Book Image

PostgreSQL 16 Administration Cookbook

By : Gianni Ciolli, Boriss Mejías, Jimmy Angelakos, Vibhor Kumar, Simon Riggs
5 (1)
Book Image

PostgreSQL 16 Administration Cookbook

5 (1)
By: Gianni Ciolli, Boriss Mejías, Jimmy Angelakos, Vibhor Kumar, Simon Riggs

Overview of this book

PostgreSQL has seen a huge increase in its customer base in the past few years and is becoming one of the go-to solutions for anyone who has a database-specific challenge. This PostgreSQL book touches on all the fundamentals of Database Administration in a problem-solution format. It is intended to be the perfect desk reference guide. This new edition focuses on recipes based on the new PostgreSQL 16 release. The additions include handling complex batch loading scenarios with the SQL MERGE statement, security improvements, running Postgres on Kubernetes or with TPA and Ansible, and more. This edition also focuses on certain performance gains, such as query optimization, and the acceleration of specific operations, such as sort. It will help you understand roles, ensuring high availability, concurrency, and replication. It also draws your attention to aspects like validating backups, recovery, monitoring, and scaling aspects. This book will act as a one-stop solution to all your real-world database administration challenges. By the end of this book, you will be able to manage, monitor, and replicate your PostgreSQL 16 database for efficient administration and maintenance with the best practices from experts.
Table of Contents (15 chapters)
13
Other Books You May Enjoy
14
Index

Speeding up queries without rewriting them

Often, you either can’t or don’t want to rewrite a query. However, you can still try and speed it up through any of the techniques we will discuss here.

How to do it…

By now, we assume that you’ve looked at various problems already, so the following are more advanced ideas for you to try.

Increasing work_mem

For queries involving large sorts or for join queries, it may be useful to increase the amount of working memory that can be used for query execution. Try setting the following:

SET work_mem = '1TB';

Then, run EXPLAIN (not EXPLAIN ANALYZE). If EXPLAIN changes for the query, then it may benefit from more memory. I’m guessing that you don’t have access to 1 terabyte (TB) of RAM; the previous setting was only used to prove that the query plan is dependent on available memory. Now, issue the following command:

RESET work_mem;

Now, choose a more appropriate...