Terminology

• ElasticSearch is a near real time search platform build on top of Apache Lucene™

• ElasticSearch supports ACID for a single document only

• Cluster is a collection of one or more nodes storing, indexing, and searching data

• Node is a single server in a cluster, nodes discover each other with unicast and by cluster name defined in /etc/elasticsearch/elasticsearch.yml

• Node can be configured as master-eligible, data, client, tribe, or left default: master-eligible and data node

• Index is a collection of documents and equivalent of Rdbms (Relational Database Management System) database or schema

• Type is a leftover from previous version as collection of documents of a specific type inside index, v6 supports one type per index only

• Document is a basic unit of operation for indexing, replications, and searching. An approximate equivalent of a row/record in Rdbms. Document can be searched for in an index, but when being indexed must be assigned to a type