Cluster Setup (N-Node)

This guide explains how to configure a JMS Bridge cluster with N broker nodes using static connectors in ActiveMQ Artemis. This approach is ideal for controlled deployments with known broker endpoints.

Why Cluster?

A cluster enables:

• Load balancing of messages across broker nodes
• Failover handling if a node goes down
• Horizontal scalability for high-throughput systems
• Message redistribution for active consumers

Architecture (N Nodes)

Each broker node connects to every other node via TCP connectors.

  Broker-1  <---->  Broker-2
     |               |
     |               |
  Broker-3  <---->  ...  <---->  Broker-N

This forms a full-mesh topology using static TCP connectors.

---

Configuration

1. Define Connectors (per node)

Each node must define a self connector and static connectors for the other N–1 nodes.

Example for Broker-1:

<connectors>
  <connector name="self">tcp://broker1-host:61616</connector>
  <connector name="broker2">tcp://broker2-host:61617</connector>
  <connector name="broker3">tcp://broker3-host:61618</connector>
  <!-- Add additional connectors up to broker-N -->
</connectors>

Do not include self in the static connectors list.

2. Define Cluster Connection

Use a <cluster-connection> to link nodes. Example for Broker-1:

<cluster-connections>
  <cluster-connection name="jms-cluster">
    <connector-ref>self</connector-ref>
    <retry-interval>1000</retry-interval>
    <message-load-balancing>STRICT</message-load-balancing>
    <max-hops>1</max-hops>
    <static-connectors>
      <connector-ref>broker2</connector-ref>
      <connector-ref>broker3</connector-ref>
      <!-- Up to broker-N -->
    </static-connectors>
  </cluster-connection>
</cluster-connections>

Repeat on every broker, changing the self connector and static-connectors accordingly.

---

Enable Message Redistribution

Set redistribution so messages are forwarded to nodes with active consumers:

<address-settings>
  <address-setting match="#">
    <redistribution-delay>0</redistribution-delay>
  </address-setting>
</address-settings>

---

Testing the Cluster

1. Start all broker nodes.
2. Send a message to any node.
3. Consume it from another node connected to the same queue.

---

Best Practices

• Use unique ports for each broker (e.g., 61616, 61617, …).
• Use meaningful connector names for clarity.
• Keep cluster topology symmetric (every node knows all others).
• Use message-load-balancing=STRICT to optimize routing.

---

Sample Connector Generator (Optional)

To quickly generate <connector-ref> blocks for N nodes:

for i in $(seq 2 20); do
  echo "<connector-ref>broker$i</connector-ref>"
done

---