Deploying Zookeeper and ActiveMQ Cluster on Three Servers

This article describes how to deploy a Zookeeper ensemble and an ActiveMQ cluster across three servers (192.168.20.40, 192.168.20.43, 192.168.20.39) and start the services.

Restart Zookeeper on each server

cd /data/zookeeper/bin/

ls

./zkServer.sh start

Output shows "ZooKeeper JMX enabled by default", the config file used, and "Starting zookeeper ... STARTED" for each node.

Install and configure ActiveMQ on the master node

tar -zxvf apache-activemq-5.15.14-bin.tar.gz

mv apache-activemq-5.15.14 /data/activemq

cd /data/activemq/conf/

vim activemq.xml

In activemq.xml replace the default persistence adapter with a replicated LevelDB configuration:

<broker xmlns="http://activemq.apache.org/schema/core" brokerName="cluster-mq" dataDirectory="${activemq.data}">
  <!--
  <persistenceAdapter>
    <kahaDB directory="${activemq.data}/kahadb"/>
  </persistenceAdapter>
  -->
  <persistenceAdapter>
    <replicatedLevelDB
      directory="${activemq.data}/leveldb"
      replicas="3"
      bind="tcp://0.0.0.0:0"
      zkAddress="192.168.20.40:2181,192.168.20.43:2181,192.168.20.39:2181"
      zkPassword=""
      hostname="192.168.20.40"
sync="local_disk"
      zkPath="/activemq/leveldb-stores"/>
  </persistenceAdapter>
  <transportConnector name="openwire" uri="tcp://192.168.20.40:61616?maximumConnections=1000&wireFormat.maxFrameSize=104857600"/>
</broker>

After saving the file, start ActiveMQ:

/data/activemq/bin/activemq start

The console logs show loading of environment, Java path, PID creation, and successful startup.

Result

The Zookeeper ensemble provides leader election, so only one node acts as the master while the other two remain in standby mode. Consequently, only the master node exposes the ActiveMQ ports (61616 and 8161); the standby nodes have those ports closed.

All steps are verified with screenshots (omitted here) and the deployment is considered complete.