Date: Thu, 24 Feb 2022 13:35:20 -0600
From: Justin Pryzby <pryzby@telsasoft.com>
To: Jaime Casanova <jcasanov@systemguards.com.ec>
Cc: pgsql-hackers@postgresql.org, Alvaro Herrera <alvherre@alvh.no-ip.org>
Subject: Re: document that brin's autosummarize parameter is off by default
Message-ID: <20220224193520.GY9008@telsasoft.com>
References: <20210406040254.GB2768@ahch-to>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <20210406040254.GB2768@ahch-to>
User-Agent: Mutt/1.9.4 (2018-02-28)
Archived-At: 
 <https://www.postgresql.org/message-id/20220224193520.GY9008%40telsasoft.com>
Precedence: bulk

Ten months ago, Jaime Casanova wrote:
> Hi everyone,
> 
> Just noted that the default value of autosummarize reloption for brin
> indexes is not documented, or at least not well documented.
> 
> I added the default value in create_index.sgml where other options
> mention their own defaults, also made a little change in brin.sgml to 
> make it more clear that is disabled by default (at least the way it 
> was written made no sense for me, but it could be that my english is 
> not that good).

It seems like "This last trigger" in the current text is intended to mean "The
second condition".  Your change improves that.

Should we also consider enabling autosummarize by default ?
It was added in v10, after BRIN was added in v9.5.  For us, brin wasn't usable
without autosummarize.

Also, note that vacuums are now triggered by insertions, since v13, so it might
be that autosummarize is needed much less.

-- 
Justin

PS. I hope there's a faster response to your pg_upgrade patch.