PGCon 2023, day 3


This is the last time PGCon will be held in Ottawa! Next year it’ll be in another city. My first time at PGCon is the end of an era!

From yesterday:

Automating index selection using Constraint Programming: One of the speakers was Lukas Fittl of PGAnalyze. He’s as good a speaker in person as he is on the PGAnalyze webinars. Lots of theory and research behind this feature. A simplified version is available as an open-source tool. Looking forward to trying this feature.

The journey towards active-active replication in PostgreSQL: Another good talk. PG 16 has bidirectional replication, the roadmap for PG 17 and beyond includes: DDL replication, replication of sequences, synchronization of replication slots to allow failover, upgrading logical replication nodes, and time-delayed replication.

PostgreSQL is nowhere near ready – 3 new directions to develop PostgreSQL: Discussion of three features that the speaker believes ought to be in Postgres.

Continuous Aggregates: A use case of Materialized views: Discussion of TimescaleDB’s Continuous Aggregates, which are like materialized views but are continuously and incrementally updated by use of update and insert pointers and watermarks. Table data must be partitioned by time.

PgBouncer: Present and Future: A disappointing talk that was aimed at beginners. It had a low density of information but a useful nugget was the PGBouncer roadmap: Prepared statements in transaction pooling mode, supporting LDAP authentication, FIPS compliance, fixing “multiple entries for the same user” bugs, OpenSSL support overhaul, additional per-pool settings, and maybe multi-threading.

Leave a comment

This site uses Akismet to reduce spam. Learn how your comment data is processed.