Configuration Guide Vol. 1

29.8 Notes on using Ring Protocol

<Structure of this section>

(1) Configuration changes during operation

Take care not to create a loop configuration when changing Ring Protocol configurations by performing the following operations:

We recommend changing such configurations as follows:

  1. Use the shutdown command or other means to take down the ring port of the device whose configuration is to be changed, or the secondary port of the master node.

  2. Stop the Ring Protocol functionality for the device whose configuration is to be changed (disable command).

  3. Change the configuration.

  4. Clear the stop on the Ring Protocol functionality (no disable command).

  5. Bring previously downed ring ports back up (such as by clearing the shutdown command).

(2) Coexistence with other functions

[See "24.3 Coexistence of Layer 2 Switch Function and Other Functions"

(3) VLAN for Sending and Receiving Control Frames

(4) About Ring VLAN Status of Transit Node

For transit nodes, when a fault occurs for a device or ring port, and recovery succeeds, the ring VLAN status of the ring port is set to Blocking to prevent loops from occurring. One of the ways in which this Blocking status is cleared is when the reception hold time (forwarding-shift-time) for flush control frames times out. If the flush control frame receive hold time (forwarding-shift-time) is shorter than the master node's health check send interval, the transit node's ring port may enter the forwarding state before the master node detects the recovery of the ring failure and changes the secondary port to the blocking state, which may cause a loop. Therefore, set the flush control frame receive hold time (forwarding-shift-time) to a value greater than the master node's health check send interval.

If a member switch fails on a master node in a stack configuration, add a member switch to the stack when the failure recovers. If the flush control frame reception hold time (forwarding-shift-time) of the adjacent transit node is shorter than the time when the addition of a member switch is completed, the ring port of the transit node may be in the forwarding status and a loop may occur. For this reason, set the flush control frame receive hold time (forwarding-shift-time) of the transit node adjacent to the master node in the stack configuration to 60 seconds or more.

(5) About networking when using Ring Protocol

Networks using the Ring Protocol have looped configurations. Therefore, build such networks as follows to avoid loops.

  1. Beforehand, use shutdown or another command to take the ring port (physical port or channel group) of the ring configuration node down.

  2. Set the Ring Protocol configuration, or copy (copy command) the configuration file including the Ring Protocol settings, to enable the Ring Protocol.

  3. Bring the ring port back up (such as by clearing the shutdown command) when the Ring Protocol is set for all devices in the network.

(6) About the interval for sending health check frames and failure monitoring time

Set the fault monitoring time (health-check holdtime) to a value greater than the sending interval (health-check interval). If the time is set to a value less than the sending interval, a reception timeout will occur and a fault will be mistakenly detected. Also, when setting the fault monitoring time and sending interval, make sure to take the network configuration and operation environment into account. We recommend setting a fault monitoring time of around three times the sending interval. If a value below the recommended value is set, a fault may be erroneously detected if a delay occurs due to network load or CPU load on the equipment.

(7) Interoperability

Ring Protocol is our proprietary feature. It cannot be used interoperably with third-party switches.

(8) About the equipment that constitutes the ring

(9) Master node failure

When the master node cannot communicate because of a device fault or other reason, ring network fault monitoring is not performed. Therefore, communication continues as is between transit nodes other than the master node, without being switched to an alternate path. Also, when the master node has recovered from a device fault, it sends a flush control frame to the transit nodes in the ring. Therefore, communication might stop temporarily.

(10) Multiple Failures in a Network

When multiple faults occur between different nodes in the same ring (a multi-fault), because the master node was already performing fault detection for the first fault, the second and subsequent faults are not detected. Also, because health check frames sent by the master node cannot be received until recovery from the last fault in a multi-fault restoration detection situation occurs, recovery cannot be detected. As a result, communication might be temporarily impossible for multi-faults when a partial fault is restored (when a fault remains for the ring).

Note that when the multi-fault monitoring functionality is applied, multi-faults might be able to be detected, depending on the combination of faults. For information about the multiple failure monitoring feature, see Multiple Failure Monitoring Feature in 29.6 Ring Protocol.

(11) Switching routes in the event of a failure due to a VLAN failure

When a downed link or other fault occurs on the primary port of the master node, VLANs set in the data transfer VLAN group might go down temporarily. In cases like this, it might take some time to restore communication by path switching.

Note that the VLAN Debounce functionality can sometimes be used to avoid downed VLANs. For more information about VLAN debounce feature, see 27.9 VLAN debounce Feature Description.

(12) Number of times flush control frames are transmitted

Adjust the number of times that flush control frames are sent by the master node, based on configurations including the VLAN count and VLAN mapping count applying to the ring network.

If 64 or more VLAN mappings are used for a single ring port, set a sending count to 4 times or more. If the count is less than 4 times, the MAC address table entries cannot be cleared, and it might take some time to perform path switching.

(13) Configuration Commands with VLAN Down Settings

If no configuration commands pertaining to the Ring Protocol have been set, when the first configuration command pertaining to the Ring Protocol (one of the following commands) is set, all VLANs will go down temporarily. Therefore, when a ring network that uses the Ring Protocol is built, we recommend setting the following configuration commands ahead of time.

Note that for VLAN mapping (the axrp vlan-mapping command), the VLANs associated with the VLAN mapping will go down temporarily, even for new settings. The VLAN mappings already set and the other VLANs to which they are associated are not affected.

(14) Transmission/reception of flush control frames at master node startup

When the master node device restarts, and the transit node detects link-up for the ring port connected to the master node later than the master node does, the transit node might not be able to receive the flush control frames sent by the master node during initial operation. Here, the status of the ring port for the transit node unable to receive flush control frames will be Blocking. The corresponding ring port changes to the Forwarding status and communication is restored after the reception hold time for flush control frames (forwarding-shift-time) elapses.

When flush control frames cannot be received on neighboring transit nodes, and the sending count for flush control frames from the master node is controlled, reception might be possible. Also, to shorten the time to cut communication due to flush control frames not yet received, shorten the reception hold time for flush control frames sent by the transit node (initial value: 10 seconds).

The same applies when the following operations are performed on the master node:

(15) Specifying the Flash control frame receive wait protection time when the path switchback suppression facility is used

When using the path switch-back suppression functionality, either specify infinity for the reception hold time (forwarding-shift-time) for flush control frames for the transit node, or specify a value greater than the path switch-back suppression time (preempt-delay). If the reception hold time for flush control frames for the transit node times out during path switch-back suppression, and logical block on the corresponding ring port is cleared, because the master node clears the logical block on the secondary port, a loop might occur.

(16) About the monitoring start timing of the multiple failure monitoring function

After the multi-fault monitoring functionality is applied to a shared node, multi-fault monitoring is started when the first multi-fault monitoring frame sent from the opposing shared node is received. As such, when the multi-fault monitoring functionality is set and a fault occurs for a ring network, multi-fault monitoring cannot be started. Set the multi-fault monitoring functionality when the status of the ring network is normal.

(17) Communication during partial recovery from multiple failures

Because the master node does not detect ring restoration during partial multi-fault recovery, the transit node ring port is logically blocked until the reception hold time (forwarding-shift-time) for flush control frames elapses. To clear the logical block status, either shorten the reception hold time for flush control frames (initial value: 10 seconds), or recover remaining link faults to have the master node detect ring restoration. Also, when setting the reception hold time for flush control frames, set a value greater than the sending interval for multi-fault monitoring frames (using the multi-fault-detection interval configuration command). When a small value is set, loops might occur temporarily.

(18) Combining the multiple failure monitoring facility with the path switchback suppression facility

When the path switch-back suppression functionality is set for a shared link non-monitoring ring and recovery for a multi-fault succeeds, because the Forwarding status is maintained until the restoration suppression status is cleared for the secondary port, this might result in a loop configuration. When using the multi-fault monitoring functionality and path switch-back suppression functionality together, perform any of the following operations:

(19) Down link aggregation specified for a ring port

When nodes in a ring network are connected via link aggregation (static mode or LACP mode), make sure to shut down all physical ports belonging to the channel group by using the shutdown command before shutting down the channel group of the link aggregation by using the shutdown command.

If you bring up the channel group using the no shutdown command, make sure to shut down all physical ports belonging to the channel group by using the shutdown command.

(20) Understanding Stacked Node Application

In a stacked configuration, link aggregation that connects to multiple member switches cannot be used as a ring port.

Stacked nodes do not support the use of Ring Protocol with Spanning Tree. If you add a stacked node to a ring network that uses Spanning Tree Protocols together, the intended topology is not built because a virtual link cannot be built. which might cause loops to occur. If you use a stacked node, stop Spanning Tree on all nodes that make up the ring.

When adding the Switch to a stand-alone node and building a stacked node, note the following:

(21) Executing restart Commands

If the following operation command is executed on a transit node, the ring port VLAN goes down, so the master node erroneously detects a ring failure and forwards the secondary port. Because the ring port of the transit node is in a transient down state, a loop occurs until the master node detects the recovery of the ring failure.

When executing these commands on a transit node, perform the following steps to prevent loops:

  1. Use shutdown command to set the ring port to the down status.

  2. Run the above restart commandI will do it.

  3. In step 1, set the ring port in the down state to the up state (release shutdown command, etc.).

[Product name label]

All Rights Reserved, Copyright(C), 2022, 2023, ALAXALA Networks, Corp.