Feedback-ID: i0fe9450f:Fastmail
Date: Sat, 12 Aug 2023 09:20:32 +0900
From: Michael Paquier <michael@paquier.xyz>
To: Thomas Munro <thomas.munro@gmail.com>
Cc: Noah Misch <noah@leadboat.com>,
	Kyotaro Horiguchi <horikyota.ntt@gmail.com>, exclusion@gmail.com,
	pgsql-bugs@lists.postgresql.org
Subject: Re: BUG #17928: Standby fails to decode WAL on termination of primary
Message-ID: <ZNbQUDcZK0ELn1o9@paquier.xyz>
References: <72bd036d-4f2a-8d50-b56e-6b1e3b9ba0a9@gmail.com>
 <CA+hUKG+t8gfYBS2TRMW6rHaWUBEC1Cy+p-0pgPBx8ag8TBeoKw@mail.gmail.com>
 <20230710200012.GB3476234@rfd.leadboat.com>
 <20230711.155414.939136339819647605.horikyota.ntt@gmail.com>
 <20230717004905.GE3826297@rfd.leadboat.com>
 <ZNSVlR/v8gTDIMsU@paquier.xyz>
 <20230811025808.GA2261232@rfd.leadboat.com>
 <ZNXQWvnmt1Fpt6vu@paquier.xyz>
 <20230811140008.GB2261449@rfd.leadboat.com>
 <CA+hUKGLEF7-bK3=KFYfaDMOh8wy2XdMOncbUrn=JrhmjX0_vPQ@mail.gmail.com>
MIME-Version: 1.0
Content-Type: multipart/signed; micalg=pgp-sha512;
	protocol="application/pgp-signature"; boundary="6fHT8cPsYwcSnxyP"
Content-Disposition: inline
In-Reply-To: 
 <CA+hUKGLEF7-bK3=KFYfaDMOh8wy2XdMOncbUrn=JrhmjX0_vPQ@mail.gmail.com>
Archived-At: 
 <https://www.postgresql.org/message-id/ZNbQUDcZK0ELn1o9%40paquier.xyz>
Precedence: bulk


--6fHT8cPsYwcSnxyP
Content-Type: text/plain; charset=utf-8
Content-Disposition: inline
Content-Transfer-Encoding: quoted-printable

On Sat, Aug 12, 2023 at 11:56:45AM +1200, Thomas Munro wrote:
> On Sat, Aug 12, 2023 at 2:00=E2=80=AFAM Noah Misch <noah@leadboat.com> wr=
ote:
>> On Fri, Aug 11, 2023 at 03:08:26PM +0900, Michael Paquier wrote:
>> I recall earlier messages theorizing that it was just harder to hit in v=
14, so
>> I'm disinclined to stop at v15.  I think the main choice is whether to s=
top at
>> v11 (normal choice) or v12 (worry about breaking the last v11 point rele=
ase).
>> I don't have a strong opinion between those.

Okay.  I wouldn't be inclined to patch v11 for that, FWIW, as this
code path is touched by recovery and more.  At least it does not seem=20
worth taking any risk compared to the potential gain.

> Thanks for working on this.
>=20
> I wonder if we need a more explicit way to construct pages with the
> right bits to reach interesting test cases and get full enough
> coverage...  (Cf throwing SQL at the WAL to see if it sticks.)

You mean SQL functions that write an arbitrary set of bytes at a given
LSN, to trigger some expected behavior on a follow-up crash recovery?
--
Michael

--6fHT8cPsYwcSnxyP
Content-Type: application/pgp-signature; name="signature.asc"

-----BEGIN PGP SIGNATURE-----
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=ymZP
-----END PGP SIGNATURE-----

--6fHT8cPsYwcSnxyP--