Re: Read-only connection mode for AI workflows.

From: Andrei Lepikhov <lepihov(at)gmail(dot)com>
To: Pavel Stehule <pavel(dot)stehule(at)gmail(dot)com>
Cc: Jack Bonatakis <jack(at)bonatak(dot)is>, pgsql-hackers <pgsql-hackers(at)postgresql(dot)org>, Bruce Momjian <bruce(dot)momjian(at)enterprisedb(dot)com>, Andres Freund <andres(at)anarazel(dot)de>
Subject: Re: Read-only connection mode for AI workflows.
Date: 2026-03-19 08:40:15
Message-ID: 42ca6abe-774b-404b-b4a2-27d8792b23fb@gmail.com
Views: Whole Thread | Raw Message | Download mbox | Resend email
Thread:
Lists: pgsql-hackers

On 19/3/26 08:53, Pavel Stehule wrote:
> We could improve it by restricting manual calls to specific utility
> operations, such as VACUUM or REINDEX. However, we would need some
> specifications first.
>
>
> It doesn't cover  possibility to set GUC by set_config function

Can you explain it? I added a test for the set_config() call.
This extension is so tiny because it exploits the rule: no ro -> rw
switch after a snapshot has been taken (but rw -> ro is possible). The
set_config can’t overcome this rule.

--
regards, Andrei Lepikhov,
pgEdge

In response to

Responses

Browse pgsql-hackers by date

  From Date Subject
Next Message JoongHyuk Shin 2026-03-19 09:12:37 Re: pg_dump: fix NOT NULL constraint name comparison using makeObjectName
Previous Message Chao Li 2026-03-19 08:16:15 Re: Release postmaster working memory context in slotsync worker