gtpm1m3t | TPF V4R1 Migration Guide: 3.1 to 4.1 |
The performance of the TPF 4.1 system must be monitored to organize system resources properly for peak operating efficiency. The TPF 4.1 system provides data collection and reduction programs to measure system performance. These programs provide operational data about activities such as:
With this information and the transaction history, you can determine how efficiently the TPF 4.1 system is running, where bottlenecks occur, and what changes can improve system performance.
The TPF 4.1 system is enhanced to provide statistical data about each ECB. To aid performance monitoring and capacity planning, this data includes:
The preceding data is available when the ECB exits through the EXI user exit.
In addition, intercept hooks were added for selective collection of statistics on Enter/Back, VFA, and SON macros. To better control the amount of data collected the DATACO (SIP) macro now has a SKIP parameter that specifies the number of entries between samplings (the default is 99). The ZMEAS command no longer supports the T parameter.
The data collection and reduction programs were rewritten substantially. Consequently, any modification that you have made to these programs will require additional effort. In addition, data collection no longer references the JDCD tape. Rather, it now uses a real-time tape named RTC.
Data collection and reduction provides:
The new blocks and structures for the TPF 4.1 system require a variety of changes in the reports:
The following counters are maintained for VFA:
The new data collection hooks allow a variety of macros and messages to be individually tracked. These include:
Collection of this information allows a finer focus on the activities of an individual ECB. In addition new fields were added describing CPU and file usage for each ECB. The dispatch and end times are recorded, as are the number of finds and files, and the number of primary memory requests (GETFCs). A field allows a user-specified identifier to be associated with the ECB.
See the TPF System Performance and Measurement Reference for more information about data collection and performance monitoring.
Data collection can be run in continuous mode or sampling mode, allowing multiple types of data to be captured while avoiding significant interference with message processing. All data collection programs write the capture data to an online tape. No attempt is made to analyze the data online, as this would have a negative impact on the system that is being measured.
The three basic techniques used for collecting data are:
All data reduction is performed on an IBM Multiple Virtual Storage (IBM MVS) system. The data reduction reports are intended for use by an analyst familiar with the TPF 4.1 system. Frequency distribution reports including means, standard deviations, and variances of many parameters are available.
The aim of the initial analysis phase of a working system is to establish the normal limits for each of the key factors affecting performance. Once these limits are set and agreed to be realistic, a periodic system check becomes routine.
The analysis of performance data must always start with summary reports. These reports provide key data required for history and trend analysis. When investigating a problem area, the more detailed plot reports or the specialized reports of the DASD and message reduction programs are used. The plot reports, which show the value of each parameter sample in chronological order, are very effective for analyzing the cause-and-effect relationship between parameters.
In the TPF 4.1 system, the input tape processed by data reduction changed significantly. The changes follow here.
The following tables show the changes to the output reports for data reduction, and detail the impact of data collection and reduction on each report in the TPF 4.1 system.
Table 48. General Reduction Report Changes
General Reduction Report | Changes |
---|---|
Data Collection and Reduction Environment Summary Report | This report is now named and the input tape name is now listed. The ZSTRC settings for the online system execution modes are listed and the online intercept Skip factor is listed. |
Application Summary Report | A total line was added. |
Pushbutton Application Summary Report | There are no changes. |
Table 49. System Reduction Report Changes
System Reduction Report | Changes |
---|---|
System Section of the System Summary Report | The IPTE and PTLB counts were added. |
I-Stream Section of the System Summary Report | There are no changes. |
Core Pool Section of the System Summary Report | This report was converted to report on the physical block types that are defined in the TPF 4.1 system (for example, LIOCB, LECB, LSWB, LFRAME, and LCOMMON). |
Shutdown Section of the System Summary Report | This report was converted to report shutdown conditions based on the physical block types defined in the TPF 4.1 system. |
Frame Usage Histogram Report | This is a new report that produces a histogram of the frames held at ECB
exit.
|
Heap Frame Usage Histogram Report | This is a new report that produces a histogram of the Heap frames held at
ECB exit.
|
System Plot Reports | There are no changes. |
System Distribution Reports | There are no changes. |
Interprocessor Communication Summary Report | No longer supported. The data previously contained in this report is now found in the Interprocessor Communication MPIF Summary Report. |
System Pools Summary Report | There are no changes. |
MPIF Configuration Summary Report | There are no changes. |
Interprocessor Communication MPIF Summary Report | There are no changes. |
MPIF Path Activity Report | There are no changes. |
Table 50. File Reduction Report Changes
File Reduction Report | Changes |
---|---|
Subsystem Summary Report | There are no changes. |
File Access Summary Report | The input/output (I/O) response times were added. |
VFA Summary Report | In the TPF 4.1 system, the virtual file access (VFA) summary data
was moved from the Random File Access Summary Report and reformatted to look
like the ZVFAC IND command output. In addition, the VFA hit rate was
added.
|
Random File Access Summary Report | The virtual file access (VFA) summary data was moved to the new VFA Summary Report and the input/output (I/O) response time was added. |
File Plot Reports | There are no changes. |
File Distribution Reports | There are no changes. |
Channel Path ID Status Report | There are no changes. |
Channel Path Activity Summary Report | There are no changes. |
SDA Activity Summary Report | There are no changes. |
Sequential File Access Summary Report | Change - The mean tape queue size will be added. |
File Accesses Per Record ID Report | The virtual file access (VFA) fast path accesses were added as a column in this report when the device type is VFA. |
File Accesses Per Cylinder Report | There are no changes. |
File Comparison Plot Report | There are no changes. |
3880 Cache Detail Report | There are no changes. |
3990 Cache Detail Report | There are no changes. |
3990 Cache Device Detail Report | There are no changes. |
3990 Cache Summary Report | There are no changes. |
3990 Device Summary Report | There are no changes. |
Table 51. Program Reduction Report Changes
Program Reduction Report | Changes |
---|---|
Program Names and Enters Report | There are no changes. |
Program Details Report | The columns indicating the nesting level were deleted since that data is no longer available. |
Cumulative Program Enters Report | There are no changes. |
Cumulative Program On-File Enters Report | There are no changes. |
Program-Category Statistics Report | There are no changes. |
Package Details Report | The columns indicating the nesting level were deleted since that data is no longer available. |
Cumulative Package Enters Report | There are no changes. |
Cumulative Package On-File Enters Report | There are no changes. |
Subsystem Details Report | This report is no longer supported. There is no replacement. |
Cumulative Subsystem Enters Report | This report is no longer supported. There is no replacement. |
Cumulative Subsystem On-File Enters Report | This report is no longer supported. There is no replacement. |
Program Subsystem Counts Spread Report | This report is no longer supported. There is no replacement. |
Program Subsystem Rates Spread Report | This report is no longer supported. There is no replacement. |
Program Subsystem Percentages Spread Report | This report is no longer supported. There is no replacement. |
Program Categories Report | References to coreslow and corefast were changed to core resident. Categories that no longer exist were removed. |
Program Legend Report | References to the columns for nesting levels were removed. |
Weighted Message Rate Report | There are no changes. |
Table 52. Message Reduction Report Changes
Message Reduction Report | Changes |
---|---|
Message Summary Report | There are no changes. |
Action Code Summary Report | There are no changes. |
Application Detail Report | There are no changes. |
City Summary Report | There are no changes. |
Terminal Activity Report | There are no changes. |
Logical Unit Activity Report | There are no changes. |
Message Stream Report | There are no changes. |