Date: Fri, 2 Oct 2026 09:00:13 +0200 (CEST)
From: rahul@rhyadav.dev
To: shihao zhong <zhong950419@gmail.com>
Cc: Kirill Reshke <reshkekirill@gmail.com>, Nktpro <nktpro@gmail.com>,
	Pgsql Bugs <pgsql-bugs@lists.postgresql.org>,
	Melanie Plageman <melanieplageman@gmail.com>
Message-ID: <P2v9EE4--F-9@rhyadav.dev>
In-Reply-To: 
 <CAGRkXqSZXoazu1x48pz2SuW_xHYQDdRcbp2HNCfRNm_WLF8aYA@mail.gmail.com-P2v1i8Y----9>
References: 
 <CAL4mQLAp562c1rCgg2Dqx6TBSdOk6vOvaLFjwFG9E6k4uwvJJw@mail.gmail.com>
 <CALdSSPjEM90Y_6RQGox7OFJC+oAurWy5YSuXaH8Mr7m-TMOyWA@mail.gmail.com>
 <CAGRkXqRZva+Wiw6jb9-owhLjGOtoDn-btoFjJmejmJYOMSEbzg@mail.gmail.com>
 <P2s-NV0--F-9@rhyadav.dev>
 <CAGRkXqSZXoazu1x48pz2SuW_xHYQDdRcbp2HNCfRNm_WLF8aYA@mail.gmail.com-P2v1i8Y----9>
Subject: Re: PostgreSQL 18.6/17.11: standby PANIC on restart after VM
 truncation
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: quoted-printable
Feedback-ID: 
 01f8c5108c48fea26a092a077590a8fcdbfcdbb733bc38e3db5db9540fc242b99227bb05256e786b1fc713cdb05f073377d63f8aec9e32cbf77d26f42c5d89118b:TurnOnPrivacy!:tutamail
Archived-At: 
 <https://www.postgresql.org/message-id/P2v9EE4--F-9%40rhyadav.dev>
Precedence: bulk

Hi Shihao,

On Fri, 2 Oct 2026, shihao zhong wrote:
> 0002 is your diff. It had no commit message, so I put one together
> from your mail. Please correct it if it is wrong.

Thanks for putting v3 together.=C2=A0 The message is right for master and
19.=C2=A0 On 17 and 18 the function that already does this is
heap_xlog_visible(), so in those versions the sentence would be:

=C2=A0 heap_xlog_visible() already initializes a VM page that was read as
=C2=A0 zeros.

0002 could also carry the same "Backpatch-through: 17" as 0001.

I ran my reproducer, which follows the same steps as 0003, against v3
on master and the nocfbot versions on 19, 18 and 17.=C2=A0 Unpatched
master and 17 fail to restart with the invalid pages PANIC.=C2=A0 With v3
the standby restarts cleanly on all four branches, in each of these
setups:

=C2=A0 - full_page_writes =3D off
=C2=A0 - full_page_writes =3D off, wal_consistency_checking =3D all
=C2=A0 - full_page_writes =3D on, wal_consistency_checking =3D all

0003 itself passes here too on all four branches, and fails on
master without 0001 and 0002.

Regards,
Rahul Yadav