Can't connect to my Elasticsearch instance: NoNodeAvailableException

Question

Can't connect to my Elasticsearch started with:

docker-compose up

Running command:

curl -XGET http://localhost:9200/_nodes/http?pretty

Output:

{
  "cluster_name" : "elasticsearch",
  "nodes" : {
    "qD18rHzhQaexExUw5sBgXg" : {
      "name" : "Scanner",
      "transport_address" : "172.19.0.3:9300",
      "host" : "172.19.0.3",
      "ip" : "172.19.0.3",
      "version" : "6.2.3",
      "build" : "fcbb46d",
      "http_address" : "172.19.0.3:9200",
      "http" : {
        "bound_address" : [ "0.0.0.0:9200" ],
        "publish_address" : "172.19.0.3:9200",
        "max_content_length_in_bytes" : 104857600
      }
    }
  }
}

Connect with transport client:

public Client client() throws Exception {
    Settings settings = Settings.builder()
        .put("spring.data.elasticsearch.cluster-nodes", "localhost:9300")
        .build();

    TransportClient client = TransportClient.builder()
        .settings(settings)
        .build()
        .addTransportAddress(new InetSocketTransportAddress(InetAddress.getByName("localhost"), 9300));
    return client;
}

Get error:

Caused by: NoNodeAvailableException[None of the configured nodes are available: [{#transport#-1}{localhost}{127.0.0.1:9300}]]
    at org.elasticsearch.client.transport.TransportClientNodesService.ensureNodesAreAvailable(TransportClientNodesService.java:326)
    at org.elasticsearch.client.transport.TransportClientNodesService.execute(TransportClientNodesService.java:223)

docker-compose.yml

services:
  elasticsearch:
    image: elasticsearch
    ports:
      - '9200:9200'
      - '9300:9300'
  kibana:
    image: kibana
    ports:
      - '5601:5601'
    environment:
      - ELASTICSEARCH_URL=http://elasticsearch:9200

Dockerfile

FROM elasticsearch, kibana
EXPOSE 9200
EXPOSE 9300

(running a local installed Elasticsearch without Docker it works!)

What am I doing wrong? Any idea how to fix it?

Answer 1

Transport endpoint is bound on localhost inside elasticsearch container. That's why it is only accessible from the container itself.

Transport endpoint should be bound to 0.0.0.0 ( "transport_address" : "0.0.0.0:9300" ) in elasticsearch container and it will be possible to access it via localhost on a host machine by localhost:9300 .

So, you can achieve that with the following docker-compose.yaml :

version: "3"
services:
  elasticsearch:
    image: elasticsearch
    command: "-Etransport.host=0.0.0.0"
    ports:
      - '9200:9200'
      - '9300:9300'
  kibana:
    image: kibana
    ports:
      - '5601:5601'
    environment:
      - ELASTICSEARCH_URL=http://elasticsearch:9200

Update:

If you just use the above docker-compose.yaml , elasticsearch container fails with the following error when it starts:

elasticsearch_1  | [2018-03-26T07:44:13,475][INFO ][o.e.n.Node               ] [rzYPgrJ] starting ...
elasticsearch_1  | [2018-03-26T07:44:13,663][INFO ][o.e.t.TransportService   ] [rzYPgrJ] publish_address {172.17.0.2:9300}, bound_addresses {0.0.0.0:9300}
elasticsearch_1  | [2018-03-26T07:44:13,688][INFO ][o.e.b.BootstrapChecks    ] [rzYPgrJ] bound or publishing to a non-loopback address, enforcing bootstrap checks
elasticsearch_1  | ERROR: [1] bootstrap checks failed
elasticsearch_1  | [1]: max virtual memory areas vm.max_map_count [65530] is too low, increase to at least [262144]
elasticsearch_1  | [2018-03-26T07:44:13,696][INFO ][o.e.n.Node               ] [rzYPgrJ] stopping ...
elasticsearch_1  | [2018-03-26T07:44:13,747][INFO ][o.e.n.Node               ] [rzYPgrJ] stopped
elasticsearch_1  | [2018-03-26T07:44:13,747][INFO ][o.e.n.Node               ] [rzYPgrJ] closing ...
elasticsearch_1  | [2018-03-26T07:44:13,762][INFO ][o.e.n.Node               ] [rzYPgrJ] closed

That's because we changed transport endpoint be listened on 0.0.0.0 .

In order to fix it, it is necessary to increase vm.max_map_count sysctl parameter on a host mashine at least up to 262144 :

sudo sysctl -w vm.max_map_count=262144

Now you can successfully start both containers by docker-compose up .

Answer 2

When downgrading to elasticsearch:2.4.4 it works, I suspect it has to do with SpringData not being compatible with latest Elasticsearch.

   ...
    services:
      elasticsearch:
        image: elasticsearch:2.4.4
        ports:
          - '9200:9200'
          - '9300:9300'
      kibana:
        image: kibana:4.6.1
    ...

Charts:

https://github.com/spring-projects/spring-data-elasticsearch

https://www.elastic.co/support/matrix