From: Antonin Houska <ah@cybertec.at>
To: shihao zhong <zhong950419@gmail.com>
cc: Robert Treat <rob@xzilla.net>,
    Masahiko Sawada <sawada.mshk@gmail.com>,
    Manu <manuelreyesbravo@gmail.com>, Thom Brown <thom@linux.com>,
    pgsql-hackers@lists.postgresql.org
Subject: Re: REPACK (CONCURRENTLY) can silently lose updates when the toast
 table is rewritten
In-reply-to: 
 <CAGRkXqR+jco=86v10goPL5jQbKJJcK_dwWB_4dFh4CEDnN7jpg@mail.gmail.com>
References: 
 <CAA-aLv5MF6BLL+BWvix2Yw+CBardtH43AofPReQunhDZPNBtuA@mail.gmail.com>
 <179012413951.1850281.5077495683381671561@gmail.com>
 <CAGRkXqRYLtBRaMzdH+e7PMO-BRaWPPo37gvOx3C=jQ1uP4Cx7w@mail.gmail.com>
 <47479.1790180571@localhost>
 <CAD21AoCYvsp5HFOXLx9BpSz6Lcyw7H9-RXtq-b=AA8joP8=-XA@mail.gmail.com>
 <CABV9wwMQ7-hQnOFc4yYo6oJv8eLE9OV0khFrpespwAkOKugsqw@mail.gmail.com>
 <10459.1790239287@localhost>
 <CAGRkXqR+jco=86v10goPL5jQbKJJcK_dwWB_4dFh4CEDnN7jpg@mail.gmail.com>
Comments: In-reply-to shihao zhong <zhong950419@gmail.com>
   message dated "Fri, 25 Sep 2026 00:32:11 -0400."
MIME-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Content-ID: <4323.1790317455.1@localhost>
Content-Transfer-Encoding: quoted-printable
Date: Fri, 25 Sep 2026 08:24:15 +0200
Message-ID: <4324.1790317455@localhost>
Archived-At: 
 <https://www.postgresql.org/message-id/4324.1790317455%40localhost>
Precedence: bulk

shihao zhong <zhong950419@gmail.com> wrote:

> > (What I said does not mean that I'm in favor of restarting the decodin=
g worker
> > either. I still prefer locking the TOAST relation early, as I noted el=
sewhere
> > in the thread.)
> =

> OK. v3 locks the TOAST relation before the worker starts, as Sawada-san
> first suggested. A rewrite of the TOAST relation now waits for REPACK,
> which I think is also what Robert asked for.

Thanks for the patch. I'm just not sure this is the best place to lock the
TOAST table: note that copy_table_data() locks it again.

I'd prefer locking it close to the place we lock the main table (perhaps i=
n
cluster_rel(), after all the checks have been done?) and replace the locki=
ng
statements (both in the copy_table_data() and in your patch) with
Assert(CheckRelationLockedByMe(...)).

-- =

Antonin Houska
Web: https://www.cybertec-postgresql.com