RabbitMQ Cluster Installation and Configuration Guide

RabbitMQ Overview

RabbitMQ is an open‑source AMQP implementation written in Erlang. It supports many client libraries (Python, Ruby, .NET, Java, etc.) and is used for storing and forwarding messages in distributed systems, offering ease of use, scalability and high availability.

AMQP (Advanced Message Queuing Protocol) is an open standard application‑layer protocol for message‑oriented middleware, enabling decoupled communication between components.

Key AMQP features include messaging, queues, routing (point‑to‑point and publish/subscribe), reliability and security.

Cluster Overview

RabbitMQ clusters are built using Erlang’s distributed capabilities; each node is a peer providing service to clients. Nodes replicate queue structures via HA (mirrored) queues. In this guide a three‑node cluster of disk nodes is deployed, where any node can serve clients.

Environment

VMware version: 12.0.0

CentOS version: 7.3.1611

RabbitMQ version: 3.6

VM IPs: 192.168.252.101, 192.168.252.102, 192.168.252.103

Installation

Single‑machine installation

Install Erlang and RabbitMQ Server on all three hosts.

Cluster Configuration

Modify hostname

Temporarily set hostname on each node:

hostname <new hostname>

Example commands:

192.168.252.101 $ hostname node1
192.168.252.102 $ hostname node2
192.168.252.103 $ hostname node3

Edit /etc/hosts

Add the following lines on each machine:

192.168.252.101 node1
192.168.252.102 node2
192.168.252.103 node3

Set Erlang cookie

Copy the .erlang.cookie file from node1 to node2 and node3, adjust permissions to 400 after copying, and set correct ownership.

chmod 777 /var/lib/rabbitmq/.erlang.cookie   # temporary
scp /var/lib/rabbitmq/.erlang.cookie node2:/var/lib/rabbitmq/
scp /var/lib/rabbitmq/.erlang.cookie node3:/var/lib/rabbitmq/
chmod 400 /var/lib/rabbitmq/.erlang.cookie
chown rabbitmq /var/lib/rabbitmq/.erlang.cookie
chgrp rabbitmq /var/lib/rabbitmq/.erlang.cookie

Stop RabbitMQ service

service rabbitmq-server stop

Join nodes to the cluster

node2 $ rabbitmqctl stop_app
node2 $ rabbitmqctl join_cluster rabbit@node1
node2 $ rabbitmqctl start_app

node3 $ rabbitmqctl stop_app
node3 $ rabbitmqctl join_cluster rabbit@node1
node3 $ rabbitmqctl start_app

Configure RAM or disk nodes

By default nodes are disk nodes. To make a node a RAM node:

node2 $ rabbitmqctl change_cluster_node_type ram

To revert to a disk node:

node2 $ rabbitmqctl change_cluster_node_type disc

Check cluster status

rabbitmqctl cluster_status

The output lists disc nodes, running nodes, partitions and alarms.

Mirrored queue policy

Enable a policy that mirrors all queues matching a pattern (e.g., names starting with “message”):

rabbitmqctl set_policy -p / ha-allqueue "^message" '{"ha-mode":"all"}'

Adjust the pattern to match the desired queues.

Cluster restart

Force‑boot the first node, then start the RabbitMQ service on the remaining nodes:

rabbitmqctl force_boot
service rabbitmq-server start   # on other nodes

Verify the cluster status on each node after restart.

Contact

Author: Peng Lei

Source: http://www.ymq.io/2017/08/17/rabbit-install-cluster

Email: [email protected]