PostgreSQL's developments for high volumes processing

Since few years, postgreSQL received several enhancements to process high volumes databases.

Logical replication internals

This post explains how logical replication works. Especially, difference beharvoirs depending of workloads

Postgres and Netdata : apply to autovacuum and pg_stat_bgwriter

Introduction of autovacuum and bgwriter charts dedicated to PostgreSQL in Netdata

PostgreSQL - JSONB and Statistics

This post shows how PostgreSQL use statistics and the restrictions with JSONB type

PostgreSQL 10 : ICU & Abbreviated Keys

Almost everyone has heard of partitioning and logical replication in PostgreSQL 10. Have you heard about the support of ICU collations (International Components for Unicode)? This article will present what this new feature is but also the possible gains by exploiting abbreviated keys.

PostgreSQL 10 : Performances improvements

PostgreSQL 10 is coming soon, it is scheduled for tomorrow : See this commit This release includes expected features : Logical replication Native partitioning Better parallelism support Multi-column statistics … For an exhaustive list see: Releases notes Wiki page about v10 In this article I will expose you performance improvements that are not listed in the releases notes! Surprisingly, the community does not list these kinds of improvements: they do not represent a significant change from user experience.

PGDay : How does Full Text Search works?

During last PGDay I gave a presentation how Full Text Search works in PostgreSQL. This feature is unfortunately not well known. I see several reasons for this: Complexity: The FTS uses unknown notions from DBA: stemming, vector representation of a document … The tendency to use a dedicated tool for full-text search: ElasticSearch, SOLR … PostgreSQL’s advanced features are not known. However, there are several advantages to use the PostgreSQL FTS:

PostgreSQL 10 and Logical replication - Setup

This article is the result of a series of articles on logical replication in PostgreSQL 10 This one will focus on the implementation of logical replication.

PostgreSQL 10 and Logical replication - Overview

Next PostgreSQL version is approaching. This version comes with an impressive feature list : Native partionning Sorts and aggregation improvements Better parallelism support : parallel index scan, parallel hash join, parallelism for subquery Extended statistics ICU collation: enable use of “abbreviated keys”, disabled in 9.5.2 due to libc bug. Abbreviated keys brings sort improvements (arround 20-30%). It is usefull when a query need a sort or for index creation. … look at wiki page New in Postgres 10 or releases notes. Another attended feature is logical replication. I will present it in a serie of articles.

PostgreSQL : Deferrable constraints

Differ constraints verification Note: This article was written during my activity at Dalibo Postgres respects ACID properties, so it guarantees the consistency of the database: a transaction will bring the database from one valid state to another. The data in the different tables is not independent but obeys semantic rules put in place when designing the conceptual model. The main purpose of integrity constraints is to ensure the consistency of the data between them, and therefore to ensure that they respect these semantic rules. If an insert, an update or a delete violates these rules, the transaction is purely and simply canceled. Postgres performs constraint verification on each change (when constraints have been defined). It is also possible to delay the checking of the constraints at the end of the transaction, at the time of the commit. Thus, the verifications will only be produced on the effective changes between the delete, update and insert operations of the whole transaction.