LAN, this bridge MAC address must be unique for proper
communication.
To avoid duplicate bridge MAC addresses, an IRF virtual device can change its bridge MAC address
automatically after its master leaves, but the change can cause temporary service interruption.
Depending on your network condition, you can enable the IRF virtual device to preserve or change its
bridge MAC address after the master leaves. The following lists available options:
•irf mac-address persistent timer—Preserves the bridge MAC address for 6 minutes after the master
leaves. If the master does not come back before the timer expires, the IRF virtual device uses the
bridge MAC address of the newly elected master as its bridge MAC address. This option avoids
unnecessary switching of bridge MAC address due to a device 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.