Configuring MCT Cluster
Summary
This document provides a detailed guide on configuring a Multi-Chassis Trunking (MCT) cluster using two ICX switches to operate as a logical cluster for enhanced redundancy and resiliency. It includes topology details, VLAN configurations, and step-by-step commands for setting up the cluster peers and verifying the configuration. •MCT cluster overview: MCT enables two independent ICX switches to function as a single logical cluster, providing active-active forwarding, link redundancy, and device-level resiliency for downstream devices through synchronized cluster peers connected by an Inter-Chassis Link (ICL). •Topology and VLAN roles: The setup involves two ICX 7850-48F switches as peers with specific interfaces assigned for the ICL, Keep Alive VLAN, and client connections. The ICL is a static LAG for cluster communication, and the MCT Session VLAN is dedicated for control traffic, separate from client VLANs.Question
Customer Environment
Peer A: ICX 7850-48F Peer B: ICX 7850-48F Client A: ICX 7550-24P Firmware Version (all devices): 10.0.10g_cd6Symptoms
Root Cause
This KB was created to provide configuration guidance for deploying an MCT cluster and downstream MCT client, including the ICL, Session VLAN, Keepalive VLAN, client/member VLAN, and verification steps.Show more linesTroubleshooting Steps
Configuring Multi-Chassis Trunk MCT
Purpose
This KB explains how to configure the initial Multi-Chassis Trunking MCT Reference :
MCT allows two independent ICX switches to operate as a logical cluster for downstream devices, providing active-active forwarding, link redundancy, and device-level resiliency. The RUCKUS MCT design guide explains that MCT provides an active-active topology where client devices can connect to both cluster peers using LAGs, while the two cluster peers synchronize through the ICL.
2. Topology used
Based on the provided topology:
|
Device |
Role |
Interfaces Used |
|
Peer-A |
MCT Cluster Peer A |
1/1/1 and 1/1/2 for ICL |
|
Peer-A |
Keep Alive Vlan |
1/1/3 |
|
Peer-A |
Client |
1/1/5 |
|
Peer-B |
Keep Alive Vlan |
1/1/3 |
|
Peer-B |
MCT Cluster Peer B |
1/1/1 and 1/1/2 for ICL |
|
Peer-B |
Client |
1/1/6 and 1/1/8 |
Topology Information:
- Peer A: ICX 7850-48F
- Peer B: ICX 7850-48F
- Client A: ICX 7550-24P
- Firmware Version (all devices): 10.0.10g_cd6
3. What Is the ICL?
The Inter-Chassis Link ICL is the connection between the two MCT peer switches. This link is required for the peers to exchange MCT control information and maintain the cluster state.
The RUCKUS MCT design guide explains that the two cluster switches are connected using a single link or LAG called the Inter-Chassis Link ICL, and that CCP runs on the ICL to establish the MCT cluster.
Why it is needed
The ICL is used for:
- Cluster communication between Peer-A and Peer-B.
- Synchronization between both MCT peers.
- Maintaining the correct forwarding behavior when downstream clients are attached.
- Allowing both switches to operate as one logical MCT cluster from the client perspective.
Requirement
The ICL must be configured as a static LAG.
4. What Is the MCT Session VLAN?
The MCT Session VLAN is the VLAN used for MCT control communication between the two cluster peers.
This VLAN provides the control path for packet exchanges between the two MCT cluster switches.
Important note The Session VLAN should be dedicated to MCT control communication. It should not be used as a normal client VLAN or production user VLAN.
5. Important Design Notes Before Configuration
Before configuring MCT, validate the following:
- Both switches should be running a compatible FastIron version.
- The same MCT-related VLANs must exist on both peers.
- The ICL must be a static LAG.
- The Session VLAN must be tagged on the ICL LAG.
- Each peer must have a unique RBridge ID.
- The peer IP address must point to the remote peer MCT session interface.
- The cluster should only be deployed after both sides are configured correctly.
5. MCT Keepalive VLAN
The Keepalive VLAN is a separate backup peer-reachability path between both MCT switches the Keepalive VLAN uses the dedicated physical link on port between the peers
6. Configuration
Peer A
6.1 Configure the static ICL LAG
Peer-A(config)#lag MCT-ICL static id 2
Peer-A(config-lag-MCT-ICL)#ports ethernet 1/1/1 to 1/1/2
LAG MCT-ICL deployed successfully!
6.2 Session VLAN
Peer-A(config)#vlan 3000 name MCT-SESSION by port
Peer-A(config-vlan-3000)#tagged lag 2
Added tagged port(s) lag lg2 to port-vlan 3000.
6.3 Session VE Interface
Peer-A(config)#interface ve 3000
Peer-A(config-vif-3000)#ip address 1.1.1.1 255.255.255.252
6.4 Keepalive VLAN
Peer-A(config)#vlan 3001 name MCT-KEEPALIVE by port
Peer-A(config-vlan-3001)#untagged ethernet 1/1/3
Added untagged port(s) ethe 1/1/3 to port-vlan 3001.
Peer B
7.1 Configure the static ICL LAG
Peer-B(config)# lag MCT-ICL static id 2
Peer-B(config-lag-MCT-ICL)#ports ethernet 1/1/1 to 1/1/2
LAG MCT-ICL deployed successfully!
7.2 Session VLAN
Peer-B(config)#vlan 3000 name MCT-SESSION by port
Peer-B(config-vlan-3000)#tagged lag 2
Added tagged port(s) lag lg2 to port-vlan 3000.
7.3 Session VE Interface
Peer-B(config)#interface ve 3000
Peer-B(config-vif-3000)#ip address 1.1.1.2 255.255.255.252
7.4 Keepalive VLAN
Peer-B(config)#vlan 3001 name MCT-KEEPALIVE by port
Peer-B(config-vlan-3001)#untagged ethernet 1/1/3
Added untagged port(s) ethe 1/1/3 to port-vlan 3001.
Important Note: On both peers, VLAN 1 must be removed from the ICL LAG configuration because it is the default VLAN on the switch and is not required for MCT control communication. The ICL should only carry the dedicated MCT Session VLAN, VLAN 3000, as a tagged VLAN.
Peer A
Peer-A(config)#vlan 1
Peer-A(config-vlan-1)#no untagged lag 2
Peer B
Peer-B(config)#vlan 1
Peer-B(config-vlan-1)#no untagged lag 2
8. Cluster Configuration
8.1 Peer A
Peer-A(config)#cluster MCT-CLUSTER 1
dPeer-A(config-cluster-MCT-CLUSTER)# rbridge-id 1
Peer-A(config-cluster-MCT-CLUSTER)# session-vlan 3000
Info - IP addresses configured on VE 3000 will not be advertised as connected route by any routing protocol as it is configured as session-vlan.
Peer-A(config-cluster-MCT-CLUSTER)# keep-alive-vlan 3001
Peer-A(config-cluster-MCT-CLUSTER)# icl MCT-ICL lag 2
1.1.1.2 rbridge-id 2 icl MCT-ICL
deployPeer-A(config-cluster-MCT-CLUSTER)# peer 1.1.1.2 rbridge-id 2 icl MCT-ICL
Peer-A(config-cluster-MCT-CLUSTER)# deploy
8.2 Peer B
Peer-B(config)#cluster MCT-CLUSTER 1
sePeer-B(config-cluster-MCT-CLUSTER)# rbridge-id 2
Peer-B(config-cluster-MCT-CLUSTER)# session-vlan 3000
Info - IP addresses configured on VE 3000 will not be advertised as connected route by any routing protocol as it is configured as session-vlan.
Peer-B(config-cluster-MCT-CLUSTER)# keep-alive-vlan 3001
Peer-B(config-cluster-MCT-CLUSTER)# icl MCT-ICL lag 2
l MCT-ICL
deployPeer-B(config-cluster-MCT-CLUSTER)# peer 1.1.1.1 rbridge-id 1 icl MCT-ICL
Peer-B(config-cluster-MCT-CLUSTER)# deploy
Important note: We already execute deploy command in the cluster, so the cluster is currently active. If we need to make any configuration changes or add new settings, we first need to access the cluster configuration then run the no deploy command. After that, apply the required changes and then perform a deploy again. Otherwise, the cluster will not allow configuration changes while it is in the active state.
9. Verification
You can use the following commands to verify the MCT cluster status:
show cluster MCT-CLUSTER config
show cluster MCT-CLUSTER ccp peer
show cluster MCT-CLUSTER peer
show cluster
Once the MCT cluster is deployed successfully, the cluster status should look similar to the following:
Configuring MCT Client-A on RUCKUS ICX
This section explains how to configure the downstream MCT client after the MCT cluster has already been created between Peer-A and Peer-B.
The client device connects to both MCT peers using a LAG. From the client perspective, several physical links are bundled into one logical LAG. From the MCT cluster perspective, the LAG is split across both peers to provide redundancy and active-active forwarding.
1. Topology Reference for Client-A
Based on the topology, Client-A is connected as follows:
|
Device |
Interface |
Connected To |
Purpose |
|
Peer-A |
1/1/5 |
Client-A 1/1/6 |
MCT client LAG member |
|
Peer-B |
1/1/6 |
Client-A 1/1/5 |
MCT client LAG member |
|
Peer-B |
1/1/8 |
Client-A 1/1/8 |
MCT client LAG member |
Important Design Notes Before Configuring the Client
Before adding Client-A to the MCT cluster, verify the following:
- The MCT cluster must already be deployed and stable.
- The ICL LAG must be up between Peer-A and Peer-B.
- The Session VLAN must be working over the ICL.
- The Keepalive VLAN must be configured and reachable.
- The client-facing LAG should use LACP/dynamic LAG.
- The same client LAG ID should be referenced under the cluster on both peers.
- The same client VLANs should be configured on both peers.
- VLAN 1 should not be carried unnecessarily across the ICL or client uplinks unless the design specifically requires it.
RUCKUS documentation explains that the ICL supports only static ports, while client LAGs can be static or dynamic
2. Peer-A Client Configuration
2.1 Client LAG on Peer-A
Peer-A(config)#lag CLIENT-A-LAG dynamic id 21
Peer-A(config-lag-CLIENT-A-LAG)#ports ethernet 1/1/5
LAG CLIENT-A-LAG deployed successfully!
2.2 Client VLAN
Peer-A(config)#vlan 10 name MCT-CLIENT-VLAN by port
Peer-A(config-vlan-10)# tagged lag 21
Added tagged port(s) lag lg21 to port-vlan 10.
Peer-A(config-vlan-10)# tagged lag 2
Added tagged port(s) lag lg2 to port-vlan 10.
Reference:
- lag 21, which is the client-facing LAG
- lag 2, which is the ICL LAG
Also remove the LAG from VLAN 1
Peer A
Peer-A(config)#vlan 1
Peer-A(config-vlan-1)#no untagged lag 21
Note : Client/Member VLAN Tagging Requirement: The client/member VLAN must be added to the ICL LAG and to the client-facing LAG on both MCT peers. This is required because the MCT member VLAN is associated with the VLANs carried across the ICL. On the client side, the VLAN can be configured as either tagged or untagged depending on the design requirements and how the connected device expects to receive the traffic.
3.3 For this example, an IP address will be assigned to the client VLAN 10. This allows VLAN 10 to be used as a Layer 3 interface for validation, management, or routing purposes, depending on the network design.
Peer-A(config)#interface ve 10
Peer-A(config-vif-10)#ip address 192.168.10.2/24
3.4 Client Deployment Under the Cluster
Peer-A(config-cluster-MCT-CLUSTER)#cluster MCT-CLUSTER 1
rbridgePeer-A(config-cluster-MCT-CLUSTER)# client Client-A
Peer-A(config-cluster-MCT-CLUSTER-client-Client-A)# rbridge-id 3
Peer-A(config-cluster-MCT-CLUSTER-client-Client-A)# client-interface lag 21
Peer-A(config-cluster-MCT-CLUSTER-client-Client-A)# deploy
Note : The client RBridge ID identifies that downstream client inside the MCT cluster, so Peer-A and Peer-B must reference the same client’s name and the same client RBridge ID when they are configuring the same Client-A.
4. Peer-B Client Configuration
4.1 Client LAG on Peer-B
Peer-B(config)#lag CLIENT-A-LAG dynamic id 21
ports ethernet 1/1/6
ports ethernet 1/1/8Peer-B(config-lag-CLIENT-A-LAG)# ports ethernet 1/1/6
LAG CLIENT-A-LAG deployed successfully!
Peer-B(config-lag-CLIENT-A-LAG)# ports ethernet 1/1/8
5. Client VLAN
Peer-B(config)#vlan 10 name MCT-CLIENT-VLAN by port
Peer-B(config-vlan-10)#tagged lag 21
Added tagged port(s) lag lg21 to port-vlan 10.
Peer-B(config-vlan-10)#tagged lag 2
Added tagged port(s) lag lg2 to port-vlan 10.
Reference:
- lag 21, which is the client-facing LAG
- lag 2, which is the ICL LAG
Also remove the LAG from VLAN 1
Peer B
Peer-B(config)#vlan 1
Peer-B(config-vlan-1)#no untagged lag 21
5.2 For this example, an IP address will be assigned to the client VLAN 10. This allows VLAN 10 to be used as a Layer 3 interface for validation, management, or routing purposes, depending on the network design.
Peer-B(config)#interface ve 10
Peer-B(config-vif-10)#ip address 192.168.10.1/24
5.3 Client Deployment Under the Cluster
Peer-B(config)#cluster MCT-CLUSTER 1
rbridge-Peer-B(config-cluster-MCT-CLUSTER)# client Client-A
Peer-B(config-cluster-MCT-CLUSTER-client-Client-A)# rbridge-id 3
Peer-B(config-cluster-MCT-CLUSTER-client-Client-A)# client-interface lag 21
Peer-B(config-cluster-MCT-CLUSTER-client-Client-A)# deploy
6. Client-A Local Configuration
On Client-A, all uplinks going to Peer-A and Peer-B must be bundled into the same LACP LAG.
Client-A port mapping:
|
Client-A Port |
Connected To |
|
1/1/6 |
Peer-A 1/1/5 |
|
1/1/5 |
Peer-B 1/1/6 |
|
1/1/8 |
Peer-B 1/1/8 |
6.1 Client-Facing LAG on the MCT Peers
Client-A(config)#lag MCT-UPLINK dynamic id 21
Client-A(config-lag-MCT-UPLINK)#ports ethernet 1/1/5 ethernet 1/1/6 ethernet 1/1/8
LAG MCT-UPLINK deployed successfully!
6.2 Client VLAN
Client-A(config)#vlan 10 name MCT-CLIENT-VLAN by port
Client-A(config-vlan-10)# tagged lag 21
Added tagged port(s) lag lg21 to port-vlan 10.
Also remove the LAG from VLAN 1
Client-A(config)#vlan 1
Client-A(config-vlan-1)#no untagged lag 21
6.3 Create a VE interface on the client VLAN to validate connectivity.
Client-A(config)#interface ve 10
Client-A(config-vif-10)#ip address 192.168.10.10/24
Client-A#ping 192.168.10.1
Sending 1, 16-byte ICMP Echo to 192.168.10.1, timeout 5000 msec, TTL 64
Type Control-c to abort
Reply from 192.168.10.1 : bytes=16 time=6ms TTL=64
Success rate is 100 percent (1/1), round-trip min/avg/max=6/6/6 ms.
Client-A#ping 192.168.10.2
Sending 1, 16-byte ICMP Echo to 192.168.10.2, timeout 5000 msec, TTL 64
Type Control-c to abort
Reply from 192.168.10.2 : bytes=16 time=3ms TTL=64
Success rate is 100 percent (1/1), round-trip min/avg/max=3/3/3 ms.
6.4 For reference, all client-facing ports should be in an up state, as shown in the screenshot below.
Download
(95.8 KB)
Attachment 2
Client-Artf
Download
(63.5 KB)
Attachment 3
Peer-Brtf
Download
(98.6 KB)
Article Number:
000015488
Updated:
July 31, 2026 01:16 PM (about 1 month ago)
Answer Attachment 1
Peer-A
rtf
Download
(95.8 KB)
Answer Attachment 2
Client-A
rtf
Download
(63.5 KB)
Answer Attachment 3
Peer-B
rtf
Download
(98.6 KB)
Tags:
Configuration, Ruckus ICX Switches
Votes:
0
This article is:
helpful
not helpful
Working...Please wait
This is here to prevent you from accidentally submitting twice.
The page will automatically refresh.