rabbitmq-cluster

Rabbitmq Cluster using Docker.

Status

In Testing

Requirements

• Docker 1.2 +
• Etcd 0.4.6 +

Features

• Automatic seed node detection using etcd (Seed node will initialize the settings for cluster).
• rabbitmq.conf file for automatic cluster initialization using etcd discovery.
• hosts file initialization using etcd discovery.
• Centralized logging using syslog.
• Random Erlang Cookie for the cluster.

Creating cluster (different machines)

Assuming that there are 2 machines (machine-1, machine-2), with data directory at /data (you may choose different path). You may deploy all nodes at once using command:

RABBITMQ_USER=<rabbitmq_user>
RABBITMQ_PASSWORD=<rabbitmq_password>
sudo docker run --rm -P -p 5672:5672 -p 25672:25672 -p 4369:4369 -p 44001:44001 --name node1 -e HOST_IP=$COREOS_PRIVATE_IPV4 -e RABBITMQ_USER=$RABBITMQ_USER -e RABBITMQ_PASSWORD=$RABBITMQ_PASSWORD -v /dev/log:/dev/log -v /data:/var/lib/rabbitmq  totem/rabbitmq-cluster

where $COREOS_PRIVATE_IPV4 is the private IP address for the host. (If using ec2, ensure that machines can talk to each other on ports: 4369, 5672, 35197) 35197)

Proxying Cluster using Yoda

If you are using dynamic proxy like yoda (that supports tcp) for port 5672, you may need to add additional port binding in Yoda proxy. e.g.: -p 5673:5672

To configure yoda with tcp listeners and sidekick discovery, see Yoda Proxy and Yoda Discover