Ubuntu 14.04上的ElasticSearch Java Transport Client NoNodeAvailableException

Question

I am running Elasticsearch v. 2.3.2 , using Java 7. Following is the printout from curl http://172.31.11.83:9200 :

{
  "name" : "ip-172-31-11-83",
  "cluster_name" : "elasticsearch",
  "version" : {
      "number" : "2.3.2",
      "build_hash" : "b9e4a6acad4008027e4038f6abed7f7dba346f94",
      "build_timestamp" : "2016-04-21T16:03:47Z",
      "build_snapshot" : false,
      "lucene_version" : "5.5.0"
    },
  "tagline" : "You Know, for Search"
}

... and I am using the following in my Java code:

<dependency>
    <groupId>org.elasticsearch</groupId>
    <artifactId>elasticsearch</artifactId>
    <version>2.3.2</version>
</dependency>

I have ports 9200 and 9300 open in my firewall rules for my ES server, and can successfully execute said Java code from my laptop (Mac OSX). Following is the code snippet that starts off the process (this works fine):

Settings settings = Settings.settingsBuilder()
                .put("cluster.name", "elasticsearch").build();
        esClient =TransportClient.builder().settings(settings).build().addTransportAddress(new 
                InetSocketTransportAddress(new InetSocketAddress(InetAddress.getByName("172.31.11.83"), 9300)));

Then later, I try to issue an index request (this fails when I run the code on Ubuntu 14.04 :

adminClient = esClient.admin().indices();
IndicesExistsResponse response = adminClient.exists(request).actionGet();

My elasticsearch.yml file contains the following network settings:

network.bind_host: 0
network.publish_host: 172.31.11.83
transport.tcp.port: 9300
http.port: 9200

I have also tried with network.bind_host: 172.31.11.83 to no avail. Using curl , I can get to port 9200 from all machines. The cluster name reported by curl is "elasticsearch" .

When I start ES, I see the following in the elasticsearch.log :

publish_address {172.31.11.83:9300}, bound_addresses {[::]:9300}

And yet, the exception I get is as follows:

NoNodeAvailableException[None of the configured nodes are available: [{#transport#-1}{172.31.11.83}{172.31.11.83:9300}]]
at org.elasticsearch.client.transport.TransportClientNodesService.ensureNodesAreAvailable(TransportClientNodesService.java:290)
at org.elasticsearch.client.transport.TransportClientNodesService.execute(TransportClientNodesService.java:207)
at org.elasticsearch.client.transport.support.TransportProxyClient.execute(TransportProxyClient.java:55)
at org.elasticsearch.client.transport.TransportClient.doExecute(TransportClient.java:283)
at org.elasticsearch.client.support.AbstractClient.execute(AbstractClient.java:347)
at org.elasticsearch.client.support.AbstractClient.execute(AbstractClient.java:336)
at org.elasticsearch.client.support.AbstractClient$IndicesAdmin.execute(AbstractClient.java:1178)
at org.elasticsearch.client.support.AbstractClient$IndicesAdmin.exists(AbstractClient.java:1198)

Again, this exact code works from my local machine. Any thoughts?

Answer 1

Having identical issue..

Upgraded elastic from 1.7 to 2.3.2 on same AWS kit

Ubunto 14.0.4

Elastic binding transport on 9300 as before

Security group has port open (not changed)

Now remote clients cannot connect via transport layer - same error as above.

The only thing that has changed in my setup is the version of Elasticsearch

Answer 2

ok I solved this. It appears 2.3.2 doesn't default TCP bind in same way as 1.7.0

I had to set this in my elasticsearch.yml :

network.bind_host: {AWS private IP address)