Marvel does not work after upgrading to Elastic 2.3.2

monitoring

(Edalford11) #1

I've just upgraded our elastic cluster from 2.1 to 2.3.2. Everything is working great except now marvel is constantly saying "No marvel data found". We have 4 elastic nodes in our cluster, all with

marvel.agent.exporters:
id1:
type: http
host: ["http://10.0.68.51:9200"]

Where that host is a separate node that serves as our monitor node, which has marvel 2.3 installed into Kibana 4.5. The data is coming into kibana just fine and the .marvel-es-data index is still there from before the upgrade. What is wrong here?


(Chris Earle) #2

Hi,

If it's complaining that No marvel data is found at the top level of the UI, then it generally means that the cluster-level details are not coming to the monitoring index.

Can you verify that the elected master node is updated and running Marvel 2.3? Please run this from the monitored cluster and let me know what it comes back with:

$ curl -XGET host:9200/_cat/plugins?v

(Edalford11) #3

Thanks pickypg. I made sure marvel agent was updated and installed before starting the nodes back up after upgrading them. Now somehow the marvel agent is completely missing from every single node. Not sure how that happened but I'm sure thats the problem. I'm sure getting them installed again will fix this problem.


(Chris Earle) #4

It happens.

One important note for anyone that comes across this post (it may not apply to you):

  • Any plugin that modifies the way that Elasticsearch works requires that Elasticsearch be restarted for it to take effect because Java plugins are only loaded at startup. So, if you install the plugin after restarting a node, then it won't take effect until the next restart.

Good luck!


(system) #5