MIME-Version: 1.0
References: <DB533C25-C6BA-4C0F-8046-96168E9CDD72@gmail.com>
 <aXhruFBS8ZZgKolq@paquier.xyz>
 <0B38B9DC-A53C-437C-B9BD-807F30A032BA@gmail.com>
 <F42E6546-E46B-46FE-9C21-30BD289259AE@gmail.com>
 <424A3799-D7FD-4691-BD10-6CBDACB14938@gmail.com>
 <43825F3D-6156-4030-B5E1-9A0841969022@gmail.com>
 <260E4616-43B7-4C24-BF53-352DC8DEB7AA@gmail.com>
 <CABPTF7U_eJ6yeiGQcBPr5SVQS8wvGoo6j7b+6=_Kq3VTk3wSeg@mail.gmail.com>
 <F478AC49-37D3-4256-AA68-952E55539B8C@gmail.com>
 <CABPTF7X-J4VCZbZfsCujhh4OE3tR73xws43ur2R1PFSMd21tpg@mail.gmail.com>
 <8F791DCB-0307-4A8C-B5AF-1EF0CF678AC2@gmail.com>
In-Reply-To: <8F791DCB-0307-4A8C-B5AF-1EF0CF678AC2@gmail.com>
From: Xuneng Zhou <xunengzhou@gmail.com>
Date: Mon, 6 Apr 2026 14:04:14 +0800
Message-ID: 
 <CABPTF7X6F4q5zE7DSwRTo-xSdApbemE2CV8UJ8y=i8WvTtLc1Q@mail.gmail.com>
Subject: Re: tablecmds: fix bug where index rebuild loses replica identity on
 partitions
To: Chao Li <li.evan.chao@gmail.com>
Cc: Postgres hackers <pgsql-hackers@lists.postgresql.org>
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable
Archived-At: 
 <https://www.postgresql.org/message-id/CABPTF7X6F4q5zE7DSwRTo-xSdApbemE2CV8UJ8y%3Di8WvTtLc1Q%40mail.gmail.com>
Precedence: bulk

On Tue, Mar 24, 2026 at 3:26=E2=80=AFPM Chao Li <li.evan.chao@gmail.com> wr=
ote:
>
>
>
> > On Mar 23, 2026, at 16:41, Xuneng Zhou <xunengzhou@gmail.com> wrote:
> >
> > Hi,
> >
> > On Mon, Mar 23, 2026 at 3:57=E2=80=AFPM Chao Li <li.evan.chao@gmail.com=
> wrote:
> > >
> > >
> > >
> > > > On Mar 21, 2026, at 18:29, Xuneng Zhou <xunengzhou@gmail.com> wrote=
:
> > > >
> > > > On Thu, Mar 19, 2026 at 1:07=E2=80=AFPM Chao Li <li.evan.chao@gmail=
.com> wrote:
> > > >>
> > > >>
> > > >>
> > > >>> On Feb 26, 2026, at 14:59, Chao Li <li.evan.chao@gmail.com> wrote=
:
> > > >>>
> > > >>>
> > > >>>
> > > >>>> On Jan 28, 2026, at 10:49, Chao Li <li.evan.chao@gmail.com> wrot=
e:
> > > >>>>
> > > >>>>
> > > >>>>
> > > >>>>> On Jan 27, 2026, at 16:30, Chao Li <li.evan.chao@gmail.com> wro=
te:
> > > >>>>>
> > > >>>>>
> > > >>>>>
> > > >>>>>> On Jan 27, 2026, at 15:59, Chao Li <li.evan.chao@gmail.com> wr=
ote:
> > > >>>>>>
> > > >>>>>>
> > > >>>>>>
> > > >>>>>>> On Jan 27, 2026, at 15:39, Michael Paquier <michael@paquier.x=
yz> wrote:
> > > >>>>>>>
> > > >>>>>>> On Tue, Jan 27, 2026 at 01:13:32PM +0800, Chao Li wrote:
> > > >>>>>>>> I found this bug while working on a related patch [1].
> > > >>>>>>>>
> > > >>>>>>>> When ALTER TABLE ... ALTER COLUMN TYPE causes an index rebui=
ld, and
> > > >>>>>>>> that index is used as REPLICA IDENTITY on a partitioned tabl=
e, the
> > > >>>>>>>> replica identity marking on partitions can be silently lost =
after the
> > > >>>>>>>> rebuild.
> > > >>>>>>>
> > > >>>>>>> I am slightly confused by the tests included in the proposed =
patch.
> > > >>>>>>> On HEAD, if I undo the proposed changes of tablecmds.c, the t=
ests
> > > >>>>>>> pass.  If I run the tests of the patch with the changes of
> > > >>>>>>> tablecmds.c, the tests also pass.
> > > >>>>>>
> > > >>>>>> Oops, that isn=E2=80=99t supposed to be so. I=E2=80=99ll check=
 the test.
> > > >>>>>>
> > > >>>>>
> > > >>>>> Okay, I see the problem is here:
> > > >>>>> ```
> > > >>>>> +CREATE UNIQUE INDEX test_replica_identity_partitioned_pkey ON =
test_replica_identity_partitioned (id);
> > > >>>>> ```
> > > >>>>>
> > > >>>>> I missed to add column =E2=80=9Cval=E2=80=9D into the index, so=
 that alter type of val didn=E2=80=99t cause index rebuild.
> > > >>>>>
> > > >>>>> Ideally, it=E2=80=99s better to also verify that index OIDs sho=
uld have changed before and after alter column type, but I haven=E2=80=99t =
figured out how to do so. Do you have an idea?
> > > >>>>
> > > >>>> I just updated the test to store index OIDs before and after reb=
uild into 2 temp tables, so that we can compare the OIDs to verify rebuild =
happens and replica identity preserved.
> > > >>>>
> > > >>>> I tried to port the test to master branch, and the test failed. =