Message-ID: <97160e4b549c726bd409c8059a1bfaf07e7a6d04.camel@cybertec.at>
Subject: Re: Disabling vacuum truncate for autovacuum
From: Laurenz Albe <laurenz.albe@cybertec.at>
To: Nathan Bossart <nathandbossart@gmail.com>, Robert Haas
	 <robertmhaas@gmail.com>
Cc: Gurjeet Singh <gurjeet@singh.im>, Postgres Hackers
	 <pgsql-hackers@postgresql.org>, Will Storey <will@summercat.com>
Date: Fri, 28 Feb 2025 07:19:32 +0100
In-Reply-To: <Z8EvFyf1D1AMxjOg@nathan>
References: <Z2DE4lDX4tHqNGZt@dev.null>
	 <fa83a77bbca5da4d4864a49c8a27df323f23d28b.camel@cybertec.at>
	 <CABwTF4U3xkF=ZRi2pztUDxohoN8h6XL10=QmTtuTXoMjzu5-zg@mail.gmail.com>
	 <f82fc7fbb69a8c8000ad74ced50381257fe57838.camel@cybertec.at>
	 <CA+TgmoZMXN19eorKdeiiFCv3AJFVaUAfkzRuamnt8A9U8uJSqg@mail.gmail.com>
	 <Z7Tl2d7HrG1AQEOc@nathan> <Z8EvFyf1D1AMxjOg@nathan>
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable
User-Agent: Evolution 3.54.3 (3.54.3-1.fc41) 
MIME-Version: 1.0
Archived-At: <https://www.postgresql.org/message-id/97160e4b549c726bd409c8059a1bfaf07e7a6d04.camel%40cybertec.at>
Precedence: bulk

On Thu, 2025-02-27 at 21:35 -0600, Nathan Bossart wrote:
> I spent some time on this one today.=C2=A0 A couple of notes:
>=20
> * Since the reloption is a Boolean, there isn't a good way to tell whethe=
r
> =C2=A0 it is actually set for the table or if it just inherited the defau=
lt
> =C2=A0 value.=C2=A0 This is important to know because we want the relopti=
on to
> =C2=A0 override the GUC.

I agree with that, without being able to think of a better solution.

> * I didn't see a good GUC category for vacuum_truncate.=C2=A0 I suppose w=
e could
> =C2=A0 create a new one, but for now I've just stashed it into the autova=
cuum
> =C2=A0 one.=C2=A0 Bikeshedding welcome.

Why not use "Client Connection Defaults / Statement Behavior", just like fo=
r
"vacuum_freeze_min_age"?  I don't think that "autovacuum" is appropriate,
since it applies to manual VACUUM as well.

Yours,
Laurenz Albe