From mboxrd@z Thu Jan 1 00:00:00 1970 From: Christoph Hellwig Subject: Re: [Bug 18632] "INFO: task" dpkg "blocked for more than 120 seconds. Date: Thu, 9 Jun 2011 08:17:42 -0400 Message-ID: <20110609121742.GA29569@infradead.org> References: <201106082138.p58Lchgj002615@demeter2.kernel.org> <20110608150241.8412a63d.akpm@linux-foundation.org> <20110609033217.GA10741@localhost> <20110609035426.GA12061@localhost> <20110609082718.GA10335@infradead.org> <20110609090906.GA19186@localhost> <20110609110214.GA9017@infradead.org> <20110609121117.GA5768@localhost> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Cc: Christoph Hellwig , Dave Chinner , Andrew Morton , Jan Kara , "linux-fsdevel@vger.kernel.org" , "bugzilla-daemon@bugzilla.kernel.org" , "daaugusto@gmail.com" , "kernel-bugzilla@cygnusx-1.org" , "listposter@gmail.com" , "justincase@yopmail.com" , "clopez@igalia.com" , Tao Ma , Jens Axboe , Vivek Goyal To: Wu Fengguang Return-path: Received: from 173-166-109-252-newengland.hfc.comcastbusiness.net ([173.166.109.252]:59223 "EHLO bombadil.infradead.org" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1754355Ab1FIMR5 (ORCPT ); Thu, 9 Jun 2011 08:17:57 -0400 Content-Disposition: inline In-Reply-To: <20110609121117.GA5768@localhost> Sender: linux-fsdevel-owner@vger.kernel.org List-ID: On Thu, Jun 09, 2011 at 08:11:17PM +0800, Wu Fengguang wrote: > > This waits for the I/O completion to actually arrive - something that > > XFS does correctly in both sync and fsync, but ext4 only does for fsync. > > Will it benefit to flush the disk _once_ at the end of sync? > (perhaps it's not be as easy in complex storage setups or whatever) XFS currently does it's own passes over data and metadata in sync. To some extent they are overkill, but things like the iowait that's prominent in your trace actually is needed to catch all in-flight direct I/O. > It's definitely a problem that cfq delays async writes too much. > However in Carlos's report, > > https://bugzilla.kernel.org/attachment.cgi?id=61222 > > there are no sync(1) or fsync running at all. So it may be indicating > a different problem. If you have a heavy read load we still need to write back timestamps, and it kinda sounds like he's hitting that. No guarantee that actuall is the cause, though.