full_page_writes

full_page_writes — Writes full pages to WAL when first modified after a checkpoint. Observed in PG9.0–19 Beta 3; its last measured boot default is on in PG19 Beta 3, with sighup context. This is a beta-snapshot fact and can change before PostgreSQL 19 GA.

Note

Fact — official short description: “Writes full pages to WAL when first modified after a checkpoint.”

Identity

Type , Valuebool: Upstream pg_settings type
Context , Valuesighup: Takes effect after configuration reload
Unit , Value—: Raw unit
Range , Value—: Raw limits in the last observed version
Enum values , Value—: — for non-enum types
Category , ValueWrite-Ahead Log / Settings: Upstream classification
Latest boot value , Valueon: on

Lifecycle

Fact	Value
First observed	PG9.0 (research boundary)
Present in	PG9.0–19 Beta 3
Removed in	No
Introduction commit	Not asserted: predates the PG9.0 research boundary
Commit date	—
Discussion	—

Default history

Measured PG9.0–19 Beta 3 boot defaults
Versions	Raw `boot_val`	Unit	Human value
PG9.0–19 Beta 3	`on`	—	on

How it works

Writes full pages to WAL when first modified after a checkpoint. A configuration reload applies a new value; existing work already in flight is not retroactively changed.

After each checkpoint, the first change to a data page logs a complete page image, preventing torn-page recovery from combining old and new sectors. The extra WAL is concentrated after checkpoints and is affected by wal_compression; disabling it is unsafe unless the storage stack provides an equivalent atomic-page guarantee.

Monitor and change full_page_writes together with data_sync_retry, restart_after_crash, recovery_init_sync_method. Validate on the relevant server role and real workload, then use its sighup context to choose session change, reload, or restart; a historical boot default is not the current effective value.

Tuning advice

Tip

Advice. These are workload-specific starting points and must be validated with measurements.

Workload	Guidance
OLTP	Keep full_page_writes=on for persistent production data. Address storage latency, checkpoints, WAL compression, and batching instead of exchanging crash safety for throughput.
OLAP	Bulk loads still need recoverability. Use explicitly rebuildable UNLOGGED/temporary staging to narrow the durability scope rather than disabling protection cluster-wide.
Small nodes	Keep it on for persistent small instances too. Only a clearly isolated, disposable cluster may make an exception after explicitly accepting rebuild risk.

Pigsty

Values use the fixed 8-vCPU, 32-GiB, 100-GiB SSD fixture and render the current Pigsty templates for PG19 Beta 3; this does not assert current Pigsty support for that historical or beta release.

Template	Effective value	Versus upstream boot	Source expression
OLTP	Unmodified	—	—
OLAP	Unmodified	—	—
CRIT	Unmodified	—	—
TINY	Unmodified	—	—

Caution

Advice — pending human review. Fact from the current Pigsty template projection: OLTP: PG9.0–19 Beta 3 unmodified; OLAP: PG9.0–19 Beta 3 unmodified; CRIT: PG9.0–19 Beta 3 unmodified; TINY: PG9.0–19 Beta 3 unmodified. No Pigsty-specific rationale is inferred from an absent override.

Common pitfalls

Disabling it without an end-to-end atomic-page guarantee.
Confusing its WAL-volume cost with logging every page on every change.
Ignoring the post-checkpoint full-page-image burst.
Assuming a reload retroactively protects WAL records that were generated while full_page_writes was off.
Benchmarking with it disabled without an end-to-end crash, recovery, and torn-page test.

data_sync_retry · restart_after_crash · recovery_init_sync_method · fsync · wal_sync_method · synchronous_commit