Configuring Redundancy

A single chassis with two supervisor engines (redundant supervisor engines)---Each supervisor engine must be equipped with a Policy Feature Card (PFC) and an MSFC. Each MSFC must be running the same release of Cisco IOS software.
Two chassis with a supervisor engine in each---You must have at least one supervisor engine in each chassis. Each supervisor engine must be equipped with a PFC and an MSFC.

Und erstanding How MSFC Redun dancy Works

These sections provide an overview of MSFC redundancy:

Layer 3 Redundancy for Single Chassis

In a single Catalyst 6000 family chassis, you can have redundant supervisor engines, each with an MSFC. You can configure HSRP on the MSFCs to provide automatic routing backup for your network. HSRP is compatible with IPX and AppleTalk. If one MSFC fails, HSRP allows one MSFC (router) to assume the function automatically of the other. Combined with the high availability feature of supervisor engine software release 5.4(1), this configuration provides an added level of redundancy for your network.

You must configure both MSFCs identically. Table 3-1 summarizes the identical requirements and the exceptions for Layer 3 redundancy for a single switch chassis.

Table 3-1: Require ments for Single Chassis Layer 3-Redundancy

Identical Requirements--- Global and Interface Levels
Both MSFCs must have the following:

Same routing protocols

Same VLAN interfaces

Same IOS ACLs¹ ²

All interfaces must have the same administrative status

Exceptions---Interface Level
HSRP standby commands

IP address commands

IP broadcast address

IPX helper address

Exceptions---Global Level
IP default-gateway

IP route prefix mask

IPX internal-network

IPX route

¹Dynamic and reflexive ACLs, which are based on actual data flow, may be programmed by either MSFC.
²In addition to defining the same ACLs on both MSFCs, you must also apply the ACLs to the same VLAN interfaces, in the same direction, on both MSFCs.

Note The Catalyst 6000 family FlexW AN module is supported in a redundant supervisor engine and dual MSFC configuration. However, we do not recommend using the FlexWAN module in a redundant high availability configuration.

The FlexWAN module's WAN interfaces are present only on the designated MSFC (the desi gnated MSFC is the first MSFC to come online, or the MSFC that has been online the longest). If you reload the designated MSFC, the WAN interfaces appear on the newly designated MSFC with no configuration.

You can manually copy the configuration from a TFTP server into NVRAM and then enter the configure memory command; however, if you accidently enter the write memory command before you enter the configure memory command, you will overwrite the startup file.

Note Redundant supervisor engines must have identical hardware (MSFC and PFC). See the "Hardware and Software Requirements" section for more information.

Note A 128-MB DRAM SIMM upgrade is available. This upgrade allows for more packet buffer and larger routing tables (the MSFC can hold over 90K routes). In most cases, the default 64-MB configuration is sufficient. We recommend this upgrade for service provider environments as well as enterprise customers with very large routing tables.

Routing Protocol P eering

In a redundant supervisor engine and dual MSFC configuration, one supervisor engine is fully operational (active) and the other supervisor engine is in standby mode; however, both MSFCs are operational (in terms of programming the Layer 3 cache in the PFC on the active supervisor engine) and act as independent routers. Only the PFC on the active supervisor engine switches the packets using Multilayer Switching (MLS).

Both MSFCs are operational from a routing protocol peering perspective. For example, if you have two MSFCs in a single Catalyst 6000 family switch chassis, each configured with interface VLAN 10 and VLAN 21, the MSFCs are peered to each other over these VLANs. Combined with a dual chassis and dual MSFC design for the same VLANs, each MSFC has 6 peers: its peer in the same chassis as well as the two MSFCs in the second chassis (3 in VLAN 10 and 3 in VLAN 21). See Figure 3-1.

Figure 3-1: Dual Chassis and Dual MSFC Peering

Although the MSFCs (from a peering perspective) act as independent routers, the two MSFCs in the chassis operate at the same time, have the same interfaces, and run the same routing protocols.

If you combine the high availability feature on the supervisor engines with HSRP on the MSFCs, you have the following Layer 2 and Layer 3 redundancy mechanisms:

Layer 2 redundancy for the supervisor engines (one active and one in standby)---If the active supervisor engine fails (the MSFC installed on it will also fail), both Layer 2 and Layer 3 functions roll over to the redundant supervisor engine and MSFC combination.
Layer 3 redundancy and load sharing for the two MSFCs---If one MSFC fails, the other MSFC takes over almost immediately (using HSRP) without any Layer 2 disruption (the active supervisor engine continues to forward Layer 2 traffic).

The Layer 3 entries programmed by the failed MSFC on the active supervisor engine are used until they are gracefully aged out and replaced by the Layer 3 entries populated by the newly active MSFC. This graceful aging occurs over a four-minute period, and allows the newly active MSFC to repopulate the MLS entries using its XTAG value, while concurrently hardware-switching flows yet to be aged. In addition, this process prevents a newly active MSFC from being overwhelmed with initial flow traffic.

Note Each MSFC has its own XTAG value to identify itself as the MLS Route Processor. MSFC-1 (on the active supervisor engine) has an XTAG of 1, and MSFC-2 (on the standby supervisor engine) has an XTAG of 2.

For same-chassis Layer 3 redundancy to function as expected, the configuration on each MSFC must be the same (see Table 3-1).

Note Table 3-1 lists configuration exceptions. For example, in Figure 3-1, there are 4 MSFCs on VLAN 10; therefore, each MSFC has different IP addresses and HSRP priorities.

Access Control List Conf iguration

If you use IOS ACLs on the MSFC, you must configure the ACLs on both MSFCs identically, globally and at the interface level. Only the designated MSFC (the MSFC to come online first, or the MSFC that has been online the longest) programs the ACL ASIC. The active supervisor engine's PFC multilayer switches packets after consulting with its ACL ASIC to determine whether a packet is forwarded or not, depending on the IOS ACL configured. If a designated MSFC fails, the new designated MSFC must reprogram the ACL ASIC for static ACLs. For consistent results, both MSFCs must have identical static ACL configurations.

Note In addition to defining the same ACLs on both MSFCs, you must also apply the ACLs to the same VLAN interfaces on both MSFCs.

Note Dynamic and reflexi ve ACLs, which are based on actual data flow, may be programmed by either MSFC.

Note You can determine the status of the designated MSFC by entering the show fm features command. This example shows that the MSFC on the supervisor engine in slot 2 is the designated MSFC:

Router# show fm features
Designated MSFC: 2 Non-designated MSFC:1
Redundancy Status: designated

Dual MSFC Operational Model for Redundancy and Load Sharing

Figure 3-2 shows a typical access and distribution layer building block with multiple VLANs in an access layer switch. Because there is no Layer 2 loop, HSRP is used for convergence and load sharing. Switches S1 and S2 have a supervisor engine with an MSFC in slot 1 (Sup1/MSFC-1) and in slot 2 (Sup2/MSFC-2). Sup1 is active and Sup2 is in standby mode in both switches. High availability is enabled on the supervisor engines. The supervisor engines automatically perform image and configuration synchronization; you must manually synchronize the images and configurations on the MSFCs.

Figure 3-2: Dual MSFC Operational Model for Redundancy and Load Sharing---VLANs 10 and 21

In Figure 3-2, you should configure redundancy and load sharing as follows:

VLAN 10 (even-numbered VLANs)---Configure MSFC-1 in Switch S1 as the primary HSRP router (priority 110) and configure MSFC-2 as the standby (priority 109).
VLAN 21 (odd-numbered VLANs)---Configure MSFC-1 in Switch S2 as the primary HSRP router (priority 110) and configure MSFC-2 as the standby (priority 109).

Load sharing is achieved by having the even-numbered VLANs routed by Switch S1 and the odd-numbered VLANs by Switch S2. In a complete switch failure, the remaining switch would service both even and odd VLANs.

You can achieve further load sharing by using MSFC-2 in Switch S1 as the primary HSRP router for VLAN 12 and MSFC-2 as the primary HSRP router in Switch S2 for VLAN 23 (see Figure 3-3).

Figure 3-3: Dual MSFC Operational Model for Redundancy and Load Sharing---
VLANs 10, 12, 21, and 23

Only the active HSRP router for a VLAN will respond with the HSRP MAC address for ARP requests to the HSRP IP address. The active HSRP router will in turn ARP for the end stations' MAC address and populate its ARP cache. By using both MSFCs in a single chassis to share HSRP duties for even VLANs, you can share the control plane ARP traffic. In an MSFC failure, only the ARP entries on the affected VLAN would need to be relearned.

The tradeoff for this level of redundancy and load sharing is the added complexity of keeping track of the even and odd VLANs on the MSFCs within a Catalyst 6000 family switch chassis.

MLS entries are created for packets arriving at the HSRP MAC addresses as well as those arriving with the router's real MAC addresses. HSRP is used for unicast traffic first hop redundancy; for traffic received via another router attached to VLAN 10, for example, the actual MAC address of Sup1/MSFC-1 is used.

Understanding Failure Scenar ios

The five examples in this section describe possible failure scenarios within a single chassis with dual supervisor engines and dual MSFCs (see Figure 3-4) when you enable high availability. The designated MSFC refers to the MSFC that is used to program the ACL ASIC for static ACLs.

Figure 3-4: Single Chassis with Dual Supervisor Engines and Dual MSFCs

Failure Case 1: Designated MSFC-1 Fails

This sequence occurs when the designated MSFC-1 fails:

1. MLS entries for MSFC-1 gracefully age out of the Sup1 Layer-3 cache, while MSFC-2 takes temporary ownership of these MLS entries using its XTAG value.

2. MLS entries for MSFC-2 are not affected.

3. MSFC-2 removes all dynamic and reflexive ACLs programmed in hardware by MSFC-1.

4. MSFC-2 reprograms the static ACLs in the Sup1 ACL ASIC because it is now the designated MSFC.

Failure Case 2: Nondesignated MSFC-2 Fails

This sequence occurs when the nondesignated MSFC-2 fails:

1. MLS entries for MSFC-2 gracefully age out of the Sup1 Layer 3 cache, while MSFC-1 takes temporary ownership of these MLS entries using its XTAG value.

2. MLS entries from MSFC-1 are not affected.

3. MSFC-1 removes all dynamic and reflexive ACLs programmed in hardware by MSFC-2.

4. MSFC-1 remains the designated MSFC.

Failure Case 3: Active Sup1 Fails

This sequence occurs when the active supervisor engine (Sup1) fails:

1. Because the Layer 3 state is maintained, MLS entries of MSFC-1 gracefully age out of the Sup2 Layer 3 cache while MSFC-2 takes temporary ownership of these MLS entries using its XTAG value.

2. The standby supervisor engine maintains the Layer 2 state so there is no Layer 2 convergence time.

3. MSFC-2 removes all dynamic and reflexive ACLs programmed in hardware by MSFC-1.

4. MSFC-2 reprograms the static ACLs in the Sup2 ACL ASIC. MSFC-2 is now the designated MSFC.

Failure Case 4: Standby Sup2 Fails

This sequence occurs when the standby supervisor engine (Sup2) fails:

1. MLS entries for MSFC-2 gracefully age out of the Sup1 Layer 3 cache while MSFC-1 takes temporary ownership of these MLS entries using its XTAG value.

2. MLS entries from MSFC-1 are not affected.

3. MSFC-1 removes all dynamic and reflexive ACLs programmed in hardware by MSFC-2. MSFC-1 remains the designated MSFC.

Failure Case 5: New or Previously Failed Supervisor Comes Back Online

This sequence occurs when the previously failed supervisor engine (Sup2) comes online:

1. Sup1 continues to be the active supervisor engine.

2. Sup2 synchronizes its image and configuration with Sup1 (unless high availability versioning is enabled).

3. MSFC-2 (on Sup2) comes up. If the HSRP preempt for VLAN 21 is configured, then MSFC-2 will become HSRP active. The MLS entries for MSFC-1 will be purged and then relearned via MSFC-2.

4. MSFC-1 remains the designated MSFC for static ACLs.

Configuring Redunda ncy with HS RP

You can configure one or more HSRP groups on MSFC VLAN interfaces to provide automatic routing backup for your network. Each VLAN interface in an HSRP group shares a virtual IP address and MAC address. You can configure end stations and other devices to use the HSRP address as the default gateway so that if one router interface fails, service is not interrupted to those devices.

The interface with the highest HSRP priority is the active interface for that HSRP group.

To configure HSRP on an MSFC VLAN interface, perform this task in interface configuration mode:

Task Command

Step 1 Enable HSRP and specify the HSRP IP address. If you do not specify a group-number, group 0 is used.¹

Router(config-if)# standby [group-number] ip [ip-address]

Step 2 Specify the priority for the HSRP interface. Increase the priority of at least one interface in the HSRP group (the default is 100). The interface with the highest priority becomes active for that HSRP group.

Router(config-if)# standby [group-number] priority priority

Step 3 Configure the interface to preempt the current active HSRP interface and become active if the interface priority is higher than the priority of the current active interface.

Router(config-if)# standby [group-number] preempt [delay delay]

Step 4 (Optional) Set the HSRP hello timer and holdtime timer for the interface. The default values are 3 (hello) and 10 (holdtime). All interfaces in the HSRP group should use the same timer values.

Router(config-if)# standby [group-number] timers hellotime holdtime

Step 5 (Optional) Specify a clear-text HSRP authentication string for the interface. All interfaces in the HSRP group should use the same authentication string.

Router(config-if)# standby [group-number] authentication string

¹To assist in troubleshooting, configure the group number to match the VLAN number.

Task	Command
Step 1 Enable HSRP and specify the HSRP IP address. If you do not specify a group-number, group 0 is used.¹	Router(config-if)# standby [group-number] ip [ip-address]
Step 2 Specify the priority for the HSRP interface. Increase the priority of at least one interface in the HSRP group (the default is 100). The interface with the highest priority becomes active for that HSRP group.	Router(config-if)# standby [group-number] priority priority
Step 3 Configure the interface to preempt the current active HSRP interface and become active if the interface priority is higher than the priority of the current active interface.	Router(config-if)# standby [group-number] preempt [delay delay]
Step 4 (Optional) Set the HSRP hello timer and holdtime timer for the interface. The default values are 3 (hello) and 10 (holdtime). All interfaces in the HSRP group should use the same timer values.	Router(config-if)# standby [group-number] timers hellotime holdtime
Step 5 (Optional) Specify a clear-text HSRP authentication string for the interface. All interfaces in the HSRP group should use the same authentication string.	Router(config-if)# standby [group-number] authentication string

This example shows how to configure an interface as part of HSRP group 100:

Router# configure terminal
Enter configuration commands, one per line.  End with CNTL/Z.
Router(config)# interface vlan100
Router(config-if)# standby 100 ip 172.20.100.10
Router(config-if)# standby 100 priority 110
Router(config-if)# standby 100 preempt
Router(config-if)# standby 100 timers 5 15
Router(config-if)# standby 100 authentication Secret
Router(config-if)# ^Z
Router#

Configuration Examp les

This section describes three configuration options for achieving redundancy:

Note For the following examples, the designated MSFC is on the active supervisor engine. You can determine the status of the designated MSFC by entering the show fm features command. This example shows that the MSFC on the supervisor engine in slot 1 is the designated MSFC:

Router# show fm features

Designated MSFC: 1 Non-designated MSFC:2
Redundancy Status: designated

Example 1---Two Chassis with One Supervisor Engine and One MSFC Each

In the example in Figure 3-5, the high availability feature cannot be configured on the supervisor engines, but HSRP can be configured on the MSFCs.

Figure 3-5: Two Chassis with One Supervisor Engine and One MSFC Each

This example shows how to configure HSRP on the MSFC in switch S1:

Console> (enable) switch console 15
Trying Router-15...
Connected to Router-15.
Type ^C^C^C to switch back...
Router# configure terminal
Enter configuration commands, one per line.  End with CNTL/Z.
Router(config)# interface vlan10
Router(config-if)# standby 10 ip 172.20.100.10
Router(config-if)# standby 10 priority 110 
Router(config-if)# standby 10 preempt
Router(config-if)# standby 10 timers 5 15
Router(config-if)# standby 10 authentication Secret
Router(config-if)# interface vlan21
Router(config-if)# standby 21 ip 192.20.100.21
Router(config-if)# standby 21 priority 109 
Router(config-if)# standby 21 preempt
Router(config-if)# standby 21 timers 5 15
Router(config-if)# standby 21 authentication Secret
Router(config-if)# ^Z
Router# ^C^C^C

This example shows how to configure HSRP on the MSFC in switch S2:

Console> (enable) switch console 15
Trying Router-15...
Connected to Router-15.
Type ^C^C^C to switch back...
Router# configure terminal
Enter configuration commands, one per line.  End with CNTL/Z.
Router(config)# interface vlan10
Router(config-if)# standby 10 ip 172.20.100.10
Router(config-if)# standby 10 priority 109 
Router(config-if)# standby 10 preempt
Router(config-if)# standby 10 timers 5 15
Router(config-if)# standby 10 authentication Secret
Router(config-if)# interface vlan21
Router(config-if)# standby 21 ip 192.20.100.21
Router(config-if)# standby 21 priority 110 
Router(config-if)# standby 21 preempt
Router(config-if)# standby 21 timers 5 15
Router(config-if)# standby 21 authentication Secret
Router(config-if)# ^Z
Router# ^C^C^C

Example 2---Single Chassis with Dual Supervisor Engines and MSFCs

In the example in Figure 3-6, the high availability feature is configured on the supervisor engines, and HSRP is configured on the MSFCs.

Figure 3-6: Single Chassis with Redundant Supervisors and MSFCs

This example shows how to configure HSRP on the MSFC in Switch S1:

Console> (enable) switch console 15
Trying Router-15...
Connected to Router-15.
Type ^C^C^C to switch back...
Router# configure terminal
Enter configuration commands, one per line.  End with CNTL/Z.
Router(config)# interface vlan10
Router(config-if)# standby 10 ip 172.20.100.10
Router(config-if)# standby 10 priority 110 
Router(config-if)# standby 10 preempt
Router(config-if)# standby 10 timers 5 15
Router(config-if)# standby 10 authentication Secret
Router(config-if)# interface vlan21
Router(config-if)# standby 21 ip 192.20.100.21
Router(config-if)# standby 21 priority 109 
Router(config-if)# standby 21 preempt
Router(config-if)# standby 21 timers 5 15
Router(config-if)# standby 21 authentication Secret
Router(config-if)# ^Z
Router# ^C^C^C
 
Console> (enable) switch console 16
Trying Router-16...
Connected to Router-16.
Type ^C^C^C to switch back...
Router# configure terminal
Enter configuration commands, one per line.  End with CNTL/Z.
Router(config)# interface vlan10
Router(config-if)# standby 10 ip 172.20.100.10
Router(config-if)# standby 10 priority 109 
Router(config-if)# standby 10 preempt
Router(config-if)# standby 10 timers 5 15
Router(config-if)# standby 10 authentication Secret
Router(config-if)# interface vlan21
Router(config-if)# standby 21 ip 172.20.100.21
Router(config-if)# standby 21 priority 110
Router(config-if)# standby 21 preempt
Router(config-if)# standby 21 timers 5 15
Router(config-if)# standby 21 authentication Secret
Router(config-if)# ^Z
Router# ^C^C^C

Example 3---Double Chassis with Dual Supervisor Engines and MSFCs

Figure 3-7 shows two Catalyst 6000 family switches (S1 and S2), each with a supervisor engine and MSFC in slot 1 (Sup1/MSFC1) and slot 2 (Sup2/MSFC2). Because there is no Layer-2 loop, HSRP is used for convergence and load sharing. In both switches, Sup1 is active, and Sup2 is in standby.

Figure 3-7: Dual MSFC Operational Model for Redundancy and Load Sharing

This example shows how to configure HSRP on the MSFC in Switch S1:

Console> (enable) switch console 15
Trying Router-15...
Connected to Router-15.
Type ^C^C^C to switch back...
Router# configure terminal
Enter configuration commands, one per line.  End with CNTL/Z.
Router(config)# interface vlan10
Router(config-if)# standby 10 ip 172.20.100.10
Router(config-if)# standby 10 priority 110
Router(config-if)# standby 10 preempt
Router(config-if)# standby 10 timers 5 15
Router(config-if)# standby 10 authentication Secret
Router(config-if)# interface vlan21
Router(config-if)# standby 21 ip 172.20.100.21
Router(config-if)# standby 21 priority 108
Router(config-if)# standby 21 preempt
Router(config-if)# standby 21 timers 5 15
Router(config-if)# standby 21 authentication Secret
Router(config-if)# ^Z
Router# ^C^C^C
 
Console> (enable) switch console 16
Trying Router-16...
Connected to Router-16.
Type ^C^C^C to switch back...
Router# configure terminal
Enter configuration commands, one per line.  End with CNTL/Z.
Router(config)# interface vlan10
Router(config-if)# standby 10 ip 172.20.100.10
Router(config-if)# standby 10 priority 109
Router(config-if)# standby 10 preempt
Router(config-if)# standby 10 timers 5 15
Router(config-if)# standby 10 authentication Secret
Router(config-if)# interface vlan21
Router(config-if)# standby 21 ip 172.20.100.21
Router(config-if)# standby 21 priority 107
Router(config-if)# standby 21 preempt
Router(config-if)# standby 21 timers 5 15
Router(config-if)# standby 21 authentication Secret
Router(config-if)# ^Z
Router# ^C^C^C

This example shows how to configure HSRP on the MSFC in Switch S2:

Console> (enable) switch console 15
Trying Router-15...
Connected to Router-15.
Type ^C^C^C to switch back...
Router# configure terminal
Enter configuration commands, one per line.  End with CNTL/Z.
Router(config)# interface vlan10
Router(config-if)# standby 10 ip 172.20.100.10
Router(config-if)# standby 10 priority 108
Router(config-if)# standby 10 preempt
Router(config-if)# standby 10 timers 5 15
Router(config-if)# standby 10 authentication Secret
Router(config-if)# interface vlan21
Router(config-if)# standby 21 ip 172.20.100.21
Router(config-if)# standby 21 priority 110
Router(config-if)# standby 21 preempt
Router(config-if)# standby 21 timers 5 15
Router(config-if)# standby 21 authentication Secret
Router(config-if)# ^Z
Router# ^C^C^C
 
Console> (enable) switch console 16
Trying Router-16...
Connected to Router-16.
Type ^C^C^C to switch back...
Router# configure terminal
Enter configuration commands, one per line.  End with CNTL/Z.
Router(config)# interface vlan10
Router(config-if)# standby 10 ip 172.20.100.10
Router(config-if)# standby 10 priority 107
Router(config-if)# standby 10 preempt
Router(config-if)# standby 10 timers 5 15
Router(config-if)# standby 10 authentication Secret
Router(config-if)# interface vlan21
Router(config-if)# standby 21 ip 172.20.100.21
Router(config-if)# standby 21 priority 109
Router(config-if)# standby 21 preempt
Router(config-if)# standby 21 timers 5 15
Router(config-if)# standby 21 authentication Secret
Router(config-if)# ^Z
Router# ^C^C^C

Identical Requirements--- Global and Interface Levels	Both MSFCs must have the following: Same routing protocols Same VLAN interfaces Same IOS ACLs¹ ² All interfaces must have the same administrative status
Exceptions---Interface Level	HSRP standby commands IP address commands IP broadcast address IPX helper address
Exceptions---Global Level	IP default-gateway IP route prefix mask IPX internal-network IPX route

Table of Contents

Configuring Redundancy