Date: Thu, 20 Mar 2025 16:31:33 -0500
From: Nathan Bossart <nathandbossart@gmail.com>
To: "David G. Johnston" <david.g.johnston@gmail.com>
Cc: Fujii Masao <masao.fujii@oss.nttdata.com>,
	Robert Treat <rob@xzilla.net>,
	Laurenz Albe <laurenz.albe@cybertec.at>,
	Gurjeet Singh <gurjeet@singh.im>,
	Andres Freund <andres@anarazel.de>,
	Will Storey <will@summercat.com>,
	Robert Haas <robertmhaas@gmail.com>,
	Postgres Hackers <pgsql-hackers@postgresql.org>
Subject: Re: Disabling vacuum truncate for autovacuum
Message-ID: <Z9yJNZ1C4TJfU-nY@nathan>
References: <CABV9wwNagobmRMCxs0dKqWmKXaQTUQj+oOOTw=fsPAgbBfVhew@mail.gmail.com>
 <Z9g8a4HGLUcSo96v@nathan>
 <Z9ohFZtDaZnkch7u@nathan>
 <88e3b55a-8ef8-4b53-8d71-6bfde1a07bc1@oss.nttdata.com>
 <Z9rlb0jgNLm8ULAB@nathan>
 <Z9rsMX5ZOL5E0cQy@nathan>
 <Z9wxuP1AY303wAxM@nathan>
 <CAKFQuwYKtEUYKS+18gRs-xPhn0qOJgM2KGyyWVCODHuVn9F-XQ@mail.gmail.com>
 <Z9xaroJSrS44HyKG@nathan>
 <CAKFQuwZJtw7Ojr=Z0xgA0G7BW6qSDkb0fMKgyaQULdOxHg_chA@mail.gmail.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <CAKFQuwZJtw7Ojr=Z0xgA0G7BW6qSDkb0fMKgyaQULdOxHg_chA@mail.gmail.com>
Archived-At: <https://www.postgresql.org/message-id/Z9yJNZ1C4TJfU-nY%40nathan>
Precedence: bulk

On Thu, Mar 20, 2025 at 02:18:33PM -0700, David G. Johnston wrote:
> So my concern about dump/restore seems to be alleviated but then, why can
> we not just do whatever pg_dump is doing to decide whether the current
> value for vacuum_truncate is its default (and thus would not be dumped) or
> not (and would be dumped)?

pg_dump looks at the pg_class.reloptions array directly.  In the vacuum
code, we look at the pre-parsed rd_options (see RelationParseRelOptions()
in relcache.c), which will have already resolved vacuum_truncate to its
default value if it was not explicitly set.  We could probably look at
pg_class.reloptions directly in the vacuum code if we _really_ wanted to,
but I felt that putting this information into rd_options was much cleaner.

-- 
nathan