Horton

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

Bharath Rupireddy <bharath.rupireddyforpostgres@gmail.com>

View thread View on postgr.es
From: Bharath Rupireddy <bharath.rupireddyforpostgres@gmail.com>
To: Michael Paquier <michael@paquier.xyz>
Cc: Nathan Bossart <nathandbossart@gmail.com>, Thomas Munro <thomas.munro@gmail.com>, Tom Lane <tgl@sss.pgh.pa.us>, Andres Freund <andres@anarazel.de>, "Bossart, Nathan" <bossartn@amazon.com>, Julien Rouhaud <rjuju123@gmail.com>, PostgreSQL Hackers <pgsql-hackers@lists.postgresql.org>
Date: 2022-08-27T08:36:32Z
Lists: pgsql-hackers

Attachments

On Thu, Aug 25, 2022 at 5:51 PM Michael Paquier <michael@paquier.xyz> wrote:
>
> FWIW, the changes in reorderbuffer.c for
> ReorderBufferCleanupSerializedTXNs() reduce the code readability, in
> my opinion, so that's one less argument in favor of this change.

Agreed. I reverted the changes.

> The gain in ParseConfigDirectory() is kind of cool.
> pg_tzenumerate_next(), copydir(), RemoveXlogFile()
> StartupReplicationSlots(), CheckPointLogicalRewriteHeap() and
> RemovePgTempFilesInDir() seem fine, as well.  At least these avoid
> extra lstat() calls when the file type is unknown, which would be only
> a limited number of users where some of the three DT_* are missing
> (right?).

Yes, the idea is to avoid lstat() system calls when possible.

PSA v17 patch with reorderbuffer.c changes reverted.

-- 
Bharath Rupireddy
RDS Open Source Databases: https://aws.amazon.com/rds/postgresql/

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.