Date: Fri, 9 Jun 2023 12:44:03 -0400
From: Bruce Momjian <bruce@momjian.us>
To: "Fujii.Yuki@df.MitsubishiElectric.co.jp" <Fujii.Yuki@df.mitsubishielectric.co.jp>
Cc: Alexander Pyhalov <a.pyhalov@postgrespro.ru>,
	PostgreSQL-development <pgsql-hackers@postgresql.org>,
	Andres Freund <andres@anarazel.de>, Tom Lane <tgl@sss.pgh.pa.us>,
	Tomas Vondra <tomas.vondra@enterprisedb.com>,
	Julien Rouhaud <rjuju123@gmail.com>,
	Daniel Gustafsson <daniel@yesql.se>,
	Ilya Gladyshev <i.gladyshev@postgrespro.ru>
Subject: Re: Partial aggregates pushdown
Message-ID: <ZINW022iuYzf2ZwO@momjian.us>
References: <OSZPR01MB6662875933E87F37C711B4FC95969@OSZPR01MB6662.jpnprd01.prod.outlook.com>
 <ZDDIU2fJxZm/k5kj@momjian.us>
 <OS3PR01MB66606C1DA960336516D9A08395959@OS3PR01MB6660.jpnprd01.prod.outlook.com>
 <ZDedXNarI32SmN06@momjian.us>
 <ZDemPMLOyXNzKkhw@momjian.us>
 <OS3PR01MB6660EBFFFA06646F402813A395989@OS3PR01MB6660.jpnprd01.prod.outlook.com>
 <ZDj4nwJANUU+kam8@momjian.us>
 <OS3PR01MB66607DC466F7F3B3909DA3CA954EA@OS3PR01MB6660.jpnprd01.prod.outlook.com>
 <8ffedecae0d21a3cd93805baf1dcc0de@postgrespro.ru>
 <OS3PR01MB66604E404B003A621B0C03A69552A@OS3PR01MB6660.jpnprd01.prod.outlook.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <OS3PR01MB66604E404B003A621B0C03A69552A@OS3PR01MB6660.jpnprd01.prod.outlook.com>
Archived-At: <https://www.postgresql.org/message-id/ZINW022iuYzf2ZwO%40momjian.us>
Precedence: bulk

On Tue, Jun  6, 2023 at 03:08:50AM +0000, Fujii.Yuki@df.MitsubishiElectric.co.jp wrote:
> > I've seen this message. But introduction of new built-in function will break
> > requests to old servers only if this new function is used in the request (this
> > means that query changes). However, this patch changes the behavior of old
> > queries, which worked prior to update. This seems to be different to me.
> 
> You are right.
> However, for now, partial aggregates pushdown is mainly used when using built-in sharding in PostgreSQL.
> I believe when using built-in sharding in PostgreSQL, the version of the primary node server and
> the version of the remote server will usually be the same.
> So I think it would be sufficient to include in the documentation a note about such problem
> and a workaround for them.

I agree that this feature is designed for built-in sharding, but it is
possible people could be using aggregates on partitions backed by
foreign tables without sharding.  Adding a requirement for non-sharding
setups to need PG 17+ servers might be unreasonable.

Looking at previous release note incompatibilities, we don't normally
change non-administrative functions in a way that causes errors if run
on older servers.  Based on Alexander's observations, I wonder if we
need to re-add the postgres_fdw option to control partial aggregate
pushdown, and default it to enabled.

If we ever add more function breakage we might need more postgres_fdw
options.  Fortunately, such changes are rare.

Yuki, thank you for writing and updating this patch, and Alexander,
thank you for helping with this patch.

-- 
  Bruce Momjian  <bruce@momjian.us>        https://momjian.us
  EDB                                      https://enterprisedb.com

  Only you can decide what is important to you.