From: Tom Lane <tgl@sss.pgh.pa.us>
To: Andres Freund <andres@anarazel.de>
cc: Andrew Dunstan <andrew@dunslane.net>, David Rowley <dgrowleyml@gmail.com>,
        Alvaro Herrera <alvherre@alvh.no-ip.org>,
        Mladen Gogala <gogala.mladen@gmail.com>,
        pgsql-performance@lists.postgresql.org
Subject: Re: Catching up with performance & PostgreSQL 15
In-reply-to: <A1EBCE55-2229-47E4-959E-7C0BB9666AC9@anarazel.de>
References: <20221129083615.fseqfvd5svtjbrb6@alvherre.pgsql>
 <871130.1669732296@sss.pgh.pa.us>
 <CAApHDvoT4KXjNGos8FvMt5NvzvWJrkzs1oqFp6J4vAb==Q=DBA@mail.gmail.com>
 <18d8cac2-36fa-3736-fa4a-c33c290301cb@dunslane.net>
 <A1EBCE55-2229-47E4-959E-7C0BB9666AC9@anarazel.de>
Comments: In-reply-to Andres Freund <andres@anarazel.de>
	message dated "Wed, 30 Nov 2022 08:07:11 -0800"
MIME-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Content-ID: <1345395.1669826188.1@sss.pgh.pa.us>
Content-Transfer-Encoding: quoted-printable
Date: Wed, 30 Nov 2022 11:36:28 -0500
Message-ID: <1345396.1669826188@sss.pgh.pa.us>
Archived-At: 
 <https://www.postgresql.org/message-id/1345396.1669826188%40sss.pgh.pa.us>
Precedence: bulk

Andres Freund <andres@anarazel.de> writes:
> On November 30, 2022 3:47:32 AM PST, Andrew Dunstan <andrew@dunslane.net=
> wrote:
>> I think Alvaro's point is that it would have been better to work out
>> these wrinkles before turning on JIT by default. Based on anecdotal
>> reports from the field I'm inclined to agree.

> The problem is that back when it was introduced these problems didn't ex=
ist to a significant degree. JIT was developed when partitioning was very =
minimal- and the problems we're seeing are almost exclusively with queries=
 with many partitions. The problems really only started much more recently=
. It also wasn't enabled in the first release..

Well, wherever you want to pin the blame, it seems clear that we
have a problem now.  And I don't think flipping back to off-by-default
is the answer -- surely there is some population of users who will
not be happy with that.  We really need to prioritize fixing the
cost-estimation problems, and/or tweaking the default thresholds.

			regards, tom lane