X-Original-To: pgsql-performance-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id 022EDD1B495 for ; Wed, 17 Dec 2003 22:19:34 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 15132-03 for ; Wed, 17 Dec 2003 18:19:04 -0400 (AST) Received: from bayswater1.ymogen.net (host-154-240-27-217.pobox.net.uk [217.27.240.154]) by svr1.postgresql.org (Postfix) with ESMTP id 08CC0D1B45D for ; Wed, 17 Dec 2003 18:19:00 -0400 (AST) Received: from finisterre (82-68-132-233.dsl.in-addr.zen.co.uk [82.68.132.233]) by bayswater1.ymogen.net (Postfix) with SMTP id A7A4B9AE38; Wed, 17 Dec 2003 22:18:45 +0000 (GMT) From: "Matt Clark" To: , "Pgsql-Performance@Postgresql. Org" Subject: Re: Adding RAM: seeking advice & warnings of hidden "gotchas" Date: Wed, 17 Dec 2003 22:18:45 -0000 Message-ID: MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: 7bit X-Priority: 3 (Normal) X-MSMail-Priority: Normal X-Mailer: Microsoft Outlook IMO, Build 9.0.2416 (9.0.2910.0) Importance: Normal X-MimeOLE: Produced By Microsoft MimeOLE V6.00.2800.1165 In-Reply-To: X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200312/260 X-Sequence-Number: 5120 If you have 3 1.5GB tables then you might as well go for 4GB while you're at it. Make sure you've got a bigmem kernel either running or available, and boost effective_cache_size by whatever amount you increase the RAM by. We run a Quad Xeon/4GB server on Redhat 7.3 and it's solid as a rock. There is no way I know of to get indexes preferentially cached over data though. Matt > -----Original Message----- > From: pgsql-performance-owner@postgresql.org > [mailto:pgsql-performance-owner@postgresql.org]On Behalf Of Nick > Fankhauser > Sent: 17 December 2003 19:57 > To: Pgsql-Performance@Postgresql. Org > Subject: [PERFORM] Adding RAM: seeking advice & warnings of hidden > "gotchas" > > > Hi- > > After having done my best to squeeze better performance out of our > application by tuning within our existing resources, I'm falling back on > adding memory as a short-term solution while we get creative for > a long-term > fix. I'm curious about what experiences others have had with the > process of > adding big chunks of RAM. In particular, if I'm trying to encourage the OS > to cache more of my index information in RAM, what sort of configuration > should I do at both the PostgreSQL and OS level? > > In a slightly off-topic vein, I'd also like to hear about it if > anyone knows > about any gotchas at the OS level that might become a problem. > > The server is a dual processor Athlon 1.2GHz box with hardware > SCSI RAID. It > currently has 1 GB RAM, and we're planning to add one GB more for > a total of > 2GB. The OS is Debian Linux Kernel 2.4.x, and we're on PostgreSQL v7.3.2 > > My current memory related settings are: > > SHMMAX and SHMALL set to 128MB (OS setting) > shared buffers 8192 (64MB) > sort_mem 16384 (16MB) > effective_cache_size 65536 (512MB) > > > We support up to 70 active users, sharing a connection pool of 16 > connections. Most of the queries center around 3 tables that are about 1.5 > GB each. > > > Thanks. > -Nick > > --------------------------------------------------------------------- > Nick Fankhauser > > nickf@doxpop.com Phone 1.765.965.7363 Fax 1.765.962.9788 > doxpop - Court records at your fingertips - http://www.doxpop.com/ > > > > ---------------------------(end of broadcast)--------------------------- > TIP 7: don't forget to increase your free space map settings >