Configuration Guide Vol. 1


31.1.5 Notes on Using Ring Protocol with Spanning Tree Protocols

<Structure of this section>

(1) About Associating Virtual-Link VLAN with VLAN Mapping

VLANs specified for virtual link VLANs must belong (be set in the VLAN mapping and VLAN group) to the VLAN for data transfer within a ring.

(2) About Virtual-Link VLAN Configuration Scopes

(3) About Spanning Tree when virtual link VLAN is not set

If no virtual link VLAN is set, the intended topology cannot be built because virtual links cannot be built. As a result, loops might occur.

(4) About Stopping Spanning Tree Protocols by Configuring and Deleting Ring Protocol

Setting and deleting a Ring Protocol configuration may cause the running Spanning Tree Protocols to be stopped. If Spanning Tree is stopped, the applicable VLAN may be looped, which may affect the topology of Spanning Tree Protocols.

(5) Network construction when Ring Protocol and Spanning Tree Protocols are used together

The basic configuration of a network using the Ring Protocol and a Spanning Tree Protocol is a loop. Before building a Spanning Tree Protocol on an access network for an existing ring network, bring the configuration port (physical port or channel group) on the Spanning Tree network down, such as by setting the shutdown command.

(6) About Ring Protocol failure monitoring times and spanning tree BPDU sending intervals

Set the fault monitoring time for health check frames for the Ring Protocol (health-check holdtime) to a value less than the timeout detection time for Spanning Tree BPDUs (hello-time x 3 (in seconds)). If a greater value is set and a fault occurs in the ring network, the Spanning Tree Protocol detects a BPDU timeout before the Ring Protocol detects a fault, causing the topology to change, and possibly creating a loop.

(7) Response to Program Restart on Transit Node

When restarting the Ring Protocol program (restart axrp operation command), first put the configuration port on the Spanning Tree network (physical port or channel group) into the down state (for example, by setting shutdown). After restart, either wait for the reception hold time for flush control frames on the transit node (forwarding-shift-time) to time out, or after the Forwarding transition time for control VLANs (forwarding-delay-time) is used to perform path switching, clear the shutdown (for example) on the port put into the down state.

(8) Handling unidirectional link failures in a ring network

The Ring Protocol does not detect ring faults for one-way link faults. When a one-way link fault occurs on a ring network, because virtual link control frames can no longer be sent or received, the Spanning Tree Protocol might mistakenly detect a BPDU timeout. This might cause a loop that lasts until the one-way link fault is resolved.

When the Ring Protocol and the IEEE 802.3ah/UDLD functionality are used together, one-way link faults can be detected to prevent the occurrence of the loops that they cause.

(9) Recovery Procedure from Multiple Failures in a Combined Spanning Tree Environment

When faults occur in multiple places in a ring network (multi-fault), virtual link control frames can no longer be sent and received, causing topology changes for the Spanning Tree Protocol. Multi-faults include when faults occur on both ring ports for a device using both the Ring Protocol and a Spanning Tree Protocol. In these cases, perform the following to restore all faults within a ring network:

  1. Bring the configuration port of the Spanning Tree network (physical port or channel group) down such as by shutdown.

  2. Restore the faults in the ring network, to have the master node detect ring fault restoration.

  3. Clear shutdown for the configuration port of the Spanning Tree network to allow restoration.

(10) Consistency between Ring Protocol's VLAN mapping and VLAN belonging to MST instances of Multiple Spanning Tree

When a change in configuration causes the settings for VLAN mappings for the Ring Protocol and VLANs belonging to MST instances of Multiple Spanning Tree to no longer match, the unmatched VLANs might be put in Blocking status, preventing communication.