Horton

Re: Possibility to disable `ALTER SYSTEM`

Magnus Hagander <magnus@hagander.net>

View thread
From: Magnus Hagander <magnus@hagander.net>
To: Daniel Gustafsson <daniel@yesql.se>
Cc: Robert Haas <robertmhaas@gmail.com>, Jelte Fennema-Nio <postgres@jeltef.nl>, Bruce Momjian <bruce@momjian.us>, Maciek Sakrejda <m.sakrejda@gmail.com>, Tom Lane <tgl@sss.pgh.pa.us>, Joel Jacobson <joel@compiler.org>, Andrew Dunstan <andrew@dunslane.net>, Gabriele Bartolini <gabriele.bartolini@enterprisedb.com>, PostgreSQL Hackers <pgsql-hackers@postgresql.org>
Date: 2024-03-18T15:46:33Z
Lists: pgsql-hackers

Commits

Same data as JSON: GET /api/v1/messages/:b64id/commits the thread's linked commits as JSON, with link sources. API reference →

  1. Add allow_alter_system GUC.

  2. Rename COMPAT_OPTIONS_CLIENT to COMPAT_OPTIONS_OTHER.

  3. Remove support for version-0 calling conventions.

On Mon, Mar 18, 2024 at 4:44 PM Daniel Gustafsson <daniel@yesql.se> wrote:
>
> > On 18 Mar 2024, at 16:34, Magnus Hagander <magnus@hagander.net> wrote:
> >
> > On Mon, Mar 18, 2024 at 2:09 PM Daniel Gustafsson <daniel@yesql.se> wrote:
> >>
> >>> On 18 Mar 2024, at 13:57, Robert Haas <robertmhaas@gmail.com> wrote:
> >>
> >>> my proposal is something like this, taking a
> >>> bunch of text from Jelte's patch and some inspiration from Magnus's
> >>> earlier remarks:
> >>
> >> I still think any wording should clearly mention that settings in the file are
> >> still applied.  The proposed wording says to implicitly but to avoid confusion
> >> I think it should be explicit.
> >
> > I haven't kept up with the thread, but in general I'd prefer it to
> > actually turn off parsing the file as well. I think just turning off
> > the ability to change it -- including the ability to *revert* changes
> > that were made to it before -- is going to be confusing.
>
> Wouldn't that break pgBackrest which IIRC write to .auto.conf directly
> without using ALTER SYSTEM?

Ugh of course. And not only that, it would also break pg_basebackup
which does the same.

So I guess that's not a good idea. I guess nobody anticipated this
when that was done:)


> > But, if we have decided it shouldn't do that, then IMHO we should
> > consider naming it maybe enable_alter_system_command instead -- since
> > we're only disabling the alter system command, not the actual feature
> > in total.
>
> Good point.


-- 
 Magnus Hagander
 Me: https://www.hagander.net/
 Work: https://www.redpill-linpro.com/