Mastering PostgreSQL 9.6
()
About this ebook
- Your one-stop guide to mastering the advanced concepts in PostgreSQL with ease
- Master query optimization, replication, and high availability with PostgreSQL
- Extend the functionalities of PostgreSQL to suit your organizational needs with minimum effort Who This Book Is For
If you are a PostgreSQL data architect or an administrator who wants to understand how to implement advanced functionalities and master complex administrative tasks with PostgreSQL, then this book is perfect for you. Prior experience of administrating a PostgreSQL database and a working knowledge of SQL is required to make the best use of this book.
Hans-Jürgen Schönig
Hans-Jurgen Schonig has 15 years of experience with PostgreSQL.He is the CEO of a PostgreSQL consulting and support company called "Cybertec Schonig & Schonig GmbH" (www.postgresql-support.de),which has successfully served countless customers around the globe. Before founding Cybertec Schonig & Schonig GmbH in the year 2000,he worked as database developer at a private research company focusing on the Austrian labor market where he was primarily focusing on data mining and forecast models. He has written several books dealing with PostgreSQL already.
Read more from Hans Jürgen Schönig
Mastering PostgreSQL 12 - Third Edition: Advanced techniques to build and administer scalable and reliable PostgreSQL database applications, 3rd Edition Rating: 0 out of 5 stars0 ratingsPostgreSQL Administration Essentials Rating: 0 out of 5 stars0 ratingsTroubleshooting PostgreSQL Rating: 5 out of 5 stars5/5PostgreSQL Replication - Second Edition Rating: 0 out of 5 stars0 ratings
Related to Mastering PostgreSQL 9.6
Related ebooks
PostgreSQL Server Programming Rating: 0 out of 5 stars0 ratingsPostgreSQL 9 Administration Cookbook LITE: Configuration, Monitoring and Maintenance Rating: 3 out of 5 stars3/5PostgreSQL Development Essentials Rating: 5 out of 5 stars5/5PostgreSQL for Data Architects Rating: 0 out of 5 stars0 ratingsPostgreSQL 11 Administration Cookbook: Over 175 recipes for database administrators to manage enterprise databases Rating: 0 out of 5 stars0 ratingsMastering NGINX - Second Edition Rating: 0 out of 5 stars0 ratingsDistributed Computing in Java 9 Rating: 0 out of 5 stars0 ratingsAngular Services Rating: 0 out of 5 stars0 ratingsPostgreSQL 9 Administration Cookbook - Second Edition Rating: 0 out of 5 stars0 ratingsLearning PostgreSQL Rating: 1 out of 5 stars1/5PostgreSQL Administration Cookbook, 9.5/9.6 Edition Rating: 0 out of 5 stars0 ratingsPractical OneOps Rating: 0 out of 5 stars0 ratingsMastering Scala Machine Learning Rating: 0 out of 5 stars0 ratingsRobust Cloud Integration with Azure Rating: 0 out of 5 stars0 ratingsPrimeFaces Beginner's Guide Rating: 0 out of 5 stars0 ratingsFast Data Processing Systems with SMACK Stack Rating: 0 out of 5 stars0 ratingsHDInsight Essentials - Second Edition Rating: 0 out of 5 stars0 ratingsPostgreSQL 9.0 High Performance Rating: 4 out of 5 stars4/5Expert Cube Development with Microsoft SQL Server 2008 Analysis Services Rating: 5 out of 5 stars5/5PostgreSQL Server Programming - Second Edition Rating: 0 out of 5 stars0 ratingsSQL Server 2014 Development Essentials Rating: 5 out of 5 stars5/5Neo4j Cookbook Rating: 0 out of 5 stars0 ratingsMongoDB Cookbook - Second Edition Rating: 0 out of 5 stars0 ratingsDocker: Build, Test, And Deploy Applications Fast Rating: 0 out of 5 stars0 ratingsExtending Docker Rating: 0 out of 5 stars0 ratingsCreating your MySQL Database: Practical Design Tips and Techniques Rating: 3 out of 5 stars3/5Mastering Eclipse Plug-in Development Rating: 0 out of 5 stars0 ratingsHadoop in Practice Rating: 0 out of 5 stars0 ratingsHigh Availability MySQL Cookbook Rating: 0 out of 5 stars0 ratings
Databases For You
Python Projects for Everyone Rating: 0 out of 5 stars0 ratingsGrokking Algorithms: An illustrated guide for programmers and other curious people Rating: 4 out of 5 stars4/5Excel 2021 Rating: 4 out of 5 stars4/5Learn SQL in 24 Hours Rating: 5 out of 5 stars5/5ITIL 4: Digital and IT strategy: Reference and study guide Rating: 5 out of 5 stars5/5Practical Data Analysis Rating: 4 out of 5 stars4/5SQL QuickStart Guide: The Simplified Beginner's Guide to Managing, Analyzing, and Manipulating Data With SQL Rating: 4 out of 5 stars4/5Visualizing Graph Data Rating: 0 out of 5 stars0 ratingsThe AI Bible, Making Money with Artificial Intelligence: Real Case Studies and How-To's for Implementation Rating: 4 out of 5 stars4/5Data Science Strategy For Dummies Rating: 0 out of 5 stars0 ratingsMastering Blockchain Rating: 5 out of 5 stars5/5Learn SAP SD in 24 Hours Rating: 0 out of 5 stars0 ratingsStar Schema The Complete Reference Rating: 0 out of 5 stars0 ratingsJAVA for Beginner's Crash Course: Java for Beginners Guide to Program Java, jQuery, & Java Programming Rating: 4 out of 5 stars4/5Sap/ABAP Hana Programming: Learn to design and build SAP HANA applications with ABAP/4 Rating: 0 out of 5 stars0 ratingsCompTIA DataSys+ Study Guide: Exam DS0-001 Rating: 0 out of 5 stars0 ratingsPhoenix in Action Rating: 0 out of 5 stars0 ratingsSchaum’s Outline of Fundamentals of SQL Programming Rating: 3 out of 5 stars3/5Audit Culture: How Indicators and Rankings are Reshaping the World Rating: 0 out of 5 stars0 ratingsNode.js Design Patterns - Second Edition Rating: 4 out of 5 stars4/5Access 2019 For Dummies Rating: 0 out of 5 stars0 ratingsSpring in Action, Sixth Edition Rating: 5 out of 5 stars5/5Blockchain For Dummies Rating: 5 out of 5 stars5/5MDM for Customer Data: Optimizing Customer Centric Management of Your Business Rating: 0 out of 5 stars0 ratingsITIL 4: High-velocity IT: Reference and study guide Rating: 0 out of 5 stars0 ratingsChessBase Complete: 2019 Supplement Covering ChessBase 13, 14 & 15 Rating: 0 out of 5 stars0 ratings
Reviews for Mastering PostgreSQL 9.6
0 ratings0 reviews
Book preview
Mastering PostgreSQL 9.6 - Hans-Jürgen Schönig
Title Page
Mastering PostgreSQL 9.6
A comprehensive guide for PostgreSQL 9.6 developers and administrators
Hans-Jürgen Schönig
BIRMINGHAM - MUMBAI
Copyright
Mastering PostgreSQL 9.6
Copyright © 2017 Packt Publishing
All rights reserved. No part of this book may be reproduced, stored in a retrieval system, or transmitted in any form or by any means, without the prior written permission of the publisher, except in the case of brief quotations embedded in critical articles or reviews.
Every effort has been made in the preparation of this book to ensure the accuracy of the information presented. However, the information contained in this book is sold without warranty, either express or implied. Neither the author, nor Packt Publishing, and its dealers and distributors will be held liable for any damages caused or alleged to be caused directly or indirectly by this book.
Packt Publishing has endeavored to provide trademark information about all of the companies and products mentioned in this book by the appropriate use of capitals. However, Packt Publishing cannot guarantee the accuracy of this information.
First published: May 2017
Production reference: 1250517
Published by Packt Publishing Ltd.
Livery Place
35 Livery Street
Birmingham
B3 2PB, UK.
ISBN 978-1-78355-535-2
www.packtpub.com
Credits
About the Author
Hans-Jürgen Schönig has 18 years of experience with PostgreSQL. He is the CEO of a PostgreSQL consulting and support company called Cybertec Schönig & Schönig GmbH (www.postgresql-support.de). It has successfully served countless customers around the globe.
Before founding Cybertec Schönig & Schönig GmbH in 2000, he worked as a database developer at a private research company that focused on the Austrian labor market, where he primarily worked on data mining and forecast models. He has also written several books about PostgreSQL.
About the Reviewer
Shaun Thomas has been working with PostgreSQL since late 2000. From 2011 and beyond, he's been a frequent presenter at the PostgresOpen conference on topics such as handling extreme throughput, high availability, monitoring, architecture, and automation. He contributed a few PostgreSQL extensions, as well as a tool for administering massive database clusters. On occasion, he's even been known to guest lecture at the local university. His goal is to help the community make PostgreSQL a bigger, better database for everyone to enjoy.
www.PacktPub.com
For support files and downloads related to your book, please visit www.PacktPub.com.
Did you know that Packt offers eBook versions of every book published, with PDF and ePub files available? You can upgrade to the eBook version at www.PacktPub.comand as a print book customer, you are entitled to a discount on the eBook copy. Get in touch with us at [email protected] for more details.
At www.PacktPub.com, you can also read a collection of free technical articles, sign up for a range of free newsletters and receive exclusive discounts and offers on Packt books and eBooks.
https://www.packtpub.com/mapt
Get the most in-demand software skills with Mapt. Mapt gives you full access to all Packt books and video courses, as well as industry-leading tools to help you plan your personal development and advance your career.
Why subscribe?
Fully searchable across every book published by Packt
Copy and paste, print, and bookmark content
On demand and accessible via a web browser
Customer Feedback
Thanks for purchasing this Packt book. At Packt, quality is at the heart of our editorial process. To help us improve, please leave us an honest review on this book's Amazon page at https://www.amazon.com/dp/1783555351.
If you'd like to join our team of regular reviewers, you can e-mail us at [email protected]. We award our regular reviewers with free eBooks and videos in exchange for their valuable feedback. Help us be relentless in improving our products!
Table of Contents
Preface
What this book covers
What you need for this book
Who this book is for
Conventions
Reader feedback
Customer support
Errata
Piracy
Questions
PostgreSQL Overview
What is new in PostgreSQL 9.6?
Understanding new database administration functions
Killing idle sessions
Finding more detailed information in pg_stat_activity
Tracking vaccum progress
Improving vacuum speed
Digging into new SQL and developer-related functions
Using new backup and replication functionality
Streamlining wal_level and monitoring
Using multiple synchronous standby servers
Understanding performance-related features
Improving relation extensions
Checkpoint sorting and kernel interaction
Using more advanced foreign data wrappers
Introducing parallel queries
Adding snapshot too old
Summary
Understanding Transactions and Locking
Working with PostgreSQL transactions
Handling errors inside a transaction
Making use of savepoints
Transactional DDLs
Understanding basic locking
Avoiding typical mistakes and explicit locking
Considering alternative solutions
Making use of FOR SHARE and FOR UPDATE
Understanding transaction isolation levels
Considering SSI transactions
Observing deadlocks and similar issues
Utilizing advisory locks
Optimizing storage and managing cleanup
Configuring VACUUM and autovacuum
Digging into transaction wraparound-related issues
A word on VACUUM FULL
Watching VACUUM at work
Making use of snapshot too old
Summary
Making Use of Indexes
Understanding simple queries and the cost model
Making use of EXPLAIN
Digging into the PostgreSQL cost model
Deploying simple indexes
Making use of sorted output
Using more than one index at a time
Using bitmap scans effectively
Using indexes in an intelligent way
Improving speed using clustered tables
Clustering tables
Making use of index only scans
Understanding additional B-tree features
Combined indexes
Adding functional indexes
Reducing space consumption
Adding data while indexing
Introducing operator classes
Hacking up an operator class for a B-tree
Creating new operators
Creating operator classes
Testing custom operator classes
Understanding PostgreSQL index types
Hash indexes
GiST indexes
Understanding how GiST works
Extending GiST
GIN indexes
Extending GIN
SP-GiST indexes
BRIN indexes
Extending BRIN indexes
Adding additional indexes
Achieving better answers with fuzzy searching
Taking advantage of pg_trgm
Speed up LIKE queries
Handling regular expressions
Understanding full-text search - FTS
Comparing strings
Defining GIN indexes
Debugging your search
Gathering word statistics
Taking advantage of exclusion operators
Summary
Handling Advanced SQL
Introducing grouping sets
Loading some sample data
Applying grouping sets
Investigating performance
Combining grouping sets with the FILTER clause
Making use of ordered sets
Understanding hypothetical aggregates
Utilizing windowing functions and analytics
Partitioning data
Ordering data inside a window
Using sliding windows
Abstracting window clauses
Making use of onboard windowing functions
rank and dense_rank functions
ntile() function
lead() and lag() functions
first_value(), nth_value(), and last_value() functions
row_number() function
Writing your own aggregates
Creating simple aggregates
Adding support for parallel queries
Improving efficiency
Writing hypothetical aggregates
Summary
Log Files and System Statistics
Gathering runtime statistics
Working with PostgreSQL system views
Checking live traffic
Inspecting databases
Inspecting tables
Making sense of pg_stat_user_tables
Digging into indexes
Tracking the background worker
Tracking, archiving, and streaming
Checking SSL connections
Inspecting transactions in real time
Tracking vacuum progress
Using pg_stat_statements
Creating log files
Configuring postgresql.conf file
Defining log destination and rotation
Configuring syslog
Logging slow queries
Defining what and how to log
Summary
Optimizing Queries for Good Performance
Learning what the optimizer does
Optimizations by example
Evaluating join options
Nested loops
Hash joins
Merge joins
Applying transformations
Inlining the view
Flattening subselects
Applying equality constraints
Exhaustive searching
Trying it all out
Making the process fail
Constant folding
Understanding function inlining
Join pruning
Speedup set operations
Understanding execution plans
Approaching plans systematically
Making EXPLAIN more verbose
Spotting problems
Spotting changes in runtime
Inspecting estimates
Inspecting buffer usage
Fixing high buffer usage
Understanding and fixing joins
Getting joins right
Processing outer joins
Understanding the join_collapse_limit variable
Enabling and disabling optimizer settings
Understanding genetic query optimization
Partitioning data
Creating partitions
Applying table constraints
Modifying inherited structures
Moving tables in and out of partitioned structures
Cleaning up data
Adjusting parameters for good query performance
Speeding up sorting
Speedup administrative tasks
Summary
Writing Stored Procedures
Understanding stored procedure languages
The anatomy of a stored procedure
Introducing dollar quoting
Making use of anonymous code blocks
Using functions and transactions
Understanding various stored procedure languages
Introducing PL/pgSQL
Handling quoting
Managing scopes
Understanding advanced error handling
Making use of GET DIAGNOSTICS
Using cursors to fetch data in chunks
Utilizing composite types
Writing triggers in PL/pgSQL
Introducing PL/Perl
Using PL/Perl for datatype abstraction
Deciding between PL/Perl and PL/PerlU
Making use of the SPI interface
Using SPI for set returning functions
Escaping in PL/Perl and support functions
Sharing data across function calls
Writing triggers in Perl
Introducing PL/Python
Writing simple PL/Python code
Using the SPI interface
Handling errors
Improving stored procedure performance
Reducing the number of function calls
Using cached plans
Assigning costs to functions
Using stored procedures
Summary
Managing PostgreSQL Security
Managing network security
Understanding bind addresses and connections
Inspecting connections and performance
Living in a world without TCP
Managing pg_hba.conf
Handling SSL
Handling instance-level security
Creating and modifying users
Defining database-level security
Adjusting schema-level permissions
Working with tables
Handling column-level security
Configuring default privileges
Digging into row-level security - RLS
Inspecting permissions
Reassigning objects and dropping users
Summary
Handling Backup and Recovery
Performing simple dumps
Running pg_dump
Passing passwords and connection information
Using environment variables
Making use of .pgpass
Using service files
Extracting subsets of data
Handling various data formats
Replaying backups
Handling global data
Summary
Making Sense of Backups and Replication
Understanding the transaction log
Looking at the transaction log
Understanding checkpoints
Optimizing the transaction log
Transaction log archiving and recovery
Configuring for archiving
Confguring the pg_hba.conf file
Creating base backups
Reducing the bandwidth of a backup
Mapping tablespaces
Using different formats
Testing transaction log archiving
Replaying the transaction log
Finding the right timestamp
Cleaning up the transaction log archive
Setting up asynchronous replication
Performing a basic setup
Improving security
Halting and resuming replication
Checking replication to ensure availability
Performing failovers and understanding timelines
Managing conflicts
Making replication more reliable
Upgrading to synchronous replication
Adjusting durability
Making use of replication slots
Handling physical replication slots
Handling logical replication slots
Use cases of logical slots
Summary
Deciding on Useful Extensions
Understanding how extensions work
Checking for available extensions
Making use of contrib modules
Using the adminpack
Applying bloom filters
Deploying btree_gist and btree_gin
Dblink - consider phasing out
Fetching files with file_fdw
Inspecting storage using pageinspect
Investigating caching with pg_buffercache
Encrypting data with pgcrypto
Prewarming caches with pg_prewarm
Inspecting performance with pg_stat_statements
Inspecting storage with pgstattuple
Fuzzy searches with pg_trgm
Connecting to remote servers using postgres_fdw
Handling mistakes and typos
Other useful extensions
Summary
Troubleshooting PostgreSQL
Approaching an unknown database
Inspecting pg_stat_activity
Querying pg_stat_activity
Treating Hibernate statements
Figuring out where queries come from
Checking for slow queries
Inspecting individual queries
Digging deeper with perf
Inspecting the log
Checking for missing indexes
Checking for memory and I/O
Understanding noteworthy error scenarios
Facing clog corruption
Understanding checkpoint messages
Managing corrupted data pages
Careless connection management
Fighting table bloat
Summary
Migrating to PostgreSQL
Migrating SQL statements to PostgreSQL
Using lateral joins
Supporting lateral
Using grouping sets
Supporting grouping sets
Using WITH clause - common table expressions
Supporting WITH clause
Using WITH RECURSIVE clause
Supporting WITH RECURSIVE clause
Using FILTER clause
Supporting FILTER clause
Using windowing functions
Supporting windowing and analytics
Using ordered sets - WITHIN GROUP clause
Supporting WITHIN GROUP clause
Using TABLESAMPLE clause
Supporting TABLESAMPLE clause
Using limit/offset
Supporting FETCH FIRST clause
Using OFFSET
Supporting OFFSET clause
Using temporal tables
Supporting temporal tables
Matching patterns in time series
Moving from Oracle to PostgreSQL
Using the oracle_fdw extension to move data
Using ora2pg to migrate from Oracle
Common pitfalls
Moving from MySQL or MariaDB to PostgreSQL
Handling data in MySQL and MariaDB
Changing column definitions
Handling null values
Expecting problems
Migrating data and schema
Using pg_chameleon
Using foreign data wrappers
Summary
Preface
PostgreSQL is an open source database management tool used for handling large datasets (big data) and as a JSON document database. It also has applications in the software and web domains. This book will enable you to build better PostgreSQL applications and administer databases more efficiently.
What this book covers
Chapter 1, PostgreSQL Overview, will give you an overview of PostgreSQL and its features. You will learn about new stuff and new functionality available in PostgreSQL.
Chapter 2, Understanding Transactions and Locking, will cover one of the most important aspects of any database system. Proper database work is usually not possible without the existence of transactions, and understanding transactions and locking is vital to performance as well as professional work.
Chapter 3, Making Use of Indexes, covers everything you need to know about indexes. Indexes are key to performance and are therefore an important cornerstone if you want good user experience and high throughput. All important aspects of indexing will be covered.
Chapter 4, Handling Advanced SQL, will introduce some of the most important concepts of modern SQL. You will learn about windowing functions as well as other important, more modern, elements of SQL.
Chapter 5, Log Files and System Statistics, will guide you through more administrative tasks, such as log file management and monitoring. You will learn how to inspect your servers and extract runtime information from PostgreSQL.
Chapter 6, Optimizing for Good Query Performance, will tell you everything you need to know about good PostgreSQL performance. The chapter will cover SQL tuning as well as information about memory management.
Chapter 7, Writing Stored Procedures, teaches you some more advanced topics related to server-side code. The most important server-side programming languages are covered and important aspects are pointed out.
Chapter 8, Managing PostgreSQL Security, has been designed to help you improve the security of your server. The chapter features everything from user management to row-level security. Information about encryption is also included.
Chapter 9, Handling Backup and Recovery, is all about backups and data recovery. You will learn to backup your data and it will enable you to restore things in case of disaster.
Chapter 10, Making Sense of Backups and Replication, is all about redundancy. You will learn to asynchronously and synchronously replicate PostgreSQL database systems. All modern features are covered as extensively as possible.
Chapter 11, Deciding on Useful Extensions, describes widely used modules that add additional functionality to PostgreSQL. You will learn about the most common extensions.
Chapter 12, Troubleshooting PostgreSQL, offers a systematic approach to fixing problems in PostgreSQL. It will enable you to spot common problems and approach them in an organized way.
Chapter 13, Migrating to PostgreSQL, is the final chapter of this book and shows you the way from commercial databases to PostgreSQL. The most important databases migrated these days will be covered.
What you need for this book
This book has been written for a broad audience. In order to follow the examples presented in this book, it makes sense to have at least some experience with SQL and maybe even PostgreSQL in general (although this is not a hard requirement). In general, it is a good idea to be familiar with the Unix command line.
Who this book is for
This book has explicitly been written for people who want to know more about PostgreSQL and who are not satisfied with basic information. The aim is to write a book that goes a bit deeper and explains the most important stuff in a clear and easy-to-understand way.
Conventions
In this book, you will find a number of text styles that distinguish between different kinds of information. Here are some examples of these styles and an explanation of their meaning.
Code words in text, database table names, folder names, filenames, file extensions, pathnames, dummy URLs, user input, and Twitter handles are shown as follows: In this case, the \timing command will tell psql to show the runtime of a query.
Any command-line input or output is written as follows:
New terms and important words are shown in bold.
Warnings or important notes appear in a box like this.
Tips and tricks appear like this.
Reader feedback
Feedback from our readers is always welcome. Let us know what you think about this book—what you liked or disliked. Reader feedback is important for us as it helps us develop titles that you will really get the most out of.
To send us general feedback, simply e-mail [email protected], and mention the book's title in the subject of your message.
If there is a topic that you have expertise in and you are interested in either writing or contributing to a book, see our author guide at www.packtpub.com/authors.
Customer support
Now that you are the proud owner of a Packt book, we have a number of things to help you to get the most from your purchase.
Errata
Although we have taken every care to ensure the accuracy of our content, mistakes do happen. If you find a mistake in one of our books-maybe a mistake in the text or the code-we would be grateful if you could report this to us. By doing so, you can save other readers from frustration and help us improve subsequent versions of this book. If you find any errata, please report them by visiting http://www.packtpub.com/submit-errata, selecting your book, clicking on the Errata Submission Form link, and entering the details of your errata. Once your errata are verified, your submission will be accepted and the errata will be uploaded to our website or added to any list of existing errata under the Errata section of that title.
To view the previously submitted errata, go to https://www.packtpub.com/books/content/support and enter the name of the book in the search field. The required information will appear under the Errata section.
Piracy
Piracy of copyrighted material on the Internet is an ongoing problem across all media. At Packt, we take the protection of our copyright and licenses very seriously. If you come across any illegal copies of our works in any form on the Internet, please provide us with the location address or website name immediately so that we can pursue a remedy.
Please contact us at [email protected] with a link to the suspected pirated material.
We appreciate your help in protecting our authors and our ability to bring you valuable content.
Questions
If you have a problem with any aspect of this book, you can contact us at [email protected], and we will do our best to address the problem.
PostgreSQL Overview
PostgreSQL is one of the world's most advanced open source database systems and it has many features widely used by developers and system administrators alike. In this book, many of those cool features will be covered and discussed in great detail.
In this chapter, you will be introduced to PostgreSQL and the cool new features available in PostgreSQL 9.6 and beyond. All relevant new functionality will be covered in detail. Given the sheer number of changes made to the code and given the size of the PostgreSQL project, this list of features is of course not complete, so I tried to focus on the most important aspects relevant to most people.
The features outlined in this chapter will be split into the following categories:
Database administration
SQL and developer-related
Backup, recovery, and replication
Performance-related topics
What is new in PostgreSQL 9.6?
PostgreSQL 9.6 was released in late 2016 and is the last version that will still be following the old numbering scheme PostgreSQL has been using for more than a decade now. From PostgreSQL 10.0 onward, a new version numbering system will be in place. From 10.0 on, major releases will happen way more frequently.
Understanding new database administration functions
PostgreSQL 9.6 has many new features that can help the administrator to reduce work and make systems more robust.
One of those features is the idle_in_transaction_session_timeout function.
Killing idle sessions
In PostgreSQL, a session or a transaction can basically live almost forever. In some cases, this has been a problem because transactions were kept open for too long. Usually, this was due to a bug. The trouble is this: insanely long transactions can cause cleanup problems and table bloat can occur. The uncontrolled growth of a table (table bloat) naturally leads to performance problems and unhappy end users.
Starting with PostgreSQL 9.6, it is possible to limit the duration a database connection is allowed to spend inside a transaction without performing real work. Here is how it works:
Administrators and developers can set a timeout, which is 2.5 seconds in my example. As soon as a transaction is idle for too long, the connection will be terminated automatically by the server. Nasty side effects of long idle transactions can be prevented easily by adjusting this parameter.
Finding more detailed information in pg_stat_activity
The pg_stat_activity function is a system view that has been around for many years. It basically contains a list of active connections. In older versions of PostgreSQL, administrators could see that a query is waiting for somebody else—however, it was not possible to figure out why and for whom. This has changed in 9.6. Two columns have been added:
In addition to this extension, a new procedure has been added, which shows who caused whom to wait:
When the function is called, it will return a list of blocking PIDs.
Tracking vaccum progress
For many years, people have asked for a progress tracker for vacuum. Finally, PostgreSQL 9.6 makes this wish come true by introducing a new system view. Here is how it works:
PostgreSQL will provide detailed information about ongoing vacuum processes so that people can track the progress of this vital operation.
Improving vacuum speed
PostgreSQL 9.6 not only provides you with deeper insights into what vacuum does at the moment, it will also speed up the process in general. From PostgreSQL 9.6 onward, PostgreSQL will keep track of all frozen pages and avoid vacuuming those pages.
Tables that are mostly read-only will massively benefit from this change, as vacuum load is drastically reduced.
Digging into new SQL and developer-related functions
One of the most promising new features of PostgreSQL is the ability to perform phrase searching. Up to 9.5 it was only possible to search for words—phrase searching was very hard to do. 9.6 nicely removes this limitation. Here is an example of how it works:
The first query returns false because the words we are looking for do not occur in the desired order. In the second example, true is returned because there really is a proper match.
However, there is more: in 9.6 it is possible to check whether words show up in a certain order. In the following example, we want a word to be between united and nations:
The second example returns false as there is no word between united and nations.
Using new backup and replication functionality
PostgreSQL 9.6 has also seen improvements in the area of backup and recovery.
Streamlining wal_level and monitoring
The wal_level setting has always been a bit hard to understand for many people. Many were struggling with the difference between the archive and hot_standby settings. To remove this confusion altogether, both settings have been replaced with the easier-to-understand replica setting, which does the same as hot_standby.
In addition to that, the monitoring of replicated setups has been simplified. Prior to 9.6, there was only the pg_stat_replication view, which could be queried on the master to supervise the flow of data to the slave. Now it is also possible to monitor the flow of data on the slaves, by consulting the pg_stat_wal_receiver function. It is basically the slave-side mirror of the pg_stat_replication function and helps to determine the state of replication.
Using multiple synchronous standby servers
PostgreSQL has been able to perform synchronous replication for quite a while already. In PostgreSQL, it is possible to have more than just one synchronous server from 9.6 onward. Earlier, only one server had to acknowledge a commit. Now it is possible to have an entire group of servers that has to confirm a commit. This is especially important if you want to improve reliability in case of multi-node error.
The syntax to use this new feature is simple:
However, there is more to synchronous replication in PostgreSQL 9.6. Previously, PostgreSQL ensured (synchronous_commit = on) that the transaction log has reached the slave. However, this did not mean that data was actually visible. Consider an example: somebody adds a user to the master, instantly connects to the slave, and checks for the user. While the transaction log was guaranteed to be on the slave, it was not necessarily guaranteed that the data inside the log was already visible to the end user (due to replication conflicts and so on). By setting synchronous_commit = 'remote_apply', it is now possible to query the slave directly after a commit on the master, without having to worry that data might not be visible yet. The remote_apply value is slower than the on value but it allows to write more advanced applications.
Understanding performance-related features
Just like every release of PostgreSQL, there are numerous performance improvements, which can help to speedup applications. In this section, I want to focus on the most important and most powerful ones. Of course, there are many more small improvements than listed here.
Improving relation extensions
For many years PostgreSQL has extended a table (or an index) block by block. In the case of a single writer process, this was usually fine. However, in cases of high-concurrency writing, writing a block at a time was a source of contention and suboptimal performance. From 9.6 onward, PostgreSQL started to extend tables by multiple blocks at a time. The number of blocks added at a time is 20 times the number of waiting processes.
Checkpoint sorting and kernel interaction
When PostgreSQL writes changes to disk during a checkpoint, it now does so in a more orderly way to ensure that writes are more sequential than earlier. This is done by sorting blocks before sending them too. Random writes will be dramatically reduced this way, which in turn leads to higher throughput on most hardware.
Sorted checkpoints are not the only scalability thing to make it into 9.6. There are also new kernel write-back configuration options: what does this mean? In case of large caches, it could take quite a long time to write all changes out. This used to be especially nasty on systems with hundreds of gigabytes of memory because fairly intense I/O storms could happen. Of course, the operating system, level behavior of Linux could be changed using the /proc/sys/vm/dirty_background_ratio command. However, only a handful of consultants and system administrators actually knew how to do that and why. The checkpoint_flush_after, bgwriter_flush_after, and backend_flush_after functions can be used now to control the flush behavior. In general, the rule is to flush earlier. Still, as the feature is new, people are still gathering experience on how to use those settings in the most efficient way possible.
Using more advanced foreign data wrappers
Foreign data wrappers have been around for many years. Starting with PostgreSQL 9.6, the optimizer can use foreign tables way more efficiently. This includes join push down (joins can now already be performed remotely) and order push down (sorting can now happen remotely). Distributing data inside a cluster is now way more efficient due to faster remote operations.
Introducing parallel queries
Traditionally, a query had to run on a single CPU. While this was just fine in the OLTP world, it started to be a problem for analytical applications, which were bound to the speed provided by a single core. With PostgreSQL 9.6, parallel queries were introduced. Of course, implementing parallel queries was hard and so a lot of infrastructure has already been implemented over the years. All this infrastructure is now available to provide the end user with parallel sequential scans. The idea is to make many CPUs work on complicated WHERE conditions during a sequential scan. Version 9.6 also allowed for parallel aggregates and parallel joins. Of course, there is a lot of work left, but we are already looking at a major leap forward.
To control parallelism, there are two essential settings:
The first one limits the overall number of worker processes available. The second one controls the number of workers allowed per gather node.
A gather node is a new thing you will see in an execution plan. It is in charge of unifying results coming from parallel subprocesses.
In addition to those fundamental settings, there are a couple of new optimizer parameters to adjust the cost of parallel queries.
Adding snapshot too old
Those of you using Oracle would be aware of the following error message: snapshot too old. In Oracle, this message indicates that a transaction has been too long, so it has to be aborted. In PostgreSQL, transactions can run almost infinitely. However, long transactions can still be a problem, so the snapshot too old error has been added as a feature to 9.6, which allows transactions to be aborted after a certain amount of time.
The idea behind that is to prevent table bloat and to make sure that end users are aware of the fact that they might be about to do something stupid.
Summary
In PostgreSQL 9.6 and 10.0, a lot of functionality has been added, which allows