You are previewing IBM Power Systems Performance Guide: Implementing and Optimizing.
O'Reilly logo
IBM Power Systems Performance Guide: Implementing and Optimizing

Book Description

This IBM® Redbooks® publication addresses performance tuning topics to help leverage the virtualization strengths of the POWER® platform to solve clients’ system resource utilization challenges, and maximize system throughput and capacity. We examine the performance monitoring tools, utilities, documentation, and other resources available to help technical teams provide optimized business solutions and support for applications running on IBM POWER systems’ virtualized environments.

The book offers application performance examples deployed on IBM Power Systems™ utilizing performance monitoring tools to leverage the comprehensive set of POWER virtualization features: Logical Partitions (LPARs), micro-partitioning, active memory sharing, workload partitions, and more. We provide a well-defined and documented performance tuning model in a POWER system virtualized environment to help you plan a foundation for scaling, capacity, and optimization
.
This book targets technical professionals (technical consultants, technical support staff, IT Architects, and IT Specialists) responsible for providing solutions and support on IBM POWER systems, including performance tuning.

Table of Contents

  1. Front cover
  2. Notices
    1. Trademarks
  3. Preface
    1. The team who wrote this book
    2. Now you can become a published author, too!
    3. Comments welcome
    4. Stay connected to IBM Redbooks
  4. Chapter 1. IBM Power Systems and performance tuning
    1. 1.1 Introduction
    2. 1.2 IBM Power Systems
    3. 1.3 Overview of this publication
    4. 1.4 Regarding performance
  5. Chapter 2. Hardware implementation and LPAR planning
    1. 2.1 Hardware migration considerations
    2. 2.2 Performance consequences for processor and memory placement
      1. 2.2.1 Power Systems and NUMA effect
      2. 2.2.2 PowerVM logical partitioning and NUMA
      3. 2.2.3 Verifying processor memory placement
      4. 2.2.4 Optimizing the LPAR resource placement
      5. 2.2.5 Conclusion of processor and memory placement
    3. 2.3 Performance consequences for I/O mapping and adapter placement
      1. 2.3.1 POWER 740 8205-E6B logical data flow
      2. 2.3.2 POWER 740 8205-E6C logical data flow
      3. 2.3.3 Differences between the 8205-E6B and 8205-E6C
      4. 2.3.4 POWER 770 9117-MMC logical data flow
      5. 2.3.5 POWER 770 9117-MMD logical data flow
      6. 2.3.6 Expansion units
      7. 2.3.7 Conclusions
    4. 2.4 Continuous availability with CHARM
      1. 2.4.1 Hot add or upgrade
      2. 2.4.2 Hot repair
      3. 2.4.3 Prepare for Hot Repair or Upgrade utility
      4. 2.4.4 System hardware configurations
    5. 2.5 Power management
  6. Chapter 3. IBM Power Systems virtualization
    1. 3.1 Optimal logical partition (LPAR) sizing
    2. 3.2 Active Memory Expansion
      1. 3.2.1 POWER7+ compression accelerator
      2. 3.2.2 Sizing with the active memory expansion planning tool
      3. 3.2.3 Suitable workloads
      4. 3.2.4 Deployment
      5. 3.2.5 Tunables
      6. 3.2.6 Monitoring
      7. 3.2.7 Oracle batch scenario
      8. 3.2.8 Oracle OLTP scenario
      9. 3.2.9 Using amepat to suggest the correct LPAR size
      10. 3.2.10 Expectations of AME
    3. 3.3 Active Memory Sharing (AMS)
    4. 3.4 Active Memory Deduplication (AMD)
    5. 3.5 Virtual I/O Server (VIOS) sizing
      1. 3.5.1 VIOS processor assignment
      2. 3.5.2 VIOS memory assignment
      3. 3.5.3 Number of VIOS
      4. 3.5.4 VIOS updates and drivers
    6. 3.6 Using Virtual SCSI, Shared Storage Pools and N-Port Virtualization
      1. 3.6.1 Virtual SCSI
      2. 3.6.2 Shared storage pools
      3. 3.6.3 N_Port Virtualization
      4. 3.6.4 Conclusion
    7. 3.7 Optimal Shared Ethernet Adapter configuration
      1. 3.7.1 SEA failover scenario
      2. 3.7.2 SEA load sharing scenario
      3. 3.7.3 NIB with an SEA scenario
      4. 3.7.4 NIB with SEA, VLANs and multiple V-switches
      5. 3.7.5 Etherchannel configuration for NIB
      6. 3.7.6 VIO IP address assignment
      7. 3.7.7 Adapter choices
      8. 3.7.8 SEA conclusion
      9. 3.7.9 Measuring latency
      10. 3.7.10 Tuning the hypervisor LAN
      11. 3.7.11 Dealing with dropped packets on the hypervisor network
      12. 3.7.12 Tunables
    8. 3.8 PowerVM virtualization stack configuration with 10 Gbit
    9. 3.9 AIX Workload Partition implications, performance and suggestions
      1. 3.9.1 Consolidation scenario
      2. 3.9.2 WPAR storage
    10. 3.10 LPAR suspend and resume best practices
  7. Chapter 4. Optimization of an IBM AIX operating system
    1. 4.1 Processor folding, Active System Optimizer, and simultaneous multithreading
      1. 4.1.1 Active System Optimizer
      2. 4.1.2 Simultaneous multithreading (SMT)
      3. 4.1.3 Processor folding
      4. 4.1.4 Scaled throughput
    2. 4.2 Memory
      1. 4.2.1 AIX vmo settings
      2. 4.2.2 Paging space
      3. 4.2.3 One TB segment aliasing
      4. 4.2.4 Multiple page size support
    3. 4.3 I/O device tuning
      1. 4.3.1 I/O chain overview
      2. 4.3.2 Disk device tuning
      3. 4.3.3 Pbuf on AIX disk devices
      4. 4.3.4 Multipathing drivers
      5. 4.3.5 Adapter tuning
    4. 4.4 AIX LVM and file systems
      1. 4.4.1 Data layout
      2. 4.4.2 LVM best practice
      3. 4.4.3 File system best practice
      4. 4.4.4 The filemon utility
      5. 4.4.5 Scenario with SAP and DB2
    5. 4.5 Network
      1. 4.5.1 Network tuning on 10 G-E
      2. 4.5.2 Interrupt coalescing
      3. 4.5.3 10-G adapter throughput scenario
      4. 4.5.4 Link aggregation
      5. 4.5.5 Network latency scenario
      6. 4.5.6 DNS and IPv4 settings
      7. 4.5.7 Performance impact due to DNS lookups
      8. 4.5.8 TCP retransmissions
      9. 4.5.9 tcp_fastlo
      10. 4.5.10 MTU size, jumbo frames, and performance
  8. Chapter 5. Testing the environment
    1. 5.1 Understand your environment
      1. 5.1.1 Operating system consistency
      2. 5.1.2 Operating system tunable consistency
      3. 5.1.3 Size that matters
      4. 5.1.4 Application requirements
      5. 5.1.5 Different workloads require different analysis
      6. 5.1.6 Tests are valuable
    2. 5.2 Testing the environment
      1. 5.2.1 Planning the tests
      2. 5.2.2 The testing cycle
      3. 5.2.3 Start and end of tests
    3. 5.3 Testing components
      1. 5.3.1 Testing the processor
      2. 5.3.2 Testing the memory
      3. 5.3.3 Testing disk storage
      4. 5.3.4 Testing the network
    4. 5.4 Understanding processor utilization
      1. 5.4.1 Processor utilization
      2. 5.4.2 POWER7 processor utilization reporting
      3. 5.4.3 Small workload example
      4. 5.4.4 Heavy workload example
      5. 5.4.5 Processor utilization reporting in power saving modes
      6. 5.4.6 A common pitfall of shared LPAR processor utilization
    5. 5.5 Memory utilization
      1. 5.5.1 How much memory is free (dedicated memory partitions)
      2. 5.5.2 Active memory sharing partition monitoring
      3. 5.5.3 Active memory expansion partition monitoring
      4. 5.5.4 Paging space utilization
      5. 5.5.5 Memory size simulation with rmss
      6. 5.5.6 Memory leaks
    6. 5.6 Disk storage bottleneck identification
      1. 5.6.1 Performance metrics
      2. 5.6.2 Additional workload and performance implications
      3. 5.6.3 Operating system - AIX
      4. 5.6.4 Virtual I/O Server
      5. 5.6.5 SAN switch
      6. 5.6.6 External storage
    7. 5.7 Network utilization
      1. 5.7.1 Network statistics
      2. 5.7.2 Network buffers
      3. 5.7.3 Virtual I/O Server networking monitoring
      4. 5.7.4 AIX client network monitoring
    8. 5.8 Performance analysis at the CEC
    9. 5.9 VIOS performance advisor tool and the part command
      1. 5.9.1 Running the VIOS performance advisor in monitoring mode
      2. 5.9.2 Running the VIOS performance advisor in post processing mode
      3. 5.9.3 Viewing the report
    10. 5.10 Workload management
  9. Chapter 6. Application optimization
    1. 6.1 Optimizing applications with AIX features
      1. 6.1.1 Improving application memory affinity with AIX RSETs
      2. 6.1.2 IBM AIX Dynamic System Optimizer
    2. 6.2 Application side tuning
      1. 6.2.1 C/C++ applications
      2. 6.2.2 Java applications
      3. 6.2.3 Java Performance Advisor
    3. 6.3 IBM Java Support Assistant
      1. 6.3.1 IBM Monitoring and Diagnostic Tools for Java - Memory Analyzer
      2. 6.3.2 Other useful performance advisors and analyzers
  10. Appendix A. Performance monitoring tools and what they are telling us
    1. NMON
    2. lpar2rrd
    3. Trace tools and PerfPMR
  11. Appendix B. New commands and new commands flags
    1. amepat
    2. lsconf
  12. Appendix C. Workloads
    1. IBM WebSphere Message Broker
    2. Oracle SwingBench
    3. Self-developed C/C++ application
  13. Related publications
    1. IBM Redbooks
    2. Online resources
    3. Help from IBM
  14. Back cover