This blog post will explain how to create and scale a Couchbase Cluster using full armor of Docker – Docker Machine, Docker Swarm and Docker Compose. Here is what we’ll do:

  • Create a 3-node Docker Swarm Cluster using Docker Machine
  • Run a Couchbase instance on two nodes
  • Create a cluster
  • Rebalance the cluster
  • Scale and rebalance the cluster again


Docker Swarm Cluster using Consul

Create a three-node Docker Swarm cluster using Docker Machine:

Provision a Swarm cluster with Docker Machine provide more details about why and what’s done in this script. Here is a summary:

  • Create a Docker Machine and run Consul for service discovery
  • Create three Docker Machines – one for Master and two for Worker nodes.Each machine is configured to be part of a Swarm cluster using --swarm. It also uses the Consul service discovery specified using --swarm-discovery.

Couchbase Nodes on Docker Swarm

Create two instances of Couchbase using Docker Compose:

arungupta/couchbase image is used here. This image is defined at Couchbase Docker Image. It uses the Official Couchbase Docker Image add
configures it as explained:

  1. Setups memory for Index and Data
  2. Configures the Couchbase server with Index, Data, and Query service
  3. Sets up username and password credentials
  4. Loads the travel-sample bucket

Compose file uses host network. This is equivalent to using --net=host on docker run CLI. It allows the container to use the host networking stack. It also limits only a single Couchbase container to run on
a single Docker Machine. So this means that our Couchbase cluster can scale based upon the number of Docker Machines – 3 in our case. The exact command to use this Compose file is:

There are three nodes in the Docker Swarm cluster. The default scheduler strategy is spread and so the containers will be spread on different hosts. This is evident
by docker ps:

Note, one Couchbase server is running on swarm-node-01 and another on swarm-node-02. Each server is configured with an administrator username Administrator and password password. Find out IP
address of the Docker Machine:

If you have jq installed then IP address can be conveniently found as:

Couchbase Cluster on Docker Swarm

All Couchbase server nodes are created equal. This allows the Couchbase cluster to truly scale horizontally to meet your growing application demands. Independently running Couchbase nodes can be added to a cluster by invoking the server-add CLI command. This is typically a two step process. The first step is adding one or more nodes. The second step then rebalances the cluster where the data on the existing nodes is rebalanced across the updated cluster. In our case, a single Couchbase
container is running on each Docker Machine. Lets pick IP address of any one Couchbase node, and add IP address of the other node:

Couchbase Web Console for both the nodes will show a similar output:
This shows that the two nodes now form a cluster, and needs to be rebalanced.

Rebalance Couchbase Cluster

Now, lets rebalance the cluster:

Couchbase Web Console will be updated to show that rebalance is happening:
And finally you’ll see a rebalanced cluster:


Scale and Rebalance Couchbase Cluster

Scale the Couchbase cluster:

Check that the container is running on a different Docker Machine:

As mentioned earlier, scaling a Couchbase cluster is a two-step process. This is so because typically you’ll add multiple servers and then rebalance the cluster. However, in cases where you only need to add a single Couchbase node and then rebalance,
the rebalance command can be used achieve that. In our case, this is done as shown:

The rebalanced cluster now looks like:

This blog showed how you can easily create and scale a Couchbase Cluster using Docker Swarm, Machine and Compose. Enjoy! Further reading …


Posted by Arun Gupta, VP, Developer Advocacy, Couchbase

Arun Gupta is the vice president of developer advocacy at Couchbase. He has built and led developer communities for 10+ years at Sun, Oracle, and Red Hat. He has deep expertise in leading cross-functional teams to develop and execute strategy, planning and execution of content, marketing campaigns, and programs. Prior to that he led engineering teams at Sun and is a founding member of the Java EE team. Gupta has authored more than 2,000 blog posts on technology. He has extensive speaking experience in more than 40 countries on myriad topics and is a JavaOne Rock Star for three years in a row. Gupta also founded the Devoxx4Kids chapter in the US and continues to promote technology education among children. An author of several books on technology, an avid runner, a globe trotter, a Java Champion, a JUG leader, NetBeans Dream Team member, and a Docker Captain, he is easily accessible at @arungupta.

Leave a reply