Horton

Re: SyncRepLock acquired exclusively in default configuration

Michael Paquier <michael@paquier.xyz>

View thread View on postgr.es
From: Michael Paquier <michael@paquier.xyz>
To: Ashwin Agrawal <aagrawal@pivotal.io>
Cc: Masahiko Sawada <masahiko.sawada@2ndquadrant.com>, Fujii Masao <masao.fujii@oss.nttdata.com>, Andres Freund <andres@anarazel.de>, PostgreSQL-development <pgsql-hackers@postgresql.org>, Simon Riggs <simon@2ndquadrant.com>, Xin Zhang <xzhang@pivotal.io>, Asim Praveen <apraveen@pivotal.io>
Date: 2020-05-20T05:38:38Z
Lists: pgsql-hackers 
On Tue, May 19, 2020 at 08:56:13AM -0700, Ashwin Agrawal wrote:
> Sure, add it to commit fest.
> Seems though it should be backpatched to relevant branches as well.

It does not seem to be listed yet.  Are you planning to add it under
the section for bug fixes?
--
Michael

Commits

  1. Avoid unnecessary acquisition of SyncRepLock in transaction commit time.

  2. Fix race condition when changing synchronous_standby_names