Front cover -- Contents -- Figures -- Tables -- Notices -- Trademarks -- Preface -- The team that wrote this redbook -- Become a published author -- Comments welcome -- Summary of changes -- August 2003, Second Edition -- Part 1 AIX 5L performance tools -- Chapter 1. Introduction to AIX performance monitoring and tuning -- 1.1 Performance expectation -- 1.2 CPU performance -- 1.2.1 Initial advice -- 1.2.2 Processes and threads -- 1.2.3 Scheduling -- 1.2.4 SMP performance -- 1.3 Memory performance -- 1.3.1 Initial advice -- 1.3.2 Memory segments -- 1.3.3 Paging mechanism -- 1.3.4 Memory load control mechanism -- 1.3.5 Paging space allocation policies -- 1.3.6 Memory leaks -- 1.3.7 Shared memory -- 1.4 Disk I/O performance -- 1.4.1 Initial advice -- 1.4.2 Disk subsystem design approach -- 1.4.3 Bandwidth-related performance considerations -- 1.4.4 Disk design -- 1.4.5 Logical Volume Manager concepts -- 1.5 Network performance -- 1.5.1 Initial advice -- 1.5.2 TCP/IP protocols -- 1.5.3 Network tunables -- 1.6 Kernel tunables -- 1.6.1 Tunables commands -- 1.6.2 Tunable files -- 1.7 The /proc file system -- Chapter 2. Getting started -- 2.1 Tools and filesets -- 2.2 Tools by resource matrix -- 2.3 Performance tuning approach -- 2.3.1 CPU bound system -- 2.3.2 Memory bound system -- 2.3.3 Disk I/O bound system -- 2.3.4 Network I/O bound system -- Part 2 Multi-resource monitoring and tuning tools -- Chapter 3. The fdpr command -- 3.1 fdpr -- 3.1.1 Information about measurement and sampling -- 3.2 Examples for fdpr -- Chapter 4. The iostat command -- 4.1 iostat -- 4.1.1 Information about measurement and sampling -- 4.2 Examples for iostat -- 4.2.1 System throughput report -- 4.2.2 tty and CPU utilization report -- 4.2.3 Disk utilization report -- 4.2.4 Disk utilization report for MPIO -- 4.2.5 Adapter throughput report -- Chapter 5. The netpmon command.

5.1 netpmon -- 5.1.1 Information about measurement and sampling -- 5.2 Examples for netpmon -- 5.2.1 Process statistics -- 5.2.2 FLIH and SLIH CPU statistics -- 5.2.3 TCP socket call statistics -- 5.2.4 Detailed statistics -- Chapter 6. Performance Diagnostic Tool (PDT) -- 6.1 PDT -- 6.1.1 Information about measurement and sampling -- 6.2 Examples for PDT -- 6.2.1 Editing the configuration files -- 6.2.2 Using reports generated by PDT -- 6.2.3 Creating a PDT report manually -- Chapter 7. The perfpmr command -- 7.1 perfpmr -- 7.1.1 Information about measurement and sampling -- 7.1.2 Building and submitting a test case -- 7.2 Examples for perfpmr -- Chapter 8. The ps command -- 8.1 ps -- 8.1.1 Information about measurement and sampling -- 8.2 Examples for ps -- 8.2.1 Displaying the top 10 CPU-consuming processes -- 8.2.2 Displaying the top 10 memory-consuming processes -- 8.2.3 Displaying the processes in order of being penalized -- 8.2.4 Displaying the processes in order of priority -- 8.2.5 Displaying the processes in order of nice value -- 8.2.6 Displaying the processes in order of real memory use -- 8.2.7 Displaying the processes in order of I/O -- 8.2.8 Displaying WLM classes -- 8.2.9 Viewing threads -- Chapter 9. The sar command -- 9.1 sar -- 9.1.1 Information about measurement and sampling -- 9.2 Examples for sar -- 9.2.1 Monitoring one CPU at a time -- 9.2.2 Collecting statistics by using cron -- 9.2.3 Displaying access time system routines -- 9.2.4 Monitoring buffer activity for transfers, access, and caching -- 9.2.5 Monitoring system calls -- 9.2.6 Monitoring activity for each block device -- 9.2.7 Monitoring kernel process activity -- 9.2.8 Monitoring the message and semaphore activities -- 9.2.9 Monitoring the kernel scheduling queue statistics -- 9.2.10 Monitoring the paging statistics -- 9.2.11 Monitoring the processor utilization.

9.2.12 Monitoring tty device activity -- 9.2.13 Monitoring kernel tables -- 9.2.14 Monitoring system context switching activity -- Chapter 10. The schedo and schedtune commands -- 10.1 schedo -- 10.1.1 Recommendations and precautions -- 10.2 Examples for schedo -- 10.2.1 Displaying current settings -- 10.2.2 Tuning CPU parameters -- 10.2.3 Tuning memory parameters -- 10.3 schedtune -- Chapter 11. The topas command -- 11.1 topas -- 11.1.1 Information about measurement and sampling -- 11.2 Examples for topas -- 11.2.1 Common uses of the topas command -- 11.2.2 Using subcommands -- 11.2.3 Monitoring CPU usage -- 11.2.4 Monitoring disk problem -- Chapter 12. The truss command -- 12.1 truss -- 12.1.1 Information about measurement and sampling -- 12.2 Examples for truss -- 12.2.1 Using truss -- 12.2.2 Using the summary output -- 12.2.3 Monitoring running processes -- 12.2.4 Analyzing file descriptor I/O -- 12.2.5 Checking program parameters -- 12.2.6 Checking program environment variables -- 12.2.7 Tracking child processes -- 12.2.8 Checking user library call -- Chapter 13. The vmstat command -- 13.1 vmstat -- 13.1.1 Information about measurement and sampling -- 13.2 Examples for vmstat -- 13.2.1 Virtual memory activity -- 13.2.2 Forks report -- 13.2.3 Interrupts report -- 13.2.4 VMM statisics report -- 13.2.5 Sum structure report -- 13.2.6 I/O report -- Chapter 14. The vmo, ioo, and vmtune commands -- 14.1 vmo -- 14.1.1 Information about measurement and sampling -- 14.1.2 Recommendations and precautions for vmo -- 14.2 Examples for vmo -- 14.3 ioo -- 14.3.1 Information about measurement and sampling -- 14.3.2 Recommendations and precautions -- 14.4 Examples for ioo -- 14.4.1 Displaying I/O setting -- 14.4.2 Changing tunable values -- 14.4.3 Logical volume striping -- 14.4.4 Increasing write activity throughput -- 14.5 vmtune.

Chapter 15. Kernel tunables commands -- 15.1 tuncheck -- 15.1.1 Examples for tuncheck -- 15.2 tunrestore -- 15.2.1 Examples for tunrestore -- 15.3 tunsave -- 15.3.1 Examples for tunsave -- 15.4 tundefault -- 15.4.1 Examples for tundefault -- 15.5 tunchange -- 15.5.1 Examples for tunchange -- Chapter 16. Process-related commands -- 16.1 procwdx -- 16.2 procfiles -- 16.3 procflags -- 16.4 proccred -- 16.5 procmap -- 16.6 procldd -- 16.7 procsig -- 16.8 procstack -- 16.9 procstop -- 16.10 procrun -- 16.11 procwait -- 16.12 proctree -- Part 3 CPU-related performance tools -- Chapter 17. The alstat and emstat commands -- 17.1 Alignment and emulation exception -- 17.2 alstat -- 17.2.1 Information about measurement and sampling -- 17.2.2 Examples for alstat -- 17.2.3 Detecting and resolving alignment problems -- 17.3 emstat -- 17.3.1 Information about measurement and sampling -- 17.3.2 Examples for emstat -- 17.3.3 Detecting and resolving emulation problems -- Chapter 18. The bindintcpu and bindprocessor commands -- 18.1 bindintcpu -- 18.1.1 Examples for bindintcpu -- 18.2 bindprocessor -- 18.2.1 Information about measurement and sampling -- 18.2.2 Examples for bindprocessor -- Chapter 19. The gprof, pprof, prof, and tprof commands -- 19.1 CPU profiling tools -- 19.1.1 Comparison of tprof versus prof and gprof -- 19.2 gprof -- 19.2.1 Information about measurement and sampling -- 19.2.2 Profiling with the fork and exec subroutines -- 19.2.3 Examples for gprof -- 19.3 pprof -- 19.3.1 Information about measurement and sampling -- 19.3.2 Examples for pprof -- 19.4 prof -- 19.4.1 Information about measurement and sampling -- 19.4.2 Examples for prof -- 19.5 tprof -- 19.5.1 Information about measurement and sampling -- 19.5.2 Examples for tprof -- Chapter 20. The nice and renice commands -- 20.1 nice -- 20.1.1 Information about measurement and sampling.

20.2 Examples for nice -- 20.2.1 Reducing the priority of a process -- 20.2.2 Improving the priority of a process -- 20.3 renice -- 20.3.1 Information about measurement and sampling -- 20.4 Examples for renice -- Chapter 21. The time and timex commands -- 21.1 time -- 21.1.1 Information about measurement and sampling -- 21.1.2 Examples for time -- 21.2 timex -- 21.2.1 Information about measurement and sampling -- 21.2.2 Examples for timex -- Part 4 Memory-related performance tools -- Chapter 22. The ipcs command -- 22.1 ipcs -- 22.1.1 Information about measurement and sampling -- 22.1.2 Examples for ipcs -- Chapter 23. The rmss command -- 23.1 rmss -- 23.1.1 Information about measurement and sampling -- 23.1.2 Recommendations and precautions -- 23.1.3 Examples for rmss -- Chapter 24. The svmon command -- 24.1 svmon -- 24.1.1 Information about measurement and sampling -- 24.1.2 Examples for svmon -- Part 5 Disk I/O-related performance tools -- Chapter 25. The filemon command -- 25.1 filemon -- 25.1.1 Information about measurement and sampling -- 25.1.2 Examples for filemon -- Chapter 26. The fileplace command -- 26.1 fileplace -- 26.1.1 Information about measurement and sampling -- 26.1.2 Examples for fileplace -- 26.1.3 Analyzing the physical report -- Chapter 27. The lslv, lspv, and lsvg commands -- 27.1 lslv -- 27.2 lspv -- 27.3 lsvg -- 27.4 Examples for lslv, lspv, and lsvg -- 27.4.1 Using lslv -- 27.4.2 Using lspv -- 27.4.3 Using lsvg -- 27.4.4 Acquiring more disk information -- Chapter 28. The lvmstat command -- 28.1 lvmstat -- 28.1.1 Information about measurement and sampling -- 28.1.2 Examples for lvmstat -- Part 6 Network-related performance tools -- Chapter 29. atmstat, entstat, estat, fddistat, and tokstat commands -- 29.1 atmstat -- 29.1.1 Information about measurement and sampling -- 29.1.2 Examples for atmstat -- 29.2 entstat.

29.2.1 Information about measurement and sampling.