From owner-freebsd-fs@FreeBSD.ORG Sun Aug 24 06:04:42 2003 Return-Path: Delivered-To: freebsd-fs@freebsd.org Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125]) by hub.freebsd.org (Postfix) with ESMTP id 9E84216A4BF for ; Sun, 24 Aug 2003 06:04:42 -0700 (PDT) Received: from out-mx1.crosswinds.net (out-mx1.crosswinds.net [216.18.117.38]) by mx1.FreeBSD.org (Postfix) with ESMTP id DA6E943FF9 for ; Sun, 24 Aug 2003 06:04:41 -0700 (PDT) (envelope-from tony@crosswinds.net) Received: from admin.crosswinds.net (out-mx1.crosswinds.net [216.18.117.38]) by out-mx1.crosswinds.net (Postfix) with ESMTP id 5AD842C614; Sun, 24 Aug 2003 09:04:39 -0400 (EDT) Received: by admin.crosswinds.net (Postfix, from userid 1001) id 09F8C3F0B; Sun, 24 Aug 2003 09:04:39 -0400 (EDT) Date: Sun, 24 Aug 2003 09:04:39 -0400 From: Tony Holmes To: Chris Dillon Message-ID: <20030824090439.A95597@crosswinds.net> References: <20030822101702.A73928@crosswinds.net> <20030824002334.E7724@duey.wolves.k12.mo.us> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline User-Agent: Mutt/1.2.5i In-Reply-To: <20030824002334.E7724@duey.wolves.k12.mo.us>; from cdillon@wolves.k12.mo.us on Sun, Aug 24, 2003 at 12:38:21AM -0500 cc: freebsd-fs@FreeBSD.ORG Subject: Re: RAID 1+0 X-BeenThere: freebsd-fs@freebsd.org X-Mailman-Version: 2.1.1 Precedence: list List-Id: Filesystems List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Sun, 24 Aug 2003 13:04:42 -0000 On +Aug 24, Chris Dillon wrote: > > If this is a SCSI array, and I assume it is, make sure your > termination is correct. After troubleshooting somebody's white-box > 10-drive external array when it would consistently drop disks and new > replacements would again "drop dead" a few days to weeks later, I > noticed the vendor didn't properly terminate either of the two LVD > busses. I was totally surprised that it worked at all with a > half-terminated bus, but I guess LVD is relatively forgiving compared > to SE busses when it comes to termination. Once a terminator was put > on the end of each bus (the RAID controller was already terminating > the other end), a drive hasn't dropped out in several months. Aye, it's hot swap LVD scsi on an internally terminated backplane. Ot the two drives that dropped, one will no longer spin up and the other has a massive media error right in the middle of the platter. -- Tony Holmes Founder and Senior Systems Architect Crosswinds Internet Communications Inc. From owner-freebsd-fs@FreeBSD.ORG Mon Aug 25 06:29:57 2003 Return-Path: Delivered-To: freebsd-fs@freebsd.org Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125]) by hub.freebsd.org (Postfix) with ESMTP id 2F8C316A4BF for ; Mon, 25 Aug 2003 06:29:57 -0700 (PDT) Received: from mail.wolves.k12.mo.us (duey.wolves.k12.mo.us [207.160.214.9]) by mx1.FreeBSD.org (Postfix) with ESMTP id ECA1F43FB1 for ; Mon, 25 Aug 2003 06:29:55 -0700 (PDT) (envelope-from cdillon@wolves.k12.mo.us) Received: from localhost (localhost [127.0.0.1]) by mail.wolves.k12.mo.us (Postfix) with ESMTP id 1F64F1FF69; Mon, 25 Aug 2003 08:29:55 -0500 (CDT) Received: from mail.wolves.k12.mo.us ([127.0.0.1]) by localhost (duey.wolves.k12.mo.us [127.0.0.1]) (amavisd-new, port 10024) with LMTP id 19620-02-35; Mon, 25 Aug 2003 08:29:44 -0500 (CDT) Received: by mail.wolves.k12.mo.us (Postfix, from userid 1001) id 924971FEB7; Mon, 25 Aug 2003 08:29:44 -0500 (CDT) Received: from localhost (localhost [127.0.0.1]) by mail.wolves.k12.mo.us (Postfix) with ESMTP id 909281B710; Mon, 25 Aug 2003 08:29:44 -0500 (CDT) Date: Mon, 25 Aug 2003 08:29:44 -0500 (CDT) From: Chris Dillon To: Tony Holmes In-Reply-To: <20030824090439.A95597@crosswinds.net> Message-ID: <20030825080957.S19707@duey.wolves.k12.mo.us> References: <20030822101702.A73928@crosswinds.net> <20030824002334.E7724@duey.wolves.k12.mo.us> <20030824090439.A95597@crosswinds.net> MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII X-Virus-Scanned: by amavisd-new at wolves.k12.mo.us cc: freebsd-fs@FreeBSD.ORG Subject: Re: RAID 1+0 X-BeenThere: freebsd-fs@freebsd.org X-Mailman-Version: 2.1.1 Precedence: list List-Id: Filesystems List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Mon, 25 Aug 2003 13:29:57 -0000 On Sun, 24 Aug 2003, Tony Holmes wrote: > Aye, it's hot swap LVD scsi on an internally terminated backplane. Are you sure? I've never seen an internally terminated hot-swap backplane, probably because it helps prevent foot-shooting. It wouldn't surprise me if someone does make one, though. > Ot the two drives that dropped, one will no longer spin up and the > other has a massive media error right in the middle of the platter. I recall that several of the drives that fell out of the array I was looking at had become physically damaged (wouldn't spin up and/or be recognized at all). The rest of them that dropped out without any apparent reason were physically OK, but the lack of proper termination probably caused enough errors to force the RAID controller to drop the drive. Those drives are still running today and with proper termination they haven't given any trouble. I have to wonder if the lack of termination can physically damage a drive given enough time (and a heavy load, as it was), or if it was just coincidental that we had some physically damaged drives in the midst of the termination problem. -- Chris Dillon - cdillon(at)wolves.k12.mo.us FreeBSD: The fastest and most stable server OS on the planet - Available for IA32, IA64, PC98, Alpha, and UltraSPARC architectures - x86-64, PowerPC, ARM, MIPS, and S/390 under development - http://www.freebsd.org No trees were harmed in the composition of this message, although some electrons were mildly inconvenienced. From owner-freebsd-fs@FreeBSD.ORG Tue Aug 26 19:15:26 2003 Return-Path: Delivered-To: freebsd-fs@freebsd.org Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125]) by hub.freebsd.org (Postfix) with ESMTP id CE23E16A4BF; Tue, 26 Aug 2003 19:15:26 -0700 (PDT) Received: from mta10.adelphia.net (mta10.adelphia.net [68.168.78.202]) by mx1.FreeBSD.org (Postfix) with ESMTP id D866A44005; Tue, 26 Aug 2003 19:15:25 -0700 (PDT) (envelope-from wmoran@potentialtech.com) Received: from potentialtech.com ([24.53.179.151]) by mta10.adelphia.net (InterMail vM.5.01.05.32 201-253-122-126-132-20030307) with ESMTP id <20030827021526.YAEX19518.mta10.adelphia.net@potentialtech.com>; Tue, 26 Aug 2003 22:15:26 -0400 Message-ID: <3F4C143D.7000909@potentialtech.com> Date: Tue, 26 Aug 2003 22:15:25 -0400 From: Bill Moran User-Agent: Mozilla/5.0 (X11; U; FreeBSD i386; en-US; rv:1.3) Gecko/20030429 X-Accept-Language: en-us, en MIME-Version: 1.0 To: freebsd-fs@freebsd.org, freebsd-performance@freebsd.org Content-Type: text/plain; charset=us-ascii; format=flowed Content-Transfer-Encoding: 7bit Subject: Performance tests I did with FreeBSD, Linux and PostgreSQL X-BeenThere: freebsd-fs@freebsd.org X-Mailman-Version: 2.1.1 Precedence: list List-Id: Filesystems List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Wed, 27 Aug 2003 02:15:26 -0000 Like all benchmarks, I doubt these are perfect (or even close) examples of real-world use. However, in the hopes that they will be useful for improving FreeBSD, I present them to the community. http://www.potentialtech.com/wmoran/postgresql.php Intelligent comments are welcome. -- Bill Moran Potential Technologies http://www.potentialtech.com From owner-freebsd-fs@FreeBSD.ORG Tue Aug 26 19:22:31 2003 Return-Path: Delivered-To: freebsd-fs@freebsd.org Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125]) by hub.freebsd.org (Postfix) with ESMTP id 59FF216A4BF; Tue, 26 Aug 2003 19:22:31 -0700 (PDT) Received: from perrin.nxad.com (internal.ext.nxad.com [69.1.70.251]) by mx1.FreeBSD.org (Postfix) with ESMTP id C5E1943FF2; Tue, 26 Aug 2003 19:22:28 -0700 (PDT) (envelope-from sean@nxad.com) Received: by perrin.nxad.com (Postfix, from userid 1001) id 3E54720F00; Tue, 26 Aug 2003 19:22:28 -0700 (PDT) Date: Tue, 26 Aug 2003 19:22:28 -0700 From: Sean Chittenden To: Bill Moran Message-ID: <20030827022228.GA51167@perrin.nxad.com> References: <3F4C143D.7000909@potentialtech.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <3F4C143D.7000909@potentialtech.com> X-PGP-Key: finger seanc@FreeBSD.org X-PGP-Fingerprint: 3849 3760 1AFE 7B17 11A0 83A6 DD99 E31F BC84 B341 X-Web-Homepage: http://sean.chittenden.org/ User-Agent: Mutt/1.5.4i cc: freebsd-fs@freebsd.org cc: freebsd-performance@freebsd.org Subject: Re: Performance tests I did with FreeBSD, Linux and PostgreSQL X-BeenThere: freebsd-fs@freebsd.org X-Mailman-Version: 2.1.1 Precedence: list List-Id: Filesystems List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Wed, 27 Aug 2003 02:22:31 -0000 > Like all benchmarks, I doubt these are perfect (or even close) examples > of real-world use. > > However, in the hopes that they will be useful for improving FreeBSD, I > present them to the community. > > http://www.potentialtech.com/wmoran/postgresql.php > > Intelligent comments are welcome. Could you please repeat your testing with 4.X? I'm quite sure you'll see a much different picture. FreeBSD 5.X's state of locking isn't ready for production or benchmarking. If, however, you are ready to profile a 5.X kernel, those results would be very interesting to the people on this list. -sc -- Sean Chittenden From owner-freebsd-fs@FreeBSD.ORG Tue Aug 26 23:06:47 2003 Return-Path: Delivered-To: freebsd-fs@freebsd.org Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125]) by hub.freebsd.org (Postfix) with ESMTP id 4C2CB16A4BF; Tue, 26 Aug 2003 23:06:47 -0700 (PDT) Received: from HAL9000.homeunix.com (12-233-57-131.client.attbi.com [12.233.57.131]) by mx1.FreeBSD.org (Postfix) with ESMTP id 8600D43FF3; Tue, 26 Aug 2003 23:06:46 -0700 (PDT) (envelope-from das@FreeBSD.ORG) Received: from HAL9000.homeunix.com (localhost [127.0.0.1]) by HAL9000.homeunix.com (8.12.9/8.12.9) with ESMTP id h7R66iXR030383; Tue, 26 Aug 2003 23:06:44 -0700 (PDT) (envelope-from das@FreeBSD.ORG) Received: (from das@localhost) by HAL9000.homeunix.com (8.12.9/8.12.9/Submit) id h7R66i7k030382; Tue, 26 Aug 2003 23:06:44 -0700 (PDT) (envelope-from das@FreeBSD.ORG) Date: Tue, 26 Aug 2003 23:06:44 -0700 From: David Schultz To: Bill Moran Message-ID: <20030827060644.GA9321@HAL9000.homeunix.com> Mail-Followup-To: Bill Moran , freebsd-fs@freebsd.org, freebsd-performance@freebsd.org References: <3F4C143D.7000909@potentialtech.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <3F4C143D.7000909@potentialtech.com> cc: freebsd-fs@FreeBSD.ORG cc: freebsd-performance@FreeBSD.ORG Subject: Re: Performance tests I did with FreeBSD, Linux and PostgreSQL X-BeenThere: freebsd-fs@freebsd.org X-Mailman-Version: 2.1.1 Precedence: list List-Id: Filesystems List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Wed, 27 Aug 2003 06:06:47 -0000 On Tue, Aug 26, 2003, Bill Moran wrote: > Like all benchmarks, I doubt these are perfect (or even close) examples > of real-world use. > > However, in the hopes that they will be useful for improving FreeBSD, I > present them to the community. > > http://www.potentialtech.com/wmoran/postgresql.php > > Intelligent comments are welcome. Well, a little trussing revealed that pgsql uses a block size of 8K, which can be a significant pessimization when the filesystem block size is 16K, which is the default in FreeBSD! ext2fs uses 4K blocks and reiserfs uses extents, so those filesystems would not suffer from that problem. You might consider trying your test on a UFS filesystem with 8K blocks, or you could tune pgsql to use a different block size. (For pgsql, I think the appropriate tunable is BLCKSZ in include/pg_config.h. You'll have to recompile it and create a new database.) I used to think that the lack of fdatasync() in FreeBSD would significantly pessimize databases such as pgsql, but after running your script, I discovered that the fsync() overhead was only about 1% of the total execution time. Maybe that would be more significant if you tried to execute lots of little transactions, rather than a few large ones. That kind of workload is probably more common these days anyway. As previously mentioned, it would also be interesting to see what you get out of FreeBSD 4.8+softupdates, just in case the performance problems you're seeing are a result of known (or unknown) regressions in 5.X. You might also consider running your results by the pgsql people, since they understand how the database works. ;-) From owner-freebsd-fs@FreeBSD.ORG Wed Aug 27 02:42:01 2003 Return-Path: Delivered-To: freebsd-fs@freebsd.org Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125]) by hub.freebsd.org (Postfix) with ESMTP id A022216A4BF; Wed, 27 Aug 2003 02:42:01 -0700 (PDT) Received: from mailout01.sul.t-online.com (mailout01.sul.t-online.com [194.25.134.80]) by mx1.FreeBSD.org (Postfix) with ESMTP id 8A9F343FBD; Wed, 27 Aug 2003 02:42:00 -0700 (PDT) (envelope-from Alexander@Leidinger.net) Received: from fwd02.aul.t-online.de by mailout01.sul.t-online.com with smtp id 19rwoR-0001hV-00; Wed, 27 Aug 2003 11:41:59 +0200 Received: from Andro-Beta.Leidinger.net (Gn3VbeZFwetCv-kc+sWe9dqEklA6TOWM2LcaO7-SajLcj3uBjxum8k@[80.131.113.207]) by fmrl02.sul.t-online.com with esmtp id 19rwo8-1nj51c0; Wed, 27 Aug 2003 11:41:40 +0200 Received: from Magelan.Leidinger.net (Magelan [192.168.1.1]) h7R9gg9O084999; Wed, 27 Aug 2003 11:42:42 +0200 (CEST) (envelope-from Alexander@Leidinger.net) Received: from Magelan.Leidinger.net (netchild@localhost [127.0.0.1]) by Magelan.Leidinger.net (8.12.9/8.12.9) with SMTP id h7R9hGBA008201; Wed, 27 Aug 2003 11:43:16 +0200 (CEST) (envelope-from Alexander@Leidinger.net) Date: Wed, 27 Aug 2003 11:43:16 +0200 From: Alexander Leidinger To: freebsd-performance@freebsd.org Message-Id: <20030827114316.300fd455.Alexander@Leidinger.net> In-Reply-To: <3F4C143D.7000909@potentialtech.com> References: <3F4C143D.7000909@potentialtech.com> X-Mailer: Sylpheed version 0.9.3claws (GTK+ 1.2.10; i386-portbld-freebsd5.1) Mime-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit X-Seen: false X-ID: Gn3VbeZFwetCv-kc+sWe9dqEklA6TOWM2LcaO7-SajLcj3uBjxum8k@t-dialin.net cc: freebsd-fs@freebsd.org cc: Bill Moran Subject: Re: Performance tests I did with FreeBSD, Linux and PostgreSQL X-BeenThere: freebsd-fs@freebsd.org X-Mailman-Version: 2.1.1 Precedence: list List-Id: Filesystems List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Wed, 27 Aug 2003 09:42:01 -0000 On Tue, 26 Aug 2003 22:15:25 -0400 Bill Moran wrote: > http://www.potentialtech.com/wmoran/postgresql.php You don't talk about integrity on this page. The ext2fs async mount may lose the FS in the power failure case, ufs+SU doesn't. You also present only the ufs sync and async numbers, but not the noasync (default) numbers, you should either present them or add a note why you don't present them. About the "without debugging" description: did you really only remove INVARIANTS? If yes: you also need to remove WITNESS. I don't know if 5.1-RELEASE has malloc debugging enabled, but to be on the safe side you should maybe do a "ln -s aj /etc/malloc.conf" before you start PostgreSQL. Bye, Alexander. -- I believe the technical term is "Oops!" http://www.Leidinger.net Alexander @ Leidinger.net GPG fingerprint = C518 BC70 E67F 143F BE91 3365 79E2 9C60 B006 3FE7 From owner-freebsd-fs@FreeBSD.ORG Wed Aug 27 06:53:24 2003 Return-Path: Delivered-To: freebsd-fs@freebsd.org Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125]) by hub.freebsd.org (Postfix) with ESMTP id 72AE116A4BF; Wed, 27 Aug 2003 06:53:24 -0700 (PDT) Received: from mta6.adelphia.net (mta6.adelphia.net [68.168.78.190]) by mx1.FreeBSD.org (Postfix) with ESMTP id 7087B44005; Wed, 27 Aug 2003 06:53:23 -0700 (PDT) (envelope-from wmoran@potentialtech.com) Received: from potentialtech.com ([24.53.179.151]) by mta6.adelphia.net (InterMail vM.5.01.05.32 201-253-122-126-132-20030307) with ESMTP id <20030827135324.UAND18217.mta6.adelphia.net@potentialtech.com>; Wed, 27 Aug 2003 09:53:24 -0400 Message-ID: <3F4CB7D2.9000503@potentialtech.com> Date: Wed, 27 Aug 2003 09:53:22 -0400 From: Bill Moran User-Agent: Mozilla/5.0 (X11; U; FreeBSD i386; en-US; rv:1.3) Gecko/20030429 X-Accept-Language: en-us, en MIME-Version: 1.0 To: Alexander Leidinger References: <3F4C143D.7000909@potentialtech.com> <20030827114316.300fd455.Alexander@Leidinger.net> In-Reply-To: <20030827114316.300fd455.Alexander@Leidinger.net> Content-Type: text/plain; charset=us-ascii; format=flowed Content-Transfer-Encoding: 7bit cc: freebsd-fs@freebsd.org cc: freebsd-performance@freebsd.org Subject: Re: Performance tests I did with FreeBSD, Linux and PostgreSQL X-BeenThere: freebsd-fs@freebsd.org X-Mailman-Version: 2.1.1 Precedence: list List-Id: Filesystems List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Wed, 27 Aug 2003 13:53:24 -0000 Alexander Leidinger wrote: > On Tue, 26 Aug 2003 22:15:25 -0400 > Bill Moran wrote: > >>http://www.potentialtech.com/wmoran/postgresql.php > > You don't talk about integrity on this page. Yes I do. It's in the conclusion section. It's not a major part of that document because I didn't test integrity, therefore I have no real conclusions beyond the fact that ext2 is not designed with integrity in mind. > You also present > only the ufs sync and async numbers, but not the noasync (default) > numbers, you should either present them or add a note why you don't > present them. You're right. I'll correct this. > About the "without debugging" description: did you really only remove > INVARIANTS? If yes: you also need to remove WITNESS. WITNESS is not enabled in the GENERIC kernel for 5.1-RELEASE, so there was no need to disable it. > I don't know if 5.1-RELEASE has malloc debugging enabled, but to be on > the safe side you should maybe do a "ln -s aj /etc/malloc.conf" before > you start PostgreSQL. I may look into that, but it seems like the more important thing to do is test 4.8 at this point. If you look at the times, these tests are time-consuming. I probably should have reduced the number of records involved, but if I do that now, I have to rerun all the tests ... -- Bill Moran Potential Technologies http://www.potentialtech.com From owner-freebsd-fs@FreeBSD.ORG Wed Aug 27 08:51:22 2003 Return-Path: Delivered-To: freebsd-fs@freebsd.org Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125]) by hub.freebsd.org (Postfix) with ESMTP id 8699016A4BF for ; Wed, 27 Aug 2003 08:51:22 -0700 (PDT) Received: from smtp.mho.com (smtp.mho.net [64.58.4.6]) by mx1.FreeBSD.org (Postfix) with SMTP id 36DCC43FE5 for ; Wed, 27 Aug 2003 08:51:20 -0700 (PDT) (envelope-from scottl@freebsd.org) Received: (qmail 29108 invoked by uid 1002); 27 Aug 2003 15:51:19 -0000 Received: from unknown (HELO freebsd.org) (64.58.1.252) by smtp.mho.net with SMTP; 27 Aug 2003 15:51:19 -0000 Message-ID: <3F4CD36C.40604@freebsd.org> Date: Wed, 27 Aug 2003 09:51:08 -0600 From: Scott Long User-Agent: Mozilla/5.0 (X11; U; FreeBSD i386; en-US; rv:1.3) Gecko/20030425 X-Accept-Language: en-us, en MIME-Version: 1.0 To: Bill Moran References: <3F4C143D.7000909@potentialtech.com> <20030827114316.300fd455.Alexander@Leidinger.net> <3F4CB7D2.9000503@potentialtech.com> In-Reply-To: <3F4CB7D2.9000503@potentialtech.com> Content-Type: text/plain; charset=us-ascii; format=flowed Content-Transfer-Encoding: 7bit cc: freebsd-fs@freebsd.org cc: Alexander Leidinger cc: freebsd-performance@freebsd.org Subject: Re: Performance tests I did with FreeBSD, Linux and PostgreSQL X-BeenThere: freebsd-fs@freebsd.org X-Mailman-Version: 2.1.1 Precedence: list List-Id: Filesystems List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Wed, 27 Aug 2003 15:51:22 -0000 Bill Moran wrote: > Alexander Leidinger wrote: >> I don't know if 5.1-RELEASE has malloc debugging enabled, but to be on >> the safe side you should maybe do a "ln -s aj /etc/malloc.conf" before >> you start PostgreSQL. malloc debugging is disabled in 5.0-R and 5.1-R. Once we hit 5-STABLE it will be turned off for good in the branch. > > I may look into that, but it seems like the more important thing to do > is test 4.8 at this point. > > If you look at the times, these tests are time-consuming. I probably > should have reduced the number of records involved, but if I do that > now, I have to rerun all the tests ... > It would be interesting to see if the recent locking of the ATA driver has any benefit on these tests. Of course if you update to 5-CURRENT then you'll need to disable INVARIANTS, WITNESS, and malloc debugging. Scott From owner-freebsd-fs@FreeBSD.ORG Wed Aug 27 09:50:30 2003 Return-Path: Delivered-To: freebsd-fs@freebsd.org Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125]) by hub.freebsd.org (Postfix) with ESMTP id 3CD5A16A4BF; Wed, 27 Aug 2003 09:50:30 -0700 (PDT) Received: from fledge.watson.org (fledge.watson.org [204.156.12.50]) by mx1.FreeBSD.org (Postfix) with ESMTP id 5A85E43FE3; Wed, 27 Aug 2003 09:50:29 -0700 (PDT) (envelope-from robert@fledge.watson.org) Received: from fledge.watson.org (localhost [127.0.0.1]) by fledge.watson.org (8.12.9/8.12.9) with ESMTP id h7RGo5rO026285; Wed, 27 Aug 2003 12:50:05 -0400 (EDT) (envelope-from robert@fledge.watson.org) Received: from localhost (robert@localhost)h7RGo5CT026282; Wed, 27 Aug 2003 12:50:05 -0400 (EDT) Date: Wed, 27 Aug 2003 12:50:05 -0400 (EDT) From: Robert Watson X-Sender: robert@fledge.watson.org To: Bill Moran In-Reply-To: <3F4C143D.7000909@potentialtech.com> Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII cc: freebsd-fs@freebsd.org cc: freebsd-performance@freebsd.org Subject: Re: Performance tests I did with FreeBSD, Linux and PostgreSQL X-BeenThere: freebsd-fs@freebsd.org X-Mailman-Version: 2.1.1 Precedence: list List-Id: Filesystems List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Wed, 27 Aug 2003 16:50:30 -0000 On Tue, 26 Aug 2003, Bill Moran wrote: > Like all benchmarks, I doubt these are perfect (or even close) examples > of real-world use. > > However, in the hopes that they will be useful for improving FreeBSD, I > present them to the community. > > http://www.potentialtech.com/wmoran/postgresql.php > > Intelligent comments are welcome. I'd have to chime in here with two recommendations: (1) Try 4.8-STABLE (2) Try 5.1-CURRENT With a distinct preference for (1) before (2). Some parts of -CURRENT have gotten faster, others have gotten temporarily slower (new locks in place without Giant coming off them yet, which adds overhead). I'd also check your block sizes for interaction with Postgres and see if different block and fragment sizes help/hurt. It could be that postgres would benefit from different tuning parameters on FreeBSD, especially with newer UFS file systems. Robert N M Watson FreeBSD Core Team, TrustedBSD Projects robert@fledge.watson.org Network Associates Laboratories From owner-freebsd-fs@FreeBSD.ORG Wed Aug 27 10:29:20 2003 Return-Path: Delivered-To: freebsd-fs@freebsd.org Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125]) by hub.freebsd.org (Postfix) with ESMTP id 2F8FF16A4BF; Wed, 27 Aug 2003 10:29:20 -0700 (PDT) Received: from fledge.watson.org (fledge.watson.org [204.156.12.50]) by mx1.FreeBSD.org (Postfix) with ESMTP id 0F11B43FDD; Wed, 27 Aug 2003 10:29:19 -0700 (PDT) (envelope-from robert@fledge.watson.org) Received: from fledge.watson.org (localhost [127.0.0.1]) by fledge.watson.org (8.12.9/8.12.9) with ESMTP id h7RHStrO026607; Wed, 27 Aug 2003 13:28:55 -0400 (EDT) (envelope-from robert@fledge.watson.org) Received: from localhost (robert@localhost)h7RHStTV026604; Wed, 27 Aug 2003 13:28:55 -0400 (EDT) Date: Wed, 27 Aug 2003 13:28:55 -0400 (EDT) From: Robert Watson X-Sender: robert@fledge.watson.org To: David Schultz In-Reply-To: <20030827060644.GA9321@HAL9000.homeunix.com> Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII cc: freebsd-fs@FreeBSD.ORG cc: freebsd-performance@FreeBSD.ORG cc: Bill Moran Subject: Re: Performance tests I did with FreeBSD, Linux and PostgreSQL X-BeenThere: freebsd-fs@freebsd.org X-Mailman-Version: 2.1.1 Precedence: list List-Id: Filesystems List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Wed, 27 Aug 2003 17:29:20 -0000 On Tue, 26 Aug 2003, David Schultz wrote: > On Tue, Aug 26, 2003, Bill Moran wrote: > > Like all benchmarks, I doubt these are perfect (or even close) examples > > of real-world use. > > > > However, in the hopes that they will be useful for improving FreeBSD, I > > present them to the community. > > > > http://www.potentialtech.com/wmoran/postgresql.php > > > > Intelligent comments are welcome. > > Well, a little trussing revealed that pgsql uses a block size of 8K, > which can be a significant pessimization when the filesystem block size > is 16K, which is the default in FreeBSD! ext2fs uses 4K blocks and > reiserfs uses extents, so those filesystems would not suffer from that > problem. You might consider trying your test on a UFS filesystem with > 8K blocks, or you could tune pgsql to use a different block size. (For > pgsql, I think the appropriate tunable is BLCKSZ in include/pg_config.h. > You'll have to recompile it and create a new database.) I wonder if it would benefit us to do some tweaks to Postgres to tune the pgsql block size to the file system it's running on, and get it integrated back into the distribution.... As disks get bigger, a 16k block size is going to become more common, so this won't just hit FreeBSD, I expect. Robert N M Watson FreeBSD Core Team, TrustedBSD Projects robert@fledge.watson.org Network Associates Laboratories From owner-freebsd-fs@FreeBSD.ORG Wed Aug 27 18:22:33 2003 Return-Path: Delivered-To: freebsd-fs@freebsd.org Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125]) by hub.freebsd.org (Postfix) with ESMTP id C64F716A4BF; Wed, 27 Aug 2003 18:22:33 -0700 (PDT) Received: from mta6.adelphia.net (mta6.adelphia.net [68.168.78.190]) by mx1.FreeBSD.org (Postfix) with ESMTP id 2D27F43F75; Wed, 27 Aug 2003 18:22:32 -0700 (PDT) (envelope-from wmoran@potentialtech.com) Received: from potentialtech.com ([24.53.179.151]) by mta6.adelphia.net (InterMail vM.5.01.05.32 201-253-122-126-132-20030307) with ESMTP id <20030828012232.SPEO18217.mta6.adelphia.net@potentialtech.com>; Wed, 27 Aug 2003 21:22:32 -0400 Message-ID: <3F4D5957.8000204@potentialtech.com> Date: Wed, 27 Aug 2003 21:22:31 -0400 From: Bill Moran User-Agent: Mozilla/5.0 (X11; U; FreeBSD i386; en-US; rv:1.3) Gecko/20030429 X-Accept-Language: en-us, en MIME-Version: 1.0 To: freebsd-performance@freebsd.org, freebsd-database@freebsd.org, freebsd-fs@freebsd.org Content-Type: text/plain; charset=us-ascii; format=flowed Content-Transfer-Encoding: 7bit Subject: Some additional tests run on my performance testing X-BeenThere: freebsd-fs@freebsd.org X-Mailman-Version: 2.1.1 Precedence: list List-Id: Filesystems List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Thu, 28 Aug 2003 01:22:34 -0000 First off, I want to thank everyone who responded with suggestions and comments. I was worried that I might get a lot of flames about these test, and that's not what happened. The most common suggestion I received was "use FreeBSD 4.8", so I managed to make some time between the electrical storms today, and my regular work to do just that. Unfortunately (as you'll see) the results were _worse_ than with FreeBSD 5.1. I'm interested in two major things right now: 1 - can anyone suggest anything I might be doing wrong to get such lousy results? 2 - Can someone please try the tests on SCSI hardware? I don't have any SCSI systems available to test on, and I'd like to figure out if it's the filesystem or the ATA driver that's the problem. Hell, can someone try out the tests on some other brand of ATA/HDD, to make sure FreeBSD doesn't just have some grief with this particular piece of hardware? The updated test results are the same place the previous tests were: http://www.potentialtech.com/wmoran/postgresql.php I'm attaching dmesg from the test machine to this email, if anyone knows of any problems with this hardware and FreeBSD, please let me know so I can stop banging my head against this and just accept that it's a hardware problem ;) I was assuming that the "falling back to PIO mode" was because this board didn't support DMA, but I'm beginning to wonder ... -- Bill Moran Potential Technologies http://www.potentialtech.com Copyright (c) 1992-2003 The FreeBSD Project. Copyright (c) 1979, 1980, 1983, 1986, 1988, 1989, 1991, 1992, 1993, 1994 The Regents of the University of California. All rights reserved. FreeBSD 4.8-RELEASE #0: Thu Apr 3 10:53:38 GMT 2003 root@freebsd-stable.sentex.ca:/usr/obj/usr/src/sys/GENERIC Timecounter "i8254" frequency 1193182 Hz CPU: AMD-K6(tm) 3D processor (450.13-MHz 586-class CPU) Origin = "AuthenticAMD" Id = 0x58c Stepping = 12 Features=0x8021bf AMD Features=0x80000800 real memory = 58720256 (57344K bytes) avail memory = 51949568 (50732K bytes) Preloaded elf kernel "kernel" at 0xc051d000. K6-family MTRR support enabled (2 registers) md0: Malloc disk Using $PIR table, 7 entries at 0xc00fa040 npx0: on motherboard npx0: INT 16 interface pcib0: on motherboard pci0: on pcib0 agp0: mem 0x50000000-0x5fffffff at device 0.0 on pci0 atapci0: port 0x2040-0x204f,0-0x3,0-0x7,0-0x3,0-0x7 irq 0 at device 0.1 on pci0 ata0: at 0x1f0 irq 14 on atapci0 ata1: at 0x170 irq 15 on atapci0 isab0: at device 1.0 on pci0 isa0: on isab0 pci0: (vendor=0x1039, dev=0x0009) at 1.1 ohci0: irq 11 at device 1.2 on pci0 ohci0: Could not map memory device_probe_and_attach: ohci0 attach returned 6 pcib2: at device 2.0 on pci0 pci1: on pcib2 pci1: at 0.0 rl0: port 0x2400-0x24ff mem 0x41000000-0x410000ff irq 10 at device 5.0 on pci0 rl0: Ethernet address: 00:c0:ca:14:03:eb miibus0: on rl0 rlphy0: on miibus0 rlphy0: 10baseT, 10baseT-FDX, 100baseTX, 100baseTX-FDX, auto pci0: (vendor=0x125d, dev=0x1969) at 10.0 irq 5 pcib1: on motherboard pci2: on pcib1 orm0: