Other Performance Tools

1 like Updated 1/23/17, 11:57 AM by naggerTags: None

Performance - Other Tools (non AIX commands)

Missing image Missing image Missing image Missing image Missing image Missing image



HMC Scanner

  • This Java code contacts an HMC, detects all managed systems, downloads systems and LPAR configuration and finally produces a single spreadsheet with all collected data.
  • Very easy solution to document an environment even when many systems and LPARs are involved.
  • When HMC is collecting performance data, it can be included in the spreadsheet.
  • Take this link to the

    HMC Scanner


AIX memory use analyzer

Missing image The AIX memory use analyzer produces a report of AIX memory use by reading the output of svmon and other commands. The report shows how memory is used at the time command output is captured. Command output can be captured by the perfpmr data collection tool or by the showmemuse shell script. A summary report is followed by a set of detailed reports which show memory use for each process and segment.

Please see the How to summarize AIX memory use web page to learn more.

VIOS Performance Advisor

Missing image

The VIOS advisor is an application that runs within the customer's VIOS for a user specified amount of time (hours), which polls and collects key performance metrics before analyzing results and providing a health check report and proposes changes to the environment or areas to investigate further. The goal of the VIOS advisor is not to provide another monitoring tool, but instead have an expert system view performance metrics already available to the customer and make assessments and recommendations based on the expertise and experience available within the IBM systems performance group.

Now the VIOS "part" command.

Take this link to the VIOS Performance Advisor

Java Performance Advisor (JPA)

Missing image The JPA Tool provides recommendations to improve the performance of Java/WAS applications running on AIX. The recommendations are based on Java tuning best practices. JPA uses the following criteria as factors to determine the recommendations:
  • Relative importance of the Java application
  • Machine usage (Test/Production)
  • User's expertise level

    Detailed usage and command line options can be viewed with the help message ('-h' option)

    Java Performance Advisor (JPA)

PowerVM Virtualization Performance Advisor:

Missing image The PowerVM Virtualization Performance Advisor diagnoses performance issues within PowerVM partitions. The tool was developed by the IBM Power Systems Virtualization Performance team and is intended to help customers, developers, testers, performance analysts, field engineers and anyone else interested in diagnosing and addressing performance issues within the PowerVM partitions.

The advisor runs within an AIX partition (dedicated or shared), analyzes performance characteristics, and provides a health check report. The user defines how long the tool runs. The health check report includes system configuration (available resources, hardware & software levels, etc), partition (type, mode, available resources, s/w levels), partition settings and recommendations to improve performance (e.g. placement, SMT, processor folding, processor, and memory pool utilization).

PowerVM Virtualization Performance Advisor


- Nigel's AIX and Linux Performance Monitor

Missing image nmon displays all the important performance stats on one screen
  • This is a popular free tool for monitoring AIX (and Linux on POWER, Intel, AMD and Mainframe).
  • Alternatively, nmon can save the data to a file for later analysis and graphing (see the analyser below).
  • nmon was developed for benchmarking but many people use it to study production performance and write performance reports.
  • Take this link to the nmon section nmon for AIX and nmon for Linux

nmon Analyser

- Steve's Excel spread sheet to graph nmon output

Missing image This is a popular addition to nmon for further analyse of performance data and graphing.
  • All the nmon data is automatically graphed for you.
  • From the graphs you can then determine bottlenecks, understand how the resources are used and start tuning.
  • Take this link to the nmon analyser section nmon Analyser Excel shpreadshet

nmon Consolidator

- Steve's Excel spreadsheet to graph lots of nmon output from different machines/LPARs

Missing image This is a new addition to nmon and allows the monitoring of many machines at once.
  • Graph all the LPARs of one machine for the global view
  • Graph many older machines and plan the server consolidation on to POWER5 LPARs
  • Or just look at many machines at once.
  • Take this link to the nmon Consolidator section nmon Consolidator


- Bruce Spencer's Perl tools to generate web pages from multiple nmon files via rrdtool

Missing image nmon files are used to aggregate performance across a shared server and display as a web page
  • The nmon data from many AIX or Linux machines can be automatically used to generate web pages
  • The web pages include the configuration and performance graphs, non-standard AIX settings and change history
  • Setting up the nmon2web tool is trivial. However, it requires a web server with Apache, PHP and rrdtool.

Due to broken external Links and websites 4 years out of date, we decided to make a sub-page here for nmon2web


- Nigel Griffiths' C tools to show convert nmon files ready for rrdtool

Missing image This tool is written in C and the source code provided to allow user fixes and learning
  • nmon2rrd generates from a single nmon output file
    • a script to create suitable rrdtool databases
    • a script to load the rrdtool databases from reformatted nmon data
    • a script to extract the data in graph file ready for a website
    • a index.html web file which pulls in the graphs and configurations information

    This can be used as the basis for writing your own automating nmon file to a website but operates on a single file
    • You have to handle the transport of nmon files to the web server, scripting to run nmon2rrd and the web server itself.

    Take a look at nmon2web from Bruce (above), if you want more automation from the tool itself as a start point
  • Take this link to the nmon page a search for "nmon2rrd" section nmon2rrd
  • Take this link to learn more about rrdtool Note:This is a link to Open Source website not on MyDeveloperWorks

pGraph Federico's Java Performance Graph Viewer

Missing image
  • This tool can take data from nmon (AIX and Linux), vmstat, iostat or topasout from AIX
  • It is Java based to allow you to graph and analyse the data dynamically (on Windows and AIX).
  • It can read multiple data files from the same system or multiple systems.
  • No limits on the number of data records.
  • Take this link to the pGraph section

    pGraph Java Graph Viewer

"Roll Your Own" AIX Performance Monitor - sample code

Missing image Missing image A collection of tools demonstrating how to get performance data from AIX 5L
  • There are C code sample programs for you to follow (for programers only)
  • Included are pre-compiled binaries for AIX 5.3
  • Take this link to the Roll Your Own section

    Roll-Your-Own Code Samples and Tools page

Stress Test your pSeries/AIX System

Missing image

These are a number of programs to stress you CPU, memory, disks and other features:

  • Soak testing = check a new machine/disk to remove early life failures
  • Prove performance of machine upgrades or alternative disk configurations
  • Learn performance monitoring and tuning
  • Each tool uses one resource like CPU or memory or disks - so you can investigate one aspect at a time.
  • Added together these tools can keep your whole machine busy!
  • Take this link to the nstress section
  • nStress Collection of Stress Testing Tools

Ganglia for Cluster Performance Monitoring with POWER5 additions

Missing image Many large AIX High performance Computing (HPC) clusters use this excellent tool to monitor performance across large clusters of machines.

It is now a useful tool for monitoring the LPARs on a single POWER5 machine.
  • The data is displayed graphical on a Website, includes configuration and performance statistics.
  • This is also increasingly being used in commercial data centers too to monitor large groups of machines.
  • Ganglia can also be used to monitor a group of logical partitions (LPARs) on a single machine - these just look like a cluster to Ganglia.
  • Ganglia is not limited to just the AIX, which makes it even more useful in heterogeneous computer rooms.
  • Take this link to the Introduction to Ganglia section Ganglia - HowTo with POWER additions

- Monitoring for the Virtual I/O Server

Missing image The Virtual I/O Server feature allows a virtual internal Ethernet to be connected to a real physical network one POWER5 machines.
  • However, monitoring the SEA network trafic is not easy and missed by most AIX tools
  • If you put an ID address on the SEA you only get to see the traffic to and from the Virtual I/O Server and not the traffic between the virtual and physical networks.
  • Take this link to the shell script to monitor SEA network traffic section Share Ethernet Adapter Statistics Script


- CPU Cross Partition Graphs from HMC data with rrdtool

Missing image
  • The tool is capable to produce historical CPU utilization graphs of LPARs and shared CPU usage
  • all LPARs (AIX, VIOS, i5OS, Linux on POWER) and all CPU pools stats are included
  • The data is extracted from HMC via ssh and loaded into the RRDTool database
  • Gives you CPU Cross Partition Graphs based on 60secs CPU utilization averages provided by HMC (agent-less)
  • It collects complete physical (HW) and logical configuration of all managed systems and their lpars and all changes in their state and configuration.
  • LPAR2rrd Project Website
  • Note: This is a Open Source project and the Link is not on IBM DeveloperWorks.

Workload Manager (WLM) for AIX

Missing image

Effectively managing system resources in growing UNIX server environments has become a critical task.

Each workload on the server must be assured the appropriate amount of system resources without penalizing mission-critical applications.

AIX 5L Workload Manager provides a great set of tools and functionalities to efficiently accomplish this.

Performance Toolbox (PTX)

This is an extra package for AIX with adds the performance GUI tools for monitoring one machine in depth and multiple machines.

The package is extremely flexible and can save data for later processing.

See the Redbook "Customizing Performance Toolbox and Performance Toolbox Parallel Extensions for AIX" for more details and the Performance Tools Redbook.

IBM Performance Management for AIX (PM/AIX)

PM for Power Systems includes IBM i and AIX. This is an IBM product where you run a data collector that ships data securly to IBM, this is processed into reports, urgent warnings and many performance graphs. You go to the PM website and once logged in you can review the reports, warnings and historc and projected performance stats.. This product has a basic "no charge" and a "full function" offering priced per LPAR. Find out more from:

rperf - Relative Performance number

Missing image This korn shell script outputs the offical rPerf number for the current machine
  • If the rPerf number is not available for the number of CPUs in the machine or LPAR then a rough guess is made by dividing the top rPerf number by the number of CPUs
  • Note: older machines only have Relative OLTP numbers (rolpt).
  • Take this link to the rPerf Relative Performance script


- General purpose network-wide graphing framework (grapher/gatherer)

Missing image Munin is a highly flexible and powerful solution used to create graphs of virtually everything imaginable throughout your network, while still maintaining a rattling ease of installation and configuration. Munin is Free Software, actively supported and developed as an Open Source project. It is reported to work on all common Unix-like systems, including AIX and all Linux distributions.
  • Linux-plugin for graphing real (physical) cpu usage in an lpar exists.
  • Munin can be compaired to lpar2rrd and Ganglia, but while lpar2rrd is exactly that and nothing else, and Ganglia is for clusters computers, Munin is a more general purpose monitoring tool.
  • WebSite: http://munin.projects.linpro.no
  • Note:This is a link to Open Source website not on IBM DeveloperWorks


- Roland's Yet Another AIX Advanced Accounting toolset

Missing image This tool very quickly sets up AIX Advanced Accounting, cron jobs and scripts to cenrally extract data to allow you to very accurately determine CPU use. This can be a boon for:
  • Performance managing your machines
  • Used as a worked example of how to extract the information
  • Used as the basis for charging user for resources consumed
  • Take this link to the ryaacct

Galileo Performance Explorer for AIX

Missing image Galileo PE empowers you to gain a unique perspective of your infrastructure. Galileo PE provides an easy-to-use, Web-based interface designed to help pro-actively identify performance bottlenecks or under-utilized hardware. Galileo PE isn't just server monitoring software, but an intuitive SaaS that provides:
  • Agent-only installation
  • Web-based consolidated GUI
  • Automated data collection
  • In-depth performance analysis
  • Audit compliance for software licensing
  • Performance monitoring
  • Capacity planning
  • Cost and/or savings analysis
  • Virtual Frames for server consolidation
  • Visit the Galileo PE website at http://www.galileosuite.com
  • Note:This is a link to a 3rd Party Vendors website that is not on IBM DeveloperWorks

nworms used for LPM and WPAR Relocation Demonstration

Missing image nworms is a small program used to demonstrate Live Partition Mobility (LPM) or Live Application Mobility (Relocation)
  • It outputs worm updates per second, machine serial number, LPAR number and name.
  • If you change machine or LPAR the colour changes two
  • And wiggles worms across the screen.
  • nworms

Topas CEC data Analyser

Missing image topasrec can collect topas -C type data & this can be analysed by Excel for graphing

Example of use:
  • Collecting the binary data: topasrec -C -o sample -s 120 -c 720
  • Creates a file like: sample_cec_101105_1752.topas
  • Convert to nmon-ish style data with: topasout -a sample_cec_101105_1752.topas
  • Creates: sample_cec_101105_1752.topas.csv
  • Topas CEC Analyser

Virtual I/O Server Performance Tools

Missing image This is a review of every tool I know about that can help including

Missing image The postings on this site solely reflect the personal views of the authors and do not necessarily represent the views, positions, strategies or opinions of IBM or IBM management.