Neues vom PostgreSQL Planet
Jan Wieremjewicz: The PG_TDE Extension Is Now Ready for Production
Andreas 'ads' Scherbaum: Samed YILDIRIM
Ian Barwick: PgPedia Week, 2025-06-29
Housekeeping announcements:
this website's PostgreSQL installation is now on version 17 ( insert champagne emoji here ) the search function now works properly with non-ASCII characters ( there was an embarrassing oversight which went unnoticed until someone kindly pointed it out ) PostgreSQL 18 changes this weekThis week there have been a couple of renamings:
Tomas Vondra: How often is the query plan optimal?
The basic promise of a query optimizer is that it picks the “optimal” query plan. But there’s a catch - the plan selection relies on cost estimates, calculated from selectivity estimates and cost of basic resources (I/O, CPU, …). So the question is, how often do we actually pick the “fastest” plan? And the truth is we actually make mistakes quite often.
Andreas Scherbaum: Performance Test for Percona Transparent Data Encryption (TDE)
Ian Barwick: PgPedia Week, 2025-06-22
PgPedia Week has been delayed this week due to malaise and other personal circumstances.
semab tariq: Choosing the Right Barman Backup Type and Mode for Your PostgreSQL Highly Available Cluster
When running a PostgreSQL database in a High Availability (HA) cluster, it’s easy to assume that having multiple nodes means your data is safe. But HA is not a replacement for backups. If someone accidentally deletes important data or runs a wrong update query, that change will quickly spread to all nodes in the cluster. Without proper safeguards, that data is gone everywhere. In these cases, only a backup can help you restore what was lost.
Dave Stokes: Entity Relationship Maps
Even the most experienced database professionals are known to feel a little anxious when peering into an unfamiliar database. Hopefully, they inspect to see how the data is normalized and how the various tables are combined to answer complex queries. Entity Relationship Maps (ERM) provide a visual overview of how tables are related and can document the structure of the data.
Gabriele Bartolini: CNPG Recipe 20 – Finer Control of Postgres Clusters with Readiness Probes
Explore the new readiness probe introduced in CloudNativePG 1.26, which advances Kubernetes-native lifecycle management for PostgreSQL. Building on the improved probing infrastructure discussed in my previous article, this piece focuses on how readiness probes ensure that only fully synchronised and healthy instances—particularly replicas—are eligible to serve traffic or be promoted to primary.
Magnus Hagander: PGConf.EU 2025 - Registration opened
The registration for PGConf.EU 2025, which will take place on 21-24 October in Riga, is now open.
We have a limited number of tickets available for purchase with the discount code EARLYBIRD.
This year, the first day of training sessions has been replaced with a Community Events Day. This day has a more limited space, and can be booked as part of the conference registration process or added later, as long as seats last.
Radim Marek: Beyond the Basics of Logical Replication
With First Steps with Logical Replication we set up a basic working replication between a publisher and a subscriber and were introduced to the fundamental concepts. In this article, we're going to expand on the practical aspects of logical replication operational management, monitoring, and dive deep into the foundations of logical decoding.
Cornelia Biacsics: Contributions for the week of 2025-06-09 (Week 24)
On June, 9th, Andrzej Nowicki held a talk about “From Queries to Pints: Building a Beer Recommendation System with pgvector” at the Malmö PostgreSQL User Group.
On June, 3rd the 5. PostgreSQL User Group NRW MeetUp took place in Germany.
Peter Eisentraut: Waiting for SQL:202y: Vectors
It’s been a while since SQL:2023 was published, and work on the SQL standard continues. Nowadays, everyone in the database field wants vectors, and SQL now has them, too.
(I’m using the term “SQL:202y” for “the next SQL standard after SQL:2023”. I took this naming convention from the C standard, but it’s not an official term for the SQL standard. The current schedule suggests a new release in 2028, but we’ll see.)
Andreas 'ads' Scherbaum: Suraj Kharage
Dave Page: PGDay UK 2025 - Registration open
PGDay UK 2025 will be held in London, England, on Tuesday, September 9, 2025 at the Cavendish Conference Centre.
It features a full day with a track of PostgreSQL presentations from global PostgreSQL experts. It will cover a wide range of topics of interest.
Registration has now opened. Seats are limited so we recommend that you register early if you are interested! There are 20 Early bird discounted tickets available until the 31th of July 2025; grab yours before they run out or the campaign ends.
Ahmet Gedemenli: pgstream v0.7.1: JSON transformers, progress tracking and wildcard support for snapshots
semab tariq: Which PostgreSQL HA Solution Fits Your Needs: Pgpool or Patroni?
When designing a highly available PostgreSQL cluster, two popular tools often come into the conversation: Pgpool-II and Patroni. Both are widely used in production environments, offer solid performance, and aim to improve resilience and reduce downtime; however, they take different approaches to achieving this goal.
Alexander Korotkov: The differences between OrioleDB and Neon
In a recent Hacker News discussion, there was some confusion about the differences between OrioleDB and Neon. Both look alike at first glance. Both promise a "next‑gen Postgres". Both have support for cloud‑native storage.
This post explains how the two projects differ in practice. And importantly, OrioleDB is more than an undo log for PostgreSQL.
Robert Haas: PostgreSQL Hacking + Patch Review Workshops for July 2025
Jan Wieremjewicz: PostgreSQL active-active replication, do you really need it?
Active-active, also referred to as multi-primary, is a setup where multiple database nodes can accept writes at the same time and propagate those changes to the others. In comparison, regular streaming replication in PostgreSQL allows only one node (the primary) to accept writes. All other nodes (replicas) are read-only and follow changes.
In an active-active setup: