MIME-Version: 1.0 (Mimestream 1.8.3)
From: David Barsky <me@davidbarsky.com>
Date: Wed, 8 Oct 2025 12:39:57 -0700
Message-ID: <CAPR5_5psNqCm1OQr5LL4v6z02Sh=2YHYUg8NY+3xMzfGHfKs+w@mail.gmail.com>
Subject: Option on `postgres` CLI to shutdown when there are no more active connections?
To: pgsql-general@lists.postgresql.org
Content-Type: multipart/alternative; boundary="000000000000844f980640aad9cb"
Archived-At: <https://www.postgresql.org/message-id/CAPR5_5psNqCm1OQr5LL4v6z02Sh%3D2YHYUg8NY%2B3xMzfGHfKs%2Bw%40mail.gmail.com>
Precedence: bulk

--000000000000844f980640aad9cb
Content-Type: text/plain; charset="UTF-8"

Hiya folks,

I'm a bit of a newcomer when it comes to PostgreSQL, so I apologize if this
is
the wrong mailing list. Anyways, my two questions:

1. Is there any interest in adding a command line option to the `postgres`
CLI
   that shuts down the PostgreSQL instance once (and optionally cleans up
the
   data directory) once all connections have disconnected?
   a. Alternatively, I wouldn't mind accomplishing this via the single-user
      mode if it could accept the binary/wire protocol in addition to the
      current text protocol.
2. Are there plans for having any additional table access methods beyond
`HEAP`
   shipping as part of Postgres? I'd love to have something that's purely
   in-memory to bypass the tempdir dance that I'm currently doing.
For context, I'm trying to make it easier to test our application against a
live, actual PostgreSQL instance and make the experience feel a lot like
sqlite's embedded/in-memory workflow. Today, we've gotten really great
latencies via test transactions, but I'd also like to ensure that there
aren't
any orphaned Postgres processes at the end of a test run or without
requiring
the user to start an instance of Postgres prior to running the tests.

Warmest regards,
David

--000000000000844f980640aad9cb
Content-Type: text/html; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable

<html><body><div>Hiya folks,<br></div><div><br></div><div>I&#39;m a bit of =
a newcomer when it comes to PostgreSQL, so I apologize if this is</div><div=
>the wrong mailing list. Anyways, my two questions:</div><div><br></div><di=
v>1. Is there any interest in adding a command line option to the `postgres=
` CLI</div><div>=C2=A0 =C2=A0that shuts down the PostgreSQL instance once (=
and optionally cleans up the</div><div>=C2=A0 =C2=A0data directory) once al=
l connections have disconnected?</div><div>=C2=A0 =C2=A0a. Alternatively, I=
 wouldn&#39;t mind accomplishing this via the single-user</div><div>=C2=A0 =
=C2=A0 =C2=A0 mode if it could accept the binary/wire protocol in addition =
to the</div><div>=C2=A0 =C2=A0 =C2=A0 current text protocol.</div><div>2. A=
re there plans for having any additional table access methods beyond `HEAP`=
</div><div>=C2=A0 =C2=A0shipping as part of Postgres? I&#39;d love to have =
something that&#39;s purely</div><div>=C2=A0 =C2=A0in-memory to bypass the =
tempdir dance that I&#39;m currently doing.</div><div>For context, I&#39;m =
trying to make it easier to test our application against a</div><div>live, =
actual PostgreSQL instance and make the experience feel a lot like</div><di=
v>sqlite&#39;s embedded/in-memory workflow. Today, we&#39;ve gotten really =
great</div><div>latencies via test transactions, but I&#39;d also like to e=
nsure that there aren&#39;t</div><div>any orphaned Postgres processes at th=
e end of a test run or without requiring</div><div>the user to start an ins=
tance of Postgres prior to running the tests.</div><div><br></div><div>Warm=
est regards,</div><div>David</div></body></html>

--000000000000844f980640aad9cb--