Discovery.zen.ping.unicast.hosts setting on tribe node


(Yu Watanabe) #1

I'd like to confirm with the setting for discovery.zen.ping.unicast.hosts on tribe node.

I have configured tribe node as below, 10.1.0.15 is the IP address of the tribe node itself.

# Tribe Node Setting
# This setting will launch node client per tribe setting
tribe:
  t1:
    cluster.name: japan-east
    discovery.zen.ping.unicast.hosts: ["10.1.0.12", "10.1.0.13", "10.1.0.14","10.1.0.15"]

  t2:
    cluster.name: japan-west
    discovery.zen.ping.unicast.hosts: ["10.4.0.10", "10.4.0.11","10.4.0.15","10.1.0.15"]

I usually set all the IP address of nodes on each node which consists the cluster. But for tribe node when I set the IP for the tribe node in discovery.zen.ping.unicast.hosts, I get below error on the tribe node.

[2017-01-13T17:41:57,447][WARN ][o.e.d.z.UnicastZenPing   ] [node-client/t1] [1] failed send ping to {#zen_unicast_4#}{7pVBPtbcQ1GDqF7mFLB-OA}{10.1.0.15}{10.1.0.15:9300}
java.lang.IllegalStateException: handshake failed, mismatched cluster name [Cluster [elasticsearch]] - {#zen_unicast_4#}{7pVBPtbcQ1GDqF7mFLB-OA}{10.1.0.15}{10.1.0.15:9300}

If I exclude it's own IP address on tribe node , tribe node starts up and starts the node discovery properly.
Do you usually exclude own IP address on tribe node from discovery.zen.ping.unicast.hosts?


(Mark Walkom) #2

What version are you on?


(Yu Watanabe) #3

I use v 5.1.2.


(Mark Walkom) #4

You may want to check that ^


(Yu Watanabe) #5

I found the reason. I was sending the zen ping to the wrong port in tribe node.
After I clearly specify the host with port number, the error is gone.

10.1.0.15:9301


(system) #6

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