From: Greg Freemyer Subject: Re: Ext4 performance regression: Post 2.6.30 Date: Wed, 31 Mar 2010 18:06:28 -0400 Message-ID: References: <1269843935.13369.52.camel@keith-laptop> <87f94c371003290810u58f64ce6uc20be6bbac420e73@mail.gmail.com> <1270000565.7193.14.camel@keith-laptop> <4BB2CA41.5050406@redhat.com> <1270072944.7193.29.camel@keith-laptop> Mime-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: QUOTED-PRINTABLE Cc: Eric Sandeen , linux-ext4 To: Keith Mannthey Return-path: Received: from mail-iw0-f202.google.com ([209.85.223.202]:55463 "EHLO mail-iw0-f202.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1754968Ab0CaWG2 convert rfc822-to-8bit (ORCPT ); Wed, 31 Mar 2010 18:06:28 -0400 Received: by iwn40 with SMTP id 40so491998iwn.1 for ; Wed, 31 Mar 2010 15:06:28 -0700 (PDT) In-Reply-To: <1270072944.7193.29.camel@keith-laptop> Sender: linux-ext4-owner@vger.kernel.org List-ID: On Wed, Mar 31, 2010 at 6:02 PM, Keith Mannthey wr= ote: > On Tue, 2010-03-30 at 23:06 -0500, Eric Sandeen wrote: >> Keith Mannthey wrote: >> > On Mon, 2010-03-29 at 11:10 -0400, Greg Freemyer wrote: >> >> On Mon, Mar 29, 2010 at 2:25 AM, Keith Mannthey wrote: >> >>> >> >>> After 2.6.30 I am seeing large performance regressions on a raid= setup. >> >>> I am working to publish a larger amount of data but I wanted to = get some >> >>> quick data out about what I am seeing. >> >>> >> >> Is mdraid involved? >> >> >> >> They added barrier support for some configs after 2.6.30 I believ= e. >> >> It can cause a drastic perf change, but it increases reliability = and >> >> is "correct". >> > >> > lvm and device mapper are is involved. =A0The git bisect just took= me to: >> > >> > 374bf7e7f6cc38b0483351a2029a97910eadde1b is first bad commit >> > commit 374bf7e7f6cc38b0483351a2029a97910eadde1b >> > Author: Mikulas Patocka >> > Date: =A0 Mon Jun 22 10:12:22 2009 +0100 >> > >> > =A0 =A0 dm: stripe support flush >> > >> > =A0 =A0 Flush support for the stripe target. >> > >> > =A0 =A0 This sets ti->num_flush_requests to the number of stripes = and >> > =A0 =A0 remaps individual flush requests to the appropriate stripe= devices. >> > >> > =A0 =A0 Signed-off-by: Mikulas Patocka >> > =A0 =A0 Signed-off-by: Alasdair G Kergon >> > >> > :040000 040000 542f4b9b442d1371c6534f333b7e00714ef98609 d490479b66= 0139fc1b6b0ecd17bb58c9e00e597e M =A0drivers >> > >> > >> > This may be correct behavior but the performance penalty in this t= est >> > case is pretty high. >> > >> > I am going to move back to current kernels and starting looking in= to >> > ext4/dm flushing. >> >> It would probably be interesting to do a mount -o nobarrier to see i= f >> that makes the regression go away. > > -o nobarrier takes the regression away with 2.6.34-rc3: > > Default mount: ~27500 > > -o nobarrier: ~12500 > > Barriers on this setup cost ALOT during writes. > > Interestingly as well the "mailserver" workload regression is also > removed by mounting with "-o nobarrier". > > I am going to see what impact is seen on a single disk setup. > > Thanks, > =A0Keith Mannthey > =A0LTC FS-Dev I'm curious if your using an internal or external journal? I'd guess the cost of barriers is much greater with an internal journal, but I don't recall seeing any benchmarks one way or the other. Greg -- To unsubscribe from this list: send the line "unsubscribe linux-ext4" i= n the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html