Date: Sat, 12 Mar 2022 14:33:32 -0800
From: Nathan Bossart <nathandbossart@gmail.com>
To: Andres Freund <andres@anarazel.de>
Cc: Bharath Rupireddy <bharath.rupireddyforpostgres@gmail.com>,
	Kyotaro Horiguchi <horikyota.ntt@gmail.com>,
	SATYANARAYANA NARLAPURAM <satyanarlapuram@gmail.com>,
	PostgreSQL Hackers <pgsql-hackers@lists.postgresql.org>
Subject: Re: Allow async standbys wait for sync replication
Message-ID: <20220312223332.GA921069@nathanxps13>
References: <20220301.163431.1826638724406024793.horikyota.ntt@gmail.com>
 <20220301170537.GA1031413@nathanxps13>
 <CALj2ACUdNd=vVTOdz=UMuWxTVJgzsCUhOwJ6mMsXK150Ec-QOQ@mail.gmail.com>
 <20220301212700.GB1033258@nathanxps13>
 <CALj2ACULc_2TbC44fdFbpPCJ-AmvixdLY9z=LgMfz07QXx9-bg@mail.gmail.com>
 <20220304195602.GA1184024@nathanxps13>
 <CALj2ACUpRdtM_6RpumzNbO6D_XbSjVMfnNp5yVOnkKUnS4Z+xA@mail.gmail.com>
 <20220305202752.4oll4hgztlgdfykl@alap3.anarazel.de>
 <CALj2ACUpgZgiPRYq545XiALZMtGJ-vfQGbUGZpCJ7i3QuAUdUg@mail.gmail.com>
 <20220309020123.sneaoijlg3rszvst@alap3.anarazel.de>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <20220309020123.sneaoijlg3rszvst@alap3.anarazel.de>
Archived-At: <https://www.postgresql.org/message-id/20220312223332.GA921069%40nathanxps13>
Precedence: bulk

On Tue, Mar 08, 2022 at 06:01:23PM -0800, Andres Freund wrote:
> To me it's architecturally the completely wrong direction. We should move in
> the *other* direction, i.e. allow WAL to be sent to standbys before the
> primary has finished flushing it locally. Which requires similar
> infrastructure to what we're discussing here.

I think this is a good point.  After all, WALRead() has the following
comment:

 * XXX probably this should be improved to suck data directly from the
 * WAL buffers when possible.

Once you have all the infrastructure for that, holding back WAL replay on
async standbys based on synchronous replication might be relatively easy.

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