Marvel multiple node error


(Kenneth Mroz) #1

i currently have two nodes running. I ran into this issue when trying to start a third but I have errors that is not allowing elasticsearch to start. I tried declaring one of my nodes as master.node : true then I started to recieve this error and seem to not get rid of it. " [marvel.agent.exporter ] [des_4] skipping exporter [default_local] as it isn't ready yet" Any idea what this can be?

(Steve Kearns) #2

What version are you running?

(Kenneth Mroz) #3

kibana 4.3
elasticsearch 2.1

(Kenneth Mroz) #4

Do you have ideas on what this error can be? I have not been able to resolve this issue?

(Steve Kearns) #5

Hi Kenneth,

Hm. There isn't a lot to go on here.

Is the issue the log message you showed, or is the issue that Elasticsearch won't start? Can you provide more details on what the issue is (including more of the log messages) , what configuration/settings cause it, and whether having/not having the Marvel-agent installed impacts things?


(Kenneth Mroz) #6

so i just realized I am missing the configurations for marvel.agent.exporter.. can you explain in more detail how I am suppose to go about setting this up . My current config is set up as so but is failing.


type: http
host: ["http://des_4:9300", "http://des_5:9300", "http://des_3:9300"]

(Steve Kearns) #7

If you just want to store Marvel data locally, you don't actually need to configure the exporters - it defaults to storing the data locally.

When you say "it is failing" please provide a clear description of the failed behavior (e.g. how do you know it's failing) and the log messages that correspond to the failure.

(Kenneth Mroz) #8

[marvel.agent.exporter ] [des_4] skipping exporter [default_local] as it isn't ready yet" is the origional error and then when setting up the exporters i get this error,
error connecting to [http://des_5:9200] [dvo_es_5]
[2016-01-04 15:09:07,390][ERROR][marvel.agent.exporter.http] exception when checking remote cluster version on host [http://des_5:9200]
ElasticsearchException[unable to check remote cluster version: no available connection for host [http://des_5:9200]]

(Steve Kearns) #9

Do these errors continue to show up after the 3 nodes are all up and running? These might not indicate a real issue unless they persist..

Does the Marvel UI show data from all 3 nodes, once they're all up and running for a few minutes?

(Tanguy) #10

Hi kmroz,

Can you check that there are spaces (no tabs) before id1, type and host? It is not rendered well in your message but I'd like to be sure that the spaces are present in your config file.

Also, you set port 9300 which is usually used by elasticsearch for node to node communication. The REST API is usually exposed using the built-in web server that listens on 9200. Maybe changing 9300 to 9200 will help in your case.

(Kenneth Mroz) #11

that was the issue .. Thanks!!

(system) #12