HP A7503, A7510, A7506-V, A7506 Configuration Manual
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
© 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.
Contents
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
i
Index ··········································································································································································· 56
ii
IRF configuration
g
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
1
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
2
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.
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
4
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
n
/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.
5
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.
6
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.
7
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.
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.
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.
8
Master election
r
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
9
topology database to determine whether the leaving switch is the master. If yes, the member switch starts
A
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
.
.
10
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
11
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
12
Task
Configuring a description for a member switch
Configuring load sharing criteria for IRF links
13
Task
c
e
3
g
e
g
r
f
e
g
o
l
e
A
e
d
n
o
a
C
n
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
14
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
15
Optional
The priority of the switch defaults
to 1.
Loading...