v5 connect panic: no Elasticsearch node available

[zplzpl]

Please use the following questions as a guideline to help me answer
your issue/question without further inquiry. Thank you.

Which version of Elastic are you using?

[ ] elastic.v5 (for Elasticsearch 5.x)

Please describe the expected behavior

panic: no Elasticsearch node available

Please describe the actual behavior

	// Create a client
	client, err := elastic.NewClient(elastic.SetURL("http://192.168.33.10:9200"))
	if err != nil {
		panic(err)
	}

curl -s -XGET 'http://192.168.33.10:9200/_nodes/http?pretty=1'
{
  "_nodes" : {
    "total" : 1,
    "successful" : 1,
    "failed" : 0
  },
  "cluster_name" : "es-search",
  "nodes" : {
    "qoa_gnFlTtSPjUiQWH_I9g" : {
      "name" : "node-1",
      "transport_address" : "10.0.2.15:9300",
      "host" : "10.0.2.15",
      "ip" : "10.0.2.15",
      "version" : "5.0.0",
      "build_hash" : "253032b",
      "roles" : [
        "master",
        "data",
        "ingest"
      ],
      "http" : {
        "bound_address" : [
          "[::]:9200"
        ],
        "publish_address" : "10.0.2.15:9200",
        "max_content_length_in_bytes" : 104857600
      }
    }
  }
}

Any steps to reproduce the behavior?

Run the code above

[olivere]

One word: Sniffing. The client will try to find other nodes in the cluster by default (as all official Elasticsearch clients do). It will do so by finding the publish_address in the response you sent. Can you ping 10.0.2.15:9200 from your node?

Either disable sniffing (elastic.SetSniff(false)) or make sure all ES nodes return an address that is accessible by your application.

[kuma-guy]

I followed Connection-Problems on Wiki, but still get "no Elasticsearch node available"

Which version of Elastic are you using?

• elastic.v5 (for Elasticsearch 5.x)

Please describe the expected behavior

panic: no Elasticsearch node available

Please describe the actual behavior

// Create a client
i.ElasticClient, err = elastic.NewClient(elastic.SetSniff(false), elastic.SetURL("http://127.0.0.1:9200"))
if err != nil {
    // Handle error
    panic(err)
}

curl -s -XGET 'http://127.0.0.1:9200/_nodes/http?pretty=1'
{
  "_nodes" : {
    "total" : 1,
    "successful" : 1,
    "failed" : 0
  },
  "cluster_name" : "elasticsearch_kumaguy",
  "nodes" : {
    "FBkUSwVqSNG0lIFnBcVAww" : {
      "name" : "FBkUSwV",
      "transport_address" : "127.0.0.1:9300",
      "host" : "127.0.0.1",
      "ip" : "127.0.0.1",
      "version" : "5.2.2",
      "build_hash" : "f9d9b74",
      "roles" : [
        "master",
        "data",
        "ingest"
      ],
      "http" : {
        "bound_address" : [
          "[fe80::1]:9200",
          "[::1]:9200",
          "127.0.0.1:9200"
        ],
        "publish_address" : "127.0.0.1:9200",
        "max_content_length_in_bytes" : 104857600
      }
    }
  }
}

{
  "name" : "FBkUSwV",
  "cluster_name" : "elasticsearch_kumaguy",
  "cluster_uuid" : "1tc7ObqOTIKg0LbYauf2pw",
  "version" : {
    "number" : "5.2.2",
    "build_hash" : "f9d9b74",
    "build_date" : "2017-02-24T17:26:45.835Z",
    "build_snapshot" : false,
    "lucene_version" : "6.4.1"
  },
  "tagline" : "You Know, for Search"
}

Any steps to reproduce the behavior?

Run the code above

[olivere]

Works fine here.

I set up an example here a few months ago: https://gist.github.com/olivere/d49cae8646122d163103e4c12cc2d66a. That works fine here.

You need to make sure that the server (the part which runs the Elastic code) can connect to 127.0.0.1:9200. That's all that's required.

[olivere]

I also added a simple connect example.

$ ./connect -h
Usage of ./connect:
  -sniff
    	Enable or disable sniffing (default true)
  -url string
    	Elasticsearch URL (default "http://localhost:9200")

$ ./connect -sniff=true
Connection succeeded

$ ./connect -sniff=false
Connection succeeded

[kuma-guy]

Thanks for quick response.

I tested your example and simple connect.
Both totally works fine here too...

Do I need something if it works on top of Google App Engine?

[olivere]

@kuma-guy There's a wiki page for that.

[kuma-guy]

@olivere it works! sorry for taking your time and thanks for your help!

[478682649]

elastic.SetSniff(false)

[olivere]

@478682649 Thinking of making this the official slogan of the library: "elastic.SetSniff(false) is a Go library for Elasticsearch" ;-)

[ZChristine]

I try to use the following command to have a quick start of carley
cayley.exe http -a ./data/disease_crawled_6212.nq -d elastic http://10.190.188.45:9200
but it return a error of "health check timeout: no Elasticsearch node available" @olivere

[olivere]

@ZChristine You can disable health checks alongside with sniffing. Something like this:

client, err := elastic.NewClient(...,
   elastic.SetSniff(false),
   elastic.SetHealthcheck(false),
   ...
)

If you want to know how to configure Elasticsearch in a way that you don't need to do this, see https://github.com/olivere/elastic-with-docker.

[olivere]

@ZChristine Maybe this is an issue for Cayley, as they don't seem to support passing parameters into Elastic.

I wonder if they could use the github.com/olivere/elastic/config package to make Elastic automatically pick up connection parameters from the URL. It's up to the authors of Cayley though.

[ZChristine]

thanks for timely replying . I got the wrong way of command prompt before , and I fixed it later. And download the nq file into local elastic . But the loading speed is pretty slow, was this because of the way that es storage the data from cayley ? If the data was like triple model, what does it look like in es?

[olivere]

Elasticsearch Bulk API is pretty fast. It depends on various factors though, e.g. the size of the documents or the I/O of the cluster. You should really talk to the authors of Cayley for those things.

BTW I opened cayleygraph/cayley#730 to address the original issue you mentioned.

[ZChristine]

thanks so much

[li-keli]

I have the same problem, I added elastic.SetSniff(false) solved this problem
thanks so much