Received: from malur.postgresql.org ([217.196.149.56]) by arkaria.postgresql.org with esmtps (TLS1.3:ECDHE_RSA_AES_256_GCM_SHA384:256) (Exim 4.92) (envelope-from ) id 1nQtSX-0007hK-BR for pgsql-general@arkaria.postgresql.org; Sun, 06 Mar 2022 16:10:13 +0000 Received: from localhost ([127.0.0.1] helo=malur.postgresql.org) by malur.postgresql.org with esmtp (Exim 4.92) (envelope-from ) id 1nQtSV-0003Cv-V1 for pgsql-general@arkaria.postgresql.org; Sun, 06 Mar 2022 16:10:11 +0000 Received: from magus.postgresql.org ([2a02:c0:301:0:ffff::29]) by malur.postgresql.org with esmtps (TLS1.3:ECDHE_RSA_AES_256_GCM_SHA384:256) (Exim 4.92) (envelope-from ) id 1nQtSV-0003Cl-Ke for pgsql-general@lists.postgresql.org; Sun, 06 Mar 2022 16:10:11 +0000 Received: from sss.pgh.pa.us ([66.207.139.130]) by magus.postgresql.org with esmtps (TLS1.3:ECDHE_RSA_AES_256_GCM_SHA384:256) (Exim 4.92) (envelope-from ) id 1nQtST-0007iO-Hl for pgsql-general@lists.postgresql.org; Sun, 06 Mar 2022 16:10:10 +0000 Received: from sss1.sss.pgh.pa.us (localhost [127.0.0.1]) by sss.pgh.pa.us (8.15.2/8.15.2) with ESMTP id 226GA0rd002046; Sun, 6 Mar 2022 11:10:00 -0500 From: Tom Lane To: depesz@depesz.com cc: PostgreSQL General , Andres Freund Subject: Re: Interesting fail when migrating Pg from Ubuntu Bionic to Focal In-reply-to: <20220306140402.GA6360@depesz.com> References: <20220303150428.GA26036@depesz.com> <20220303151156.GB26036@depesz.com> <20220303163921.GA10776@depesz.com> <20220303164401.GA13056@depesz.com> <3969790.1646420992@sss.pgh.pa.us> <20220304204510.GA21340@depesz.com> <3984384.1646431394@sss.pgh.pa.us> <20220306140402.GA6360@depesz.com> Comments: In-reply-to hubert depesz lubaczewski message dated "Sun, 06 Mar 2022 15:04:02 +0100" MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-ID: <2044.1646583000.1@sss.pgh.pa.us> Content-Transfer-Encoding: quoted-printable Date: Sun, 06 Mar 2022 11:10:00 -0500 Message-ID: <2045.1646583000@sss.pgh.pa.us> List-Id: List-Help: List-Subscribe: List-Post: List-Owner: List-Archive: Archived-At: Precedence: bulk hubert depesz lubaczewski writes: > On Fri, Mar 04, 2022 at 05:03:14PM -0500, Tom Lane wrote: >> Mmm ... it might have just been that the planner chose not to use >> JIT when it thought there were fewer rows involved. Did you check >> with EXPLAIN that these cut-down cases still used JIT? > I tore these boxes down, so can't check immediately, but I think > I remember that you're right - single-row queries didn't use JIT. FWIW, I went to the trouble of installing Ubuntu Focal on my raspberry pi to see if I could duplicate this, and I couldn't. However, what you get from a fresh install now is $ dpkg -l | grep libllvm = ii libllvm10:arm64 1:10.0.0-4ubuntu1 a= rm64 Modular compiler and toolchain technologies, runtime library not 9.0.1. I also found that Fedora 31/aarch64 is still downloadable, and that does contain LLVM 9 ... and I could not reproduce it there either. So I've run out of things to try. It still seems like a bug in a specific point release of LLVM is a possible explanation, especially given that Debian went so far as to replace that release in an LTS distro. (The gcc-related packages are all exactly the same versions as you show.) I'd believe that data corruption could be a contributing factor too, except that you were able to read all the rows without JIT. regards, tom lane