MIME-Version: 1.0
References: <E1uLe9X-000zsY-2g@gemulon.postgresql.org>
 <ed2296f1-1a6b-4932-b870-5bb18c2591ae@oss.nttdata.com>
 <aD0bg6Km5IigzUkC@paquier.xyz>
 <CAPmGK17Vf4fWt=+kCZ8V3JP7jO6iMrUze7TjOVtGzZwNbevnwg@mail.gmail.com>
 <b269f7ad-90a7-4217-a58e-038899148718@oss.nttdata.com>
In-Reply-To: <b269f7ad-90a7-4217-a58e-038899148718@oss.nttdata.com>
From: Etsuro Fujita <etsuro.fujita@gmail.com>
Date: Sun, 15 Feb 2026 17:40:13 +0900
Message-ID: 
 <CAPmGK14yYoO2iBg7SoT3WdnAKoCpWy_LPdoTDepJ21-Yf8TiKA@mail.gmail.com>
Subject: Re: pgsql: postgres_fdw: Inherit the local transaction's
 access/deferrable
To: Fujii Masao <masao.fujii@oss.nttdata.com>
Cc: Michael Paquier <michael@paquier.xyz>,
 Etsuro Fujita <efujita@postgresql.org>,
	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/CAPmGK14yYoO2iBg7SoT3WdnAKoCpWy_LPdoTDepJ21-Yf8TiKA%40mail.gmail.com>
Precedence: bulk

On Wed, Jun 4, 2025 at 1:15=E2=80=AFPM Fujii Masao <masao.fujii@oss.nttdata=
.com> wrote:
> I agree this could be considered a fix if the new behavior has been
> clearly explained in the documentation from before or based on
> standards like SQL/MED. But if that's not the case, it seems more
> like a behavior change. In that case, I think it should wait for v19
> and be applied only after reaching consensus. Some systems might
> rely on the previous behavior.
>
> By the way, if a read-only transaction on the local server is meant
> to block all write operations on the remote server, this patch alone
> might not be sufficient, for example, that read-only transaction can
> invoke a login trigger on the remote server and it could still
> perform writes.

This patch 1) modifies postgres_fdw so that it opens remote
transactions in read-only mode if the corresponding local transaction
is read-only, as noted in the documentation, but 2) keeps the existing
behavior of login triggers that they can write even if the invoking
transaction is read-only.  So declaring a transaction as read-only on
the local side doesn't mean it blocks all write operations on the
remote side; it still allows login triggers invoked on the remote side
to write.  Considering typical use-cases of such triggers, this seems
reasonable to me.  I think it might be a good idea to add a note about
it to the documentation, though.

I'd like to re-propose this patch for v19, as mentioned in this thread.

Thanks!

Best regards,
Etsuro Fujita