Re: Official Windows Installer and Documentation

From: Julien Rouhaud <rjuju123(at)gmail(dot)com>
To: Thomas Kellerer <shammat(at)gmx(dot)net>
Cc: pgsql-hackers(at)lists(dot)postgresql(dot)org
Subject: Re: Official Windows Installer and Documentation
Date: 2022-07-28 01:42:19
Message-ID: 20220728014219.6sxz3ql4xwymhpua@jrouhaud
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-hackers

Hi,

On Wed, Jul 27, 2022 at 11:36:11PM +0200, Thomas Kellerer wrote:
> David G. Johnston schrieb am 27.07.2022 um 21:21:
> > And then there is the issue of file ownership.
> >
> > Assuming we want better documentation for this specific issue for
> > back-patching what would that look like?
> >
> > Going forward should our installer be creating the postgres user for
> > consistency with other platforms or not?
>
> Didn't the installer used to do that in earlier releases and that
> was removed when Postgres was able to "drop privileges" when the
> service is started?
>
> I remember a lot of problems around the specific Postgres service
> account when that still was the case.

Note that there's no "official" Windows installer, and companies providing one
are free to implement it the way they want, which can contradict the official
documentation. The download section of the website clearly says that this is a
third-party installer.

For now there's only the EDB installer that remains, but I think that some time
ago there was 2 or 3 different providers.

For the EDB installer, I'm not sure why or when it was changed, but it indeed
used to have a dedicated local account and now relies on "Local System Account"
or something like that. But IIRC, when it used to create a local account the
name could be configured, so there was no guarantee of a local "postgres"
account by then either.

In response to

Responses

Browse pgsql-hackers by date

  From Date Subject
Next Message Robert Haas 2022-07-28 01:47:19 Re: collect_corrupt_items_vacuum.patch
Previous Message Masahiko Sawada 2022-07-28 01:07:36 Re: Introduce wait_for_subscription_sync for TAP tests