From john_re at fastmail.us Sat Nov 1 05:46:06 2008 From: john_re at fastmail.us (john_re) Date: Sat, 01 Nov 2008 06:46:06 -0700 Subject: [svlug] BerkeleyTIP Today Nov 1 New Videos. VOIP Online & in Berkeley Message-ID: <1225547166.15167.1282435057@webmail.messagingengine.com> Join with the global GNU(Linux) BSD FreeSW monthly meeting. - Online from your home or anywhere, by yourself or with friends - Use a microphone & headset for VOIP OR - Come to the local Berkeley Meeting. == Great new VIDEOS - See BTG page for times DebConf - Welcome talk - introductory session DebConf - Debian - 15 years and counting Akedemy - The KDE e.V.: Foundation for the Community SciPy - Scientific computing Python workshop - Intro Linux - Kexec/Kdump: Managing Linux crash dumps DISCUSSION: New: UBUNTU 8.10, Kubuntu w/ KDE4, OPEN OFFICE 3 PROGRAMMING PARTY: VOIP conferencing. == BerkeleyTIP-Global & Local meets Nov 1 Satruday 12N-6PM PDST Monthly Global meeting: FreeSW GNU(Linux) BSD Distros Programming DBs Internet Hardware Applications Join us online from home, or anywhere - IRC Freenode #BerkeleyTIP & using VOIP, Ekiga - (under development) Or come to the physical meeting at UC Berkeley. == See the Google Groups for the latest details: http://groups.google.com/group/BerkTIPGlobal http://groups.google.com/group/BerkTIP ================================================= CONTENTS: 1) MEETING Nov 1 Sat 10A-6P PDST - Berkeley 12N-6P PDST - Global 2) Join ONLINE using IRC & maybe VOIP-Ekiga 3) VIDEOS - Get them here 4) PROGRAMMING PARTY - VOIP conferencing 5) DISCUSSION: New UBUNTU 8.10, OPEN OFFICE 3 6) No Potluck this meeting ================================================= 1) Meeting Nov 1 Sat 10A-6P PDST - Online everywhere, & in Berkeley Berkeley Meeting starts 10AM Global Online meeting starts during 11AM PDST, IRC up by 12N ADJUST FOR YOUR TIME ZONE: Ex: EDST 3P = PDST 12N See the Google Groups for the hourly schedule. -- Free Speech Cafe in Moffitt Undergrad Library -- WATCH WEB PAGE IN CASE OF LOCATION CHANGE -- See BerkTIP Google Group web page for locations details. PARKING: Football game - Might need to park 5-10 blocks away for free street parking. Garages might charge $5-20. BART, CALTRAIN, BUS, CARPOOL - GOOD ALTERNATIVES TO DRIVING. 2) Join Online using IRC & maybe VOIP-Ekiga Join us online IRC Freenode #BerkeleyTIP & using VOIP, Ekiga - (under development) VOIP-Ekiga - Get your own VOIP client working Ekiga - do a loopback test with mic & headset http://groups.google.com/group/BerkTIPGlobal/web/irc-voip 3) Videos - Get them here http://groups.google.com/group/BerkTIPGlobal/web/videos 4) Programming Party - VOIP conferencing Work on getting VOIP conferencing going Get your own VOIP client working http://wiki.ekiga.org/index.php/Fun_Numbers Work on gettin a conference call going. 5) DISCUSSION: New UBUNTU 8.10, OPEN OFFICE 3 6) No Potluck this meeting See you online, or in Berkeley :) From oldestgeek at gmail.com Sat Nov 1 09:08:18 2008 From: oldestgeek at gmail.com (Paul Cubbage) Date: Sat, 1 Nov 2008 10:08:18 -0700 Subject: [svlug] Wonderfest - Cool free event today and tomorrow Message-ID: Today at Stanford. Tomorrow at UC Berkeley. Bay Area Festival of Science http://www.wonderfest.org/wp/ Paul Cubbage From rick at linuxmafia.com Sun Nov 2 15:00:48 2008 From: rick at linuxmafia.com (Rick Moen) Date: Sun, 2 Nov 2008 15:00:48 -0800 Subject: [svlug] [svlug-announce] SVLUG Nov. 4 meeting: Question and Answer w/ Dr. James Gosling Message-ID: <20081102230048.GA17154@linuxmafia.com> The Silicon Valley Linux Users Group (SVLUG) meets this Wednesday: WHEN: Wednesday, November 5, 2008 7pm-9pm MAIN PRESENTATION: TOPIC: Question and Answer PRESENTED BY: James Gosling, VP and Sun Fellow, Sun Microsystems, Inc. TOPIC SUMMARY: Everything you wanted to know, but were afraid to ask. Dr. Gosling has been one of the best-known software developers of the current age. Come to get his views on multiprocessor support and virtual machines, in which areas he was a pioneer, the Java programming language, which he invented, and more. ABOUT THE PRESENTER: James Gosling received a BSc in Computer Science from the University of Calgary, Canada in 1977. He received a Ph.D. in Computer Science from Carnegie-Mellon University in 1983. The title of his thesis was "The Algebraic Manipulation of Constraints". He has built satellite data acquisition systems, a multiprocessor version of Unix, several compilers, mail systems, and window managers. He has also built a WYSIWYG text editor, a constraint based drawing editor and a port of the 'Emacs' text editor to Unix systems. At Sun, his early activity was as lead engineer of the NeWS window system. He did the original design of the Java programming language and implemented its original compiler and virtual machine. LOCATION: Symantec (formerly Veritas Software Corp.) VCAFE Facility 350 Ellis Street (near E. Middlefield Road) Mountain View, CA 94043 Directions on how to get there are listed at: http://www.svlug.org/directions/veritas.php We've tried our very best for these directions to be accurate. If you have any improvements to make, please let SVLUG's volunteers know! webmaster at svlug.org POST-MEETING GATHERING: If you just can't get enough, a smaller group usually goes to a local restaurant/diner after the meeting. We'll announce the restaurant selection at the meeting. We look forward to seeing you there! _______________________________________________ svlug-announce mailing list svlug-announce at lists.svlug.org http://lists.svlug.org/lists/listinfo/svlug-announce From james at mortoncpa.com Mon Nov 3 14:16:23 2008 From: james at mortoncpa.com (james@mortoncpa.com) Date: Mon, 3 Nov 2008 22:16:23 +0000 Subject: [svlug] Free Equipment Available - In Palo Alto Message-ID: <20081103221624.46083.qmail@mortoncpa.com> Hi, My name is James Leone and I work for a company that has just relocated to some smaller office space and we have some extra equipment that, if anyone needs or wants, can have for gratis. None of the items are new - some are in fact quite old - but who knows might have some historical value. Some might be useful - but not the latest and greatest. I'll be the first to confess that I'm underinformed as to what might be significant - for a handful of items. So I apologize if none of the items I list are not of much value and have no historical signifigance. I'll start with items that predate my knowledge - may have some value I don't understand - hopefully its not comedic. a.) 1 Zenith Laptop (I have not used this and have to look it over before I give away to make sure it has no data on it) b.) 1 HP LXE Pro 200 - I believe it has a Pentium Pro Processor and it has several slots for hard disks which are in it. It is on rollers, hums when powered on and I'd say about the size of a small lamptable. Has Mandrake Linux 6 or 7 on it. c.) 1 Compaq Proliant 2000, Series 3110 - 486/50 - This server stands about waist high. It has two disks and I think has an old version of Windows NT 3. This list includes things that may have some use - but is not the latest technology. Not used in a while - I can't say it works - but no reason to believe they are: a.) 1 Plextor CD Rom Tower x 4. b.) 1 3-om Office Connect 16 port Modem c.) 1 HP Laserjet III Printer This list includes older monitors that I'm comfortable can be used. None of the monitors are flat screen: 12" or 13" diagonal measurement of screen: a.) 3 HP 1024 Low Emissions Color Monitors 15" diagonal measurement of screen: a.) 1 Dell M781P Color Monitor b.) 1 Dell D1025TM Trinitron Color Monitor c.) 1 Viewsonic G771 Color Monitor Items I will replace - but were useable last I checked, albeit having dead batteries: a.) 2 Belkin Power Supplies Other Accessories: a.) 5 3 1/2" floppy disk containers of various size b.) Assorted Keyboards in various conditions of appearance - but functional as far as I know. (But some have a particular connection type that should be given with the older machines - if anyone wants them.) c.) 3 13" to 15" Glare Guards Computer of Various Possiblities of Interest - that I will have to format the hard disks before giving away: a.) 1 Vanilla PIII computer that I believe needs a new power supply. The computer will hum audibly when turned on, but if powered off and on - it will act normally. b.) IBM PC - probably a 8086 - grey in color with big red switch on side. It has a 5 1/2 inch floppy drive. Let me know if you are interested. James Leone From einfeldt at gmail.com Mon Nov 3 15:08:13 2008 From: einfeldt at gmail.com (Christian Einfeldt) Date: Mon, 3 Nov 2008 15:08:13 -0800 Subject: [svlug] Free Equipment Available - In Palo Alto In-Reply-To: <20081103221624.46083.qmail@mortoncpa.com> References: <20081103221624.46083.qmail@mortoncpa.com> Message-ID: <4b5781040811031508p6ee0c2f5ofa191a8d5cda66c2@mail.gmail.com> If no one goes for this stuff, it would be great if you could give it to ACCRC.org. They have given away 16,000 Linux computers to non-profits and needy individuals. Your donation to them will help them in that mission. -------------- next part -------------- An HTML attachment was scrubbed... URL: http://lists.svlug.org/archives/svlug/attachments/20081103/ab95cc46/attachment.htm From nbs at sonic.net Tue Nov 4 23:57:43 2008 From: nbs at sonic.net (nbs) Date: Tue, 4 Nov 2008 23:57:43 -0800 Subject: [svlug] Linux Installfest workshop in Davis - Saturday, November 8th Message-ID: <200811050757.mA57vhMI012052@bolt.sonic.net> The Linux Users' Group of Davis will be holding a free "Linux Installfest" workshop in Davis, California. When: Saturday, November 8th, 2008 11:00am - 6:00pm Where: Zinfandel Lounge Experimental College 260 South Silo UC Davis Davis, California Maps/directions at: http://www.lugod.org/if/directions.php Special Topic this Month: Plone Content Management System Members of the the Plone Users Group of Davis (www.zugod.org) will be running a workshop at the installfest to teach people about Plone server software for managing collaborative web content. The community is welcome to come ask questions, see demonstrations and get help installing and customizing Plone. What: Linux is a completely free, 'open source' operating system that can run on a wide variety of computer hardware. It can act as a web-, file-, print- or game-server, run in a 'cluster' of computers to do 3D rendering or other intense math, or sit under your TV and record your favorite shows for later viewing. Many people use it as an inexpensive, stable, virus- and spyware-free alternative to commercial software, such as Microsoft Windows. It can be installed over, or alongside, Windows or Mac OS X. Members of the community are invited to bring their computers and laptops to this informal workshop, and volunteers from LUGOD will help you install and configure Linux... for FREE! How: If you wish to bring in your PC, you must RSVP beforehand to reserve a space. The RSVP form, and lots of useful information about Linux and Installfests, and how to prepare for the event, are accessible on the web at: http://www.lugod.org/if/ Help! We're always looking for volunteers for our workshops. If you'd like to help, or come watch and learn as others are helped, please feel free to drop by. Our "vox-if" mailing list is where we discuss plans and needs for these events, so we encourage you to sign up: http://www.lugod.org/mailinglists/#vox-if ...or contact the Installfest coordinators directly via email: if at lugod.org About LUGOD: The Linux Users' Group of Davis is a 501(c)7 non-profit organization dedicated to the Linux operating; system and Open Source software, and which (along with holding Installfests) holds regular meetings with guest speakers each month in Davis, CA. For details, visit: http://www.lugod.org/ -- Bill Kendrick LUGOD Public Relations Officer pr at lugod.org http://www.lugod.org/ (Your address: svlug at lists.svlug.org ) From nbs at sonic.net Wed Nov 5 00:01:19 2008 From: nbs at sonic.net (nbs) Date: Wed, 5 Nov 2008 00:01:19 -0800 Subject: [svlug] Linux Users' Group of Davis, Nov. 17th meeting: "Open Source Business Model" Message-ID: <200811050801.mA581JQv012678@bolt.sonic.net> The Linux Users' Group of Davis (LUGOD) will be holding its next meeting on: Monday November 17, 2008 7:00pm - 9:00pm The meeting will be held at our regular location: Davis Public Library Blanchard Meeting Room 315 East 14th St. Davis, California 95616 Presentation: Open Source Business Model Presented by: Cathy Malmrose, CEO, ZaRason ZaReason is a company in that builds and sells Linux-based computers. The company was founded based on experience the founders had working with a local computer recycling center, building Linux systems for the community. At this talk Cathy Malmrose, CEO of zaReason and "Linux hardware chick", dives into the many dynamic facets of living in LinuxLand. Explore the tension between business and volunteerism, focusing on the best aspects of both. Take a peek at hardware creation, copyright law, software support, our favorite subject, software as a service. Influenced by the Open Source Business Resource and various luminaries in the IP field, Cathy ends on the evolution of the open source mindset, discussing her upcoming book: 'The Open Source Child' which explores the next gen's approach to product creation. Give-aways: Cathy will be bringing some give-aways, and holding a pop-quiz. For details on this meeting, visit: http://www.lugod.org/meeting/ For maps, directions, public transportation schedules, etc., visit: http://www.lugod.org/meeting/library/ About LUGOD: The Linux Users' Group of Davis is a 501(c)7 non-profit organization dedicated to the Linux computer operating system and other Open Source and Free Software. LUGOD holds regular meetings with guest speakers in Davis, California, as well as other events in Davis and the greater Sacramento region. Events are always free and open to the public. Please visit our website for more details: http://www.lugod.org/ -- Bill Kendrick pr at lugod.org Public Relations Officer Linux Users' Group of Davis http://www.lugod.org/ (Your address: svlug at lists.svlug.org ) From romain at kzsu.stanford.edu Wed Nov 5 18:10:46 2008 From: romain at kzsu.stanford.edu (Romain Kang) Date: Wed, 5 Nov 2008 18:10:46 -0800 Subject: [svlug] Free Equipment Available - In Palo Alto In-Reply-To: <4b5781040811031508p6ee0c2f5ofa191a8d5cda66c2@mail.gmail.com> References: <20081103221624.46083.qmail@mortoncpa.com> <4b5781040811031508p6ee0c2f5ofa191a8d5cda66c2@mail.gmail.com> Message-ID: <20081106021046.GA56720@kzsu.stanford.edu> Folks who have stuff that might be of historical interest could check out the Computer History Museum, especially their wish list. See: http://www.computerhistory.org/collections/donateArtifact/faq.php Romain Kang Disclaimer: I speak for myself alone, romain at kzsu.stanford.edu except when indicated otherwise. From einfeldt at gmail.com Thu Nov 6 17:37:51 2008 From: einfeldt at gmail.com (Christian Einfeldt) Date: Thu, 6 Nov 2008 17:37:51 -0800 Subject: [svlug] Save the date! Sat Dec 6 tech fair at St. Anthony Message-ID: <4b5781040811061737g403eafdcp106261f00772410e@mail.gmail.com> hi, Please save Saturday, Dec 6 from 9 am to 2 pm for the St. Anthony tech fair, to be held at 150 Golden Gate at Jones in SF. This event will be pretty much the same as past events held at St. Anthony. We will be demonstrating Linux and installing Linux on hardware brought in by low income clients and we will be giving away distros. This is a great opportunity to embed ourselves and Linux in one of the most well-run, well-financed charities in Northern California, and reach out to low income individuals and non-profits. I will have more info and a link to a web page soon. c u -------------- next part -------------- An HTML attachment was scrubbed... URL: http://lists.svlug.org/archives/svlug/attachments/20081106/1a0eb738/attachment-0001.htm From bill at wards.net Thu Nov 6 12:00:12 2008 From: bill at wards.net (bill@wards.net) Date: Thu, 06 Nov 2008 12:00:12 -0800 Subject: [svlug] NEXT WEEK: PenLUG meeting 11/13/2008 Message-ID: PENINSULA LINUX USERS' GROUP (PenLUG) PRESENTS: +---------------------------------------------------------------+ |Date: |Thursday, November 13th, 2008 | |---------+-----------------------------------------------------| |Time: |meeting 7:00 - 9:00 PM, social/networking until 10 PM| |---------+-----------------------------------------------------| | |Bayshore Technology Park | |Location:|1300 Island Drive | | |Redwood City, CA 94065 | | |Suite 106 - Training Room | +---------------------------------------------------------------+ Dan Dennedy, Kino - Non-linear DV video editor Dan Dennedy, the lead developer, will talk about the history of Kino and give a tour of its features. He will demonstrate capturing, importing, editing, exporting, and publishing to a web service. He will also talk a little about his newer projects to resume where Kino leaves off. Dan Dennedy, lead developer, Kino Dan Dennedy has been an open source and Linux developer for around 7 years. He has worked on IEEE 1394 (FireWire) in the Linux kernel and various 1394- and DV-related libraries and applications. He moved to the San Francisco Bay Area around 3 years ago with his wife and 3 kids. He works full time for MobiTV for the past 4 years as lead software engineer in the areas of digital media and streaming technology. RSVP Although it is not required, we like to have an idea of how many people to expect, so if possible please email rsvp at penlug.org if you are planning to attend. GETTING THERE For information on getting to the meeting, please see: http://maps.google.com/maps?q=1300+Island+Drive,+Redwood+City,+CA http://www.penlug.org/twiki/bin/view/Home/DrivingDirectionsQualys http://www.penlug.org/twiki/bin/view/Home/TransitDirectionsQualys Traffic on 101 can be pretty bad in the evening, so we encourage you to check traffic conditions before driving by dialing 5-1-1 on your phone or visiting www.511.org, and if possible to take public transit (best bet: bicycle via Caltrain) or carpool to this meeting. MORE INFORMATION See www.penlug.org for more information. This notice is being sent to the following mailing lists: members at penlug.org announce at penlug.org sf-lug at linuxmafia.com balug-talk at lists.balug.org svlug at lists.svlug.org svevents at yahoogroups.com vox at lists.lugod.org Please reply to suggest any additions or other changes. From walterv at gbbservices.com Mon Nov 10 09:47:03 2008 From: walterv at gbbservices.com (Walter Vannini) Date: Mon, 10 Nov 2008 09:47:03 -0800 Subject: [svlug] ACCU meeting on Wednesday 'iPhone Application Development' Message-ID: <49187397.5040000@gbbservices.com> When: Wednesday, November 12, 2008 Topic: iPhone Application Development Speaker: Walter Vannini Time: 6:30pm doors open 7:00pm meeting begins Where: Symantec VCAFE building 350 Ellis Street (near E. Middlefield Road) Mountain View, CA 94043 Map: Directions: VCAFE is accessible from the semicircular courtyard between Symantec buildings Cost: Free More Info: This talk will be an introduction to developing native applications for the iPhone by using Apple's recently released SDK. No knowledge of Objective-C, the Cocoa frameworks, or Apple's development environment will be assumed. Walter Vannini is a programmer based in Silicon Valley. As an application developer, he's seen his targeted machines get smaller over the years: starting with VAX and HP and SGI workstations, then desktop systems and laptops running various versions of Windows NT, and now iPhones and iPod Touches. For the last six months Walter has been immersed in iPhone development. His website is at http://www.gbbservices.com Meetings are open to the public and are free of charge. ---- The ACCU meets monthly. Meetings are always open to the public and are free of charge. To suggest topics and speakers please email Walter Vannini via walterv at gbbservices.com From walterv at gbbservices.com Wed Nov 12 11:27:28 2008 From: walterv at gbbservices.com (Walter Vannini) Date: Wed, 12 Nov 2008 11:27:28 -0800 Subject: [svlug] ACCU meeting tonight 'iPhone Application Development' Message-ID: <491B2E20.1010801@gbbservices.com> When: Wednesday, November 12, 2008 Topic: iPhone Application Development Speaker: Walter Vannini Time: 6:30pm doors open 7:00pm meeting begins Where: Symantec VCAFE building 350 Ellis Street (near E. Middlefield Road) Mountain View, CA 94043 Map: Directions: VCAFE is accessible from the semicircular courtyard between Symantec buildings Cost: Free More Info: This talk will be an introduction to developing native applications for the iPhone by using Apple's recently released SDK. No knowledge of Objective-C, the Cocoa frameworks, or Apple's development environment will be assumed. Walter Vannini is a programmer based in Silicon Valley. As an application developer, he's seen his targeted machines get smaller over the years: starting with VAX and HP and SGI workstations, then desktop systems and laptops running various versions of Windows NT, and now iPhones and iPod Touches. For the last six months Walter has been immersed in iPhone development. His website is at http://www.gbbservices.com Meetings are open to the public and are free of charge. ---- The ACCU meets monthly. Meetings are always open to the public and are free of charge. To suggest topics and speakers please email Walter Vannini via walterv at gbbservices.com From AFife at untangle.com Wed Nov 12 13:41:36 2008 From: AFife at untangle.com (Andrew Fife) Date: Wed, 12 Nov 2008 13:41:36 -0800 Subject: [svlug] OpenMoko CEO @ BALUG (Tuesday Nov 18) Message-ID: <9347C2AED406124287DD2A007EB15B130183949C8D@EXCHANGE.Untangle.local> Hi Folks: Sean Moss-Pultz will be speaking at BALUG on Tuesday (11/18). Sean is the CEO and founder of The OpenMoko Project, which makes a open source linux based cell phone called the NEO FreeRunner. Sean is a self described San Diego surfer that landed in Taiwan as a software developer working on embedded systems, which ultimately led to the founding of OpenMoko. More info on Sean, OpenMoko and the Neo FreeRunner at these links: http://moss-pultz.com/ http://www.openmoko.com/ http://www.openmoko.com/product.html We're expecting a great crowd, so if you'd like to join us please RSVP: RSVP at balug.org **Why RSVP??** Well, don't worry we won't turn you away, but the RSVPs really help the Four Seas Restaurant plan the meal and they ensure that we're able to eat upstairs in the private banquet room. Bottomline: if we don't have 40 RSVPs the day before the meeting, we don't get the banquet room. Meeting Details... 6:30pm November 18th, 2008 (Next Tuesday) Four Seas Restaurant 731 Grant Ave. San Francisco, CA 94108 Easy $5 PARKING: Portsmouth Square Garage at 733 Kearny Cost: The meetings are always free, but dinner is $13 -- Andrew Fife Untangle - The Open Source Network Gateway www.untangle.com/download 650.425.3327 desk 415.806.6028 cell -------------- next part -------------- An HTML attachment was scrubbed... URL: http://lists.svlug.org/archives/svlug/attachments/20081112/b850fd05/attachment.htm From jim at well.com Wed Nov 12 17:28:36 2008 From: jim at well.com (jim) Date: Wed, 12 Nov 2008 17:28:36 -0800 Subject: [svlug] BayPIGgies meeting Thursday November 13, 2008: NumPy and SciPy Message-ID: <1226539716.6358.792.camel@ubuntu> BayPIGgies meeting Thursday November 13, 2008: NumPy and SciPy By Fernando Perez NumPy and SciPy: how python is becoming the dominant high-level language for scientific work, who's using it, what we do with it, what the key projects are, where we're headed, what challenges we have, and where others can contribute. Tonight's Newbie Nugget is... DocTest by Wesley Chun Documentation and testing are traditionally two areas that programmers shy away from. With one simple change, namely doctest integration, you can get both in a simplistic manner and improve the usability of your code as well as get free testing. That is what makes doctest so cool! Location: Google Campus Building 40, the Seville room (check in at the lobby in bldg 43) bayPIGgies meeting information: http://baypiggies.net/new/plone * Please sign up in advance to have your google access badge ready: http://wiki.python.org/moin/BayPiggiesGoogleMeetings (no later than close of business on Wednesday.) Agenda ..... 7:30 PM ........................... General hubbub, inventory end-of-meeting announcements, any first-minute announcements. ..... 7:35 PM to 7:45 PM ................ Newbie Nugget: Doctest by Wesley Chun ..... 7:45 PM to 8:45 PM ................ Talk by Whom ..... 8:45 PM to 9:00 PM -- After The Talk ................ Mapping and Random Access Mapping is a rapid-fire audience announcement of topics the announcers are interested in. Random Access follows immediately to allow follow up individually on the announcements and other topics of interest. From sarah at cloudmade.com Thu Nov 13 11:46:07 2008 From: sarah at cloudmade.com (Sarah Manley) Date: Thu, 13 Nov 2008 11:46:07 -0800 Subject: [svlug] Free open source mapping event in Palo Alto Message-ID: Dear All, Are you interested in learning about open source geodata? If so, you should attend the upcoming South Bay mapping party on November 22nd and 23rd. The event details can be found here: http://wiki.openstreetmap.org/index.php/South_bay You can RSVP here: http://upcoming.yahoo.com/event/1327805 I am hosting this event for OpenStreetMap. OpenStreetMap is an open source project to collaboratively map the entire world. There are over 70,000 participants in OpenStreetMap worldwide, and we are working to increase participation here in the US. Mapping parties are events where anyone can come and participate in the OpenStreetMap project. They are social events where experienced and new mappers can meet to share and learn more about the project as well as get out and map. Please contact me with any questions you may have, as well as suggestions for locations for future mapping events. Cheers, Sarah Manley Bay Area Community Ambassador for OpenStreetMap Sarah at cloudmade.com From AFife at untangle.com Mon Nov 17 16:10:11 2008 From: AFife at untangle.com (Andrew Fife) Date: Mon, 17 Nov 2008 16:10:11 -0800 Subject: [svlug] OpenMoko @ BALUG Tomorrow Message-ID: <9347C2AED406124287DD2A007EB15B13018394A23A@EXCHANGE.Untangle.local> Hi Folks: Sean Moss-Pultz is speaking at BALUG on tomorrow about the OpenMoko Project, which he is founder & CEO of. The OpenMoko Project makes an open source, linux-based cell phone called the NEO FreeRunner. More info on Sean, OpenMoko and the Neo FreeRunner at these links: http://moss-pultz.com/ http://www.openmoko.com/ http://www.openmoko.com/product.html We're expecting a great crowd, so if you'd like to join us please RSVP: RSVP at balug.org **Why RSVP??** Well, don't worry we won't turn you away, but the RSVPs really help the Four Seas Restaurant plan the meal and they ensure that we're able to eat upstairs in the private banquet room. Bottomline: if we don't have 40 RSVPs the day before the meeting, we don't get the banquet room. Meeting Details... 6:30pm November 18th, 2008 (Tomorrow) Four Seas Restaurant 731 Grant Ave. San Francisco, CA 94108 Easy $5 PARKING: Portsmouth Square Garage at 733 Kearny Cost: The meetings are always free, but dinner is $13 -- Andrew Fife Untangle - The Open Source Network Gateway www.untangle.com/download 650.425.3327 desk 415.806.6028 cell -------------- next part -------------- An HTML attachment was scrubbed... URL: http://lists.svlug.org/archives/svlug/attachments/20081117/589d3f3f/attachment-0001.htm From mohammad.afroze at gmail.com Mon Nov 17 19:56:26 2008 From: mohammad.afroze at gmail.com (Afroze Mohammad) Date: Mon, 17 Nov 2008 19:56:26 -0800 Subject: [svlug] RPC timed out error when I use rsh Message-ID: <49223CEA.9020001@gmail.com> Hi, When I try to rsh to our execution server with the command below, it throws out RPC timed out error. Can someone please help on how to get rid of this. We use NIS and have hosts: files dns rsh -l afroze hostname do_ypcall: clnt_call: RPC: Timed out -- Cheers Afroze From john_re at fastmail.us Tue Nov 18 20:03:11 2008 From: john_re at fastmail.us (john_re) Date: Tue, 18 Nov 2008 20:03:11 -0800 Subject: [svlug] Tonight- Join VOIP BALUG conference online, OpenMoko Neo Phone topic Message-ID: <1227067391.15769.1285561755@webmail.messagingengine.com> You might want to listen in live to the BALUG.org meeting for the next hour or so. We are working now on the first ever VOIP conference for a BALUG meeting. You can listen in two ways: 1) using a VOIP phone like Ekiga from your computer over the internet, 2) by dialing into the voip server with a regular phone there is no cost for the server. If you have free calling to 605 area code you can connect for free also. Join us on IRC freenode.net #BALUG for the full details. The speaker will likely talk for about 1 hour, beginning about 8PM or 815PM. This is a "first time, best efforts, experimetal basis, #1 learn something, #2 maybe it will work" basis effort. Hope you can join us in this next hour or so. From drew at drewb.com Thu Nov 20 23:38:02 2008 From: drew at drewb.com (Drew Bertola) Date: Thu, 20 Nov 2008 23:38:02 -0800 Subject: [svlug] Noob needs help connecting via pptp (pppd) to Cisco PIX... Message-ID: <4926655A.50601@drewb.com> Hey all, I'm trying to get a box back east to work w/ a cisco PIX running pptp. The connection seems to be established, but no packets seem to make it to the network inside the PIX. Once connected I add a route: route add -net 10.10.20.0 netmask 255.255.255.0 dev ppp0 I've also tried the same, but adding "gw 10.10.19.140" which is the connection IP of ppp0 issued by the PIX. I can ssh to 10.10.19.140, but I can't do anything on the 10.10.20.0 network. Anyone have any ideas? thanks, -- Drew Bertola ------------------------------------------------- * PHP/LAMP Consultant, ZCE-1000 * * * * Tel: 408-966-6671 * * * * current resume: * * http://drewb.com/blog/about/resume/ * ------------------------------------------------- From james at linuxrebel.org Fri Nov 21 12:08:58 2008 From: james at linuxrebel.org (James Sparenberg) Date: Fri, 21 Nov 2008 12:08:58 -0800 Subject: [svlug] Noob needs help connecting via pptp (pppd) to Cisco PIX... In-Reply-To: <4926655A.50601@drewb.com> References: <4926655A.50601@drewb.com> Message-ID: <200811211208.58931.james@linuxrebel.org> On Thursday 20 November 2008 23:38:02 Drew Bertola wrote: > Hey all, > > I'm trying to get a box back east to work w/ a cisco PIX running pptp. > The connection seems to be established, but no packets seem to make it > to the network inside the PIX. > > Once connected I add a route: > > route add -net 10.10.20.0 netmask 255.255.255.0 dev ppp0 > > I've also tried the same, but adding "gw 10.10.19.140" which is the > connection IP of ppp0 issued by the PIX. > > I can ssh to 10.10.19.140, but I can't do anything on the 10.10.20.0 > network. > > Anyone have any ideas? > > thanks, Not a PIX person myself but having setup a pptp vpn through my routers/firewalls to my house. Do you have the rules setup to allow GRE James From drew at drewb.com Fri Nov 21 19:08:04 2008 From: drew at drewb.com (Drew Bertola) Date: Fri, 21 Nov 2008 19:08:04 -0800 Subject: [svlug] Noob needs help connecting via pptp (pppd) to Cisco PIX... In-Reply-To: <200811211208.58931.james@linuxrebel.org> References: <4926655A.50601@drewb.com> <200811211208.58931.james@linuxrebel.org> Message-ID: <49277794.3010500@drewb.com> James Sparenberg wrote: > On Thursday 20 November 2008 23:38:02 Drew Bertola wrote: > >> Hey all, >> >> I'm trying to get a box back east to work w/ a cisco PIX running pptp. >> The connection seems to be established, but no packets seem to make it >> to the network inside the PIX. >> >> Once connected I add a route: >> >> route add -net 10.10.20.0 netmask 255.255.255.0 dev ppp0 >> >> I've also tried the same, but adding "gw 10.10.19.140" which is the >> connection IP of ppp0 issued by the PIX. >> >> I can ssh to 10.10.19.140, but I can't do anything on the 10.10.20.0 >> network. > > Not a PIX person myself but having setup a pptp vpn through my > routers/firewalls to my house. Do you have the rules setup to allow GRE > Thanks, James. I did try flushing all iptables rules and had the same problem. Once I have it working w/o the firewall variable, I'll make sure to look at that. -- Drew Bertola ------------------------------------------------- * PHP/LAMP Consultant, ZCE-1000 * * * * Tel: 408-966-6671 * * * * current resume: * * http://drewb.com/blog/about/resume/ * ------------------------------------------------- From james at linuxrebel.org Mon Nov 24 17:32:40 2008 From: james at linuxrebel.org (James Sparenberg) Date: Mon, 24 Nov 2008 17:32:40 -0800 Subject: [svlug] Sync command slowdown Message-ID: <200811241732.40649.james@linuxrebel.org> All, I've got 4 new DB servers exhibiting a rather problematic set of actions with disk sync. Running the sync command (as is done when logs are rotated.) takes 4-6 seconds (not micro, full seconds) Other than this these boxes are running 10x faster than the servers they replaced and making our customers much happier. Further it's only happening on my 4 new servers not on the older systems (I have them in QA, etc). CPU load is at best 15% per core, normally down in the dirt, as the primary CPU function is to manage disk I/0 (most queries are the same so they are cached, and writes are limited.) RAM is fine and the box never touches it's swap partition. I'm personally suspecting that it is Kernel module vs LSI raid controller, on these boxes. Additionally our DBA has discovered that it seems to work faster on smaller partitions than the larger one we have for our DBs. The only tweak I'm currently using beyond defaults in fstab are noatime (RHEL 5 mount command doesn't recognize relatime) however changing out the noatime has no affect on the sync command. Has anyone out there seen anything at all similar? I'm hoping to avoid having to go to software raid only because it will mean a fairly extensive rebuild of the system, and from now until mid February I'm not going to be able to take that time. For now the mid level solution is going with smaller binlogs and rotating a lot more often. Details on the systems. Manf Penguin Computing (And yes they are providing some assistance on this, they really are proactive.) OS: CentOS 5.2 x86_64 (fully updated) running 2.6.18-92 CPU: Single Quad Core Intel Xeon RAM: 16GB HDD LSI Raid 1 controller lspci: 00:00.0 Host bridge: Intel Corporation 5000P Chipset Memory Controller Hub (rev b1) 00:02.0 PCI bridge: Intel Corporation 5000 Series Chipset PCI Express x8 Port 2-3 (rev b1) 00:03.0 PCI bridge: Intel Corporation 5000 Series Chipset PCI Express x4 Port 3 (rev b1) 00:04.0 PCI bridge: Intel Corporation 5000 Series Chipset PCI Express x8 Port 4-5 (rev b1) 00:05.0 PCI bridge: Intel Corporation 5000 Series Chipset PCI Express x4 Port 5 (rev b1) 00:06.0 PCI bridge: Intel Corporation 5000 Series Chipset PCI Express x8 Port 6-7 (rev b1) 00:07.0 PCI bridge: Intel Corporation 5000 Series Chipset PCI Express x4 Port 7 (rev b1) 00:08.0 System peripheral: Intel Corporation 5000 Series Chipset DMA Engine (rev b1) 00:10.0 Host bridge: Intel Corporation 5000 Series Chipset FSB Registers (rev b1) 00:10.1 Host bridge: Intel Corporation 5000 Series Chipset FSB Registers (rev b1) 00:10.2 Host bridge: Intel Corporation 5000 Series Chipset FSB Registers (rev b1) 00:11.0 Host bridge: Intel Corporation 5000 Series Chipset Reserved Registers (rev b1) 00:13.0 Host bridge: Intel Corporation 5000 Series Chipset Reserved Registers (rev b1) 00:15.0 Host bridge: Intel Corporation 5000 Series Chipset FBD Registers (rev b1) 00:16.0 Host bridge: Intel Corporation 5000 Series Chipset FBD Registers (rev b1) 00:1c.0 PCI bridge: Intel Corporation 631xESB/632xESB/3100 Chipset PCI Express Root Port 1 (rev 09) 00:1d.0 USB Controller: Intel Corporation 631xESB/632xESB/3100 Chipset UHCI USB Controller #1 (rev 09) 00:1d.1 USB Controller: Intel Corporation 631xESB/632xESB/3100 Chipset UHCI USB Controller #2 (rev 09) 00:1d.2 USB Controller: Intel Corporation 631xESB/632xESB/3100 Chipset UHCI USB Controller #3 (rev 09) 00:1d.3 USB Controller: Intel Corporation 631xESB/632xESB/3100 Chipset UHCI USB Controller #4 (rev 09) 00:1d.7 USB Controller: Intel Corporation 631xESB/632xESB/3100 Chipset EHCI USB2 Controller (rev 09) 00:1e.0 PCI bridge: Intel Corporation 82801 PCI Bridge (rev d9) 00:1f.0 ISA bridge: Intel Corporation 631xESB/632xESB/3100 Chipset LPC Interface Controller (rev 09) 00:1f.1 IDE interface: Intel Corporation 631xESB/632xESB IDE Controller (rev 09) 00:1f.2 IDE interface: Intel Corporation 631xESB/632xESB/3100 Chipset SATA IDE Controller (rev 09) 00:1f.3 SMBus: Intel Corporation 631xESB/632xESB/3100 Chipset SMBus Controller (rev 09) 01:00.0 PCI bridge: Intel Corporation 6311ESB/6321ESB PCI Express Upstream Port (rev 01) 01:00.3 PCI bridge: Intel Corporation 6311ESB/6321ESB PCI Express to PCI-X Bridge (rev 01) 02:00.0 PCI bridge: Intel Corporation 6311ESB/6321ESB PCI Express Downstream Port E1 (rev 01) 02:02.0 PCI bridge: Intel Corporation 6311ESB/6321ESB PCI Express Downstream Port E3 (rev 01) 03:00.0 SCSI storage controller: LSI Logic / Symbios Logic SAS1064ET PCI-Express Fusion-MPT SAS (rev 04) 04:00.0 Ethernet controller: Intel Corporation 80003ES2LAN Gigabit Ethernet Controller (Copper) (rev 01) 04:00.1 Ethernet controller: Intel Corporation 80003ES2LAN Gigabit Ethernet Controller (Copper) (rev 01) 07:00.0 Ethernet controller: Intel Corporation 82571EB Gigabit Ethernet Controller (rev 06) 07:00.1 Ethernet controller: Intel Corporation 82571EB Gigabit Ethernet Controller (rev 06) 0b:00.0 VGA compatible controller: Matrox Graphics, Inc. MGA G200e [Pilot] ServerEngines (SEP1) (rev 02) Stack Trace execve("/bin/sync", ["sync"], [/* 16 vars */]) = 0 brk(0) = 0x1ce91000 mmap(NULL, 4096, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_ANONYMOUS, -1, 0) = 0x2aaaaaaab000 uname({sys="Linux", node="my.server.name", ...}) = 0 access("/etc/ld.so.preload", R_OK) = -1 ENOENT (No such file or directory) open("/etc/ld.so.cache", O_RDONLY) = 3 fstat(3, {st_mode=S_IFREG|0644, st_size=23796, ...}) = 0 mmap(NULL, 23796, PROT_READ, MAP_PRIVATE, 3, 0) = 0x2aaaaaaac000 close(3) = 0 open("/lib64/libc.so.6", O_RDONLY) = 3 read(3, "\177ELF\2\1\1\0\0\0\0\0\0\0\0\0\3\0>\0\1\0\0\0\260\331! \2773\0\0\0"..., 832) = 832 fstat(3, {st_mode=S_IFREG|0755, st_size=1699880, ...}) = 0 mmap(NULL, 4096, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_ANONYMOUS, -1, 0) = 0x2aaaaaab2000 mmap(0x33bf200000, 3481848, PROT_READ|PROT_EXEC, MAP_PRIVATE| MAP_DENYWRITE, 3, 0) = 0x33bf200000 mprotect(0x33bf34a000, 2093056, PROT_NONE) = 0 mmap(0x33bf549000, 20480, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_FIXED| MAP_DENYWRITE, 3, 0x149000) = 0x33bf549000 mmap(0x33bf54e000, 16632, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_FIXED| MAP_ANONYMOUS, -1, 0) = 0x33bf54e000 close(3) = 0 mmap(NULL, 4096, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_ANONYMOUS, -1, 0) = 0x2aaaaaab3000 arch_prctl(ARCH_SET_FS, 0x2aaaaaab3250) = 0 mprotect(0x33bf549000, 16384, PROT_READ) = 0 mprotect(0x33bf01a000, 4096, PROT_READ) = 0 munmap(0x2aaaaaaac000, 23796) = 0 brk(0) = 0x1ce91000 brk(0x1ceb2000) = 0x1ceb2000 open("/usr/lib/locale/locale-archive", O_RDONLY) = 3 fstat(3, {st_mode=S_IFREG|0644, st_size=56470832, ...}) = 0 mmap(NULL, 56470832, PROT_READ, MAP_PRIVATE, 3, 0) = 0x2aaaaaab4000 close(3) = 0 sync() = 0 ###NOTE: This is where it pauses ### close(1) = 0 exit_group(0) = ? From lindahl at pbm.com Mon Nov 24 18:08:30 2008 From: lindahl at pbm.com (Greg Lindahl) Date: Mon, 24 Nov 2008 18:08:30 -0800 Subject: [svlug] Sync command slowdown Message-ID: <20081125020830.GA29446@bx9> On Mon, Nov 24, 2008 at 05:32:40PM -0800, James Sparenberg wrote: > I've got 4 new DB servers exhibiting a rather problematic set of > actions with disk sync. Running the sync command (as is done when logs > are rotated.) takes 4-6 seconds (not micro, full seconds) Other than > this these boxes are running 10x faster than the servers they replaced > and making our customers much happier. I'd bet that you're ending up with many more dirty pages on the new servers because they have more memory than the old ones. You can watch this value over time with % grep Dirty /proc/meminfo sync is a big club: it syncs all the files in the system instead of just the ones you care about. And it can kill performance for other processes if there are a lot of pages to be written out. Doctor, it hurts when I do this. -- greg From lsc at prgmr.com Mon Nov 24 18:17:15 2008 From: lsc at prgmr.com (Luke S Crawford) Date: 24 Nov 2008 21:17:15 -0500 Subject: [svlug] Sync command slowdown In-Reply-To: <200811241732.40649.james@linuxrebel.org> References: <200811241732.40649.james@linuxrebel.org> Message-ID: James Sparenberg writes: > I've got 4 new DB servers exhibiting a rather problematic set of > actions with disk sync. Running the sync command (as is done when logs > are rotated.) takes 4-6 seconds (not micro, full seconds) Other than > this these boxes are running 10x faster than the servers they replaced > and making our customers much happier. Are you using ext3? (or something else that does write-back caching?) I saw something very similar the other night when I was setting up my new ridiculous[1] servers (I switched to 32GB ram/dual socket quad-core systems for my Xen hosting setup.) - See, normally I limit the Dom0 to having 512M or a gigabyte of ram, but these were new servers and I was just playing with them in the garage. I hadn't even installed the xen kernel yet, so the OS instance had a whole 32GB ram to play with. I ran some disk tests, a dd command, bonnie, - the system screamed. Sync, on the other hand, took north of 120 seconds to return. but on another console, iostat showed that my sata disks were still chugging away, even though the disk tests finished some time ago. As far as I can tell, it was just the expected ext3 write caching, just, you know, bigger. it takes a long time to flush 32GB of data to SATA disks. But yeah, that'd be my guess; you just have lots of dirty buffers. More ram means that you can buffer more, and disks are still pretty slow. [1]http://prgmr.com/~lsc/luke_opterons.jpg From bjt23 at cornell.edu Mon Nov 24 18:30:56 2008 From: bjt23 at cornell.edu (Brian J. Tarricone) Date: Mon, 24 Nov 2008 18:30:56 -0800 Subject: [svlug] Sync command slowdown In-Reply-To: <20081125020830.GA29446@bx9> References: <20081125020830.GA29446@bx9> Message-ID: <492B6360.6080801@cornell.edu> Greg Lindahl wrote: > sync is a big club: it syncs all the files in the system instead of > just the ones you care about. And it can kill performance for other > processes if there are a lot of pages to be written out. Though, if you're using ext3, a whole-disk sync() and a single-file fsync() are sadly equivalent with certain (common?) journaling modes. (See the not-so-recent Firefox 3 bug[1] relating to fsync() when committing various sqlite databases to disk.) -brian [1] https://bugzilla.mozilla.org/show_bug.cgi?id=421482 From dmarti at zgp.org Mon Nov 24 18:43:24 2008 From: dmarti at zgp.org (Don Marti) Date: Mon, 24 Nov 2008 18:43:24 -0800 Subject: [svlug] Sync command slowdown In-Reply-To: References: <200811241732.40649.james@linuxrebel.org> Message-ID: <20081125024323.GA28264@zgp.org> begin Luke S Crawford quotation of Mon, Nov 24, 2008 at 09:17:15PM -0500: > But yeah, that'd be my guess; you just have lots of dirty buffers. More ram > means that you can buffer more, and disks are still pretty slow. Here's the consultant who wants to make his plane home answer...set up another cron job that runs 1 minute before the logrotate cron job and does a sleep 50 && sync so you only have a few seconds of dirty buffers to sync from the real cron job. What does /proc/sys/vm/dirty_writeback_centisecs say? You might be able to get a faster sync time by making it smaller. -- Don Marti +1 510-814-0932 http://zgp.org/~dmarti/ +1 510-332-1587 mobile dmarti at zgp.org From xbmodder at gmail.com Mon Nov 24 18:13:45 2008 From: xbmodder at gmail.com (Sargun Dhillon) Date: Mon, 24 Nov 2008 18:13:45 -0800 Subject: [svlug] Sync command slowdown In-Reply-To: <20081125020830.GA29446@bx9> References: <20081125020830.GA29446@bx9> Message-ID: <7c9d57ea0811241813q5d4654dayff969f023b0853ec@mail.gmail.com> Why not mount the partition -o sync entirely? ;-P Also, you should be using the write-back cache on your controller, so sync should just be a "soft" sync. See if you can turn on the right back cache on your controller. On Mon, Nov 24, 2008 at 6:08 PM, Greg Lindahl wrote: > On Mon, Nov 24, 2008 at 05:32:40PM -0800, James Sparenberg wrote: > >> I've got 4 new DB servers exhibiting a rather problematic set of >> actions with disk sync. Running the sync command (as is done when logs >> are rotated.) takes 4-6 seconds (not micro, full seconds) Other than >> this these boxes are running 10x faster than the servers they replaced >> and making our customers much happier. > > I'd bet that you're ending up with many more dirty pages on the new > servers because they have more memory than the old ones. You can watch > this value over time with > > % grep Dirty /proc/meminfo > > sync is a big club: it syncs all the files in the system instead of > just the ones you care about. And it can kill performance for other > processes if there are a lot of pages to be written out. > > Doctor, it hurts when I do this. > > -- greg > > > _______________________________________________ > svlug mailing list > svlug at lists.svlug.org > http://lists.svlug.org/lists/listinfo/svlug > From james at linuxrebel.org Mon Nov 24 19:15:22 2008 From: james at linuxrebel.org (James Sparenberg) Date: Mon, 24 Nov 2008 19:15:22 -0800 Subject: [svlug] Sync command slowdown In-Reply-To: <492B6360.6080801@cornell.edu> References: <20081125020830.GA29446@bx9> <492B6360.6080801@cornell.edu> Message-ID: <200811241915.22782.james@linuxrebel.org> On Monday 24 November 2008 18:30:56 Brian J. Tarricone wrote: > Greg Lindahl wrote: > > > sync is a big club: it syncs all the files in the system instead of > > just the ones you care about. And it can kill performance for other > > processes if there are a lot of pages to be written out. > > Though, if you're using ext3, a whole-disk sync() and a single-file > fsync() are sadly equivalent with certain (common?) journaling modes. > > (See the not-so-recent Firefox 3 bug[1] relating to fsync() when > committing various sqlite databases to disk.) > > -brian > > [1] https://bugzilla.mozilla.org/show_bug.cgi?id=421482 > > Quick note to both who have replied. Thanks this is useful info to chew on. So far I've used both the default and the the writeback journaling methods. James From james at linuxrebel.org Mon Nov 24 19:46:59 2008 From: james at linuxrebel.org (James Sparenberg) Date: Mon, 24 Nov 2008 19:46:59 -0800 Subject: [svlug] Sync command slowdown In-Reply-To: <492B6360.6080801@cornell.edu> References: <20081125020830.GA29446@bx9> <492B6360.6080801@cornell.edu> Message-ID: <200811241946.59396.james@linuxrebel.org> On Monday 24 November 2008 18:30:56 Brian J. Tarricone wrote: > Greg Lindahl wrote: > > > sync is a big club: it syncs all the files in the system instead of > > just the ones you care about. And it can kill performance for other > > processes if there are a lot of pages to be written out. > > Though, if you're using ext3, a whole-disk sync() and a single-file > fsync() are sadly equivalent with certain (common?) journaling modes. > > (See the not-so-recent Firefox 3 bug[1] relating to fsync() when > committing various sqlite databases to disk.) > > -brian > > [1] https://bugzilla.mozilla.org/show_bug.cgi?id=421482 > > > _______________________________________________ > svlug mailing list > svlug at lists.svlug.org > http://lists.svlug.org/lists/listinfo/svlug > Ok, Following bugs and links I found this bug in the kernel http://bugzilla.kernel.org/show_bug.cgi?id=9546 This seems to match identically what I've got going on. I've now setup the test system to where all partitions are running in the writeback mode, for the journal instead of the default ordered mode. Finger crossed. Again all thanks! James From james at linuxrebel.org Mon Nov 24 20:12:22 2008 From: james at linuxrebel.org (James Sparenberg) Date: Mon, 24 Nov 2008 20:12:22 -0800 Subject: [svlug] Sync command slowdown In-Reply-To: <7c9d57ea0811241813q5d4654dayff969f023b0853ec@mail.gmail.com> References: <20081125020830.GA29446@bx9> <7c9d57ea0811241813q5d4654dayff969f023b0853ec@mail.gmail.com> Message-ID: <200811242012.22438.james@linuxrebel.org> On Monday 24 November 2008 18:13:45 Sargun Dhillon wrote: > Why not mount the partition -o sync entirely? ;-P > Also, you should be using the write-back cache on your controller, so > sync should just be a "soft" sync. See if you can turn on the right > back cache on your controller. > > On Mon, Nov 24, 2008 at 6:08 PM, Greg Lindahl wrote: > > On Mon, Nov 24, 2008 at 05:32:40PM -0800, James Sparenberg wrote: > > > >> I've got 4 new DB servers exhibiting a rather problematic set of > >> actions with disk sync. Running the sync command (as is done when logs > >> are rotated.) takes 4-6 seconds (not micro, full seconds) Other than > >> this these boxes are running 10x faster than the servers they replaced > >> and making our customers much happier. > > > > I'd bet that you're ending up with many more dirty pages on the new > > servers because they have more memory than the old ones. You can watch > > this value over time with > > > > % grep Dirty /proc/meminfo > > > > sync is a big club: it syncs all the files in the system instead of > > just the ones you care about. And it can kill performance for other > > processes if there are a lot of pages to be written out. > > > > Doctor, it hurts when I do this. > > > > -- greg > > > > The default in RHEL is sync IIRC (i'll need to check) but as far as the hw goes the writeback is the setting from Penguin. Thanks! James From james at linuxrebel.org Mon Nov 24 20:20:43 2008 From: james at linuxrebel.org (James Sparenberg) Date: Mon, 24 Nov 2008 20:20:43 -0800 Subject: [svlug] Sync command slowdown In-Reply-To: References: <200811241732.40649.james@linuxrebel.org> Message-ID: <200811242020.43317.james@linuxrebel.org> On Monday 24 November 2008 18:17:15 Luke S Crawford wrote: > James Sparenberg writes: > > I've got 4 new DB servers exhibiting a rather problematic set of > > actions with disk sync. Running the sync command (as is done when logs > > are rotated.) takes 4-6 seconds (not micro, full seconds) Other than > > this these boxes are running 10x faster than the servers they replaced > > and making our customers much happier. > > Are you using ext3? (or something else that does write-back caching?) > I saw something very similar the other night when I was setting up my new > ridiculous[1] servers (I switched to 32GB ram/dual socket quad-core systems > for my Xen hosting setup.) - See, normally I limit the Dom0 to having > 512M or a gigabyte of ram, but these were new servers and I was just playing > with them in the garage. I hadn't even installed the xen kernel yet, so > the OS instance had a whole 32GB ram to play with. I ran some disk tests, > a dd command, bonnie, - the system screamed. Sync, on the other hand, > took north of 120 seconds to return. > > but on another console, iostat showed that my sata disks were still chugging > away, even though the disk tests finished some time ago. As far as I can > tell, it was just the expected ext3 write caching, just, you know, bigger. > it takes a long time to flush 32GB of data to SATA disks. > > But yeah, that'd be my guess; you just have lots of dirty buffers. More ram > means that you can buffer more, and disks are still pretty slow. > > > [1]http://prgmr.com/~lsc/luke_opterons.jpg > Yes I am using ext3 and in the process I also found this at kerneltrap. http://bugzilla.kernel.org/show_bug.cgi?id=9546 James From james at linuxrebel.org Mon Nov 24 20:39:19 2008 From: james at linuxrebel.org (James Sparenberg) Date: Mon, 24 Nov 2008 20:39:19 -0800 Subject: [svlug] Sync command slowdown In-Reply-To: <20081125024323.GA28264@zgp.org> References: <200811241732.40649.james@linuxrebel.org> <20081125024323.GA28264@zgp.org> Message-ID: <200811242039.19170.james@linuxrebel.org> On Monday 24 November 2008 18:43:24 Don Marti wrote: > begin Luke S Crawford quotation of Mon, Nov 24, 2008 at 09:17:15PM -0500: > > > But yeah, that'd be my guess; you just have lots of dirty buffers. More ram > > means that you can buffer more, and disks are still pretty slow. > > Here's the consultant who wants to make his plane > home answer...set up another cron job that runs 1 > minute before the logrotate cron job and does a > sleep 50 && sync > so you only have a few seconds of dirty buffers > to sync from the real cron job. > > What does > /proc/sys/vm/dirty_writeback_centisecs > say? You might be able to get a faster sync time by > making it smaller. Don, Only hassle is that mysql, though it uses gnu commands to accomplish it, runs the rotation on it's own, not via cron. Additionals we have found 1. small partitions have less of a problem than large ones. (if I send the binlogs to a partiion that is really to small to hold them for long they rotate quickly) 2. small logs rotate faster. James (thanks!) From james at linuxrebel.org Tue Nov 25 15:55:48 2008 From: james at linuxrebel.org (James Sparenberg) Date: Tue, 25 Nov 2008 15:55:48 -0800 Subject: [svlug] Sync command slowdown In-Reply-To: <200811241732.40649.james@linuxrebel.org> References: <200811241732.40649.james@linuxrebel.org> Message-ID: <200811251555.48353.james@linuxrebel.org> On Monday 24 November 2008 17:32:40 James Sparenberg wrote: > All, > > I've got 4 new DB servers exhibiting a rather problematic set of > actions with disk sync. Running the sync command (as is done when logs > are rotated.) takes 4-6 seconds (not micro, full seconds) Other than > this these boxes are running 10x faster than the servers they replaced > and making our customers much happier. > > Further it's only happening on my 4 new servers not on the older > systems (I have them in QA, etc). CPU load is at best 15% per core, > normally down in the dirt, as the primary CPU function is to manage disk > I/0 (most queries are the same so they are cached, and writes are > limited.) RAM is fine and the box never touches it's swap partition. > I'm personally suspecting that it is Kernel module vs LSI raid > controller, on these boxes. > > Additionally our DBA has discovered that it seems to work faster on > smaller partitions than the larger one we have for our DBs. The only > tweak I'm currently using beyond defaults in fstab are noatime (RHEL 5 > mount command doesn't recognize relatime) however changing out the > noatime has no affect on the sync command. > > Has anyone out there seen anything at all similar? I'm hoping to > avoid having to go to software raid only because it will mean a fairly > extensive rebuild of the system, and from now until mid February I'm not > going to be able to take that time. For now the mid level solution is > going with smaller binlogs and rotating a lot more often. > > Details on the systems. > > Manf > Penguin Computing (And yes they are providing some assistance on > this, they really are proactive.) > > OS: > CentOS 5.2 x86_64 (fully updated) running 2.6.18-92 > > CPU: > Single Quad Core Intel Xeon > > RAM: > 16GB > > HDD > LSI Raid 1 controller > > lspci: > > 00:00.0 Host bridge: Intel Corporation 5000P Chipset Memory Controller > Hub (rev b1) > 00:02.0 PCI bridge: Intel Corporation 5000 Series Chipset PCI Express x8 > Port 2-3 (rev b1) > 00:03.0 PCI bridge: Intel Corporation 5000 Series Chipset PCI Express x4 > Port 3 (rev b1) > 00:04.0 PCI bridge: Intel Corporation 5000 Series Chipset PCI Express x8 > Port 4-5 (rev b1) > 00:05.0 PCI bridge: Intel Corporation 5000 Series Chipset PCI Express x4 > Port 5 (rev b1) > 00:06.0 PCI bridge: Intel Corporation 5000 Series Chipset PCI Express x8 > Port 6-7 (rev b1) > 00:07.0 PCI bridge: Intel Corporation 5000 Series Chipset PCI Express x4 > Port 7 (rev b1) > 00:08.0 System peripheral: Intel Corporation 5000 Series Chipset DMA > Engine (rev b1) > 00:10.0 Host bridge: Intel Corporation 5000 Series Chipset FSB Registers > (rev b1) > 00:10.1 Host bridge: Intel Corporation 5000 Series Chipset FSB Registers > (rev b1) > 00:10.2 Host bridge: Intel Corporation 5000 Series Chipset FSB Registers > (rev b1) > 00:11.0 Host bridge: Intel Corporation 5000 Series Chipset Reserved > Registers (rev b1) > 00:13.0 Host bridge: Intel Corporation 5000 Series Chipset Reserved > Registers (rev b1) > 00:15.0 Host bridge: Intel Corporation 5000 Series Chipset FBD Registers > (rev b1) > 00:16.0 Host bridge: Intel Corporation 5000 Series Chipset FBD Registers > (rev b1) > 00:1c.0 PCI bridge: Intel Corporation 631xESB/632xESB/3100 Chipset PCI > Express Root Port 1 (rev 09) > 00:1d.0 USB Controller: Intel Corporation 631xESB/632xESB/3100 Chipset > UHCI USB Controller #1 (rev 09) > 00:1d.1 USB Controller: Intel Corporation 631xESB/632xESB/3100 Chipset > UHCI USB Controller #2 (rev 09) > 00:1d.2 USB Controller: Intel Corporation 631xESB/632xESB/3100 Chipset > UHCI USB Controller #3 (rev 09) > 00:1d.3 USB Controller: Intel Corporation 631xESB/632xESB/3100 Chipset > UHCI USB Controller #4 (rev 09) > 00:1d.7 USB Controller: Intel Corporation 631xESB/632xESB/3100 Chipset > EHCI USB2 Controller (rev 09) > 00:1e.0 PCI bridge: Intel Corporation 82801 PCI Bridge (rev d9) > 00:1f.0 ISA bridge: Intel Corporation 631xESB/632xESB/3100 Chipset LPC > Interface Controller (rev 09) > 00:1f.1 IDE interface: Intel Corporation 631xESB/632xESB IDE Controller > (rev 09) > 00:1f.2 IDE interface: Intel Corporation 631xESB/632xESB/3100 Chipset > SATA IDE Controller (rev 09) > 00:1f.3 SMBus: Intel Corporation 631xESB/632xESB/3100 Chipset SMBus > Controller (rev 09) > 01:00.0 PCI bridge: Intel Corporation 6311ESB/6321ESB PCI Express > Upstream Port (rev 01) > 01:00.3 PCI bridge: Intel Corporation 6311ESB/6321ESB PCI Express to > PCI-X Bridge (rev 01) > 02:00.0 PCI bridge: Intel Corporation 6311ESB/6321ESB PCI Express > Downstream Port E1 (rev 01) > 02:02.0 PCI bridge: Intel Corporation 6311ESB/6321ESB PCI Express > Downstream Port E3 (rev 01) > 03:00.0 SCSI storage controller: LSI Logic / Symbios Logic SAS1064ET > PCI-Express Fusion-MPT SAS (rev 04) > 04:00.0 Ethernet controller: Intel Corporation 80003ES2LAN Gigabit > Ethernet Controller (Copper) (rev 01) > 04:00.1 Ethernet controller: Intel Corporation 80003ES2LAN Gigabit > Ethernet Controller (Copper) (rev 01) > 07:00.0 Ethernet controller: Intel Corporation 82571EB Gigabit Ethernet > Controller (rev 06) > 07:00.1 Ethernet controller: Intel Corporation 82571EB Gigabit Ethernet > Controller (rev 06) > 0b:00.0 VGA compatible controller: Matrox Graphics, Inc. MGA G200e > [Pilot] ServerEngines (SEP1) (rev 02) > > Stack Trace > > execve("/bin/sync", ["sync"], [/* 16 vars */]) = 0 > brk(0) = 0x1ce91000 > mmap(NULL, 4096, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_ANONYMOUS, -1, 0) > = 0x2aaaaaaab000 > uname({sys="Linux", node="my.server.name", ...}) = 0 > access("/etc/ld.so.preload", R_OK) = -1 ENOENT (No such file or > directory) > open("/etc/ld.so.cache", O_RDONLY) = 3 > fstat(3, {st_mode=S_IFREG|0644, st_size=23796, ...}) = 0 > mmap(NULL, 23796, PROT_READ, MAP_PRIVATE, 3, 0) = 0x2aaaaaaac000 > close(3) = 0 > open("/lib64/libc.so.6", O_RDONLY) = 3 > read(3, "\177ELF\2\1\1\0\0\0\0\0\0\0\0\0\3\0>\0\1\0\0\0\260\331! > \2773\0\0\0"..., 832) = 832 > fstat(3, {st_mode=S_IFREG|0755, st_size=1699880, ...}) = 0 > mmap(NULL, 4096, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_ANONYMOUS, -1, 0) > = 0x2aaaaaab2000 > mmap(0x33bf200000, 3481848, PROT_READ|PROT_EXEC, MAP_PRIVATE| > MAP_DENYWRITE, 3, 0) = 0x33bf200000 > mprotect(0x33bf34a000, 2093056, PROT_NONE) = 0 > mmap(0x33bf549000, 20480, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_FIXED| > MAP_DENYWRITE, 3, 0x149000) = 0x33bf549000 > mmap(0x33bf54e000, 16632, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_FIXED| > MAP_ANONYMOUS, -1, 0) = 0x33bf54e000 > close(3) = 0 > mmap(NULL, 4096, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_ANONYMOUS, -1, 0) > = 0x2aaaaaab3000 > arch_prctl(ARCH_SET_FS, 0x2aaaaaab3250) = 0 > mprotect(0x33bf549000, 16384, PROT_READ) = 0 > mprotect(0x33bf01a000, 4096, PROT_READ) = 0 > munmap(0x2aaaaaaac000, 23796) = 0 > brk(0) = 0x1ce91000 > brk(0x1ceb2000) = 0x1ceb2000 > open("/usr/lib/locale/locale-archive", O_RDONLY) = 3 > fstat(3, {st_mode=S_IFREG|0644, st_size=56470832, ...}) = 0 > mmap(NULL, 56470832, PROT_READ, MAP_PRIVATE, 3, 0) = 0x2aaaaaab4000 > close(3) = 0 > sync() = 0 > ###NOTE: This is where it pauses ### > close(1) = 0 > exit_group(0) = ? > > > _______________________________________________ > svlug mailing list > svlug at lists.svlug.org > http://lists.svlug.org/lists/listinfo/svlug > Well 24 hours later (so that we could monitor) and the conclusion seem to be that yes it does appear to be related somehow to the ext3 FS and no changing to writeback doesn't improve the speed. It seems that the problem has to do with the size of the ext3 FS, in that over some (as yet) unknow size there is a problem with the speed of the sync command. James From john_re at fastmail.us Wed Nov 26 00:29:22 2008 From: john_re at fastmail.us (john_re) Date: Wed, 26 Nov 2008 00:29:22 -0800 Subject: [svlug] Best value Quad core sweet spot, holiday sales, processor mboard overclock retailers Message-ID: <1227688162.8417.1286833687@webmail.messagingengine.com> What is the current high end of the quad core price _sweet spot_? CONTENTS: 1) BACKGROUND - HOLIDAY SALES, DESIRED PROCESSOR & MOTHERBOARD 2) SYSTEM PURPOSE, DESIRED PROPERTIES, 8+ GB RAM, GRAPHICS COMPUTE CARDS 3) YOUR RECOMMENDATIONS - HW, INFO, WEBSITES 4) APPENDIX - LINKS DEALS, PRICE COMPARISON, HW REVIEW, OVERCLOCKING, ONLINE STORES MANUFACTURERS: CPU, MOTHERBOARD, PROGRAMMABLE GRAPHICS CARD == 1) BACKGROUND - SALES, DESIRED PROCESSOR & MOTHERBOARD Have you bought or designed a quad core system recently? The holiday sales will happen soon. I thought it would be useful for a thread on this list about what the quad core value sweet spot is - what to look for, & where to look. For anyone interested in getting a new desktop processor/motherboard in the next several months. There's a lot to be aware of, for "What is the current high end of the quad core sweet spot?" - Overclocking, processor cache sizes, bus speeds, processor architectures, GPU graphics card (bus, space & power), power supply & thermal requirements. I think over/upclocking is likely crucial for getting the best value. True?? This thread is about: All the latest technology used in getting an optimal system - particularly details of overclocking regarding processors & motherboard requirements. Following is a proposed system for an engineer, something at the high end of the performance/price _sweet spot_. (Which currently precludes i7 - prices seem way out of the sweet spot, especially RAM.) (Maybe we could do a separate i7 system design, on a separate email list thread.) Although I've included info about many system components below, I'm primarily interested in finding out about only the processor & motherboard combo. Basically, I think my questions boil down to: 1) Phenom or Core 2 Quad? 2) Overclock, or not? 3) What motherboard is best for 3a) overclock (clocking BIOS & HW settings, & power issues) 3b) periherals (video/GPU cards primarily) == 2) SYSTEM PURPOSE, DESIRED PROPERTIES: 8+ GB RAM, GPU (GRAPHICS COMPUTE CARDS) Purpose: Run custom/engineering software - doing lots of computation on GPU (Graphics Processing Unit) card(s) & the system CPU, not much disk or network IO. Suitable for SW development, & as an engineering compute/CAD system. MAIN REQUIREMENTS: 1a) PROCESSOR: Quad core: Intel Core 2 Quad or AMD Phenom (not i7) 1b) UPCLOCK: Able to upclock in order to maximize performance/price. 1c) CPU COOLING: Quiet fan or water cooled for upclock - 200-300 Watts capable? 2) OS: 64 bit Linux, perhaps with virtual 2nd OS capable 3) RAM: 8-16 GB DDR2 - what speed? 4) VIDEO: for 2 high resolution simultaneous displays 4a) High res text for programming (on MB or add in card) 4b) High res text/graphics - GPU accelerated 4c) Non video out graphics GPU card 5) GPU ADD IN CARDS: Able to hold 1-2 programmable graphics cards (Nvidia CUDA, or ATI's comparable solution [SLI, CrossFire]). Preferentially, these will only be used for computing, not video out. So, it would be nice if the motherboard has on board video that can work while the graphics-compute cards are in place. 6) POWER SUPPLY: For Proc + 1-2 Graphics board, 400-600 W? Quiet 7) HARD DISK: 1 1.5 TB Drive, ability to add 1 more drive later, RAID capabilities = ? == 3) YOUR RECOMMENDATIONS - HW, INFO, WEBSITES Since there is so much info to sort through, I'd love to hear from someone who is up to date on this kind of hardware. Have you bought or designed a quad core system recently? What hardware would you recommend? What are the recommended retailers? Would good values be as likely to be obtained from local retailers as online or mail order? [& no time wasting "mail in" rebates.] What web sites have great info about value sweet spot quad core HW for use running Linux? ===================================================================== ===== 4) APPENDIX - LINKS DEALS, PRICE COMPARISON, HW REVIEW, OVERCLOCKING, ONLINE STORES MANUFACTURERS: CPU, MOTHERBOARD, PROGRAMMABLE GRAPHICS CARD LINUX SPECIFIC Below are some deal sites, price comparison sites, HW review sites, overclocking sites, & online stores. What is the best linux site - with info about what HW is compatible / easy-to-use with Linux? DEALS dealnews.?? the computer hardware link. http://dealnews.com/memory/prices/PC2-8500-DDR2-1066-MHz/67/4GB.html Computer Deals http://dealnews.com/categories/Computer/39.html Upgrades / Components Deals http://dealnews.com/categories/Computer/Upgrades-Components/93.html PRICE COMPARISON http://www.pricewatch.com/motherboard_combos/ Processors - CPUs http://www.pricescan.com/Computer-Processors-CPUs/co/Search01120900.html http://www.pricewatch.com/system_memory/ http://www.pricewatch.com/computer_systems_no_os/ HW REVIEW AMD Phenom vs i7 hardware reviews http://www.amdzone.com/phpbb3/viewtopic.php?f=52&t=135736 http://pclab.pl/art34180-9.html Core i7 Multi-GPU SLI Crossfire Game performance review Core 2 Quad QX 9770 versus Core i7 965 http://www.guru3d.com/article/core-i7-multigpu-sli-crossfire-game-performance-review/19 April 9, 2008 Quad Core Showdown: AMD Phenom X4 9850 vs. Intel Core 2 Quad Q9300 http://www.extremetech.com/article2/0,1697,2282395,00.asp 2008 November A Phenom II X4 managed 4GHz with air-cooling http://www.engadget.com/2008/11/21/amd-overclocks-the-snot-out-of-phenom-ii-processors/ OVERCLOCKING http://en.wikipedia.org/wiki/Overclocking http://www.overclock.net/ http://forums.extremeoverclocking.com/ http://www.ocforums.com/ http://www.ocia.net/ http://www.overclockersclub.com/pages/overclock_faq/ ONLINE STORES (Nice motherboard option display on Portatech) Phenom X4 9950+ (4 x 2.6GHz - 4000FSB - 4MB Cache) $214.97 combo http://www.portatech.com/catalog/viewitem.asp?ID=25354&O=26528 Core 2 Quad Q6600 (4 x 2.4GHz - 1066FSB - 8MB Cache) $214.98 combo http://www.portatech.com/catalog/viewitem.asp?ID=25355&O=26139 newegg.com AMD Phenom 9600 Agena 2.3GHz Quad-Core Socket AM2+ Processor, model no. HD960ZWCGDBOX, Asus M3A-H Socket AM2+ ATX Motherboard $149.98 http://www.newegg.com/Product/ComboDealDetails.aspx?ItemList=Combo.144174 Combo w/ Intel Core 2 Quad Q6600 (2.4GHz - 1066FSB - 8MB Cache) http://www.smksuperstore.com/catalog/viewitem.asp?ID=31157 CPU MANUFACTURERS AMD, Intel MOTHERBOARD MANUFACTURERS Asus, Gigabyte, Intel, ?? PROGRAMMABLE GRAPHICS CARD HW MANUFACTURERS ATI/AMD, NVidia LINUX SPECIFIC www.phoronix.com - Lots of info +++ :) Processors - Not much info on Quad cores :( http://www.phoronix.com/scan.php?page=category&item=Processors Motherboards - Many, lots of recent articles +++ http://www.phoronix.com/scan.php?page=category&item=Motherboards Graphics Cards - Lots of recent info. +++ :) http://www.linux.com/forums/forum/2 Hardware - the most active forum. :) Hardware Compatibility List - active http://www.linuxquestions.org/hcl/ Processors + :) http://www.linuxquestions.org/hcl/index.php/cat/6 MotherBoards Asus Gigabyte Intel MSI +++ :) http://www.linuxquestions.org/hcl/index.php/cat/8 -- http://www.linuxhardware.org/ Nothing here. http://www.linuxhardware.org/forums/index.php Nothing here http://hardware4linux.info Eh. Some info, but no=poor indexing. http://www.linuxjournal.com/taxonomy/term/23 Hardware - nothing useful http://linuxmagazine.com/hardware Multicore, desktop - eh? http://www.linux.org/vendor/hardware/index.html http://www.linux.com/ 8 Cores on a Budget- Building a Better Workstation - 2 Xeon 1.6 GHz http://www.linux.com/?module=comments&func=display&cid=1198752 As Intel ships 10 millionth quad-core, AMD gets it in gear http://www.linux.com/feed/133447 http://www.linux.org/vendor/hardware/index.html -- November 18, 2008 MSI GeForce 9800GT 512MB mid-range discrete graphics card more choices NVIDIA continuing to release new stable Linux drivers AMD this year making evolutionary leaps http://www.phoronix.com/vr.php?view=13124 4 Monitors using Nvidia Quadro NVS 440 http://www.phoronix.com/forums/showthread.php?t=7603 The Truth About ATI/AMD & Linux http://www.phoronix.com/vr.php?view=10083 -- Linux Hardware Compatibility Lists & Linux Drivers http://www.linux-drivers.org/ From john_re at fastmail.us Wed Nov 26 00:55:45 2008 From: john_re at fastmail.us (john_re) Date: Wed, 26 Nov 2008 00:55:45 -0800 Subject: [svlug] Best value Quad core sweet spot, holiday sales, processor mboard overclock retailers In-Reply-To: <1227688162.8417.1286833687@webmail.messagingengine.com> References: <1227688162.8417.1286833687@webmail.messagingengine.com> Message-ID: <1227689745.13386.1286836131@webmail.messagingengine.com> For websites, include also: LOCAL STORE ADS & SUMMARY Summary of the Fry's ads. http://frys-electronics-ads.com/ Fry's ad in sjmn http://newspaperads.mercurynews.com/ROP/ads.aspx?advid=32664&adid=7110008&cat=3525 http://www.centralcomputers.com/ Are there any other local store web pages relevant to "quad core sweet spot"? PROCESSOR NEWS for both AMD & Intel http://amdandintel.blogspot.com/ On Wed, 26 Nov 2008 00:29:22 -0800, "john_re" said: > What is the current high end of the quad core price _sweet spot_? > > CONTENTS: > 1) BACKGROUND - HOLIDAY SALES, DESIRED PROCESSOR & MOTHERBOARD > 2) SYSTEM PURPOSE, DESIRED PROPERTIES, 8+ GB RAM, GRAPHICS COMPUTE CARDS > 3) YOUR RECOMMENDATIONS - HW, INFO, WEBSITES > 4) APPENDIX - LINKS > DEALS, PRICE COMPARISON, HW REVIEW, OVERCLOCKING, ONLINE STORES > MANUFACTURERS: CPU, MOTHERBOARD, PROGRAMMABLE GRAPHICS CARD > ===================================================================== > ===== 4) APPENDIX - LINKS > DEALS, PRICE COMPARISON, HW REVIEW, OVERCLOCKING, ONLINE STORES > MANUFACTURERS: CPU, MOTHERBOARD, PROGRAMMABLE GRAPHICS CARD > LINUX SPECIFIC > > Below are some deal sites, price comparison sites, HW review sites, > overclocking sites, & online stores. > > What is the best linux site - with info about what HW is compatible / > easy-to-use with Linux? > From darose at darose.net Wed Nov 26 04:35:12 2008 From: darose at darose.net (David Rosenstrauch) Date: Wed, 26 Nov 2008 07:35:12 -0500 Subject: [svlug] Best value Quad core sweet spot, holiday sales, processor mboard overclock retailers In-Reply-To: <1227688162.8417.1286833687@webmail.messagingengine.com> References: <1227688162.8417.1286833687@webmail.messagingengine.com> Message-ID: <492D4280.50902@darose.net> john_re wrote: > What is the current high end of the quad core price _sweet spot_? > > CONTENTS: > 1) BACKGROUND - HOLIDAY SALES, DESIRED PROCESSOR & MOTHERBOARD > 2) SYSTEM PURPOSE, DESIRED PROPERTIES, 8+ GB RAM, GRAPHICS COMPUTE CARDS > 3) YOUR RECOMMENDATIONS - HW, INFO, WEBSITES > 4) APPENDIX - LINKS > DEALS, PRICE COMPARISON, HW REVIEW, OVERCLOCKING, ONLINE STORES > MANUFACTURERS: CPU, MOTHERBOARD, PROGRAMMABLE GRAPHICS CARD > > > == 1) BACKGROUND - SALES, DESIRED PROCESSOR & MOTHERBOARD > Have you bought or designed a quad core system recently? I recently built the following quad core machine, and bought all the components at Newegg. (ZipZoomFly would be my 2nd preferred vendor.) Seagate Barracuda 7200.11 ST3500320AS 500GB 7200 RPM SATA 3.0Gb/s Hard Drive - OEM LITE-ON 20X DVD?R DVD Burner Black SATA Model iHAS120-08 - Retail LITE-ON Black SATA DVD-ROM Drive Model DH-16D3S-04 - OEM GIGABYTE GA-EP43-DS3L LGA 775 Intel P43 ATX Intel Motherboard - Retail Antec Sonata III 500 Black 0.8mm cold rolled steel ATX Mid Tower Computer Case 500W Power Supply - Retail SAPPHIRE 100233L Radeon HD 3450 256MB 64-bit GDDR2 PCI Express 2.0 x16 HDCP Ready CrossFire Supported Video Card - Retail mushkin 4GB (2 x 2GB) 240-Pin DDR2 SDRAM DDR2 800 (PC2 6400) Dual Channel Kit Desktop Memory Model 996558 - Retail Intel Core 2 Quad Q6600 Kentsfield 2.4GHz LGA 775 Quad-Core Processor Model BX80562Q6600 - Retail I'm happy with all the components except for the mobo. It gave me some memory compatibility grief, plus I have some issues between the audio front panel headers and the mobo re: sound that I haven't quite worked out yet. MSI has an equivalent mobo, and I used an MSI mobo in a previous build that I've been very happy with, so in retrospect I'd probably recommend that instead. Didn't thoroughly read through the rest of your message. Email back if you've got additional questions. HTH, DR From mark at cosmicpenguin.com Wed Nov 26 05:59:40 2008 From: mark at cosmicpenguin.com (Mark S Bilk) Date: Wed, 26 Nov 2008 05:59:40 -0800 Subject: [svlug] Best value Quad core sweet spot, holiday sales, processor mboard overclock retailers In-Reply-To: <1227688162.8417.1286833687@webmail.messagingengine.com> References: <1227688162.8417.1286833687@webmail.messagingengine.com> Message-ID: <20081126135940.GG4980@Isis> Overclocking is a bad idea. Chips are specified with a particular clock rate because above that rate errors and/or chip damage may occur. The same goes for using a power voltage beyond the spec. And since the increase in speed is barely noticeable, there really isn't any reason to do it. As to quad core, my dual core system very rarely uses the second core (as shown by ksysguard). I'd be interested in reports from others on this point. I've always used ASUS motherboards and found them very reliable. It might be worthwhile to look at the Playstation 3 with its Cell CPU and tremendous computing power. Apparently it runs Linux well. The limitations that I know of are the small amount of memory in the box, and the fact that it can't run x86 code without interpretation, so WINE and virtual machine hosting probably won't work. Finally, I would not use an Intel processor, because that company has written the firmware for Microsoft's "Palladium" system for total government censorship of the Internet. Intel has also engaged in dishonest and coercive business practices against AMD. On Wed, Nov 26, 2008 at 12:29:22AM -0800, john_re wrote: >What is the current high end of the quad core price _sweet spot_? > >CONTENTS: >1) BACKGROUND - HOLIDAY SALES, DESIRED PROCESSOR & MOTHERBOARD >2) SYSTEM PURPOSE, DESIRED PROPERTIES, 8+ GB RAM, GRAPHICS COMPUTE CARDS >3) YOUR RECOMMENDATIONS - HW, INFO, WEBSITES >4) APPENDIX - LINKS > DEALS, PRICE COMPARISON, HW REVIEW, OVERCLOCKING, ONLINE STORES > MANUFACTURERS: CPU, MOTHERBOARD, PROGRAMMABLE GRAPHICS CARD > > >== 1) BACKGROUND - SALES, DESIRED PROCESSOR & MOTHERBOARD >Have you bought or designed a quad core system recently? > >The holiday sales will happen soon. I thought it would be useful for a >thread on this list about what the quad core value sweet spot is - what >to look for, & where to look. > >For anyone interested in getting a new desktop processor/motherboard in >the next several months. > >There's a lot to be aware of, for "What is the current high end of the >quad core sweet spot?" - Overclocking, processor cache sizes, bus >speeds, processor architectures, GPU graphics card (bus, space & power), >power supply & thermal requirements. I think over/upclocking is likely >crucial for getting the best value. True?? > >This thread is about: All the latest technology used in getting an >optimal system - particularly details of overclocking regarding >processors & motherboard requirements. > >Following is a proposed system for an engineer, something at the high >end of the performance/price _sweet spot_. (Which currently precludes >i7 - prices seem way out of the sweet spot, especially RAM.) (Maybe we >could do a separate i7 system design, on a separate email list thread.) > >Although I've included info about many system components below, I'm >primarily interested in finding out about only the processor & >motherboard combo. > >Basically, I think my questions boil down to: >1) Phenom or Core 2 Quad? >2) Overclock, or not? >3) What motherboard is best for >3a) overclock (clocking BIOS & HW settings, & power issues) >3b) periherals (video/GPU cards primarily) > > >== 2) SYSTEM PURPOSE, DESIRED PROPERTIES: > 8+ GB RAM, GPU (GRAPHICS COMPUTE CARDS) >Purpose: Run custom/engineering software - doing lots of computation on >GPU (Graphics Processing Unit) card(s) & the system CPU, >not much disk or network IO. >Suitable for SW development, & as an engineering compute/CAD system. > >MAIN REQUIREMENTS: >1a) PROCESSOR: Quad core: Intel Core 2 Quad or AMD Phenom (not i7) >1b) UPCLOCK: Able to upclock in order to maximize performance/price. >1c) CPU COOLING: Quiet fan or water cooled for upclock - > 200-300 Watts capable? >2) OS: 64 bit Linux, perhaps with virtual 2nd OS capable >3) RAM: 8-16 GB DDR2 - what speed? >4) VIDEO: for 2 high resolution simultaneous displays >4a) High res text for programming (on MB or add in card) >4b) High res text/graphics - GPU accelerated >4c) Non video out graphics GPU card >5) GPU ADD IN CARDS: Able to hold 1-2 programmable graphics cards >(Nvidia CUDA, or ATI's comparable solution [SLI, CrossFire]). >Preferentially, these will only be used for computing, not video out. >So, it would be nice if the motherboard has on board video that can work >while the graphics-compute cards are in place. >6) POWER SUPPLY: For Proc + 1-2 Graphics board, 400-600 W? Quiet >7) HARD DISK: 1 1.5 TB Drive, ability to add 1 more drive later, > RAID capabilities = ? > > > >== 3) YOUR RECOMMENDATIONS - HW, INFO, WEBSITES >Since there is so much info to sort through, I'd love to hear from >someone who is up to date on this kind of hardware. > >Have you bought or designed a quad core system recently? >What hardware would you recommend? > >What are the recommended retailers? Would good values be as likely to >be obtained from local retailers as online or mail order? >[& no time wasting "mail in" rebates.] > >What web sites have great info about value sweet spot quad core HW for >use running Linux? > > > >===================================================================== >===== 4) APPENDIX - LINKS >DEALS, PRICE COMPARISON, HW REVIEW, OVERCLOCKING, ONLINE STORES >MANUFACTURERS: CPU, MOTHERBOARD, PROGRAMMABLE GRAPHICS CARD >LINUX SPECIFIC > >Below are some deal sites, price comparison sites, HW review sites, >overclocking sites, & online stores. > >What is the best linux site - with info about what HW is compatible / >easy-to-use with Linux? > > >DEALS >dealnews.?? the computer hardware link. >http://dealnews.com/memory/prices/PC2-8500-DDR2-1066-MHz/67/4GB.html >Computer Deals >http://dealnews.com/categories/Computer/39.html >Upgrades / Components Deals >http://dealnews.com/categories/Computer/Upgrades-Components/93.html > > >PRICE COMPARISON >http://www.pricewatch.com/motherboard_combos/ >Processors - CPUs >http://www.pricescan.com/Computer-Processors-CPUs/co/Search01120900.html >http://www.pricewatch.com/system_memory/ >http://www.pricewatch.com/computer_systems_no_os/ > >HW REVIEW >AMD Phenom vs i7 hardware reviews >http://www.amdzone.com/phpbb3/viewtopic.php?f=52&t=135736 >http://pclab.pl/art34180-9.html > >Core i7 Multi-GPU SLI Crossfire Game performance review >Core 2 Quad QX 9770 versus Core i7 965 >http://www.guru3d.com/article/core-i7-multigpu-sli-crossfire-game-performance-review/19 > >April 9, 2008 Quad Core Showdown: AMD Phenom X4 9850 vs. Intel Core 2 >Quad Q9300 >http://www.extremetech.com/article2/0,1697,2282395,00.asp > > >2008 November A Phenom II X4 managed 4GHz with air-cooling >http://www.engadget.com/2008/11/21/amd-overclocks-the-snot-out-of-phenom-ii-processors/ > > >OVERCLOCKING >http://en.wikipedia.org/wiki/Overclocking >http://www.overclock.net/ >http://forums.extremeoverclocking.com/ >http://www.ocforums.com/ >http://www.ocia.net/ >http://www.overclockersclub.com/pages/overclock_faq/ > > >ONLINE STORES > >(Nice motherboard option display on Portatech) > >Phenom X4 9950+ (4 x 2.6GHz - 4000FSB - 4MB Cache) $214.97 combo >http://www.portatech.com/catalog/viewitem.asp?ID=25354&O=26528 > >Core 2 Quad Q6600 (4 x 2.4GHz - 1066FSB - 8MB Cache) $214.98 combo >http://www.portatech.com/catalog/viewitem.asp?ID=25355&O=26139 > > >newegg.com AMD Phenom 9600 Agena 2.3GHz Quad-Core Socket AM2+ Processor, >model no. HD960ZWCGDBOX, Asus M3A-H Socket AM2+ ATX Motherboard $149.98 >http://www.newegg.com/Product/ComboDealDetails.aspx?ItemList=Combo.144174 > >Combo w/ Intel Core 2 Quad Q6600 (2.4GHz - 1066FSB - 8MB Cache) >http://www.smksuperstore.com/catalog/viewitem.asp?ID=31157 > > > >CPU MANUFACTURERS >AMD, Intel > >MOTHERBOARD MANUFACTURERS >Asus, Gigabyte, Intel, ?? > >PROGRAMMABLE GRAPHICS CARD HW MANUFACTURERS >ATI/AMD, NVidia > > >LINUX SPECIFIC >www.phoronix.com - Lots of info +++ :) >Processors - Not much info on Quad cores :( >http://www.phoronix.com/scan.php?page=category&item=Processors >Motherboards - Many, lots of recent articles +++ >http://www.phoronix.com/scan.php?page=category&item=Motherboards >Graphics Cards - Lots of recent info. +++ :) > >http://www.linux.com/forums/forum/2 Hardware - the most active forum. >:) > >Hardware Compatibility List - active >http://www.linuxquestions.org/hcl/ >Processors + :) >http://www.linuxquestions.org/hcl/index.php/cat/6 >MotherBoards Asus Gigabyte Intel MSI +++ :) >http://www.linuxquestions.org/hcl/index.php/cat/8 > > >-- >http://www.linuxhardware.org/ Nothing here. >http://www.linuxhardware.org/forums/index.php Nothing here >http://hardware4linux.info Eh. Some info, but no=poor indexing. > >http://www.linuxjournal.com/taxonomy/term/23 Hardware - nothing useful >http://linuxmagazine.com/hardware Multicore, desktop - eh? > > >http://www.linux.org/vendor/hardware/index.html >http://www.linux.com/ >8 Cores on a Budget- Building a Better Workstation - 2 Xeon 1.6 GHz >http://www.linux.com/?module=comments&func=display&cid=1198752 >As Intel ships 10 millionth quad-core, AMD gets it in gear >http://www.linux.com/feed/133447 > >http://www.linux.org/vendor/hardware/index.html > >-- >November 18, 2008 MSI GeForce 9800GT 512MB mid-range discrete >graphics card more choices NVIDIA continuing to release new stable >Linux drivers AMD this year making evolutionary leaps >http://www.phoronix.com/vr.php?view=13124 > >4 Monitors using Nvidia Quadro NVS 440 >http://www.phoronix.com/forums/showthread.php?t=7603 >The Truth About ATI/AMD & Linux >http://www.phoronix.com/vr.php?view=10083 > >-- >Linux Hardware Compatibility Lists & Linux Drivers >http://www.linux-drivers.org/ > From john_re at fastmail.us Wed Nov 26 06:20:02 2008 From: john_re at fastmail.us (john_re) Date: Wed, 26 Nov 2008 06:20:02 -0800 Subject: [svlug] Best value Quad core sweet spot, holiday sales, processor mboard overclock retailers In-Reply-To: <492D4280.50902@darose.net> References: <1227688162.8417.1286833687@webmail.messagingengine.com> <492D4280.50902@darose.net> Message-ID: <1227709202.9650.1286881415@webmail.messagingengine.com> Thnaks for the Q6600 data. :) 0) Why did you choose the Q6600 vs amd, 0b) why the q6600 vs other intel processors? 1) you don't overclock, correct? Considered/planning overclocking? 2) what OS do you use? 3) what do you use the system for? games, home, engineering, SW dev, business? 4) what MSI MB would you recommend? Why? 5) why do you like newegg & zipzoomfly? 6) How are the Radeon drivers? On Wed, 26 Nov 2008 07:35:12 -0500, "David Rosenstrauch" said: > john_re wrote: > > What is the current high end of the quad core price _sweet spot_? > > == 1) BACKGROUND - SALES, DESIRED PROCESSOR & MOTHERBOARD > > Have you bought or designed a quad core system recently? > > I recently built the following quad core machine, and bought all the > components at Newegg. (ZipZoomFly would be my 2nd preferred vendor.) > > Seagate Barracuda 7200.11 ST3500320AS 500GB 7200 RPM SATA 3.0Gb/s Hard > Drive - OEM > > > LITE-ON 20X DVD?R DVD Burner Black SATA Model iHAS120-08 - Retail > > > LITE-ON Black SATA DVD-ROM Drive Model DH-16D3S-04 - OEM > > > GIGABYTE GA-EP43-DS3L LGA 775 Intel P43 ATX Intel Motherboard - Retail > > > Antec Sonata III 500 Black 0.8mm cold rolled steel ATX Mid Tower > Computer Case 500W Power Supply - Retail > > > SAPPHIRE 100233L Radeon HD 3450 256MB 64-bit GDDR2 PCI Express 2.0 x16 > HDCP Ready CrossFire Supported Video Card - Retail > > > mushkin 4GB (2 x 2GB) 240-Pin DDR2 SDRAM DDR2 800 (PC2 6400) Dual > Channel Kit Desktop Memory Model 996558 - Retail > > > Intel Core 2 Quad Q6600 Kentsfield 2.4GHz LGA 775 Quad-Core Processor > Model BX80562Q6600 - Retail > > > I'm happy with all the components except for the mobo. It gave me some > memory compatibility grief, plus I have some issues between the audio > front panel headers and the mobo re: sound that I haven't quite worked > out yet. MSI has an equivalent mobo, and I used an MSI mobo in a > previous build that I've been very happy with, so in retrospect I'd > probably recommend that instead. > > Didn't thoroughly read through the rest of your message. Email back if > you've got additional questions. From darose at darose.net Wed Nov 26 08:28:30 2008 From: darose at darose.net (David Rosenstrauch) Date: Wed, 26 Nov 2008 11:28:30 -0500 Subject: [svlug] Best value Quad core sweet spot, holiday sales, processor mboard overclock retailers In-Reply-To: <1227709202.9650.1286881415@webmail.messagingengine.com> References: <1227688162.8417.1286833687@webmail.messagingengine.com> <492D4280.50902@darose.net> <1227709202.9650.1286881415@webmail.messagingengine.com> Message-ID: <492D792E.1060006@darose.net> john_re wrote: > Thnaks for the Q6600 data. :) > > 0) Why did you choose the Q6600 vs amd, > 0b) why the q6600 vs other intel processors? Saw it recommended for good balance of price/performance in a Tom's Hardware article. Never really considered the AMD this time around, mostly due to reputation. AMD had a good rep circa the Athlon chip, but most of the top hardware sites say the advantage has rather definitively swung back to Intel with the Core 2. > 1) you don't overclock, correct? Considered/planning overclocking? No o/c - now or planned. > 2) what OS do you use? Currently, WinXP. (Mostly because this is our main desktop and the wife needs Win, MS Word, etc.) Eventually this machine will get converted become my server though, at which point it'll run Linux (specifically, Arch, my favorite distro) - as happened to the last desktop I built once I upgraded to this machine. > 3) what do you use the system for? games, home, engineering, SW dev, > business? Our main home desktop (web surfing, office stuff). But I also do software dev. on it. > 4) what MSI MB would you recommend? Why? Can't recall specifically. Go to the MSI site and look for their boards for the P43 and P45 chipset, which are the competitors to the Gigabyte board I bought. > 5) why do you like newegg & zipzoomfly? Cheap, fast, reliable, good web sites. Newegg, in particular, since their site does a great job of showing customer review ratings and comments for each product. > 6) How are the Radeon drivers? I've not run into any problems yet, but a) I'm not at all a gamer, and b) as I said, this is a Win box, so I'm using the Win drivers. HTH, DR From rick at linuxmafia.com Wed Nov 26 10:30:52 2008 From: rick at linuxmafia.com (Rick Moen) Date: Wed, 26 Nov 2008 10:30:52 -0800 Subject: [svlug] Best value Quad core sweet spot, holiday sales, processor mboard overclock retailers In-Reply-To: <20081126135940.GG4980@Isis> References: <1227688162.8417.1286833687@webmail.messagingengine.com> <20081126135940.GG4980@Isis> Message-ID: <20081126183051.GB832@linuxmafia.com> Quoting Mark S Bilk (mark at cosmicpenguin.com): > Finally, I would not use an Intel processor, because that company > has written the firmware for Microsoft's "Palladium" system for > total government censorship of the Internet. What's interesting, here, is not the fact that this is bullshit, but rather _why_ it is bullshit. TPM chips implemented in general-purpose computers (and, by the way, your boycott list should logically also include Atmel, Broadcom, Infineon, Sinosun, STMicroelectronics, and NuvoTon (ex-Winbond)) have not prevented the computer in question from running unauthenticated operating systems and applications -- such as, for example, Linux. When running such a system, the TPM's functions, which are to decrypt data with its TPM endorsement key when the OS and application have validated themselves, are inaccessible by design, even locally. (Making them accessible under an unsigned OS would defeat the purpose of the hardware.) If general-purpose computing hardware emerges that refuses to boot OSes that don't satisfy its crypto-hash requirements, _that_ would be unacceptable -- not on account of "censorship of the Internet" exactly, but rather dictating of allowed operating system and applications. You can experience that particular joy by buying an iPhone. Obvious application of TPM chips _with_ authenticating operating systems providing a matching NEXUS layer, etc., would be by Our Lords in Hollywood, and similar. (There's also proprietary full-disk encryption, et al.) Schneier on TCPA 1.2 in year 2002: http://www.schneier.com/crypto-gram-0208.html#1 ObSVLUG reference: Seth Schoen gave SVLUG a rather scholarly lecture about (in part) Microsoft's Palladium software and TCPA, in 2003. http://lists.svlug.org/pipermail/svlug-announce/2003-February/000101.html From lsc at prgmr.com Wed Nov 26 11:43:04 2008 From: lsc at prgmr.com (Luke S Crawford) Date: 26 Nov 2008 14:43:04 -0500 Subject: [svlug] Best value Quad core sweet spot, holiday sales, processor mboard overclock retailers In-Reply-To: <1227709202.9650.1286881415@webmail.messagingengine.com> References: <1227688162.8417.1286833687@webmail.messagingengine.com> <492D4280.50902@darose.net> <1227709202.9650.1286881415@webmail.messagingengine.com> Message-ID: "john_re" writes: > Thnaks for the Q6600 data. :) A word of warning on the q6600: they run hot (meaning they draw lots of power) I used q6600s for my xen hosting setup for a while, and I moved off them, mostly because they are pigs when it comes to power, and power is ridicioulously expensive at my co-lo. 135w tdp. compare that to 50w tdp for the low power xeons. So yeah... careful with the q6600. It's probably fine for a desktop, but in a power-constrained environment, you are better off with a low-power xeon, if you have the cash and need the CPU power, or a celeron -L if you don't need CPU power. (35w TDP for the celleron-L) AMD rates its power misleadingly... they use 'average' power or something... the low-power quad-core opterons have a 55w rating, even though their TDP is somewhat higher. So if I had more capital I'd be buying Xeons, partly 'cause they have much better power usage, and partly because I find misleading marketing irritating. I went from a single core2quad in one of these mb/chassis combos: http://serverconfigurator.intel.com/details.aspx?id=1334 with 8GB unbuffered ecc, to a supermicro 1u twin: http://supermicro.com/Aplus/system/1U/1021/AS-1021TM-T+.cfm with 2x1.9Ghz quad-core opteron HEs and 32GB ram in each motherboard. illustration: http://prgmr.com/~lsc/luke_opterons.jpg the total cost per gigabyte is much lower, and it has hot-swap drives. (never again will I buy a case without hot-swap drive bays for production. that cost me my SLA, and a months revinue on a server. Ouch) but you aren't going to find one of those at the 'black friday' sales, and it's so loud I can hear it in the garage from my bedroom (It will be at the co-lo pretty soon.) I did find an awesomely good deal at newegg on tyan brand opteron server boards: http://www.newegg.com/Product/Product.aspx?Item=N82E16813151070 (there was some deal where you buy that board and get $40 off the 1.9Ghz opteron HE cpus I needed for my 1u twin monster, so I have 3 of those puppies sitting out in the garage right now... as far as I am concerned, I paid $20 plus shipping for them) you flash them with the latest tyan bios http://book.xen.prgmr.com/mediawiki/index.php/Booting_DOS_from_a_USB_thumbdrive (oy, that took way longer than it should have) and they support quad-core CPUs and the super-cheap quad-rank memory modules. I've got one of those puppies with 32G ram and 2x1.8Ghz quad-core opteron HE CPUs running right now. It's been solid. The motherboard even has a POST diagnostic LED on it. Oh, and a tip on ram prices. Go to kingston.com. click the 'shop' link after you find your ram, they almost always have a discount going, but they don't show the discount until after you click 'shop' I bought 64G worth of 2GB modules of registered ecc ddr2 667. each pair of two set me back like $76, I think. Also, they sent me a bunch of free thumbdrives. (you can only order 4 at a time, and they give you a thumbdrive with every order.) From lordsauronthegreat at gmail.com Wed Nov 26 12:03:11 2008 From: lordsauronthegreat at gmail.com (Chris Miller) Date: Wed, 26 Nov 2008 12:03:11 -0800 Subject: [svlug] Best value Quad core sweet spot, holiday sales, processor mboard overclock retailers In-Reply-To: <20081126135940.GG4980@Isis> References: <1227688162.8417.1286833687@webmail.messagingengine.com> <20081126135940.GG4980@Isis> Message-ID: On Wed, Nov 26, 2008 at 5:59 AM, Mark S Bilk wrote: > Overclocking is a bad idea. Chips are specified with a particular > clock rate because above that rate errors and/or chip damage may > occur. The same goes for using a power voltage beyond the spec. > And since the increase in speed is barely noticeable, there really > isn't any reason to do it. There are a few chips (the Intel Core 2 Duo Extreme Editions) which are shipped with an unlocked clock multiplier. Depending on the hardware, you can get away with a significant amount of overclocking. I've heard of gamers overclocking a 3.2 GhZ chip to 4.00 GhZ with reasonable levels of success (given that they have a freakishly massive fan attached to their new heat engine, otherwise they experience what I like to call Microchernobyl syndrome). I wouldn't recommend it per the capacity for memory timing errors. When a gamer's machine randomly reboots (al la memory timing error due to overclocking) it's a stream of cuss words and otherwise the world keeps on turning. In my experience Linux users place a much higher premium on stability. Having your system randomly quit does not seem like it would endear itself to the Linux-user's mindset. > As to quad core, my dual core system very rarely uses the second > core (as shown by ksysguard). I'd be interested in reports from > others on this point. I have noticed that Windows better spreads single-process jobs across the cores better on my dual-core, but I haven't paid too much attention to notice. > I've always used ASUS motherboards and found them very reliable. I've had large amounts of success with ECS Elitegroup boards. ATM I'm using an Abit board with great success. Of course the layout kind of sucks... bought this beastly NVIDIA GeForce 9800GTX+ to drive my visual needs, and it covers three of six SATA300 ports! What's up with that! If I got another for SLI, I'd have no more SATA ports to use! YIKES! > total government censorship of the Internet. Intel has also > engaged in dishonest and coercive business practices against AMD. They did, but now they're beating the hell out of my poor AMD in a more or less fair fight. AMD still has one trick up their sleeve: their chips are hundreds of dollars cheaper, and still deliver within ~17% or so of their Intel counterpart's performance. That's a hundred dollars or more you could donate to free software or get yourself a bigger hard drive or more RAM or mail it off to Microsoft for no apparent reason. ;-) -- Registered Linux Addict #431495 http://profile.xfire.com/mrstalinman | John 3:16! http://www.fsdev.net/ | http://lordsauron.wordpress.com/ From lindahl at pbm.com Wed Nov 26 18:20:40 2008 From: lindahl at pbm.com (Greg Lindahl) Date: Wed, 26 Nov 2008 18:20:40 -0800 Subject: [svlug] Sync command slowdown In-Reply-To: <200811251555.48353.james@linuxrebel.org> References: <200811241732.40649.james@linuxrebel.org> <200811251555.48353.james@linuxrebel.org> Message-ID: <20081127022039.GG13643@bx9> On Tue, Nov 25, 2008 at 03:55:48PM -0800, James Sparenberg wrote: > Well 24 hours later (so that we could monitor) and the conclusion seem > to be that yes it does appear to be related somehow to the ext3 FS and > no changing to writeback doesn't improve the speed. It seems that the > problem has to do with the size of the ext3 FS, in that over some (as > yet) unknow size there is a problem with the speed of the sync command. You never posted anything about my suggestion of checking the number of Dirty pages in /proc/meminfo. That should have been a big clue. -- greg From lindahl at pbm.com Wed Nov 26 18:22:57 2008 From: lindahl at pbm.com (Greg Lindahl) Date: Wed, 26 Nov 2008 18:22:57 -0800 Subject: [svlug] Best value Quad core sweet spot, holiday sales, processor mboard overclock retailers In-Reply-To: References: <1227688162.8417.1286833687@webmail.messagingengine.com> <492D4280.50902@darose.net> <1227709202.9650.1286881415@webmail.messagingengine.com> Message-ID: <20081127022257.GH13643@bx9> On Wed, Nov 26, 2008 at 02:43:04PM -0500, Luke S Crawford wrote: > AMD rates its power misleadingly... they use 'average' power or something... > the low-power quad-core opterons have a 55w rating, even though their TDP > is somewhat higher. So if I had more capital I'd be buying Xeons, > partly 'cause they have much better power usage, and partly because I find > misleading marketing irritating. I found it misleading that Intel trumpeted the efficiency of the Core2 and then paired it with FBDIMMs in servers, which eat power. YMMV. The AMD servers that I have (normal Opterons, DDR2) use less power than similar-performance Intel servers (low power quad-cores, FB-DIMMs.) -- greg From lordsauronthegreat at gmail.com Wed Nov 26 18:40:35 2008 From: lordsauronthegreat at gmail.com (Chris Miller) Date: Wed, 26 Nov 2008 18:40:35 -0800 Subject: [svlug] Best value Quad core sweet spot, holiday sales, processor mboard overclock retailers In-Reply-To: <20081127022257.GH13643@bx9> References: <1227688162.8417.1286833687@webmail.messagingengine.com> <492D4280.50902@darose.net> <1227709202.9650.1286881415@webmail.messagingengine.com> <20081127022257.GH13643@bx9> Message-ID: On Wed, Nov 26, 2008 at 6:22 PM, Greg Lindahl wrote: > On Wed, Nov 26, 2008 at 02:43:04PM -0500, Luke S Crawford wrote: > >> AMD rates its power misleadingly... they use 'average' power or something... >> the low-power quad-core opterons have a 55w rating, even though their TDP >> is somewhat higher. So if I had more capital I'd be buying Xeons, >> partly 'cause they have much better power usage, and partly because I find >> misleading marketing irritating. > > I found it misleading that Intel trumpeted the efficiency of the Core2 > and then paired it with FBDIMMs in servers, which eat power. YMMV. The > AMD servers that I have (normal Opterons, DDR2) use less power than > similar-performance Intel servers (low power quad-cores, FB-DIMMs.) I honestly didn't think anyone here still made the mistake of listening to marketing. I just go straight for the third-party evaluations. -- Registered Linux Addict #431495 http://profile.xfire.com/mrstalinman | John 3:16! http://www.fsdev.net/ | http://lordsauron.wordpress.com/ From james at linuxrebel.org Wed Nov 26 18:47:10 2008 From: james at linuxrebel.org (James Sparenberg) Date: Wed, 26 Nov 2008 18:47:10 -0800 Subject: [svlug] Sync command slowdown In-Reply-To: <20081127022039.GG13643@bx9> References: <200811241732.40649.james@linuxrebel.org> <200811251555.48353.james@linuxrebel.org> <20081127022039.GG13643@bx9> Message-ID: <200811261847.10182.james@linuxrebel.org> On Wednesday 26 November 2008 18:20:40 Greg Lindahl wrote: > On Tue, Nov 25, 2008 at 03:55:48PM -0800, James Sparenberg wrote: > > > Well 24 hours later (so that we could monitor) and the conclusion seem > > to be that yes it does appear to be related somehow to the ext3 FS and > > no changing to writeback doesn't improve the speed. It seems that the > > problem has to do with the size of the ext3 FS, in that over some (as > > yet) unknow size there is a problem with the speed of the sync command. > > You never posted anything about my suggestion of checking the number of > Dirty pages in /proc/meminfo. That should have been a big clue. > > -- greg > > Your right I didn't and it's sitting right now on one of the systems with 2232 dirty pages. and a few seconds later it's showing 2200 James From lindahl at pbm.com Wed Nov 26 19:20:43 2008 From: lindahl at pbm.com (Greg Lindahl) Date: Wed, 26 Nov 2008 19:20:43 -0800 Subject: [svlug] Sync command slowdown In-Reply-To: <200811261847.10182.james@linuxrebel.org> References: <200811241732.40649.james@linuxrebel.org> <200811251555.48353.james@linuxrebel.org> <20081127022039.GG13643@bx9> <200811261847.10182.james@linuxrebel.org> Message-ID: <20081127032043.GB19757@bx9> On Wed, Nov 26, 2008 at 06:47:10PM -0800, James Sparenberg wrote: > Your right I didn't and it's sitting right now on one of the systems > with 2232 dirty pages. and a few seconds later it's showing 2200 That's a lot -- 90 megabytes. You might be able to write out 50 MB/s to a single drive, if the pages were contiguous. If they're scattered all over the drive, it could take much longer. Dirty pages get written according to age, or exceeding a % of memory, or if free pages are low. If these new servers have a lot more memory than the old ones, then it's easy to see why a sync could take much more time than before. If I recall your original posting, it was never clear to me why you were doing the sync, and why you cared how fast it is. If your logrotate script does it once a day, then you can simply remove that sync, plus you shouldn't care if it's slow. -- greg From lindahl at pbm.com Wed Nov 26 19:45:56 2008 From: lindahl at pbm.com (Greg Lindahl) Date: Wed, 26 Nov 2008 19:45:56 -0800 Subject: [svlug] Sync command slowdown In-Reply-To: <20081127032043.GB19757@bx9> References: <200811241732.40649.james@linuxrebel.org> <200811251555.48353.james@linuxrebel.org> <20081127022039.GG13643@bx9> <200811261847.10182.james@linuxrebel.org> <20081127032043.GB19757@bx9> Message-ID: <20081127034556.GA21444@bx9> On Wed, Nov 26, 2008 at 07:20:43PM -0800, Greg Lindahl wrote: > That's a lot -- 90 megabytes. It would help if I could multiply. 2200 x 4000 = 9 megabytes. Still, if they're scattered over the disk, it could take 2,000 seeks at 6ms each to write those pages out. -- greg From shaeffer at neuralscape.com Wed Nov 26 20:30:10 2008 From: shaeffer at neuralscape.com (Karen Shaeffer) Date: Wed, 26 Nov 2008 20:30:10 -0800 Subject: [svlug] Sync command slowdown In-Reply-To: <20081127032043.GB19757@bx9> References: <200811241732.40649.james@linuxrebel.org> <200811251555.48353.james@linuxrebel.org> <20081127022039.GG13643@bx9> <200811261847.10182.james@linuxrebel.org> <20081127032043.GB19757@bx9> Message-ID: <20081127043010.GA12524@synapse.neuralscape.com> On Wed, Nov 26, 2008 at 07:20:43PM -0800, Greg Lindahl wrote: > On Wed, Nov 26, 2008 at 06:47:10PM -0800, James Sparenberg wrote: > > > Your right I didn't and it's sitting right now on one of the systems > > with 2232 dirty pages. and a few seconds later it's showing 2200 > > That's a lot -- 90 megabytes. You might be able to write out 50 MB/s > to a single drive, if the pages were contiguous. If they're scattered > all over the drive, it could take much longer. > > Dirty pages get written according to age, or exceeding a % of memory, > or if free pages are low. If these new servers have a lot more memory > than the old ones, then it's easy to see why a sync could take much > more time than before. Hi, Another issue to consider. In the block layer, the IO scheduler has a strong preference for completing reads before writes. If there is a backlog of read requests, then the time to completion for writes will go way up. Thanks, Karen -- Karen Shaeffer Neuralscape, Palo Alto, Ca. 94306 shaeffer at neuralscape.com http://www.neuralscape.com From dmarti at zgp.org Wed Nov 26 22:29:32 2008 From: dmarti at zgp.org (Don Marti) Date: Wed, 26 Nov 2008 22:29:32 -0800 Subject: [svlug] Best value Quad core sweet spot, holiday sales, processor mboard overclock retailers In-Reply-To: <20081126183051.GB832@linuxmafia.com> References: <1227688162.8417.1286833687@webmail.messagingengine.com> <20081126135940.GG4980@Isis> <20081126183051.GB832@linuxmafia.com> Message-ID: <20081127062932.GA26206@zgp.org> begin Rick Moen quotation of Wed, Nov 26, 2008 at 10:30:52AM -0800: > Schneier on TCPA 1.2 in year 2002: > http://www.schneier.com/crypto-gram-0208.html#1 > > ObSVLUG reference: Seth Schoen gave SVLUG a rather scholarly lecture > about (in part) Microsoft's Palladium software and TCPA, in 2003. > http://lists.svlug.org/pipermail/svlug-announce/2003-February/000101.html A couple more: "Why TCPA" by David Safford http://domino.research.ibm.com/comm/research_projects.nsf/pages/gsal.TCG.html/$FILE/why_tcpa.pdf "The TCPA chip is not particularly suited to DRM. While it does have the ability to report signed PCR information, and this information could be used to prevent playback unless a trusted operating system and application were in use, this type of scheme would be a nightmare for content providers to manage. Any change to the BIOS, the operating system, or the application would change the reported values." "The bottom line is that the physical owner of the machine could easily recover any DRM secrets from the chip....we simply are not concerned with threats based on the user attacking the chip." "Take Control of TCPA" by David Safford, Jeff Kravitz and Leendert van Doorn http://www.linuxjournal.com/article/6633 -- Don Marti +1 510-814-0932 http://zgp.org/~dmarti/ +1 510-332-1587 mobile dmarti at zgp.org From james at linuxrebel.org Sat Nov 29 20:40:11 2008 From: james at linuxrebel.org (James Sparenberg) Date: Sat, 29 Nov 2008 20:40:11 -0800 Subject: [svlug] Sync command slowdown In-Reply-To: <20081127043010.GA12524@synapse.neuralscape.com> References: <200811241732.40649.james@linuxrebel.org> <20081127032043.GB19757@bx9> <20081127043010.GA12524@synapse.neuralscape.com> Message-ID: <200811292040.11422.james@linuxrebel.org> On Wednesday 26 November 2008 20:30:10 Karen Shaeffer wrote: > On Wed, Nov 26, 2008 at 07:20:43PM -0800, Greg Lindahl wrote: > > On Wed, Nov 26, 2008 at 06:47:10PM -0800, James Sparenberg wrote: > > > > > Your right I didn't and it's sitting right now on one of the systems > > > with 2232 dirty pages. and a few seconds later it's showing 2200 > > > > That's a lot -- 90 megabytes. You might be able to write out 50 MB/s > > to a single drive, if the pages were contiguous. If they're scattered > > all over the drive, it could take much longer. > > > > Dirty pages get written according to age, or exceeding a % of memory, > > or if free pages are low. If these new servers have a lot more memory > > than the old ones, then it's easy to see why a sync could take much > > more time than before. > > Hi, > Another issue to consider. In the block layer, the IO scheduler has a > strong preference for completing reads before writes. If there is a > backlog of read requests, then the time to completion for writes will > go way up. > > Thanks, > Karen > -- Karen, Thanks this makes sense I just need now to figure out how to work around this. James From james at linuxrebel.org Sat Nov 29 21:12:02 2008 From: james at linuxrebel.org (James Sparenberg) Date: Sat, 29 Nov 2008 21:12:02 -0800 Subject: [svlug] Sync command slowdown In-Reply-To: <20081127034556.GA21444@bx9> References: <200811241732.40649.james@linuxrebel.org> <20081127032043.GB19757@bx9> <20081127034556.GA21444@bx9> Message-ID: <200811292112.02867.james@linuxrebel.org> On Wednesday 26 November 2008 19:45:56 Greg Lindahl wrote: > On Wed, Nov 26, 2008 at 07:20:43PM -0800, Greg Lindahl wrote: > > > That's a lot -- 90 megabytes. > > It would help if I could multiply. 2200 x 4000 = 9 megabytes. > > Still, if they're scattered over the disk, it could take 2,000 seeks > at 6ms each to write those pages out. > > -- greg Greg, A couple of points. Sync is simply my test mule. MySQL doesn't use logrotate, (Linux version) but instead uses fsync(sync) to sync it's binary logs before it decides to rotate them itself. What I have determined is a few things. 1. That a tiny (10mb) log on a Large file system is fast (Large defined as +100GB partition single spindle) 2. That a small or large log on a small partition (<100GB) is a fast rotate. 3. That a larger log file (btw defined as >10MB) is extremely slow and at 500MB the log file can end up locking the file system for as much as 16 seconds. (range 14-16 seconds) This does not make my DBA's at all happy let me assure you :) 4. Currently our DB sits at around 30GB of raw data and we currently grow at a rate of around 10GB every 3 months with an additional 40-60GB of support (in the form of logs.) I'm no big fan of ext3 myself (as I've had the pain of forced by OS despite configs saying no, fsck of a 3TB file system that carped all over itself). However since it is the only option other than ext2 offered by RH in default install I need to make it sing. I've run it with an without writeback, with and without noatime (RH's mount doesn't understand relatime) and so far it does seem that the ext3 file system on RHEL5.2 breaks down when file systems get large enough. BTW I have a DB server in our office running RHEL5.2, with a smaller file system and it can easily sync. To demonstrate here are some numbers you might find interesting large file system box (114GB) dirty listed as 1836 #time sync real 0m3.065s user 0m0.000s sys 0m0.005s small file system box (80GB) dirty listed as 2940 #time sync real 0m0.076s user 0m0.000s sys 0m0.006s James Again to all, thanks for the help, it really is appreciated. From john_re at fastmail.us Sun Nov 30 01:02:30 2008 From: john_re at fastmail.us (john_re) Date: Sun, 30 Nov 2008 01:02:30 -0800 Subject: [svlug] Do you overclock any hw working under linux? Message-ID: <1228035750.21684.1287414259@webmail.messagingengine.com> If so, what equipment, what parameters, what distro? From lordsauronthegreat at gmail.com Sun Nov 30 09:29:14 2008 From: lordsauronthegreat at gmail.com (Chris Miller) Date: Sun, 30 Nov 2008 09:29:14 -0800 Subject: [svlug] Do you overclock any hw working under linux? In-Reply-To: <1228035750.21684.1287414259@webmail.messagingengine.com> References: <1228035750.21684.1287414259@webmail.messagingengine.com> Message-ID: On Sun, Nov 30, 2008 at 1:02 AM, john_re wrote: > If so, what equipment, what parameters, what distro? Absolutely not. The memory timing errors could potentially ruin all the stability I migrated to Linux for. -- Registered Linux Addict #431495 http://profile.xfire.com/mrstalinman | John 3:16! http://www.fsdev.net/ | http://lordsauron.wordpress.com/ From lordsauronthegreat at gmail.com Sun Nov 30 09:29:14 2008 From: lordsauronthegreat at gmail.com (Chris Miller) Date: Sun, 30 Nov 2008 09:29:14 -0800 Subject: [svlug] Do you overclock any hw working under linux? In-Reply-To: <1228035750.21684.1287414259@webmail.messagingengine.com> References: <1228035750.21684.1287414259@webmail.messagingengine.com> Message-ID: On Sun, Nov 30, 2008 at 1:02 AM, john_re wrote: > If so, what equipment, what parameters, what distro? Absolutely not. The memory timing errors could potentially ruin all the stability I migrated to Linux for. -- Registered Linux Addict #431495 http://profile.xfire.com/mrstalinman | John 3:16! http://www.fsdev.net/ | http://lordsauron.wordpress.com/ From einfeldt at gmail.com Sun Nov 30 21:51:31 2008 From: einfeldt at gmail.com (Christian Einfeldt) Date: Sun, 30 Nov 2008 21:51:31 -0800 Subject: [svlug] Save the date! Sat Dec 6 tech fair at St. Anthony Message-ID: <4b5781040811302151n55bb51c2rcef51e044ad4c345@mail.gmail.com> Please save Saturday, Dec 6 from 9 am to 2 pm for the St. Anthony tech fair, to be held at 150 Golden Gate at Jones in SF. This event will be pretty much the same as past events held at St. Anthony. We will be demonstrating Linux and installing Linux on hardware brought in by low income clients and we will be giving away distros. This is a great opportunity to embed ourselves and Linux in one of the most well-run, well-financed charities in Northern California, and reach out to low income individuals and non-profits. -------------- next part -------------- An HTML attachment was scrubbed... URL: http://lists.svlug.org/archives/svlug/attachments/20081130/77305f94/attachment.htm