Horton

Re: IO wait events for COPY FROM/TO PROGRAM or file

Michael Paquier <michael@paquier.xyz>

View thread View on postgr.es
From: Michael Paquier <michael@paquier.xyz>
To: Nikolay Samokhvalov <nik@postgres.ai>
Cc: Sami Imseih <samimseih@gmail.com>, pgsql-hackers <pgsql-hackers@lists.postgresql.org>
Date: 2026-02-03T03:27:49Z
Lists: pgsql-hackers 
On Sat, Jan 31, 2026 at 11:51:39AM -0800, Nikolay Samokhvalov wrote:
> v2 attached with the rename per feedback: COPY_DATA_READ/WRITE →
> COPY_FROM_READ/COPY_TO_WRITE.
> 
> Given how small this patch is, any chance it could still make PG19?

While double-checking the code, one thing felt incorrect in the
description of these two new wait events, as copyto.c and
copyfromparse.c can set copy_file under copy_dest=COPY_FILE for three
cases, not two contrary to what is said upthread:
- PROGRESS_COPY_TYPE_FILE
- PROGRESS_COPY_TYPE_PROGRAM
- PROGRESS_COPY_TYPE_PIPE, for a non-DestRemote.

The COPY code is designed to be pluggable depending on what extension
code passes down as option, and not mentioning the pipe case would be
incorrect.  Added a mention about this part, and called it a day.
--
Michael

Commits

  1. Add two IO wait events for COPY FROM/TO on a pipe/file/program