Elasticsearch 2.x

[CanisHelix]

I'm currently running Rancher 1.0.2 and trying out the Elasticsearch 2.x stack, however it's not working and I'm lost as to why.

All the services load up and report healthy.But navigating to kopf I get an error connecting http://192.168.11.40:8081/es/. But if I click the red warning, or go to it directly I get the following

{
  "name" : "Elasticsearch_elasticsearch-clients_1",
  "cluster_name" : "ShiroiES",
  "version" : {
    "number" : "2.3.3",
    "build_hash" : "218bdf10790eef486ff2c41a3df5cfa32dadcfde",
    "build_timestamp" : "2016-05-17T15:40:04Z",
    "build_snapshot" : false,
    "lucene_version" : "5.5.0"
  },
  "tagline" : "You Know, for Search"
}

I'm not sure what to do, I'm not even progressing with Kibana4 until I get Elasticsearch 2.x working.

[2016-07-06 00:19:16] Elasticseach connection: [2016-07-06 00:19:16] {"host":"http://192.168.11.40:8081/es/","withCredentials":false} [2016-07-06 00:19:17] Error executing request: [2016-07-06 00:19:17] {"method":"GET","transformRequest":[null],"transformResponse":[null],"url":"http://192.168.11.40:8081/es//","data":{},"params":{},"headers":{"Accept":"application/json, text/plain, /"}} [2016-07-06 00:19:17] REST API output: [2016-07-06 00:19:17] {"name":"Elasticsearch_elasticsearch-clients_1","cluster_name":"ShiroiES","version":{"number":"2.3.3","build_hash":"218bdf10790eef486ff2c41a3df5cfa32dadcfde","build_timestamp":"2016-05-17T15:40:04Z","build_snapshot":false,"lucene_version":"5.5.0"},"tagline":"You Know, for Search"}

I'm not using the default port 80 as that is already in use, but with the 'question' I picked 8081 as that's free. I've tried 2.2.1, I've used 2.3.3 template from a PR, the same error happens regardless of Elasticsearch's version. Any idea what's going on?

[fjoesne]

+1 same issue, on both ES1 and ES2 from the catalog

I have tried specifying host:port on the container through the rancher ui but this causes an odd behavior. example tells me i can specify ip:port, however if i do this it ends up running exposing the container on a port from the specified ip address' first number.

Example/suggestion result from entering "192.168.1.28:80":

I guess this may be an issue that concerns rancher more than this catalog item alone though.

Edit: not quite the same issue as i get a 502 Bad Gateway whereas you seem to get a proper elasticsearch connection.

[ddmng]

@ShiroiKuma I've just started Elastic 2.x stack with the same cluster name and 8081 port and it works fine. Even though I'm using Rancher 1.1.0 I don't think it depends on it because I'm sure I have tested it in the past with your same version. Just to test it in another way: could you stop the kopf container and add another kopf service inside the running elastic stack on the same port? I've just done it and it works: if it works for you maybe we could find an explanation... or at least you have a kopf working. For this test use the original https://github.com/lmenezes/elasticsearch-kopf/tree/master/docker image and set KOPF_SERVER_NAME and KOPF_ES_SERVERS accordingly.

@fjoesne on my Rancher 1.1.0 this has been fixed because it correctly took and mapped IP:port i specified.

[sanel]

Same problem here on Rancher 1.2.0. It's definitely not up to Rancher - I've manually added old docker-compose.yml and rancher-compose.yml from this repo, and kopf is working without problems.

[rawmind0]

Hi guys,

have you already tried with the recently updated version of elasticsearch-2?? I tried it before merge and was working ok....

[sanel]

@rawmind0 how much recent? I've tried to setup ES with elasticsearch-2 two days ago, from refreshed list and got above failures.

[rawmind0]

Tested and working with elasticsearch-2 version 2.4.3-rancher1

Accessing kopf without any problem. Kopf is showing all es nodes and cluster info.