From oracle-l-bounce@freelists.org Fri Jul 9 15:17:16 2004 Return-Path: Received: from air189.startdedicated.com (root@localhost) by orafaq.com (8.11.6/8.11.6) with ESMTP id i69KGpe04956 for ; Fri, 9 Jul 2004 15:17:01 -0500 X-ClientAddr: 206.53.239.180 Received: from turing.freelists.org (freelists-180.iquest.net [206.53.239.180]) by air189.startdedicated.com (8.11.6/8.11.6) with ESMTP id i69KGU604899 for ; Fri, 9 Jul 2004 15:16:50 -0500 Received: from localhost (localhost [127.0.0.1]) by turing.freelists.org (Avenir Technologies Mail Multiplex) with ESMTP id C522172C9B4; Fri, 9 Jul 2004 14:57:34 -0500 (EST) Received: from turing.freelists.org ([127.0.0.1]) by localhost (turing [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id 20837-66; Fri, 9 Jul 2004 14:57:34 -0500 (EST) Received: from turing (localhost [127.0.0.1]) by turing.freelists.org (Avenir Technologies Mail Multiplex) with ESMTP id 192C972C9A1; Fri, 9 Jul 2004 14:57:34 -0500 (EST) Received: with ECARTIS (v1.0.0; list oracle-l); Fri, 09 Jul 2004 14:56:08 -0500 (EST) X-Original-To: oracle-l@freelists.org Delivered-To: oracle-l@freelists.org Received: from localhost (localhost [127.0.0.1]) by turing.freelists.org (Avenir Technologies Mail Multiplex) with ESMTP id B06B472C5FB for ; Fri, 9 Jul 2004 14:56:07 -0500 (EST) Received: from turing.freelists.org ([127.0.0.1]) by localhost (turing [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id 20837-40 for ; Fri, 9 Jul 2004 14:56:07 -0500 (EST) Received: from puzzle.pobox.com (unknown [207.8.226.4]) by turing.freelists.org (Avenir Technologies Mail Multiplex) with ESMTP id 5B14472C40D for ; Fri, 9 Jul 2004 14:56:07 -0500 (EST) Received: from localhost.localdomain (localhost [127.0.0.1]) by puzzle.pobox.com (Postfix) with ESMTP id EF714139148 for ; Fri, 9 Jul 2004 16:19:38 -0400 (EDT) Received: from hackintosh (CPE-65-30-80-2.kc.rr.com [65.30.80.2]) by puzzle.pobox.com (Postfix) with ESMTP id 0BD51139145 for ; Fri, 9 Jul 2004 16:19:38 -0400 (EDT) From: "Rich Holland" To: Subject: RE: AIX and Largefiles Performance Date: Fri, 9 Jul 2004 16:18:59 -0400 Organization: Guidance Technologies, Inc. MIME-Version: 1.0 Content-type: text/plain; charset=us-ascii Content-Transfer-Encoding: 8bit X-Mailer: Microsoft Office Outlook, Build 11.0.5510 Thread-Index: AcRa9kbUXESmWf/uRo6yiRHB7s4wIAKyeZmg In-Reply-To: <40DC9022.1040101@tenure.com> X-MimeOLE: Produced By Microsoft MimeOLE V6.00.2800.1409 Message-Id: <20040709201938.0BD51139145@puzzle.pobox.com> X-Virus-Scanned: by amavisd-new at freelists.org X-archive-position: 4800 X-ecartis-version: Ecartis v1.0.0 Sender: oracle-l-bounce@freelists.org Errors-To: oracle-l-bounce@freelists.org X-original-sender: holland@guidancetech.com Precedence: normal Reply-To: oracle-l@freelists.org X-list: oracle-l X-Virus-Scanned: by amavisd-new at freelists.org Byron, Any chance you can forward the tuning paper along? I'm interested in reading a copy. I'm currently running half a dozen ERP systems on Oracle 9204, AIX 51-003 and having no problems with the bf option on native JFS. From the tiny bit of reading I've done about JFS2, I'd worry more about performance there than via JFS -- I made the decision early on to stick with JFS, and haven't been disappointed. We've survived multiple power outage crashes, fibre channel disconnections, etc. It's been solid, and performed well -- we're using a pair of FCS adapters to SAN disk on an EMC via FC switches, and to be honest I don't think the client can generate enough workload to really stress the I/O configuration here... :-) Rich -- Rich Holland (913) 645-1950 SAP Technical Consultant print unpack("u","92G5S\=\"!A;F]T:&5R(\'!E -----Original Message----- > From: oracle-l-bounce@freelists.org [mailto:oracle-l-bounce@freelists.org] On > Behalf Of Byron Pearce > Sent: Friday, June 25, 2004 3:51 PM > To: oracle-l@freelists.org > Subject: AIX and Largefiles Performance > > All: > > I am not a native of Planet IBM, but this morning one of the systems > administrators here forwarded over a tuning guide from the mother ship > regarding Oracle on AIX. There were a number of interesting things in > there that I am still working through, but one particular item piqued my > interest. > > "... Do not allocate a JFS with the large file enabled (bf) attribute. > The big file attribute increases the minimum DIO transfer size > (diocapbuf.dio_min returned by the finfo system call) from 4K to 128K, > forcing Oracle to read and write a minimum of 128K byte to exploit DIO. > Please note that 2GB is the largest file size applications (including > Oracle) can use in a JFS without the 'bf' attribute -- large file > enablement. Do not allocate a compressed JFS, which defeats DIO. > Respecting the JFS 2GB maximum file size avoids potential file level > inode locking performance issues in JFS or JFS2 (without CIO), but a > downside of small (2GB) files in a very large database (VLDB) > environment is the time required to open/close files during the > startup/shutdown." > > I have worked on most of the other major O/S variants, and this is the > first time I have ever heard the largefiles settings have a negative > impact on performance. In the interest of scientific investigation on > this topic, I am officially checking my ego at the door and opening my > mind to this possibility. I am trying to do some research on this, but, > thus far, my findings have been sparse. > > Has anyone heard this before? Or is this a performance myth along the > lines of the BCHR's (if so it is a first for me to hear it)? Or is > there a genuine benefit to be gained by doing this? > > Any comments or thoughts on this matter would be appreciated. > > -- > ==================================================================== > Byron Pearce mailto:pearceb@tenure.com > Tenure Systems, Inc. Dallas/Fort Worth, TX > > "It's hard to be a ninja when you wear a beeper." > > > > ---------------------------------------------------------------- > Please see the official ORACLE-L FAQ: http://www.orafaq.com > ---------------------------------------------------------------- > To unsubscribe send email to: oracle-l-request@freelists.org > put 'unsubscribe' in the subject line. > -- > Archives are at http://www.freelists.org/archives/oracle-l/ > FAQ is at http://www.freelists.org/help/fom-serve/cache/1.html > ----------------------------------------------------------------- ---------------------------------------------------------------- Please see the official ORACLE-L FAQ: http://www.orafaq.com ---------------------------------------------------------------- To unsubscribe send email to: oracle-l-request@freelists.org put 'unsubscribe' in the subject line. -- Archives are at http://www.freelists.org/archives/oracle-l/ FAQ is at http://www.freelists.org/help/fom-serve/cache/1.html -----------------------------------------------------------------