How to Run Multiple Elasticsearch Nodes on a Single Server (Step‑by‑Step Guide)

Usually a single server runs one Elasticsearch node, but for learning you may want multiple nodes on the same machine to simulate a distributed cluster.

Step 1: Duplicate the Elasticsearch directory

# cp -R elasticsearch-5.0.2 elasticsearch-5.0.2-node-2
$ ls -l
drwxr-xr-x 12 wuxiwei staff 408 Dec 12 16:23 elasticsearch-5.0.2
drwxr-xr-x 12 wuxiwei staff 408 Dec 12 16:27 elasticsearch-5.0.2-node-2

Step 2: Edit elasticsearch.yml for the second instance

Change node.name: node-1 to node.name: node-2 and adjust ports if needed.

Step 3: Start both nodes

# cd ~/software/elasticsearch-5.0.2
$ ./bin/elasticsearch
# cd ../elasticsearch-5.0.2-node-2
$ ./bin/elasticsearch

Verify the cluster health:

$ curl -i -XGET 'http://localhost:9200/_cluster/health?pretty'
{
  "cluster_name":"my",
  "status":"green",
  "number_of_nodes":2,
  "number_of_data_nodes":2,
  "active_primary_shards":20,
  "active_shards":40,
  "relocating_shards":0,
  "initializing_shards":0,
  "unassigned_shards":0,
  "delayed_unassigned_shards":0,
  "number_of_pending_tasks":0,
  "number_of_in_flight_fetch":0,
  "task_max_waiting_in_queue_millis":0,
  "active_shards_percent_as_number":100.0
}

You can also view the nodes in Kibana.

Common pitfalls

If you change http.port or transport.tcp.port, ensure each node uses a different value.

Node‑1 may warn about discovery.zen.minimum_master_nodes. Set node.master: false on the second node or increase the minimum master nodes setting.

Make sure the data directory of the second node is empty; otherwise the node may fail to join because it reuses the first node’s data.