ES-5.6.0 RPM installation, cluster not adding more than 5-nodes

HI,
I have setup ES-5.6.0 using RPM (on OEL-7.3), I have setup 7-nodes in cluter.

Issue: I am able to start the ES services on all 7-nodes but when I am listing cluster nodes using curl command then I am getting on 5-nodes listed. When I am restarting the ES service on all this 7-nodes then cluster is picking random 5-nodes from it, but not all 7-nodes.

[root@perf03 x-pack]# curl -XGET -u elastic 'http://172.20.90.60:9200/_cat/nodes?v'
Enter host password for user 'elastic':
ip heap.percent ram.percent cpu load_1m load_5m load_15m node.role master name
172.20.90.62 1 59 0 0.05 0.04 0.05 mdi - perf05
172.20.90.63 1 85 0 0.00 0.01 0.05 mdi * perf06
172.20.90.60 2 90 1 0.00 0.02 0.05 mdi - perf03
172.20.90.64 2 58 0 0.00 0.01 0.05 mdi - perf07
172.20.90.61 1 58 1 0.00 0.01 0.05 mdi - perf04

  • elasticsearch.yml file content:
    node.name: perf03 (set same as hostname on each node)
    cluster.name: PERFES5
    path.data: /data/es_data
    network.host: ens160:ipv4
    http.port: 9200
    transport.tcp.port: 9300
    discovery.zen.ping.unicast.hosts: ["perf01:9300", "perf02:9300", "perf03:9300"]
    discovery.zen.minimum_master_nodes: 2
    node.max_local_storage_nodes: 2
    xpack.security.audit.enabled: true

Any help appreciated!!

Regards,
Vishal

Check the logs on the nodes that aren't in the cat output, it should state if they are joining or not, and why.

You have more than 3 master eligible nodes, which means that minimum_master_nodes is incorrectly set. The host list also does not seem to contain all master eligible nodes.

This topic was automatically closed 28 days after the last reply. New replies are no longer allowed.