Re: initdb username doc bug

On Sat, Jan  6, 2024 at 08:11:14PM -0700, David G. Johnston wrote:
> On Saturday, January 6, 2024, Bruce Momjian <bruce@momjian.us> wrote:
>
>     This commit:
>
>             commit 910cab820d
>             Author: Alvaro Herrera <alvherre@alvh.no-ip.org>
>             Date:   Fri Nov 18 11:59:26 2022 +0100
>
>                 Add glossary entries related to superusers
>
>                 Extracted from a more ambitious patch.
>
>                 Author: David G. Johnston <david.g.johnston@gmail.com>
>                 Discussion: https://postgr.es/m/CAKFQuwZC4K0XYBm0bwBMDOZySBqhO
>     SekDhLuaw4vPi+ozi8gqQ@mail.gmail.com
>
>     used this text for the --username option:
>
>            -U username
>            --username=username
>                Selects the user name of the bootstrap superuser. This defaults
>                to the name of the cluster owner.
>
>     This seems wrong since the cluster owner doesn't exist until the cluster
>     is created.  The text exists in PG 16 and master.  The attached patch
>     fixes this.
>
>
> The glossary defines cluster owner as the pre-existing operating system user. 

No, it does not.

> There may be an argument that installation user is a better term but the
> existing choice isn’t wrong.  If you are going to change it you need to update
> the glossary as well.  The description in initdb uses cluster owner as well.
>
> I agree we presently use a mix of set and specify in these definitions so the
> word choice of selects is out-of-place.

The patch says:

        +  <glossentry id="glossary-cluster-owner">
        +   <glossterm>Cluster owner</glossterm>
        +   <glossdef>
        +    <para>
-->     +     The operating system user that owns the
-->     +     <glossterm linkend="glossary-data-directory">data directory</glossterm>
        +     and under which the <literal>postgres</literal> process is run.
        +     It is required that this user exist prior to creating a new
        +     <glossterm linkend="glossary-db-cluster">database cluster</glossterm>.
        +    </para>
        +    <para>
        +     On operating systems with a <literal>root</literal> user,
        +     said user is not allowed to be the cluster owner.
        +    </para>
        +   </glossdef>
        +  </glossentry>

How can you default to be the owner of something that doesn't exist before
it is created?

Also, the initdb code gets the default username from geteuid(), the
effective process owner; man geteuid says:

        geteuid() returns the effective user ID of the calling process.

so it is not the owner of the data dirctory, but the installation user.

I am open to saying that it is the owner of the data directory but we
would then need to change initdb to do that, and we aren't going to
backpatch that.