DB-14: Tales of the Bunker - 2005 Gus Björklund, Progress Software Corporation John Harlow, Bravepoint, Inc. Dan Foreman, Bravepoint, Inc. Rich Banville,

Slides:



Advertisements
Similar presentations
Hardware Lesson 3 Inside your computer.
Advertisements

Skyward Server Design Mike Bianco.
Tag line, tag line Perforce Benchmark with PAM over NFS, FCP & iSCSI Bikash R. Choudhury.
DB-7: OpenEdge® Database Performance Tuning
DB-03: A Tour of the OpenEdge™ RDBMS Storage Architecture Richard Banville Technical Fellow.
Query Processing and Optimizing on SSDs Flash Group Qingling Cao
Big Data Working with Terabytes in SQL Server Andrew Novick
T OP N P ERFORMANCE T IPS Adam Backman Partner, White Star Software.
Database Storage for Dummies Adam Backman President – White Star Software, LLC.
Numbers, We don’t need no stinkin’ numbers Adam Backman Vice President DBAppraise, Llc.
Copyright GeneGo CONFIDENTIAL »« MetaCore TM (System requirements and installation) Systems Biology for Drug Discovery.
1 How Healthy is Your Progress System? ( Progess DB Best Practices) Dan Foreman BravePoint, Inc.
1 Magnetic Disks 1956: IBM (RAMAC) first disk drive 5 Mb – Mb/in $/year 9 Kb/sec 1980: SEAGATE first 5.25’’ disk drive 5 Mb – 1.96 Mb/in2 625.
Best Practice for Performance Edward M. Kwang President.
28. Januar, Zürich-Oerlikon. Switch/Update to Team Foundation Server 2012 André Hofmann Software Engineer bbv Software Services AG.
Cacti Workshop Tony Roman Agenda What is Cacti? The Origins of Cacti Large Installation Considerations Automation The Current.
Novell Server Linux vs. windows server 2008 By: Gabe Miller.
By Mr. Abdalla A. Shaame 1. What is Computer An electronic device that stores, retrieves, and processes data, and can be programmed with instructions.
OpenEdge Replication Made Easy Adam Backman White Star Software
Tripwire Enterprise Server – Getting Started Doreen Meyer and Vincent Fox UC Davis, Information and Education Technology June 6, 2006.
Presented by Jacob Wilson SharePoint Practice Lead Bross Group 1.
VMware vCenter Server Module 4.
Mark Nesson June, 2008 Fine Tuning WebFOCUS for the IBM Mainframe (zSeries, System z9)
NovaBACKUP 10 xSP Technical Training By: Nathan Fouarge
MOVE-4: Upgrading Your Database to OpenEdge® 10 Gus Björklund Wizard, Vice President Technology.
MCITP Administrator: Microsoft SQL Server 2005 Database Server Infrastructure Design Study Guide (70-443) Chapter 1: Designing the Hardware and Software.
1 The Virtual Reality Virtualization both inside and outside of the cloud Mike Furgal Director – Managed Database Services BravePoint.
Birth, Death, Infinity Gus Björklund. Progress. Dan Foreman. BravePoint. PUG Challenge Americas, 9-12 June 2013.
Database Storage Considerations Adam Backman White Star Software DB-05:
Introduction to HP LoadRunner Getting Familiar with LoadRunner >>>>>>>>>>>>>>>>>>>>>>
Berlin SPARQL Benchmark (BSBM) Presented by: Nikhil Rajguru Christian Bizer and Andreas Schultz.
Chapter 4 COB 204. What do you need to know about hardware? 
1 Progress Software’s OpenEdge Platform Which database is right for your environment? Simon Epps.
SOA 16: Computing in the Cloud – Is it in your future? Dr. Sheldon Borkin Gus Björklund Vice Presidents, Technology I've looked at clouds from both sides.
Top Performance Enhancers Top Performance Killers in Progress Dan Foreman Progress Expert
© 2015 Progress Software Corporation. 1 abstract To paraphrase Benjamin Disraeli, there are lies, damned lies and benchmarks. Your intrepid band of benchmarkers.
 For Oracle Based Products  Windows XP Professional-SP3 / Window7-SP1 (32 Bit)  Min. Dual Processor or above  RAM 2GB or above  Internet explorer.
Profiling Grid Data Transfer Protocols and Servers George Kola, Tevfik Kosar and Miron Livny University of Wisconsin-Madison USA.
Top 10 Performance Hints Adam Backman White Star Software
Strength. Strategy. Stability.. Progress Performance Monitoring and Tuning Dan Foreman Progress Expert BravePoint BravePoint
Goodbye rows and tables, hello documents and collections.
© 2007 Hewlett-Packard Development Company, L.P. The information contained herein is subject to change without notice BAC 7.x Sizing April 2008.
PROGRESS S O F T W A R E Linux And The Progress RDBMS Gus Björklund Wizard Progress Software PUG Challenge 2002, Veldhoven, the Netherlands.
Block1 Wrapping Your Nugget Around Distributed Processing.
Kaizilege Karoma Storage Capacity + What to buy Binary Did you know that all of the information that travels through your computer is based on two commands?
CMAQ Runtime Performance as Affected by Number of Processors and NFS Writes Patricia A. Bresnahan, a * Ahmed Ibrahim b, Jesse Bash a and David Miller a.
Amy Apon, Pawel Wolinski, Dennis Reed Greg Amerson, Prathima Gorjala University of Arkansas Commercial Applications of High Performance Computing Massive.
DONE-08 Sizing and Performance Tuning N-Tier Applications Mike Furgal Performance Manager Progress Software
Progress Database Admin 1 Jeffrey A. Brown - Technical Support Consultant
© 2008 Quest Software, Inc. ALL RIGHTS RESERVED. Perfmon and Profiler 101.
Achieving Scalability, Performance and Availability on Linux with Oracle 9iR2-RAC Grant McAlister Senior Database Engineer Amazon.com Paper
Computer Guts and Operating Systems CSCI 101 Week Two.
Click here to download this powerpoint template : Colorful Networks Free Powerpoint TemplateColorful Networks Free Powerpoint Template For more : Powerpoint.
CS Operating System & Database Performance Tuning Xiaofang Zhou School of Computing, NUS Office: S URL:
1.00 Computer Basics 1.01 Classify computer component. 1 Primary sources of information and images: GCFLearnFree.org, Microsoft Digital Learning E-Learning.
Arne Wiebalck -- VM Performance: I/O
Infrastructure for Data Warehouses. Basics Of Data Access Data Store Machine Memory Buffer Memory Cache Data Store Buffer Bus Structure.
© 2015 Progress Software Corporation. 1 abstract Your intrepid band of benchmarkers returns once more, with the results of testing a recent release of.
1 Benchmarking Cloud Serving Systems with YCSB Brian F. Cooper, Adam Silberstein, Erwin Tam, Raghu Ramakrishnan and Russell Sears Yahoo! Research.
Copyright Sammamish Software Services All rights reserved. 1 Prog 140  SQL Server Performance Monitoring and Tuning.
What to look for?  Consider whether you should purchase a new or used computer.  Many people who sell computers before they are out of date allow you.
Tales of the Secret Bunker 2016 (231) Dump and Load Edition Mike Furgal – Director MDBA and Pro2 Services Gus Bjorklund - Lackey.
Information Technology (IT). Information Technology – technology used to create, store, exchange, and use information in its various forms (business data,
Understanding and Improving Server Performance
How Social is your data Fundamentals on database storage
We Have Found Nirvana with Online Dump and Load (224)
Computer Organization
Installation and database instance essentials
Software Architecture in Practice
Walking Through A Database Health Check
Presentation transcript:

DB-14: Tales of the Bunker Gus Björklund, Progress Software Corporation John Harlow, Bravepoint, Inc. Dan Foreman, Bravepoint, Inc. Rich Banville, Progress Software Corporation

© 2006 Progress Software Corporation3 DB-14 Tales of the Bunker 2005 Goals of the Bunker Test  Find the optimal way to run Progress® on Linux  Test various ideas and theories  Have fun  Bunker 2005: Pre-release 10.1A 64-bit AMD Performance of Utilities Investigate “coma” problem Network speed effects

© 2006 Progress Software Corporation4 DB-14 Tales of the Bunker 2005 Why Linux?

© 2006 Progress Software Corporation5 DB-14 Tales of the Bunker 2005 Bunker 2005 Team  Gus Björklund, Wizard, Progress Software Progress User since 1989  John Harlow, President of BravePoint Progress User since 1984  Dan Foreman Progress User since 1984  Rich Banville, Fellow, Progress Software Progress User since 1993

© 2006 Progress Software Corporation6 DB-14 Tales of the Bunker 2005 The ATM Benchmark Environment

© 2006 Progress Software Corporation7 DB-14 Tales of the Bunker 2005 The ATM Benchmark  Simulates teller machine transactions deposit or withdrawal heavy database update workload  Each transaction does 3 fetches, 3 updates, 1 create –retrieve and update account, branch, and teller rows –create a history row  Run “n” transaction generators concurrently for fixed time period count total number of transactions performed

© 2006 Progress Software Corporation8 DB-14 Tales of the Bunker 2005 Test Database (logical) TableNumber of Rows Account80,000,000 (100 bytes each) Teller80,000 (100 bytes each) Branch8,000 (100 bytes each) Historyadd 1 per transaction (50 bytes each)

© 2006 Progress Software Corporation9 DB-14 Tales of the Bunker 2005 Test Database (physical) Total size12 gigabytes (data) Data extent size2,000,000 (2 GB) Data extent count6 Data block size4096 bytes Rows per block64 Data Areas1, Type II Data cluster size512 blocks (2 megabytes)

© 2006 Progress Software Corporation10 DB-14 Tales of the Bunker 2005 Test Database (other info) Account table data9.2 gigabytes Branch table data922 kilobytes Teller table data9.3 megabytes Indexes691 megabytes RM blocks2,669,712 Index blocks178,463 Free blocks150,682

© 2006 Progress Software Corporation11 DB-14 Tales of the Bunker 2005 Equipment

© 2006 Progress Software Corporation12 DB-14 Tales of the Bunker 2005 Server 1: hostname “uniblab” Operating SystemSuSE Linux Enterprise Server 9 Cpus2 x 2.8 GHz Intel Xeon MotherboardIBM xSeries Memory4 gigabytes Disk drives on controller6 x n GB 10,000 rpm USCSI

© 2006 Progress Software Corporation13 DB-14 Tales of the Bunker 2005 Server 3: hostname “hal” Operating SystemSuSE Linux 10.0 Cpus2 x 2.4 GHz Intel Xeon Memory2 gigabytes Disk drives on motherboard1 x n GB 7200 rpm SATA Disk drives on controller6 x n GB 7200 rpm SATA

© 2006 Progress Software Corporation14 DB-14 Tales of the Bunker 2005 Server 3: hostname “jumbo” Operating SystemSuSE Linux 10.0 Cpus2 x AMD 64-bit 2.0 GHz Dual Core MotherboardAsus K8N-DL Disk controller3Ware 9500S-8 Memory6 gigabytes Disk drives on motherboard2 x 160 GB Maxtor 7200 rpm Ultra IDE Disk drives on controller8 x 164 GB Hitachi Deskstar 7200 rpm SATA II

© 2006 Progress Software Corporation15 DB-14 Tales of the Bunker 2005 AMD-64 Hyper-Transport Design

© 2006 Progress Software Corporation16 DB-14 Tales of the Bunker 2005 Bunker Network Components  3 server machines  Netgear GS 105 Gigabit switch  SMC Barricade WAP  LinkSys WVC11b “bunker cam”  Various laptops running Windoze Linux Mac OS X 10.3

© 2006 Progress Software Corporation17 DB-14 Tales of the Bunker 2005 Bunker Network (partial) haluniblabjumbo gigabit switch wapbunker Cam Internet router laptops servers

© 2006 Progress Software Corporation18 DB-14 Tales of the Bunker 2005 Other Equipment

© 2006 Progress Software Corporation19 DB-14 Tales of the Bunker 2005 Past Results

© 2006 Progress Software Corporation20 DB-14 Tales of the Bunker 2005 Lessons from Past Bunkers  Type II Data Areas are faster  Don’t use Reiser File System  Use EXT3 or XFS File System  Don’t use the Anticipatory Scheduler Deadline or CFQ is better  2.6 Kernel is faster than 2.4 Kernel  For RAID 10, the Largest Possible Stripe Size was always the fastest, both Software & Hardware Striping  Very good performance at low cost

© 2006 Progress Software Corporation21 DB-14 Tales of the Bunker 2005 This Year’s (October 2005) Bunker Results

© 2006 Progress Software Corporation22 DB-14 Tales of the Bunker 2005 Setup Results db load time30 minutes db load rate2.7 million rows per minute dbanalysis3 min 55 sec idxbuild (-threads 1)8 min 54 sec idxbuild scratch1,129,764 KB prorest (from disk)9 minutes Database is about 11 gigabytes

© 2006 Progress Software Corporation23 DB-14 Tales of the Bunker 2005 Baseline Server Configuration  Data extents on striped array, BI log on own disk  BI cluster size:16384  BI blocksize: 16  Server options: –-n 200 -L –-B –-spin –-bibufs 32  Page writers: 4  BI writer: yes  AI writer: no

© 2006 Progress Software Corporation24 DB-14 Tales of the Bunker 2005 Linux I/O schedulers - 64-bit AMD SuSE 10 What do we learn from this?

© 2006 Progress Software Corporation25 DB-14 Tales of the Bunker 2005 Dump/Load  -index 0 Option on Binary Dump Excellent Performance Improvement But order of records may not be what you want

© 2006 Progress Software Corporation26 DB-14 Tales of the Bunker 2005 Dump/Load with -RO  OpenEdge® 10 -RO: Faster than V9 V9 with -RO: faster than without  OpenEdge 10 with -RO: performance same as without  OpenEdge 10 -RO: Clients now write entries in the.lg file

© 2006 Progress Software Corporation27 DB-14 Tales of the Bunker 2005 Dump/Load  Logical Scatter Factor is very important Performance Difference of 400% to 1000%

© 2006 Progress Software Corporation28 DB-14 Tales of the Bunker 2005 Remote Clients  -Ma The Lower Value, the Better the Performance  -Mm No Negative Impact on ATM Benchmarks

© 2006 Progress Software Corporation29 DB-14 Tales of the Bunker 2005 Network Speed

© 2006 Progress Software Corporation30 DB-14 Tales of the Bunker 2005 “Coma” Problem  We have experienced this problem in every Bunker Test  We still don’t know what’s wrong  A customer on RH AS 4 Kernel: Elsmp reports problem solved - for him  There are an infinite number of things and combinations of things that can be changed in the kernel  We need to do some work with the “aggressiveness” of the APWs to help… but also more testing

© 2006 Progress Software Corporation31 DB-14 Tales of the Bunker 2005 Coma problem: -directio helps

© 2006 Progress Software Corporation32 DB-14 Tales of the Bunker 2005 We still think about this problem

© 2006 Progress Software Corporation33 DB-14 Tales of the Bunker Bit  We saw No difference in general performance between 32 and 64 bit Progress

© 2006 Progress Software Corporation34 DB-14 Tales of the Bunker 2005 Strange Problem  WEIRD PROBLEM ON ONE MACHINE:  The bigger –B, the lower the TPS rate True with both 32 and 64 bit Progress/Linux Could be caused by: –Enterprise versus Desktop version of Linux –10.1A Beta problem –SUSE Linux 10 issue (unsupported OS) –Something else –HyperTransport Effects –All the above  DID NOT OCCUR ON OTHER MACHINES  HAVE NOT SEEN AGAIN

© 2006 Progress Software Corporation35 DB-14 Tales of the Bunker 2005 TPS vs Response Time Avg Response time 0.2 seconds

© 2006 Progress Software Corporation36 DB-14 Tales of the Bunker 2005 V10.0B versus V10.1A Beta  No Difference in general performance

© 2006 Progress Software Corporation37 DB-14 Tales of the Bunker 2005 SATA versus SCSI  SCSI was faster  SATA is less expensive  Beware: desktop drives not rated for 24x7 operation

© 2006 Progress Software Corporation38 DB-14 Tales of the Bunker 2005 Online backup time workload

© 2006 Progress Software Corporation39 DB-14 Tales of the Bunker 2005 Online backup rate (approximate) workload

© 2006 Progress Software Corporation40 DB-14 Tales of the Bunker 2005 Online backup performance impact 150 user atm workload

© 2006 Progress Software Corporation41 DB-14 Tales of the Bunker 2005 Adding extent online: elapsed time Add 2 GB extent on same disk array operationno users150 users create extent18 seconds35 seconds enable extent0 seconds

© 2006 Progress Software Corporation42 DB-14 Tales of the Bunker 2005 Adding extent online: performance impact 150 user workload, add 2 GB data extent online, Extent on same striped array as other extents

© 2006 Progress Software Corporation43 DB-14 Tales of the Bunker 2005 Secret Bunker Web Pages  March 2002 & October  April  Oct

© 2006 Progress Software Corporation44 DB-14 Tales of the Bunker 2005 Join Us in the Bunker  Film

© 2006 Progress Software Corporation45 DB-14 Tales of the Bunker 2005 Also see Gus’s RDBMS Tuning Guide on conference CD Want Answers

© 2006 Progress Software Corporation46 DB-14 Tales of the Bunker 2005