Neues vom PostgreSQL Planet

Due to an unfortunate recent visitation by the Virus of the Decade (so far), I have a backlog of these which I'm trying to work through, so in the remote chance anyone is waiting with bated breath for the newest editions, my apologies. Normal service will be resumed as soon as humanly possible.

Almost a month late, but I hope you enjoy it!

PostgreSQL and NUMA, part 2 of 4

This series covers the specifics of running PostgreSQL on large systems with many processors. My experience is that people often spend months learning the basics when confronted with the problem. This series tries to dispel these difficulties by providing a clear background into the topics in question. The hope is that future generations of database engineers and administrators don’t have to spend months figuring things out through trial and error.

In one of our earlier blogs, we explored how improper volatility settings in PL/pgSQL functions — namely using IMMUTABLE, STABLE, or VOLATILE — can lead to unexpected behavior and performance issues during migrations.

As a database consulting company, we are often faced with analytics and reporting related tasks which seem to be easy on the surface but are in reality not that trivial. The number of those seemingly simple things is longer than one might think, especially in the area of reporting

I pushed the most boring change imaginable, add an index. Our CI/CD pipeline is textbook ==> spin up a fresh DB, run every migration file in one single transaction, in sequential manner. If anything hiccups, the whole thing rolls back and the change never hits main. Foolproof autotests.

Enter The Drama Queen :

As part of the new ABI compliance reporting system for the PostgreSQL Buildfarm, there has long been a question: if a Buildfarm run fails due to an ABI incompatibility, how can it be marked successful again once the ABI break is fixed? This post explains the introduction of a new file, .abi-compliance-history, in the PostgreSQL source tree and how it is used to track ABI breaks.

When using tsearch one usually, often, creates a tsvector column to put data in, and then create index on it. But, do you really need the index? I wrote once already that you don't have to, but then a person talked with me on IRC, and pointed this section of docs: One advantage of the … Continue reading "Do you really need tsvector column?"

As I explained in my talk on PostgreSQL Conference Europe 2025, data corruption can be silently present in any PostgreSQL database and will remain undetected until we physically read corrupted data. There can be many reasons why some data blocks in tables or other objects can be damaged. Even modern storage hardware is far from being infallible. Binary backups done with pg_basebackup tool – which is very common backup strategy in PostgreSQL environment – leave these problems hidden. Because they do not check data but copy whole data files as they are.