Re: Refactor recovery conflict signaling a little

Heikki Linnakangas <hlinnaka@iki.fi>

From: Heikki Linnakangas <hlinnaka@iki.fi>

To: Xuneng Zhou <xunengzhou@gmail.com>

Cc: Alexander Lakhin <exclusion@gmail.com>, "pgsql-hackers@postgresql.org" <pgsql-hackers@postgresql.org>, Chao Li <li.evan.chao@gmail.com>

Date: 2026-03-09T15:28:42Z

Lists: pgsql-hackers

Attachments

repro-sleep.patch (text/x-patch) patch

On 09/03/2026 17:02, Xuneng Zhou wrote:
> Did you use Alexander’s reproducer script? I tried reproducing with a
> 1 ms pg_usleep() added to all three functions that clear
> MyProc->pendingRecoveryConflicts, but I still couldn’t reproduce the
> issue.

I used the attached, to be precise. With that it fails every time for 
me. I'm not sure if the "if (am_walsender)" check is necessary, I added 
it just to make the test run faster.

- Heikki

Commits

Don't clear pendingRecoveryConflicts at end of transaction
- 138592d1b066 19 (unreleased) landed
Refactor ProcessRecoveryConflictInterrupt for readability
- be5257725d7f 19 (unreleased) landed
Separate RecoveryConflictReasons from procsignals
- 17f51ea81875 19 (unreleased) landed
Use ProcNumber rather than pid in ReplicationSlot
- ddc3250208bd 19 (unreleased) landed
Don't hint that you can reconnect when the database is dropped
- 57bff90160fd 19 (unreleased) landed
Remove useless errdetail_abort()
- cd375d5b6d5f 19 (unreleased) landed
Teach standby conflict resolution to use SIGUSR1
- a8ce974cddef 9.0.0 cited