capability. The channel is then added to the spanning tree as a bridge port.

Warning Dynamic VLAN ports can force a VLAN change; as a result, PAgP cannot be used to form a bundle on ports that are configured for dynamic VLANs. The VLANs must be static VLANs, meaning that the port on the switch must be assigned to a VLAN. PAgP also requires that all ports in the channel belong to the same VLAN or be configured as trunk ports.

If you have a pre−existing EtherChannel bundle, and a VLAN of a port contained in the bundle is modified, all ports in the bundle are modified to match the VLAN configuration. PAgP will not group ports running at different speeds or duplex. PAgP will change the port speed and duplex for all ports in the bundle.

All ports in a Fast EtherChannel bundle should be assigned to the same VLAN or be configured as trunked ports. You must also configure both ends of the link with the same trunking mode.

Tip You can configure the broadcast limits by percentage limit or by packets−per−second. Packets−per−second allows unicast packets to be dropped when the broadcast limit is exceeded.

Fast Convergence Components of STP

There are a number of protocols for STP that allow for fast convergence, including PortFast, UplinkFast, and BackboneFast. In the following sections I’ll discuss the functions of these protocols and components.

PortFast

PortFast is a switch function that can be used on ports where a single server or workstation is connected, to allow a port to enter the forwarding mode almost immediately. Doing so prevents the port from entering the listening and learning states.

As we’ve discussed, when a switch using STP is powered up, the ports running STP go through four states before forwarding frames through each port. In order to get to the forwarding state, the STA makes each port wait up to 50 seconds before data is allowed to be forwarded. This delay may cause problems with certain protocols and applications. By implementing PortFast, you can avoid these delays.

UplinkFast

One of the most important factors in a network is the convergence time when a link fails. By implementing the UplinkFast function, you can reduce the time it takes for the network to converge by optimizing convergence times. Because of the convergence time of STP, some end stations become inaccessible, depending on the current state of each switch port.

By decreasing convergence time, you reduce the length of the disruption. UplinkFast allows a port in a blocked state on a switch to almost immediately begin forwarding when the switch detects a link failure or a topology change. However, UplinkFast must have direct knowledge of the link failure in order to move a blocked port into a forwarding state.

Note An Uplink Group is a root port that is in a forwarding state and a set of blocked ports that does not include self−looping ports. The Uplink Group is the alternate path when the currently forwarding link fails.

The UplinkFast feature should be placed only on Access layer switches that are connected to the end−user nodes. In order to utilize UplinkFast, several criteria must be met:

∙ UplinkFast must be enabled on the switch.

211

∙The switch must have at least one blocked port.

∙The failure must be on the root port.

If a link fault occurs on the primary root link, UplinkFast transitions the blocked port to a forwarding state. UplinkFast changes the port so that it bypasses the listening and learning phases. This change occurs in three to four seconds, allowing convergence to begin immediately without waiting for the MaxAge timer to expire.

Note UplinkFast becomes a global setting on the switch. It affects all the VLANs on the switch and cannot be applied on just one VLAN. When you enable UplinkFast, it automatically increases the path cost, making it unlikely that the switch will become the root switch. If UplinkFast is not being used, you should use the Catalyst default settings.

BackboneFast

BackboneFast is a function that allows the switch to converge more quickly in the event that a redundant link fails. An inferior BPDU is sent when a link from the designated switch has been lost to the root bridge. When the root port or a blocked port on a switch receives an inferior BPDU from its designated bridge, if BackboneFast is enabled, this event in turn triggers a root Link Query.

The designated switch transmits these BPDUs with the new information that it is now the root bridge as well as the designated bridge, and the BPDUs begin arriving on a port that is blocked on the switch. The switch receiving inferior BPDUs will ignore the message until the configured MaxAge timer expires, to give the network time to overcome the network problem.

If inferior BPDU messages continue to arrive after the MaxAge timer has been used, the root port and other blocked ports on the switch become alternate paths to the root bridge. The switch will send another kind of BPDU called the root Link Query PDU if more than one link exists to the root bridge. The switch will send a root Link Query PDU out all the available alternate paths to the root bridge to determine which one will forward.

If there are no other blocked ports, the switch automatically assumes that it has lost connectivity to the root bridge, causes the maximum aging time on the root to expire, and becomes the root switch. BackboneFast must be enabled on all switches in the network in order to function properly.

Tip BackboneFast cannot be used in a Token Ring network.

Enabling STP on a Set/Clear Command−Based Switch

The Set/Clear command−based switch allows you to enable and disable STP on a per−port basis. Every port on the switch is enabled for STP by default. If STP has been disabled on the switch, you can re−enable STP from the Privileged mode prompt.

Note Cisco recommends that STP remain enabled on the switch. It is important that it remain enabled on any trunk port where the possibility exists of a bridging loop in the network.

To enable STP on a Set/Clear command−based switch, use the following Privileged mode command:

catalyst5000> (enable) set spantree enable

After using this command you must use the set spantree enable all command to enable STP on all the VLANs. An example is shown in the next section.

212

Enabling STP on a Set/Clear Command−Based Switch for All VLANs

To enable STP on all VLANs, use the following command in Privileged mode:

set spantree enable all

Disabling STP on a Set/Clear Command−Based Switch

To disable STP on a Set/Clear command−based switch, use the following Privileged mode command:

catalyst5000> (enable) set spantree disable all Spantree disabled.

Disabling STP on a Set/Clear Command−Based Switch by VLAN

To selectively disable specific ports on the switch by VLAN, use this Privileged mode command:

set spantree disable [VLAN] set spantree disable all

For example, you could use disable STP on VLAN 2 with the following command:

set spantree disable 2

Viewing the STP Configuration on a Set/Clear Command−Based Switch

To view the current configuration of STP on your switch, use this Privileged mode command:

show spantree <VLAN number>

In the following example, the VLAN number is 5:

show spantree 5

The output should look similar to the following (Table 10.6 defines each of the fields in this output):

213

The listing at the bottom of the output shows the ports in use in the spanning tree. It states the port, port−state, and priority, as well as whether Fast−Start (PortFast) is enabled.

Table 10.6: The show spantree command output fields.

Configuring STP on an IOS Command−Based Switch

Unlike the Set/Clear command−based switch, enabling the Spanning Tree Protocol on a Cisco IOS command−based switch is performed in Global Configuration mode. To enable STP, enter the following command:

spantree <VLAN list>

In the following example, 5 is the VLAN number and is considered a VLAN−list field; you can include up to 10 VLANs in the list:

spantree 5

Disabling STP on an IOS Command−Based Switch

To disable STP on a VLAN, in Global Configuration mode use this command:

no spantree <VLAN list>

In this example, 5 is the VLAN number and is considered a VLAN−list field:

no spantree 5

214