Received: from malur.postgresql.org ([217.196.149.56]) by arkaria.postgresql.org with esmtp (Exim 4.80) (envelope-from ) id 1YBuvw-0005iz-2g for pgsql-performance@arkaria.postgresql.org; Fri, 16 Jan 2015 00:38:40 +0000 Received: from localhost ([127.0.0.1] helo=postgresql.org) by malur.postgresql.org with smtp (Exim 4.80) (envelope-from ) id 1YBuvv-0001kd-I9 for pgsql-performance@arkaria.postgresql.org; Fri, 16 Jan 2015 00:38:39 +0000 Received: from magus.postgresql.org ([2a02:c0:301:0:ffff::29]) by malur.postgresql.org with esmtps (TLS1.2:DHE_RSA_AES_256_CBC_SHA256:256) (Exim 4.80) (envelope-from ) id 1YBuub-0000cu-Tm for pgsql-performance@postgresql.org; Fri, 16 Jan 2015 00:37:17 +0000 Received: from mail.catalyst.net.nz ([2404:130:0:10::40:0]) by magus.postgresql.org with esmtp (Exim 4.80) (envelope-from ) id 1YBuuW-00086V-Vu for pgsql-performance@postgresql.org; Fri, 16 Jan 2015 00:37:16 +0000 Received: from localhost (localhost [127.0.0.1]) by mail.catalyst.net.nz (Postfix) with ESMTP id C93F8322DB; Fri, 16 Jan 2015 13:37:07 +1300 (NZDT) X-Virus-Scanned: Debian amavisd-new at catalyst.net.nz Received: from mail.catalyst.net.nz ([127.0.0.1]) by localhost (bertrand.catalyst.net.nz [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id ozx0B4Ai2cPM; Fri, 16 Jan 2015 13:37:06 +1300 (NZDT) Received: from [IPv6:2404:130:0:1000:9107:c07f:1300:5b09] (unknown [IPv6:2404:130:0:1000:9107:c07f:1300:5b09]) (Authenticated sender: mark.kirkwood) by mail.catalyst.net.nz (Postfix) with ESMTPSA id D94FE3229F; Fri, 16 Jan 2015 13:37:06 +1300 (NZDT) Message-ID: <54B85D32.7090106@catalyst.net.nz> Date: Fri, 16 Jan 2015 13:37:06 +1300 From: Mark Kirkwood User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:31.0) Gecko/20100101 Thunderbird/31.3.0 MIME-Version: 1.0 To: Josh Berkus , postgres performance list Subject: Re: Strange choice of general index over partial index References: <54B83F70.7080004@agliodbs.com> In-Reply-To: <54B83F70.7080004@agliodbs.com> Content-Type: text/plain; charset=utf-8; format=flowed Content-Transfer-Encoding: 7bit X-Pg-Spam-Score: -1.9 (-) List-Archive: List-Help: List-ID: List-Owner: List-Post: List-Subscribe: List-Unsubscribe: X-Mailing-List: pgsql-performance Precedence: bulk Sender: pgsql-performance-owner@postgresql.org On 16/01/15 11:30, Josh Berkus wrote: > This is an obfuscation and mock up, but: > > table files ( > id serial pk, > filename text not null, > state varchar(20) not null > ... 18 more columns > ) > > index file_state on (state) > (35GB in size) > index file_in_flight_state (state) where state in ( > 'waiting','assigning', 'processing' ) > (600MB in size) > ... 10 more indexes > > More important facts: > * state = 'done' 95% of the time. thereform the partial index > represents only 5% of the table > * all indexes and the table are very bloated > * server has 128GB RAM > * Version 9.2. > > Given this setup, I would expect the planner to *always* choose > file_in_flight_state over file_state for this query: > > SELECT id, filename FROM files WHERE state = 'waiting'; > > ... and yet it keeps selecting file_state based on extremely small > changes to the stats. This is important because the same query, using > file_state, is 20X to 50X slower, because that index frequently gets > pushed out of memory. > > What am I missing? Or is this potentially a planner bug for costing? > Are you seeing a bitmapscan access plan? If so see if disabling it gets you a plan on the files_in_flight index. I'm seeing this scenario with a fake/generated dataset a bit like yours in 9.2 (9.5 uses the files_in_flight w/o any coercing). regards Mark -- Sent via pgsql-performance mailing list (pgsql-performance@postgresql.org) To make changes to your subscription: http://www.postgresql.org/mailpref/pgsql-performance