[ES 7.4] elasticsearch node dont see each other

Elastic Stack Elasticsearch
1

Hi
Im just start in ES about some days
I have 3 machine in network
ips: machine 1 : 10.255.255.6 ,machine 2 : 10.255.255.7, machine 3: 10.255.255.8
i want to setup a es cluster with machine 1 is a master node and two other are slave node.
I've installed ES 7.4 in 3 machine and started
but when i check by _cluster/state?pretty
i cant see all node in cluster
i find some topic with same problem but i can's resolve it
anybody can give me a solution to fix ?

49
Screen Shot 2019-10-20 at 21.27.49.png1124×648 119 KB

thank you

2

@JustinLe

you will have to configure discovery for those nodes to start communicating with each other.
See https://www.elastic.co/guide/en/elasticsearch/reference/current/modules-discovery-bootstrap-cluster.html for detailed documentation on how to set things up.

3

Thank you, Armin
I've configed ther configuration file but it does not work
this is my configuration
machine 1:

cluster.name: myCluster
node.name: node-0
node.master: true
node.data: true
network.host: 10.255.255.6
http.port: 9200
cluster.initial_master_nodes: ["node-0", "node-1", "node-2", "node-3"]
discovery.seed_hosts: ["10.255.255.6", "10.255.255.7", "10.255.255.8", "10.255.255.9"]

machine 2:

cluster.name: myCluster
node.name: node-1
node.master: true
node.data: true
network.host: 10.255.255.7
http.port: 9200
cluster.initial_master_nodes: ["node-0", "node-1", "node-2", "node-3"]
discovery.seed_hosts: ["10.255.255.6", "10.255.255.7", "10.255.255.8", "10.255.255.9"]

machine 3:

cluster.name: myCluster
node.name: node-2
node.master: true
node.data: true
network.host: 10.255.255.8
http.port: 9200
cluster.initial_master_nodes: ["node-0", "node-1", "node-2", "node-3"]
discovery.seed_hosts: ["10.255.255.6", "10.255.255.7", "10.255.255.8", "10.255.255.9"]

machine 4:

cluster.name: myCluster
node.name: node-3
node.master: true
node.data: true
network.host: 10.255.255.9
http.port: 9200
cluster.initial_master_nodes: ["node-0", "node-1", "node-2", "node-3"]
discovery.seed_hosts: ["10.255.255.6", "10.255.255.7", "10.255.255.8", "10.255.255.9"]

please help me find out the wrong config

4

try this
shutdown all, setup this two line on config of all node

cluster.initial_master_nodes: ["node-0"]
discovery.seed_hosts:["node-0", "node-1", "node-2", "node-3"]

now start.

if this does not work you might have to remove your path.data dir
default is /var/data/elasticsearch/*

5

i've try but it does not work

6

What do the logs say after this change? There should be some intelligible warnings about the discovery/master-election process in there now.
Also, did you ensure that the nodes can see each other network-wise (i.e. ping each other and the relevant network port (9300) is open?

7

the logs dont say anything
the instances was started well but nodes do not see each other
i using

telnet ip 9300

and it connected
i've change to one master and 3 data node. and it still does not work

8

@JustinLe

sorry I missed one thing here:
See this documentation: https://www.elastic.co/guide/en/elasticsearch/reference/current/modules-discovery-bootstrap-cluster.html#modules-discovery-bootstrap-cluster-joining

Since you already started all the nodes once without the proper configuration they formed individual "clusters". You'll have to follow the steps in that link to merge them into a single cluster now.

9

@Armin_Braun thank you Armin
it does work
i've spent 2 days to fix
thank you very much