Horton

Re: Avoid erroring out when unable to remove or parse logical rewrite files to save checkpoint work

Andres Freund <andres@anarazel.de>

View thread View on postgr.es
From: Andres Freund <andres@anarazel.de>
To: Nathan Bossart <nathandbossart@gmail.com>
Cc: Bharath Rupireddy <bharath.rupireddyforpostgres@gmail.com>, "Bossart, Nathan" <bossartn@amazon.com>, Tom Lane <tgl@sss.pgh.pa.us>, Julien Rouhaud <rjuju123@gmail.com>, PostgreSQL Hackers <pgsql-hackers@lists.postgresql.org>
Date: 2022-02-15T18:10:34Z
Lists: pgsql-hackers 
Hi,

On 2022-02-15 09:57:53 -0800, Nathan Bossart wrote:
> IIUC you are advocating for something more like the attached patches.

At least for patch 1 yes. Don't have the cycles just now to look at the
others.

I generally think it'd be a good exercise to go through an use
get_dirent_type() in nearly all ReadDir() style loops - it's a nice efficiency
win in general, and IIRC a particularly big one on windows.

It'd probably be good to add a reference to get_dirent_type() to
ReadDir[Extended]()'s docs.

Greetings,

Andres Freund

Commits

  1. Expand the use of get_dirent_type(), shaving a few calls to stat()/lstat()

  2. fsync pg_logical/mappings in CheckPointLogicalRewriteHeap().

  3. Introduce logical decoding.