Chapter 7. Low-Level Index Control

In the last chapter, we learned about creating a multi-node cluster and then went through index distribution architecture in Elasticsearch, including choosing the right number of shards and replicas along with a detailed discussion on how routing works, how shard allocation works, and how to alter the default shard allocation behavior. In addition to that, we also discussed what query execution preference is and how to configure it. We also saw how data stripping works and considerations for creating multiple document types inside an Elasticsearch index.

In this chapter, we will cover some granular details of Apache Lucene scoring and how you can alter them and can choose the right scoring algorithm. We will also cover Near Real Time (NRT) searching and indexing and the transaction logs. You will also get to know about segment merging and removed merge policies inside Elasticsearch 5.0. At the end of the chapter, you will also find information about IO throttling...