Re: Automatically sizing the IO worker pool
Thomas Munro <thomas.munro@gmail.com>
From: Thomas Munro <thomas.munro@gmail.com>
To: Dmitry Dolgov <9erthalion6@gmail.com>
Cc: PostgreSQL Hackers <pgsql-hackers@lists.postgresql.org>
Date: 2026-03-28T09:31:36Z
Lists: pgsql-hackers
Attachments
- v3-0001-aio-Simplify-pgaio_worker_submit.patch (text/x-patch) patch v3-0001
- v3-0002-aio-Improve-I-O-worker-behavior-on-full-queue.patch (text/x-patch) patch v3-0002
- v3-0003-aio-Adjust-I-O-worker-pool-size-automatically.patch (text/x-patch) patch v3-0003
Here is a rebase. I would like to push these shortly if there are no objections. I propose 8 as the default upper limit.
Commits
Same data as JSON:
GET /api/v1/messages/:b64id/commits
the thread's linked commits as JSON, with link sources.
API reference →
-
aio: Adjust I/O worker pool automatically.
- d1c01b79d4ae 19 (unreleased) landed
-
Convert lwlock.c to use the new shmem allocation functions
- a006bc7b1699 19 (unreleased) cited
-
aio: Simplify pgaio_worker_submit().
- fc44f106657a 19 (unreleased) landed
-
Conditional locking in pgaio_worker_submit_internal
- 29a0fb215779 19 (unreleased) cited
-
aio: Remove obsolete IO worker ID references.
- b4c19da93a08 18.0 landed
- 177c1f059338 19 (unreleased) landed
-
aio: Regularize IO worker internal naming.
- b2afb0676337 18.0 landed
- 01d618bcd782 19 (unreleased) landed