From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from mail-vk0-f69.google.com (mail-vk0-f69.google.com [209.85.213.69]) by kanga.kvack.org (Postfix) with ESMTP id 72AC06B0253 for ; Fri, 15 Apr 2016 15:31:37 -0400 (EDT) Received: by mail-vk0-f69.google.com with SMTP id h185so221093413vkg.0 for ; Fri, 15 Apr 2016 12:31:37 -0700 (PDT) Received: from mx1.redhat.com (mx1.redhat.com. [209.132.183.28]) by mx.google.com with ESMTPS id h189si18340496qhd.2.2016.04.15.12.31.36 for (version=TLS1_2 cipher=ECDHE-RSA-AES128-GCM-SHA256 bits=128/128); Fri, 15 Apr 2016 12:31:36 -0700 (PDT) Message-ID: <1460748682.25336.41.camel@redhat.com> Subject: Re: [Bug 107771] New: Single process tries to use more than 1/2 physical RAM, OS starts thrashing From: Rik van Riel Date: Fri, 15 Apr 2016 15:31:22 -0400 In-Reply-To: <20160415121549.47e404e3263c71564929884e@linux-foundation.org> References: <20160415121549.47e404e3263c71564929884e@linux-foundation.org> Content-Type: multipart/signed; micalg="pgp-sha256"; protocol="application/pgp-signature"; boundary="=-3dBXzDUI2y/coimDsBba" Mime-Version: 1.0 Sender: owner-linux-mm@kvack.org List-ID: To: Andrew Morton , linux-mm@kvack.org Cc: bugzilla-daemon@bugzilla.kernel.org, Johannes Weiner , Mel Gorman , theosib@gmail.com --=-3dBXzDUI2y/coimDsBba Content-Type: text/plain; charset="UTF-8" Content-Transfer-Encoding: quoted-printable On Fri, 2016-04-15 at 12:15 -0700, Andrew Morton wrote: > (switched to email.=C2=A0=C2=A0Please respond via emailed reply-to-all, n= ot via > the > bugzilla web interface). >=20 > This is ... interesting. First things first. What is the value of /proc/sys/vm/zone_reclaim? I am assuming this is a two socket system, with two 12-core CPUs. Am I right? > On Thu, 12 Nov 2015 18:46:35 +0000 bugzilla- > daemon@bugzilla.kernel.org wrote: >=20 > >=20 > > https://bugzilla.kernel.org/show_bug.cgi?id=3D107771 > >=20 > > =C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2= =A0Bug ID: 107771 > > =C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0Summa= ry: Single process tries to use more than 1/2 > > physical > > =C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2= =A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0RAM, OS starts thrashing > > =C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0Produ= ct: Memory Management > > =C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0Versi= on: 2.5 > > =C2=A0=C2=A0=C2=A0=C2=A0Kernel Version: 4.3.0-040300-generic (Ubuntu) > > =C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0Hardware: A= ll > > =C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2= =A0=C2=A0=C2=A0=C2=A0=C2=A0OS: Linux > > =C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2= =A0=C2=A0=C2=A0Tree: Mainline > > =C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2= =A0Status: NEW > > =C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0Severity: n= ormal > > =C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0Priority: P= 1 > > =C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0Component: Page A= llocator > > =C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0Assignee: a= kpm@linux-foundation.org > > =C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0Reporter: t= heosib@gmail.com > > =C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0Regression: No > >=20 > > I have a 24-core (48 thread) system with 64GB of RAM.=C2=A0=C2=A0 > >=20 > > When I run multiple processes, I can use all of physical RAM before > > swapping > > starts.=C2=A0=C2=A0However, if I'm running only a *single* process, the > > system will start > > swapping after I've exceeded only 1/2 of available physical > > RAM.=C2=A0=C2=A0Only after > > swap fills does it start using more of the physical RAM.=C2=A0=C2=A0 > >=20 > > I can't find any ulimit settings or anything else that would cause > > this to > > happen intentionally.=C2=A0 > >=20 > > I had originally filed this against Ubuntu, but I'm now running a > > more recent > > kernel, and the problem persists, so I think it's more appropriate > > to file > > here.=C2=A0=C2=A0There are some logs that they had me collect, so if yo= u want > > to see > > them, the are here: > >=20 > > https://bugs.launchpad.net/ubuntu/+source/linux/+bug/1513673 > >=20 > > I don't recall this problem happening with older kernels (whatever > > came with > > Ubuntu 15.04), although I may just not have noticed.=C2=A0=C2=A0By swap= ping > > early, I'm > > limited by the speed of my SSD, which is moving only about 20MB/sec > > in each > > direction, and that makes what I'm running take 10 times as long to > > complete. > >=20 --=20 All Rights Reversed. --=-3dBXzDUI2y/coimDsBba Content-Type: application/pgp-signature; name="signature.asc" Content-Description: This is a digitally signed message part Content-Transfer-Encoding: 7bit -----BEGIN PGP SIGNATURE----- Version: GnuPG v2 iQEcBAABCAAGBQJXEUGLAAoJEM553pKExN6DalEIAJglyPn2Nx+0IzCbjmipdA1J uLUjD2aIrBFWlObysRnM0m1NiLCz/luroMn1SHR2EinRDQhXlimt67/Idd3kC5yO bW4yS3GgNUYB82pFtkzhViGIfFAjCTWzaiPPkbNSRPpAmbu0GpznEO6Fwc0IIUM4 F9fhOiCRMiD1SFb4RiotrT7r9ONyCPf2ihGx30HK6Pwkz3i+ZH/Eko/51OE5c55S 0N4hqOffXssuegD35vfYfuCcyVvWGpgddlUpHYhZJ0mckrtAfS+C0UjXJ+Ag1Qad zBEBrij6FeGMuztlKX16ZGLEQgyW0wb1VdhzAjxfgzzyTMOpp944d5PnstCuOc8= =Dhd2 -----END PGP SIGNATURE----- --=-3dBXzDUI2y/coimDsBba-- -- To unsubscribe, send a message with 'unsubscribe linux-mm' in the body to majordomo@kvack.org. For more info on Linux MM, see: http://www.linux-mm.org/ . Don't email: email@kvack.org