Received: from maia.hub.org (maia-3.hub.org [200.46.204.243]) by mail.postgresql.org (Postfix) with ESMTP id D5385632BEF for ; Wed, 25 Aug 2010 13:04:46 -0300 (ADT) Received: from mail.postgresql.org ([200.46.204.86]) by maia.hub.org (mx1.hub.org [200.46.204.243]) (amavisd-maia, port 10024) with ESMTP id 85088-05 for ; Wed, 25 Aug 2010 16:04:39 +0000 (UTC) X-Greylist: delayed 00:06:01.887556 by SQLgrey-1.7.6 Received: from nabmail01.northamericanbancard.com (unknown [66.103.255.162]) by mail.postgresql.org (Postfix) with ESMTP id A0848632B6C for ; Wed, 25 Aug 2010 13:04:39 -0300 (ADT) Received: from [172.20.202.24] (unknown [172.20.202.24]) by nabmail01.northamericanbancard.com (Postfix) with ESMTP id 4A40612122 for ; Wed, 25 Aug 2010 11:58:36 -0400 (EDT) Message-ID: <4C753D98.1050705@nabancard.com> Date: Wed, 25 Aug 2010 11:58:16 -0400 From: Bob Branch User-Agent: Mozilla/5.0 (X11; U; Linux x86_64; en-US; rv:1.9.1.11) Gecko/20100713 Thunderbird/3.0.6 MIME-Version: 1.0 To: pgsql-performance@postgresql.org Subject: New servers, need suggestions for sensible tuning settings Content-Type: text/plain; charset=ISO-8859-1; format=flowed Content-Transfer-Encoding: 7bit X-Virus-Scanned: Maia Mailguard 1.0.1 X-Spam-Status: No, hits=-1.107 tagged_above=-10 required=5 tests=BAYES_00=-1.9, RDNS_NONE=0.793 X-Spam-Level: X-Archive-Number: 201008/278 X-Sequence-Number: 40121 Hello! We've just installed a couple new servers that will be our new database servers. The basic hardware stats of each are: 2 x Xeon E5620 (2.4GHz quad-core) 32GB DDR3 1333 RAM 6 x 600GB 15krpm SAS drives - RAID 10 Perc 6/i RAID controller with battery backup I've lurked on this list for a short while and gotten a small (tiny, I'm sure) set of ideas about parameters that have commonly been mentioned when performance tuning questions come up, but haven't had the freedom to do much tuning to our existing setups. Now that we've got new servers (and thus my tuning won't potentially impact production performance until we roll these out), I've begun researching how I should be tuning these to keep the speeds up in an environment where we're regularly dealing with queries handling multiple tables with 20 million or more rows, and tables that grow by half a million records per day. I've found a great many resources suggesting settings to tune and even some that suggest ways to determine the values to set for them, but I'm having trouble tracking down anything CURRENT in terms of what options to set and what values to give them. The most comprehensive lists of suggestions (as provided by a few, though not a great many minutes perusing Google results) list options that don't seem to exist anymore (eg- sort_mem) in the version of 8.4 that Ubuntu currently has for 10.04 server (8.4.4). Is there a reference I can look at that will give me a description of how to determine sensible values for settings like shared_buffers, effective_cache_size, etc. that I see discussed on this list and elsewhere? If there isn't a really good reference, would any of you mind suggesting sensible starting points for these options or methods for determining what those would be, as well as other options I might benefit from changing the values of? I'm limited a bit in my tuning by the disks available - I can keep the WAL on a separate partition in the array, but I wasn't able to allocate an entire disk or separate array to it due to the amount of the avaliable space I'll need for the databases these servers will house. I've helped it a bit by using ext2 on the partition the WAL is on (the databases themselves are on ext4), but I'm not sure what else I can do in terms of tuning the WAL without re-thinking the storage setup. Thanks in advance for any help you might be able to provide. Bob Branch