elasticsearch must restart master for new node joining in zen discovery

Question

I have an elasticsearch cluster of version 1.5.2
There is one master node and one data node in the cluster.
When I add a new data node into the cluster(with the same cluster name) with the name node.name : "Node2", the new node can't join in the cluster with no logs indicating what happens.
My main conifg file is as below:
master node config:

cluster.name: cluster1
node.name: "Master"
node.master: true
node.data: false
discovery.zen.minimum_master_nodes: 1
discovery.zen.ping.timeout: 3s
discovery.zen.ping.retries: 6
discovery.zen.ping.interval: 30s

data node config of node1:

cluster.name: cluster1
node.name: "Node1"
node.master: false
node.data: true
discovery.zen.minimum_master_nodes: 1
discovery.zen.ping.timeout: 3s
discovery.zen.ping.retries: 6
discovery.zen.ping.interval: 30s

data node config of node2:

cluster.name: cluster1
node.name: "Node2"
node.master: false
node.data: true
discovery.zen.minimum_master_nodes: 1
discovery.zen.ping.timeout: 3s
discovery.zen.ping.retries: 6
discovery.zen.ping.interval: 30s

When I restart the master node, the new node successfully join in the cluster.
Does anyone know how to join new nodes into the cluster without restarting the master??

You have the same node name for both node1 and node2: node.name: "Node1" — Andrei Stefan
@Stefan I typed the wrong node name and I have changed it to "Node2". It still failed to join the cluster — tottishi05

Steve P. Steve P. · Accepted Answer · 2015-10-14T15:03:47

If you're using unicast, you need to add:

discovery.zen.ping.unicast.hosts = [...]
discovery.zen.ping.multicast.enabled = False

If you're using multicast, you need to set the appropriate multicast settings (for external multicast).

If you're using AWS, I suggest looking at the elasticsearch-cloud-aws plugin.

PS: In the current configuration, you have no data nodes.

elasticsearch must restart master for new node joining in zen discovery

1 Answers