From: Solofo.Ramangalahy@bull.net Subject: Re: jbd/jbd2 performance improvements Date: Thu, 23 Oct 2008 14:22:03 +0200 Message-ID: <18688.27755.969223.822862@frecb006361.adech.frec.bull.fr> References: <48F62893.9060606@redhat.com> <18678.55651.556822.187508@frecb006361.adech.frec.bull.fr> <48F72E5F.2050409@redhat.com> <48F735E8.7060803@redhat.com> <18688.21800.641991.52484@frecb006361.adech.frec.bull.fr> <49006758.9020901@redhat.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Cc: Eric Sandeen , Ric Wheeler , "Theodore Ts'o" , linux-ext4@vger.kernel.org, Andrew Morton To: Arjan De Ven Return-path: Received: from ecfrec.frec.bull.fr ([129.183.4.8]:46433 "EHLO ecfrec.frec.bull.fr" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1752307AbYJWMWU (ORCPT ); Thu, 23 Oct 2008 08:22:20 -0400 In-Reply-To: <49006758.9020901@redhat.com> Sender: linux-ext4-owner@vger.kernel.org List-ID: Ric Wheeler writes: > Do you have any details on the test case that you ran that showed a > clear improvement? What kind of storage & IO pattern did you use? Is it possible to record latencytop output (like top batch mode) to capture the highest latency during a test run? Or how did you collect this: >my reproducer is sadly very simple (claws-mail is my mail client that uses maildir) >Process claws-mail (4896) Total: 2829.7 msec >EXT3: Waiting for journal access 2491.0 msec 88.4 % >Writing back inodes 160.9 msec 5.7 % >synchronous write 78.8 msec 3.0 % >is an example of such a trace (this is with patch, without patch the numbers are about 3x bigger) >Waiting for journal access is "journal_get_write_access" >Writing back inodes is "writeback_inodes" >synchronous write is "do_sync_write" -- solofo