PostgreSQL: Parallel Material Storage Reporting
Two related commits extend EXPLAIN ANALYZE maximum storage reporting for Material nodes to parallel workers, with a supporting change to the tuple store statistics interface. A startup-failure test race fix and a documentation markup cleanup round out activity with no broader pattern.
Duration: PT1M56S
Episode overview
This episode is a short developer briefing from PostgreSQL.
It explains recent repository work in plain language.
- Show: PostgreSQL
- Published: 2026-10-12T13:02:39Z
- Audio duration: PT1M56S
Transcript excerpt
This excerpt keeps the crawler page concise. Listen to the episode or use the RSS feed for the full update.
Good morning, today is October 12th, 2026, and here is your PostgreSQL briefing.
The lead story is improved visibility into parallel query execution, specifically how Material nodes use temporary storage across workers.
David Rowley extended recent work that reported maximum storage for the leader process to cover workers running below gather and gather merge nodes. Detailed in commit a241bc4, the change adds shared memory handling so each worker reports its storage use back to the leader for display in EXPLAIN ANALYZE output. For…
To enable that, a preparatory change in commit c6b8af4 reworks how the tuple store reports its statistics. Instead of returning a pointer to a text label for memory or disk, it now sets a simple flag for whether disk was used. That avoids transferring a process-specific memory address to the leader and removes…
In stability and maintenance, Heikki Linnakangas fixed a race in the shared memory startup failure test in commit 36a5622. The single-user test expects failure before processing queries, so sending prior queries could trigger a broken pipe. It now sends empty input, backpatched through version 19. A separate…
What's next is straightforward: if you…
Nearby episodes from PostgreSQL
- Weekly Recap - Correctness Fixes and Backup Trust
- Index Scan Correctness Fixes
- Planner Correctness and Startup Lifecycle Fixes
- Compressed Backup Integrity and Planner Fixes
- WAL Reliability and Failover Slot Fixes
- B-Tree Deduplication Fix and Planner Corrections
- Planner Correctness Cluster
- Safer Vacuum Pauses and Quiet Correctness Fixes