I have a cluster consisting of twelve nodes (CentOS 7). It had ES 2.3.0 and I wanted to upgrade it to 2.4.0. I followed the rolling upgrades instructions for each node and thought it all went well. But if I did a curl localhost:9200 then I saw it was still using ES 2.3.0 instead of the new one.
So I did a yum remove elasticsearch on all nodes and then installed the latest version of elasticsearch. Updated the elasticsearch.yml files (clustername, nodename, netwerkhost 0.0.0.0, unicast nodes) and did a systemctl daemon-reload and then a service elasticsearch restart on all nodes. Then I tried on the master node a curl localhost:9200 but now it says:
curl: (7) Failed connect to localhost:9200; Connection refused.
I also tried this command which I've found after googling but it did not help: sudo chown -R elasticsearch:elasticsearch /var/lib/elasticsearch/
I have no idea why I can't get the cluster up and running and I have a gut feeling that I did not do logical things but I hope you guys can point me in the right direction.