Received: from localhost (unknown [200.46.204.183]) by mail.postgresql.org (Postfix) with ESMTP id 8A3B264FDF5 for ; Thu, 27 Nov 2008 12:45:51 -0400 (AST) Received: from mail.postgresql.org ([200.46.204.86]) by localhost (mx1.hub.org [200.46.204.183]) (amavisd-maia, port 10024) with ESMTP id 07664-08 for ; Thu, 27 Nov 2008 12:45:48 -0400 (AST) X-Greylist: domain auto-whitelisted by SQLgrey-1.7.6 Received: from blu0-omc2-s36.blu0.hotmail.com (blu0-omc2-s36.blu0.hotmail.com [65.55.111.111]) by mail.postgresql.org (Postfix) with ESMTP id A71FF64FD19 for ; Thu, 27 Nov 2008 12:45:48 -0400 (AST) Received: from BLU103-W26 ([65.55.111.72]) by blu0-omc2-s36.blu0.hotmail.com with Microsoft SMTPSVC(6.0.3790.3959); Thu, 27 Nov 2008 08:45:47 -0800 Message-ID: Content-Type: multipart/alternative; boundary="_ea8c259f-be1b-4aa1-821f-dd043c119462_" X-Originating-IP: [201.147.232.185] From: Fabricio To: , Subject: Re: NUMA architecture and PostgreSQL Date: Thu, 27 Nov 2008 09:45:47 -0700 Importance: Normal In-Reply-To: <20081127092004.GC7442@easy2.in-chemnitz.de> References: <3a0028490811260543x74d02719w775a457999cfc74@mail.gmail.com> <492D5C87.5090505@usit.uio.no> <3a0028490811260700w6e8d9d86o24effbe0bea2bddb@mail.gmail.com> <3a0028490811260921l5d6aac06v54dca32aae672b97@mail.gmail.com> <492DA95D.3010303@zeut.net> <20081127092004.GC7442@easy2.in-chemnitz.de> MIME-Version: 1.0 X-OriginalArrivalTime: 27 Nov 2008 16:45:47.0459 (UTC) FILETIME=[99157930:01C950AF] X-Virus-Scanned: Maia Mailguard 1.0.1 X-Spam-Status: No, hits=0.001 tagged_above=0 required=5 tests=HTML_MESSAGE=0.001 X-Spam-Level: X-Archive-Number: 200811/165 X-Sequence-Number: 30819 --_ea8c259f-be1b-4aa1-821f-dd043c119462_ Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: quoted-printable > I suppose this has nothing to do with PostgreSQL but needs some> operati= ng system level tuning. Maybe there is a way to bind the postgres> processe= s to one node's CPUs... What operating system are you using? =20 I am using Red Hat Enterprise 5.2=2C kernel 2.6.18-92.el5 x86_64 > I know that there are ways to bind processes to a CPU on Linux=2C I'm not= > sure how to assign multiple processes to multiple CPUs. The Linux> schedu= ler should know about NUMA architectures though=2C there not that> new any = more. =20 something like this? numactl --membind=3D1 su - postgres -c "/usr/local/pgsql/bin/pg_ctl start -= D /usr/local/pgsql/data/" this bind memory to node 1 to use the shared memory on node 1 and cpus pre= ferred node 1. With this the other processes and S.O by default will use th= e node 0? =20 thanks...=20 _________________________________________________________________ Realiza b=FAsquedas en Internet y ll=E9vate hasta =A1Un Auto!=20 http://www.ganabuscando.com/Default.aspx= --_ea8c259f-be1b-4aa1-821f-dd043c119462_ Content-Type: text/html; charset="iso-8859-1" Content-Transfer-Encoding: quoted-printable  =3B
>=3B I suppose this has nothing to do with PostgreSQL but nee= ds some
>=3B operating system level tuning. Maybe there is a way to bi= nd the postgres
>=3B processes to one node's CPUs... What operating sy= stem are you using?
 =3B
I am using Red Hat Enterprise 5.2=2C kernel =3B 2.6.18-92.el5 x86_64
>=3B I know that there are ways to bind processes to a CPU on Linux= =2C I'm not
>=3B sure how to assign multiple processes to multiple CPU= s. The Linux
>=3B scheduler should know about NUMA architectures thoug= h=2C there not that
>=3B new any more.
 =3B
something like this?
numactl --membind=3D1 su - postgres -c "/usr/local/pgsql/bin/pg_ctl start -= D /usr/local/pgsql/data/"
 =3Bthis bind memory to node =3B1 to use the shared memory on node = 1 and cpus preferred node 1. =3BWith this the other processes and S.O b= y default will use the node 0?
 =3B
thanks...


=BFTienes c=E1mara digital=2C m=E1s de una cuenta= de email=2C planeas fiestas o blogueas? Entonces=2C necesitas Windows Live = --_ea8c259f-be1b-4aa1-821f-dd043c119462_--