Received: from localhost (unknown [200.46.208.211]) by mail.postgresql.org (Postfix) with ESMTP id 92C09633712 for ; Mon, 23 Mar 2009 01:05:17 -0300 (ADT) Received: from mail.postgresql.org ([200.46.204.86]) by localhost (mx1.hub.org [200.46.208.211]) (amavisd-maia, port 10024) with ESMTP id 62626-02 for ; Mon, 23 Mar 2009 01:05:10 -0300 (ADT) X-Greylist: from auto-whitelisted by SQLgrey-1.7.6 Received: from westnet.com (westnet.com [216.187.52.2]) by mail.postgresql.org (Postfix) with ESMTP id 09E8B636113 for ; Mon, 23 Mar 2009 01:04:41 -0300 (ADT) Received: from westnet.com (localhost [127.0.0.1]) by westnet.com (8.14.0/8.14.0) with ESMTP id n2N44eUx009476; Mon, 23 Mar 2009 00:04:40 -0400 (EDT) Received: from localhost (gsmith@localhost) by westnet.com (8.14.0/8.13.2/Submit) with ESMTP id n2N44elx009473; Mon, 23 Mar 2009 00:04:40 -0400 (EDT) X-Authentication-Warning: westnet.com: gsmith owned process doing -bs Date: Mon, 23 Mar 2009 00:04:40 -0400 (EDT) From: Greg Smith X-X-Sender: gsmith@westnet.com To: Laurent Wandrebeck cc: pgsql-performance@postgresql.org Subject: Re: "iowait" bug? In-Reply-To: Message-ID: References: User-Agent: Alpine 2.01 (GSO 1184 2008-12-16) MIME-Version: 1.0 Content-Type: TEXT/PLAIN; format=flowed; charset=US-ASCII X-Virus-Scanned: Maia Mailguard 1.0.1 X-Spam-Status: No, hits=0 tagged_above=0 required=5 tests=AWL=0.000 X-Spam-Level: X-Archive-Number: 200903/325 X-Sequence-Number: 33247 On Mon, 23 Mar 2009, Laurent Wandrebeck wrote: > I thought about completely fair scheduler at first, but that one came > in around 2.6.21. CFS showed up in 2.6.23. > I think that forcing the system to write down more often, and smaller > data just hides the problem, and doesn't correct it. That's one possibility. I've been considering things like whether the OS is getting bogged down managing things like the elevator sorting for outstanding writes. If there was something about that process that gets really inefficient proportionally to the size of the pending queue, that would both match the kinds of symptoms people are reporting, and would go better just reducing the maximum size of the issue by lowering the pdflush tunables. Anyway, the point I was trying to make is that there sure seem to be multiple problems mixed into that one bug report, and it's starting to look just as unmanagably messy as the older bug that had to be abandoned. It would have been nice if somebody kicked out all the diversions it wanted into to keep the focus a bit better. Anybody using a SSD device, USB, or ext4 should have been punted to somewhere else for example. Plenty of examples that don't require any of those things. -- * Greg Smith gsmith@gregsmith.com http://www.gregsmith.com Baltimore, MD