Horton

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

Nathan Bossart <nathandbossart@gmail.com>

View thread View on postgr.es
From: Nathan Bossart <nathandbossart@gmail.com>
To: Bharath Rupireddy <bharath.rupireddyforpostgres@gmail.com>
Cc: 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: 2022-08-16T20:32:33Z
Lists: pgsql-hackers 
On Wed, Aug 10, 2022 at 03:28:25PM +0530, Bharath Rupireddy wrote:
>  		snprintf(path, sizeof(path), "pg_logical/mappings/%s", mapping_de->d_name);
> -		if (lstat(path, &statbuf) == 0 && !S_ISREG(statbuf.st_mode))
> +		if (get_dirent_type(path, mapping_de, false, LOG) != PGFILETYPE_REG)
>  			continue;

Previously, failure to lstat() wouldn't lead to skipping the entry.  With
this patch, a failure to determine the file type will cause the entry to be
skipped.  This might be okay in some places (e.g., CheckPointSnapBuild())
but not in others.  For example, in CheckPointLogicalRewriteHeap(), this
could cause us to skip fsync-ing a file due to a get_dirent_type() failure,
which seems bad.

-- 
Nathan Bossart
Amazon Web Services: https://aws.amazon.com

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.