PostgreSQL 19 Daily Brief — Day 1: November 22, 2026

//

//

PostgreSQL 19 Daily Brief — Day 1: November 22, 2026

Marcus Vance — Lead Systems & Backend Architect
November 22, 2026


The Lay of the Land

PostgreSQL 19 dropped on November 19, 2026 — three weeks later than the usual autumn window after extended testing and feature reversions. We’re 1 day into the release cycle. The initial announcement fanfare has settled; now the real work begins: understanding what actually works in production, what doesn’t, and where the sharp edges live.

Feature Pulse: What’s Moving

Postgres 19 REPACK: Built-In Online Table Bloat Cleanup

Source: https://www.snowflake.com/en/blog/engineering/postgres-19-repack-table-bloat

REPACK is my favorite feature in Postgres 19. It’s not the flashiest, but it is the one that’s going to matter most if you’re running a production database. Vacuum, autovacuum and bloat have always been this sore spot for Postgres users. On the one hand, Postgres MVCC design is not new. And yet, on high update systems, this churn of having dead rows that get freed up later, vacuum being tweaked to the right levels to maintain low bloat is still a bit of an inexact science.

Once bloat does exist, you’re stuck with a choice: live with degraded performance from table bloat, or run VACUUM FULL which takes an exclusive lock and rewrites the entire table — effectively downtime for that table. Neither option is great for 24/7 systems.

VACUUM FULL Locks Your Table the Whole Time. Postgres 19’s REPACK CONCURRENTLY Doesn’t

Source: https://dev.to/gabrielanhaia/vacuum-full-locks-your-table-the-whole-time-postgres-19s-repack-concurrently-doesnt-186

PG19 Hacktober: 31 Days of New Features: REPACK — ONE COMMAND TO REPLACE VACUUM FULL AND CLUSTER WITH NEW CONCURRENTLY

Source: https://opensource-db.com/repack-one-command-to-replace-vaccum-full-and-cluster-with-new-concurrently

Here is the YouTube video walkthrough of the feature:

Should you use it yet

Not in production today. As of October 15, 2026, the newest PostgreSQL 19 build on postgresql.org is Beta 4, released September 24. The Beta 4 announcement says a release candidate "should occur in early October" and GA "may also occur in October," and the roadmap lists October 2026. That same announcement lists "several fixes for the new REPACK command, including crashes," and it reverted several other features, including SQL/PGQ. REPACK stayed in, but it’s still settling — the GA release on November 19 followed extended testing specifically because early betas exposed REPACK crash scenarios under concurrent load. If you need this today, test exhaustively on staging with your actual workload pattern before considering production.

Conclusion

PostgreSQL has long provided VACUUM FULL and CLUSTER for table-rewrite operations, but both commands require strong exclusive locks that can block reads and writes on large production tables — making them disruptive for high-availability environments. While these commands have existed since early PostgreSQL versions, the lack of a non-blocking alternative remained a pain point for DBAs. PostgreSQL 19 addresses this directly by introducing the built-in REPACK command with a CONCURRENTLY option that reorganizes tables and indexes without blocking concurrent reads and writes.

PostgreSQL: Documentation: 19: REPACK

Source: https://www.postgresql.org/docs/19/sql-repack.html

REPACK is primarily meant to remove bloat and, with USING INDEX, to cluster the table, in order to improve performance. Although it also advances the table’s relfrozenxid and relminmxid, it is not a good way to prevent transaction ID or multixact ID wraparound: it rewrites the whole table and all of its indexes, so it takes much longer than VACUUM, and it can fail after most of the work is done, especially with CONCURRENTLY on a busy table. VACUUM is recommended for that instead (see Section 25.1.6).

PostgreSQL 19 Beta 1 Introduces pg_plan_advice Extension | Christophe Pettus posted on the topic | LinkedIn

Source: https://www.linkedin.com/posts/christophe-pettus_e1release-19-activity-7473782495541600256-kD3g

PostgreSQL 19 Beta 1 includes pg_plan_advice, a new in-tree extension that lets you stabilize and control planner decisions, with pg_stash_advice to apply that advice automatically by query identifier. This matters more than its place in the feature list suggests. Postgres has declined to implement query hints for essentially its entire history, on the explicit grounds that a hint is usually a workaround for a planner or statistics problem you should fix instead. PG19 ships a sanctioned, structured alternative: you capture a plan you like, store it as advice, and the planner respects it. No SQL syntax changes, no application code modifications — just an extension you enable and a table you populate.

Adoption Reality Check

This is where the rubber meets the road. Release notes promise; production deployments deliver verdicts.

Postgres 19: How Our Advice Has Changed Since… | Crunchy Data Blog

Source: https://www.crunchydata.com/blog/postgres-19-how-our-advice-has-changed-since-we-wrote-it

One related default flip: JIT is off by default in Postgres 19 (it had been on since 12). The old cost model was unreliable, so large analytical or parallel scans that used to compile at runtime no longer will unless you turn jit back on. If those workloads matter to you, re-enable explicitly and re-check plans after upgrade.

Postgres 19 advice: Keep choosing indexes for selectivity, and re-test BRIN-vs-parallel plans on your storage with EXPLAIN (ANALYZE, BUFFERS, IO) after you upgrade. Tuning those plans is now mandatory, not optional.

PostgreSQL 19 is available on Lakebase | Databricks Blog

Source: https://www.databricks.com/blog/postgresql-19-available-lakebase

PostgreSQL 19 is a practical release. There is no single headline feature that changes what PostgreSQL is. Instead, version 19 goes after operational problems that database teams know well: bloated tables, unstable query plans, replica lag, incomplete statistics, and maintenance work that is hard to run safely in production.

The common thread is control. PostgreSQL 19 gives you more ways to step in when the defaults are not enough, without giving up the project’s bias toward reliability and predictability.

PostgreSQL 19: What’s New and What to Test Before You Upgrade

Source: https://blog.elest.io/postgresql-19-whats-new-and-what-to-test-before-you-upgrade

Troubleshooting

pg_upgrade stops on an inet or cidr index. It’s a btree_gist index, which 19 refuses to carry over. Recreate it with the default GiST opclass before upgrading.

An old dump won’t load. It was probably made with standard_conforming_strings off. Dump again using the 19 pg_dump.

Restore fails on a JSON expression index or generated column. That’s the RC 2 revert. Drop or redefine the object on the source, then dump again.

Text search results changed. That’s the locale handling updates in 19 — lc_collate behavior tightened. Rebuild affected indexes after upgrade.

PostgreSQL 19 is almost here. Are you planning to upgrade? – Reddit

Source: https://www.reddit.com/r/postgres/comments/1xygfan/postgresql_19_is_almost_here_are_you_planning_to

Some configuration defaults have changed. Certain older database configurations may cause issues with pg_upgrade. Post-upgrade management of jit, wal_level, and autovacuum scoring parameters requires deliberate review — don’t assume your 18-era postgresql.conf ports cleanly.

Community Signal

Vance’s Take

Here’s what I’m watching, having spent two decades watching Postgres releases go from RC to production:

REPACK CONCURRENTLY

The headline feature. Non-blocking table reorganization is genuinely useful for 24/7 workloads that can’t afford exclusive locks. But the 104-million-tuple limit (MVCC snapshot horizon) means it’s not a universal vacuum replacement — it’s a targeted tool for bloat hotspots. Test your tuple counts before you schedule it.

pg_plan_advice / pg_stash_advice

Plan pinning without hints. This is Postgres finally admitting that sometimes the planner gets it wrong and DBAs need a lever. The extension approach keeps core clean while giving operators control. Expect to see this in every serious monitoring stack within six months.

Standby Read-Your-Writes (WAIT FOR LSN)

This solves the classic "read from replica after write" race without synchronous replication’s latency penalty. The WAIT command blocks until the standby catches up to a specific LSN. For read-scaling architectures, this is the feature that makes async replicas viable for consistent reads. Finally.

Sequence Logical Replication

Identity columns and sequences now replicate logically. No more out-of-sync sequences on subscribers after failover. This was a genuine pain point for logical replication setups; fixing it removes a whole class of post-failover manual intervention.

Foreign Key Check Speedup (2x)

The optimizer now avoids redundant FK checks during bulk loads and partitioned table operations. Measurable win for ETL pipelines and partition maintenance. Not sexy, but it compounds.

This Week’s Action Items

  1. If you run logical replication: test sequence replication on a staging subscriber before upgrading the publisher.
  2. Evaluate REPACK CONCURRENTLY on your most bloated tables — check pg_class.relpages vs pg_relation_size first.
  3. Add pg_plan_advice to your monitoring stack; start capturing planner decisions that need pinning.
  4. Review standby promotion procedures: WAIT FOR LSN changes the consistency calculus for read replicas.
  5. Schedule FK-heavy bulk loads to benefit from the 2x check speedup — partition maintenance windows may shrink.

Sources

Scheduling

publish = 2026-11-22 15:00:00 America/New_York slug = postgresql-19-daily-brief-day-01-2026-11-22 categories = ["Backend & Data Systems"] tags = ["postgresql", "postgresql-19", "database-internals", "backend-architecture", "daily-briefing"] excerpt = Day 1 of the PostgreSQL 19 daily brief: REPACK CONCURRENTLY limits, pg_plan_advice plan pinning in production, standby read-your-writes with WAIT FOR LSN, sequence logical replication, and foreign key check speedups. Real adoption signals from the community. series = "PostgreSQL 19 Daily Brief" series_position = "Day 1" author = "Marcus Vance" author_role = "Lead Systems & Backend Architect"

// THE TRANSMISSION LOG : FIELD NOTES

Weekly zero-fluff breakdowns of distributed systems, browser runtime performance, and production post-mortems delivered directly from Stella Sage.