How to setup Three Machines with different IP Address?

[Karthik_Saravanan]

I have three machines, installed same ELK(6.2.2) version in all machine
Now, I have to connect all node, One is master and another two's are client node,

** Each machine is the different IP address

I have tried like this
Server:

            cluster.name: sever
	node.name: main-server
	node.data: true
	node.ingest: true
	node.master: true
	node.max_local_storage_nodes: 1
	path.data: E:/ELK-6.2.2/elasticsearch/data
	path.logs: E:/ELK-6.2.2/elasticsearch/logs
	network.host: 11.xx.xx.xx:9200
	http.port: 9200
	transport.tcp.port: 9300
	discovery.zen.ping.unicast.hosts: ["11.XX.XX.XX",  "12.xx.xx.xx:9200", "13.xx.xx.xx:9200"]
	discovery.zen.minimum_master_nodes: 1

Client: 1

cluster.name: client-one
	node.name: client-node-one
	node.data: true
	node.ingest: true
	node.master: false
	node.max_local_storage_nodes: 1
	path.data: E:/ELK-6.2.2/elasticsearch/data
	path.logs: E:/ELK-6.2.2/elasticsearch/logs
	network.host: 12.xx.xx.xx
	http.port: 9200
	transport.tcp.port: 9300
	discovery.zen.ping.unicast.hosts:  ["11.XX.XX.XX",  "12.xx.xx.xx:9200", "13.xx.xx.xx:9200"]
	discovery.zen.minimum_master_nodes: 1

Client:2

cluster.name: client-two
	node.name: client-node-two
	node.data: true
	node.ingest: true
	node.master: false
	node.max_local_storage_nodes: 1
	path.data: E:/ELK-6.2.2/elasticsearch/data
	path.logs: E:/ELK-6.2.2/elasticsearch/logs
	network.host: 13.xx.xx.xx
	http.port: 9200
	transport.tcp.port: 9300
	discovery.zen.ping.unicast.hosts:  ["11.XX.XX.XX",  "12.xx.xx.xx:9200", "13.xx.xx.xx:9200"]
	discovery.zen.minimum_master_nodes: 1

Please guide me how to setup these machine?

[Christian_Dahlqvist]

Elasticsearch uses port 9300 to communicate internally using the transport protocol, so leave out the port number in the discovery.zen.ping.unicast.hosts list as 9300 is the default.

[Karthik_Saravanan]

Hi Christian_Dahlqvist,
here is working sample config
Master: 54.xx.xx.xx
cluster.name: sciera-master node.name: node-153 node.data: false node.ingest: false node.master: true path.data: D:/elk/elasticsearch/data path.logs: D:/elk/elasticsearch/logs bootstrap.memory_lock: true network.host: 10.144.250.69 network.publish_host: 54.xx.xx.xx http.port: 9200 transport.tcp.port: 9300 discovery.zen.ping.unicast.hosts: ["52.xx.xx.xx:9300"]

Client: 52.xx.xx.xx
cluster.name: sciera-master node.name: node-105 node.data: true node.ingest: false node.master: false path.data: D:/elk/elasticsearch/data path.logs: D:/elk/elasticsearch/logs bootstrap.memory_lock: true network.host: 172.31.44.186 network.publish_host: 52.xx.xx.xx http.port: 9200 transport.tcp.port: 9300 discovery.zen.ping.unicast.hosts: ["54.xx.xx.xx:9300"]

[Christian_Dahlqvist]

Good to hear you managed to get it resolved. Having a dedicated master and data nodes in a two node cluster does however not necessarily make much sense. You now have a cluster in which any node failure will cause the cluster to become impaired, and since you have 2 nodes, the risk is greater than if you only had a single node with default configuration. Just because you can have dedicated node types, it doesn't mean that you should, especially for small clusters. I would recommend that you add another data node to get resiliency and make all three nodes master eligible so that you can continue operating even if a node goes down.