It seems to me we cannot use Embedded Elasticsearch anymore like what we used to start our application and build a NodeClient and join the cluster to #1. reduce network hop #2. give a name to this node client so that we can identify which application is connecting to the cluster.
The article about using TransportClient alongside the Client Node is the way to go if you want to have autodiscovery and such.
But if you know for sure some of the addresses of some nodes, then you can directly use the TransportClient.
Or better, start to use the Rest Client. It's a low level one but 5.6.0 will come with a high level one.
I started to use it and it's really great.
Apache, Apache Lucene, Apache Hadoop, Hadoop, HDFS and the yellow elephant
logo are trademarks of the
Apache Software Foundation
in the United States and/or other countries.