Feedback-ID: ia2694551:Fastmail
Date: Thu, 7 Mar 2024 18:56:46 +0100
From: Alvaro Herrera <alvherre@alvh.no-ip.org>
To: Antonin Houska <ah@cybertec.at>
Cc: Pavel Stehule <pavel.stehule@gmail.com>, 
	Michael Paquier <michael@paquier.xyz>, PostgreSQL Hackers <pgsql-hackers@postgresql.org>
Subject: Re: why there is not VACUUM FULL CONCURRENTLY?
Message-ID: <202403071756.f55kh6tj6bmf@alvherre.pgsql>
MIME-Version: 1.0
Content-Type: text/plain; charset=utf-8
Content-Disposition: inline
Content-Transfer-Encoding: 8bit
In-Reply-To: <6347.1708105409@antos>
Archived-At: <https://www.postgresql.org/message-id/202403071756.f55kh6tj6bmf%40alvherre.pgsql>
Precedence: bulk

On 2024-Feb-16, Antonin Houska wrote:

> BTW, I'm failing to understand why cluster_rel() has no argument of the
> BufferAccessStrategy type. According to buffer/README, the criterion for using
> specific strategy is that page "is unlikely to be needed again
> soon". Specifically for cluster_rel(), the page will *definitely* not be used
> again (unless the VACCUM FULL/CLUSTER command fails): BufferTag contains the
> relatin file number and the old relation file is eventually dropped.
> 
> Am I missing anything?

No, that's just an oversight.  Access strategies are newer than that
cluster code.

-- 
Álvaro Herrera               48°01'N 7°57'E  —  https://www.EnterpriseDB.com/
"Most hackers will be perfectly comfortable conceptualizing users as entropy
 sources, so let's move on."                               (Nathaniel Smith)
      https://mail.gnu.org/archive/html/monotone-devel/2007-01/msg00080.html