Fabio Coccaro
Fabio Coccaro

Reputation: 1

Kibana: Unable to revive connection: http://elastic-url:9200/

I installed on Centos8:

curl -I localhost:9200/status is ok

curl -I localhost:5601/status --> kibana is not ready yet

In machine with centos7 (.226) all is ok

screenshot 1

This is kibana log:

This is kibana log

Can somebody help me please?

Upvotes: 0

Views: 962

Answers (1)

Domin
Domin

Reputation: 1145

Elasticsearch 7.x.x requires cluster bootstrapping at first launch and Kibana won't start unless Elasticsearch is ready and each node is running Elasticsearch in version 7.x.x.

I will write steps which you would normally do on a real machine, so that anybody else could do the same. In docker it may look similarly, except that you are working in the containers.

Before we kick off, stop kibana and elasticsearch:

service kibana stop
service elasticsearch stop
killall kibana
killall elasticsearch

Make sure it's dead:

service kibana status
service elasticsearch status

Then head into /etc/elasticsearch/ and edit elasticsearch.yml file. Add at the end of the file:

cluster.initial_master_nodes:
  - master-a
  - master-b
  - master-c

Where master-* will be equal to node.name on each node. Save and exit. Start Elasticsearch and then Kibana. On machines with lower memory (~4GB and probably in Docker too, as it normally gives 4GB memory for containers) you may have to start Kibana first, let it "compile", stop it, start Elasticsearch and back Kibana.

On machines with puppet make sure that puppet or cron is not running, just in case not to start off kibana/elastic too early.

Here's source: https://www.elastic.co/guide/en/elasticsearch/reference/master/modules-discovery-bootstrap-cluster.html

Upvotes: 1

Related Questions