Intel 7xx Servers manual Performance Data Collection Services, How Collection Services works

B.1 Performance Data Collection Services

Collecting performance data with Collection Services is an operating system function designed to run continuously that collects system and job level performance data at regular intervals which can be set from 15 seconds to 1 hour. It runs a number of collection routines called probes which collect data from many system resources including jobs, disk units, IOPs, buses, pools, and communication lines. Collection Services is the replacement for the Performance Monitor function which you may have used in previous releases to collect performance data by running the STRPFRMON command. Collection Services has been available in OS/400 since V4R4. The Performance Monitor remained on the system through V4R5 to allow time to switch over to the new Collection Services function.

How Collection Services works

Collection Services has an improved method for storing the performance data that is collected. A system object called a management collection object (*MGTCOL) was created in V4R4 to store Collection Services data. The management collection object takes advantage of teraspace support to make it a more efficient way to store large quantities of performance data. Collection Services stores the data in a single collection object and supports a release independent design which allows you to move data to a system at a different release without requiring database file conversions.

A command called CRTPFRDTA (Create Performance Data) can be used to create the database files from the contents of the management collection object. The CRTPFRDTA command gives you the flexibility to generate only the database files you need to analyze a specific situation. If you decide that you always want to generate the database, you can configure Collection Services to run CRTPFRDTA as a low-priority batch job while data is being collected. Separating the collection of the data from the database generation, and running the database function at a lower priority are key reasons why Collection Services is efficient and can collect data from large quantities of jobs and threads at very frequent intervals. With Collection Services, you can collect performance data at intervals as frequent as every 15 seconds if you need that level of granularity to diagnose a performance problem. Collection Services also supports collection intervals of 30 seconds, and 1, 5, 15, 30, and 60 minutes.

The overhead associated with collecting performance data is minimal enough that Collection Services can run continuously, no matter what workload is being run on your system. If Collection Services is run continuously as designed, you will capture the data needed to analyze and solve many performance slowdowns before they turn into a serious problem.

Starting Collection Services

You can start Collection Services by using option 2 on the Performance menu (GO PERFORM), the Collection Services component of iSeries Navigator, the STRPFRCOL command, or the QYPSSTRC (Start Collector) API. For more details on these options, see Performance under the Systems Management topic in the latest version of Information Center which is available at http://www.ibm.com/eserver/iseries/infocenter .

When using the Collection Services component of iSeries Navigator, you will find that it gives you flexibility to collect only the performance data you are interested in. Collection Services data is organized into over 20 categories and you have the ability to turn on and off each category or select a

IBM i 6.1 Performance Capabilities Reference - January/April/October 2008
© Copyright IBM Corp. 2008	Appendix B - System i Platform Sizing	341