The Kubernetes service and deployment files in this repo, which are derived from pires/kunernetes-elasticsearch-cluster, handle all the environment variables exposed by the pires/docker-elasticsearch-kubernetes image. I've also added cluster start up and stop scripts to make testing easier.

For instructions on how to create an Elasticsearch cluster on Kubernetes using this repo, you can read my blog: Running Elasticsearch on Kubernetes.

• CLUSTER_NAME
• NODE_MASTER
• NODE_DATA
• NETWORK_HOST
• HTTP_ENABLE
• NUMBER_OF_MASTERS
• NUMBER_OF_SHARDS
• NUMBER_OF_REPLICAS
• ES_HEAP_SIZE
• DISCOVERY_SERVICE
• NAMESPACE

You can start the cluster with the start_es.sh and stop the clusters with stop_es.sh. The current start script create s client nodes, 4 data nodes, and 3 master nodes. Leave the number master nodes alone, 3 is what you want for high availability, but you can increase the number of client and data nodes as necessary.