From: Karsten Weiss Subject: Re: Bad ext4 sync performance on 16 TB GPT partition Date: Fri, 26 Feb 2010 16:47:04 +0100 (CET) Message-ID: References: <87ljegb4ow.fsf@openvz.org> Mime-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Cc: "linux-ext4@vger.kernel.org" To: Dmitry Monakhov Return-path: Received: from smtp2.belwue.de ([129.143.2.15]:33954 "EHLO smtp2.belwue.de" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S965085Ab0BZPrO (ORCPT ); Fri, 26 Feb 2010 10:47:14 -0500 Received: from mx3.science-computing.de (mx3.science-computing.de [193.197.16.20]) by smtp2.belwue.de with ESMTP id o1QFlChs011334 for ; Fri, 26 Feb 2010 16:47:12 +0100 (MET) env-from (prvs=6660bea79=K.Weiss@science-computing.de) In-Reply-To: <87ljegb4ow.fsf@openvz.org> Sender: linux-ext4-owner@vger.kernel.org List-ID: Hi Dmitry! On Fri, 26 Feb 2010, Dmitry Monakhov wrote: > > Kernels: > > * 2.6.18-164.11.1.el5 x86_64 (latest CentOS 5.4 kernel) > > * 2.6.18-190.el5 x86_64 (latest Red Hat EL5 test kernel I've found from > > http://people.redhat.com/jwilson/el5/ which contains an ext4 version > > which (according to the rpm's changelog) was updated from the 2.6.32 > > ext4 codebase. > Hmm.. It is hard to predict differences between vanilla tree. > This is no only ext4 related. writeback path is changed dramatically. > It is not easy to port writeback code to 2.6.18 with full performance > improvements but without introducing new issues. > > * I did not try a vanilla kernel so far. > IMHO It would be really good to know vanilla kernel's stats. I did a quick&dirty compilation of vanilla kernel 2.6.33 and repeated the test: # /usr/bin/time bash -c "dd if=/dev/zero of=/mnt/large/10GB bs=1M count=10000 && sync" 10000+0 records in 10000+0 records out 10485760000 bytes (10 GB) copied, 50.044 seconds, 210 MB/s 0.01user 13.76system 1:04.75elapsed 21%CPU (0avgtext+0avgdata 6224maxresident)k 0inputs+0outputs (0major+1049minor)pagefaults 0swaps => The problem shows only with the CentOS / Red Hat 5.4 kernels (including RH's test kernel 2.6.18-190.el5). Aadmittedly ext4 is only a technology preview in 5.4... I've also tried the latest CentOS 5.3 kernel-2.6.18-128.7.1.el5 but couldn't mount the device (with -t ext4dev). 2.6.18-164.el5 (the initial CentOS 5.4 kernel) has the bug, too. I'm willing to test patches if somebody wants to debug the problem. -- Karsten Weiss