Date: Mon, 31 Jan 2022 10:11:00 -0500
From: Stephen Frost <sfrost@snowman.net>
To: Bharath Rupireddy <bharath.rupireddyforpostgres@gmail.com>
Cc: Nathan Bossart <nathandbossart@gmail.com>,
	Kyotaro Horiguchi <horikyota.ntt@gmail.com>,
	"Bossart, Nathan" <bossartn@amazon.com>,
	Julien Rouhaud <rjuju123@gmail.com>,
	Michael Paquier <michael@paquier.xyz>,
	PostgreSQL Hackers <pgsql-hackers@lists.postgresql.org>
Subject: Re: Add checkpoint and redo LSN to LogCheckpointEnd log message
Message-ID: <20220131151100.GR10577@tamriel.snowman.net>
References: <CALj2ACUGutu6As-9p76Ta2bV59J97fbiCc9NSzaWXqxPJdMvag@mail.gmail.com>
 <CALj2ACVM4wjS+wiNZf4ssdNMhSR-XUhR99fYbV=pC9_-=5wCtw@mail.gmail.com>
 <BDAE2BB3-2F37-4B7C-A053-874250A47C21@amazon.com>
 <20220120.120029.1092463916382268628.horikyota.ntt@gmail.com>
 <CALj2ACUaGJQVbQ5Hms_C2tu+6iJX-EWkEdOiXwX1WfrxQX8pHA@mail.gmail.com>
 <20220128054620.GA654866@nathanxps13>
 <CALj2ACXWyK9DujhndkuGXb7idjL9wPJfMpzRPpM-DT1LQ6MB=Q@mail.gmail.com>
MIME-Version: 1.0
Content-Type: multipart/signed; micalg=pgp-sha512;
	protocol="application/pgp-signature"; boundary="wlFCko3yyTbBmKK8"
Content-Disposition: inline
In-Reply-To: <CALj2ACXWyK9DujhndkuGXb7idjL9wPJfMpzRPpM-DT1LQ6MB=Q@mail.gmail.com>
User-Agent: Mutt/1.5.24 (2015-08-30)
Archived-At: <https://www.postgresql.org/message-id/20220131151100.GR10577%40tamriel.snowman.net>
Precedence: bulk


--wlFCko3yyTbBmKK8
Content-Type: text/plain; charset=utf-8
Content-Disposition: inline
Content-Transfer-Encoding: quoted-printable

Greetings,

* Bharath Rupireddy (bharath.rupireddyforpostgres@gmail.com) wrote:
> On Fri, Jan 28, 2022 at 11:16 AM Nathan Bossart
> <nathandbossart@gmail.com> wrote:
> > I know I voted for "start=3D%X/%X, end=3D%X/%X," but looking at this ag=
ain, I
> > wonder if it could be misleading.  "start" is the redo location, and "e=
nd"
> > is the location of the checkpoint record, but I could understand why
> > someone might think that "start" is the location of the previous checkp=
oint
> > record and "end" is the redo location of the new one.  I think your
> > original idea of "lsn=3D%X/%X, redo lsn=3D%X/%X" could be a good altern=
ative.
> >
> > =D0=86n any case, this patch is small and otherwise looks reasonable to=
 me, so I
> > am going to mark it as ready-for-committer.
>=20
> Thanks for your review. In summary, we have these options to choose
> checkpoint LSN and last REDO LSN:
>=20
> 1)  "start=3D%X/%X, end=3D%X/%X" (ControlFile->checkPointCopy.redo,
> ControlFile->checkPoint)
> 2)  "lsn=3D%X/%X, redo lsn=3D%X/%X"
> 3) "location=3D%X/%X, REDO location=3D%X/%X" -- similar to what
> pg_controldata and pg_control_checkpoint shows currently.
> 4) "location=3D%X/%X, REDO start location=3D%X/%X"
>=20
> I will leave that decision to the committer.

I'd also vote for #2.  Regarding 3 and 4, I'd argue that those should
have been changed when we changed a number of other things from the
generic 'location' to be 'lsn' in system views and functions, and
therefore we should go change those to also specify 'lsn' rather than
just saying 'location'.

Thanks,

Stephen

--wlFCko3yyTbBmKK8
Content-Type: application/pgp-signature; name="signature.asc"

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1
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=OO/m
-----END PGP SIGNATURE-----

--wlFCko3yyTbBmKK8--