Real talk about the challenges faced by PostgreSQL developers/hackers is something that I very much welcome. Kudos for normalizing it, Robert Haas.

In today's 5mins of Postgres we take a look at the recent B-Tree Index Scan improvements committed to Postgres 17 by Peter Geoghegan together with Matthias van de Meent. If you're using IN lists or ANY= in Postgres, this will likely make your queries faster: pganalyze.com/blog/5mins-pos…

Simon Riggs | #PostgreSQL Person of the Week j.mp/3dEAlXE

Simon Riggs gave me my start as a Postgres contributor back in 2011. He saw something in me at a time when it really mattered. I'll miss him.

Simon Riggs was a visionary software developer who contributed fundamental work on the PostgreSQL database, replication and so much more. He was my business partner and mentor for a period where we wrote a pair of books together. I will miss him.

metro.co.uk/2024/03/27/pla…

Another explained result, it wasn't a perf regression in Postgres 14.

smalldatum.blogspot.com/2024/02/it-was…

AWS has dramatically increased its contributions with PostgreSQL 16 by understanding the project history and working closely with the community says core team member Jonathan S. Katz. #OSSummit #AWSCloud #opensource AWS Open Source

An example where things not yet cleaned up by MVCC GC can make queries too slow for InnoDB and MyRocks but not for Postgres.

smalldatum.blogspot.com/2023/07/myrock…

Much detail, nothing but good news from Postgres:
* CPU overhead doesn't change much from v11 to v15
* A few things are much faster in v15 (full table scan, update the same row)

Context is: small server, low concurrency, in-memory

smalldatum.blogspot.com/2023/03/search…

more metrics in pg16! pg_stat_io, and now this

Insert Benchmark vs Postgres/MySQL vs ARM/x86
* x86 did better at 1 thread, results similar with concurrency
* Postgres did best, InnoDB was next, MyRocks was worst (in-memory benchmark, expected)
* some open issues like figuring out compiler flags

smalldatum.blogspot.com/2023/01/the-in…

Christoph Engelbert / Noctarius ツ / エンゲルベルト クリス Franck Pachot The regular tables and indexes went from 76 GB to 49 GB, so around 35%.

I'll be writing a detailed blog post about it :)

Flying to Berlin for PGConf Europe tonight. Looking forward to catching up with everyone!

note: not exaggerating on this one. [postgresql.org, ~2011, around the time our work on Serializable Snapshot Isolation became available]

Fascinating disconnect I’ve seen on this one:
- academic researchers: SERIALIZABLE isn’t the default? But everyone enables it first thing, right?
- database users: SERIALIZABLE exists?