Content-Type: text/plain;
	charset=us-ascii
Mime-Version: 1.0 (Mac OS X Mail 16.0 \(3864.400.21\))
Subject: Re: Bug in MultiXact replay compat logic for older minor version
 after crash-recovery
From: Andrey Borodin <x4mmm@yandex-team.ru>
In-Reply-To: <5b7f0a04-4a60-44bb-9d2c-8917af0b10fa@iki.fi>
Date: Fri, 20 Mar 2026 20:05:53 +0300
Cc: =?utf-8?B?IuauteWdpOS7gSjliLvpn6cpIg==?= <duankunren.dkr@alibaba-inc.com>,
 pgsql-hackers <pgsql-hackers@postgresql.org>
Content-Transfer-Encoding: quoted-printable
Message-Id: <C44AF20B-7AA3-42A6-B30D-08DEFD2E3F8C@yandex-team.ru>
References: 
 <c4ef1737-8cba-458e-b6fd-4e2d6011e985.duankunren.dkr@alibaba-inc.com>
 <18a40051-abc2-4fc2-8dcc-4dc39aa3e79e@iki.fi>
 <F513DFBB-DCFE-4B99-B01E-DDB04414857C@yandex-team.ru>
 <452355ef-7c20-4d96-88a2-8fbb49737dc9@iki.fi>
 <BD37E553-A017-4D2C-8203-702378BF5882@yandex-team.ru>
 <5b7f0a04-4a60-44bb-9d2c-8917af0b10fa@iki.fi>
To: Heikki Linnakangas <hlinnaka@iki.fi>
Archived-At: 
 <https://www.postgresql.org/message-id/C44AF20B-7AA3-42A6-B30D-08DEFD2E3F8C%40yandex-team.ru>
Precedence: bulk


> On 20 Mar 2026, at 18:14, Heikki Linnakangas <hlinnaka@iki.fi> wrote:
>=20
> Zeroing the page again is dangerous because the CREATE_ID records can =
be out of order. The page might already contain some later multixids, =
and zeroing will overwrite them.

I see only cases when it's not a problem: we zeroed page, did not flush =
it, thus did not extend the file, crashed, tested FS, zeroed page once =
more, overwrote again by replaying WAL, no big deal.
We should never zero a page with offsets, that will not be replayed by =
WAL.

If the page was persisted, even partially, we will read it from disk =
without zeroing out.


Best regards, Andrey Borodin.=