To update to the current GiAPA release please
1. Click here: Download latest update
2. Download the zipped iSeries save file, unzip it on your PC
3. Use e.g. FTP to send the save file to an existing *SAVF on your iSeries
4. Terminate GiAPA data collection
5. (To be on the safe side, you may want to back GIAPALIB up first, but it shouldn’t be necessary.)
6. Restore the save file to GiAPALIB:
RSTOBJ *ALL GIAPALIB *SAVF SAVF(savfname) MBROPT(*ALL) ALWOBJDIF(*ALL)
7. Run command
8. If resulting data from GiAPA analysis is stored in other data libraries than GIAPALIB, command GIAPAINST must be used for these libraries also to convert to new data base layouts.
V06M00C released February 23rd, 2024
Only a few changes to the security code routine in connection with SQL Observer.
Vo6M00B released February 5th, 2024
Repeated appearance of page documenting Plan Cache dumps for a Job and QRO-code corrected.
Option *NONE (meaning “Select all”) added to keyword JOB in command GIAPA610 (= compliance with command ADDJWDFN).
Duplication of lines on SQL Observer’s list of user names are removed.
(A few only “cosmetic” modifications are not listed here.)
V06M00A released January 13th, 2024
A line showing values for SQL Observer (Job GIAPAJWCOL) has been added to the totals for CPU usage shown in GiAPA Menu option 16, selection 7.
Error correction: Date calculation for deletion of obsolete Plan Cache dumps could cause abnormal end around change of year.
A keyword parameter in command GIAPA610 has been updated to match all options in CL command ADDJWDFN.
V06M00 released January 2nd, 2024 – requires Op.Sys. V7R2M0 or higher
New GiAPA Feature: SQL Observer.
The intensive and increased use of SQL has made optimization of frequently used SQL statements one of the most rewarding ways of saving server resources. GiAPA never offered analysis of the efficiency of the Access Plans selected by the Query Optimizer, because this function is offered by IBM’s “SQL Performance Center” in the Database section of the Access Client Solution (ACS) tool.
However, the Plan Cache data needed as input for IBM’s performance analysis is not dumped automatically given it would consume excessive resources. An IBM performance expert recently suggested that a tool offering an automated and user-controlled dumping of Access Plans that might be wanted for analysis could be a popular option. This idea is implemented as the “SQL Observer” in GiAPA version 6.
We trust that you will find this new feature is rather powerful. However, to get the wanted result it is very important to understand the somewhat tricky “rules of the game” for collecting Plan Cache data. Therefore, please see the documentation in the User Manual describing GiAPA Menu options 61 – 63, or use this link https://www.giapa.com/images/PDF_files/SQL_Observer.pdf.
- - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
Function Type USR added: The table of function types has been updated with the new USR = User (Type of active job).
Error correction: The calculation of CPU percentage in the Job Summary and Interval Details reports was incorrect in certain cases when allocated CPU resources was less than 1.
Sorting Data Base Files: The SORTDB command (from QUSROOL) was moved to GiAPA Menu option 99.
Totals for automated analysis improved: The user selected limit for minimum run time savings to be selected is now also reflected on the page showing total savings (GiAPA Menu option 20, selection 4).
Many smaller changes: Several minor updates were made – the details are not listed here. A total of 52 programs and 16 files were modified or added.
V05M01I released August 5th, 2023
Problem fixed: Analysis of data covering only very few minutes could result in abnormal end of program GIAPA146.
Problem fixed: The last record in file GiAPA144P3 (Input for user defined graphs) might not always be written.
Data volumes are increasing: A user reported a problem because a job in one day made more than 999,999,999 physical I/Os, i.e. more than the 9-positions field could handle. The lengths of the fields involved were increased to 11 digits.
V05M01H released June 7th, 2023
Error corrected: Specification of user defined graph (GiAPA Menu option 26) could cause a duplicate key message for file GIAPA260P1.
Server table updated to include latest new models.
Job log messages were enhanced for GiAPA Data Expansion and Analyzing run.
Spelling errors in Display File texts were corrected.
Estimates for savings found by automatic analysis were adjusted.
V05M01G released February 1st, 2023
Error corrected: Generated new graphs could sometimes fail to be copied to the IFS, causing either no result to be shown, or display of a previous edition of the same graph.
Error corrected: Title line of automatic performance analysis would show an incorrect minute value for the time when performance collection was started.
V05M01F released December 15th, 2022
Fully automatic performance analysis (Menu option 20) was improved to report considerably more optimization potential for both programs and files. Please observe that option 19 still gives a more detailed and complete overview of resources used by a file access or a program since it is not limited to show cases where a certain level of run time improvements seems possible.
Analyzed call stack report for multithreaded jobs erroneously included all threads when *Initial thread was selected. This has been corrected.
A German user with GiAPA on many LPARs wanted to
- create diagrams comparing assigned number of CPUs with the avg. or max CPU usage, and
- copy templates containing the graph definitions to other LPARs.
Both suggestions were implemented:
- When a user defined graphs specifies 1 for “CPUs in LPAR” plus selects only CPU percentage (avg. and/or max), the result will now show “CPUs used” instead of “CPU %”, i.e., a shared Y-axis will be used.
- The simple 3-step procedure for copying templates to other LPARs have been included on page 82 of the GiAPA User Manual.
In rare cases GiAPA Menu option 19 could display an empty screen when analyzing a selected program, and option 20 could show too low statistics for number of jobs using that program. Both problems were fixed.
V05M01E released October 24th, 2022
Menu option 20 = Automatic Performance Analysis improved to detect probably 30 % more optimization potential.
Use of the job watch function (Menu option 12) will generate job log messages telling number of HotSpots generated and the Job-Id of the job being watched. For special cases where HotSpot data only is wanted for job(s) being watched, the CPULIMIT keyword of command GIAPA110 (current default value 4.0) was changed to accept a very high value which will suppress all “normal” HotSpots.
Error fixed: Command GIAPA055 (Generate graph from unexpanded data) could fail due to a divide-by-zero if used with data from an almost empty test LPAR.
2nd title line in graphs generated by commands GiAPA053 and 054 will now show data collection from/to date and time for input used.
GiAPA option 19, selection 2 (File analysis) can now also flag cases of redundant reads.
Occurrence count was added to command GiAPA045 (Where-used for programs).
V05M01D released August 25th, 2022
GiAPA commands can now send Emails having generated diagrams attached. The four commands GIAPA052, GIAPA054, GIAPA055 and GIAPA070, which can produce charts (also in unattended batch runs), have been extended with two new optional keywords, SNDMAILPGM and RCP (Recipients). For additional information please see https://www.giapa.com/images/PDF_files/V05M01Dnews.pdf.
Groups like *BASE. *INTERACT, etc., which in the Job Performance Summary Report contain the totals for jobs not shown individually, have incorrectly been shown with a certain pool number even when the jobs were running in different pools. Such cases have been changed to show ++ as pool number.
Changes to existing templates for graphs (Menu option 26) might fail to save the new specifications, even when save was requested. This error has been fixed.
Transfer of data from other LPARs to a “master LPAR” might fail when the defined maximum of 499 was reached. This has been solved by adding an automatic clean-up routine.
Run time for GiAPA expansion and analysis in cases where HotSpot data was collected, but not requested analyzed, have been reduced considerably.
V05M01C released June 20th, 2022
An additional F9/F21 has been added to the panel where individual threads may be selected for call stack analysis. In stead of selecting a thread, F9/F21 can be used to give a report containing the call stacks collected for all threads of the job, thus also allowing generation of a common pie chart.
Use of the “Copy to Clipboard” function for generated charts failed to include the GiAPA logo. The entire picture will now be copied.
V05M01B_ released May 11th, 2022
I/O-count exceeding 2,147,483,647 for one thread within a job caused overflow and abnormal end of data expansion and analysis. Changed to allow twice that value, and flagging any higher values with 9,999,999,999.
V05M01A released April 29th, 2022
When an interactive job generates GiAPA output in html format, GiAPA retrieves the IP-address of the server and runs the CL-command “Start PC command” to display the result immediately. Example: STRPCCMD PCCMD('start file://192.168.0.71/GIAPA/GIAPA15105.html')
For Windows PCs used as workstations and connected to the server through a company network the IP-address alone may not be sufficient for reaching the server IFS. In such cases a new command may be used to define a letter mapping the /GIAPA folder in the IFS – please see this link: https://www.giapa.com/GiAPA_html_results.pdf
For Apple computers, where the STRPCCMD was not implemented, the same link also describes how to access the GiAPA results created in html format.
- - - - - - - -
A GiAPA user wanted to see the actual values for each bar, column, etc. GiAPA Menu option 28, selection 2 has been updated to allow this. Furthermore, the user selected values for type of diagram, “Show Values” and color palette to use will now be kept and reused when a new version of an existing graph is generated.
V05M01 released February 8th, 2022 NEW GRAPHS CODED IN HTML
A Swedish GiAPA customer kindly drew our attention to the following warning given by IBM:
"Oracle has announced that Java Applet and WebStart functionality, including the Applet API, The Java plug-in, the Java Applet Viewer, JNLP and Java Web Start (containing the javaws tool) are all deprecated in JDK 9 and will be removed in a future release.”
For GiAPA this meant the end for the “GiapaNavigator” developed in JAVA.
We therefore had to produce new code in GiAPA V05M01 for generation of charts and diagrams. We changed the programs to generate HTML code and store the result in folder /GIAPA in the IFS.
To get familiar with this new feature we recommend the updated tutorial (28 minutes) available here: https://www.giapa.com/Tutorial10Graphics%20(Published)/
For most users this change will be a clear advantage, because the resulting graphs will appear immediately when a PC running Microsoft Windows is used as workstation.
The CL-command STRPCO (Start PC Organizer) is not implemented for Apple, so if you work from a MacBook the chart cannot be shown automatically. It should however still be easier than the old method requiring to start JAVA: Use Command-K within the “Finder”, and enter Smb://xxx.xxx.x.xxx/GIAPA (replace xxx.xxx.x.xxx with the server IP address). Then select the html file which contains the result – it is probably the newest.
November 2021: Incorrect logical read count may cause MCH1210 error in GIAPA192 statement 638.00. Please see this link describing the PTFs needed to avoid this error: https://www.ibm.com/support/pages/apar/MA49235
V05M00E released October 1st, 2021
The IBM Performance Collector API data should for single threaded jobs be 100 % accurate. However, the API only includes I/O counts (and overflows) for the initial thread. Multithreaded jobs often make most I/Os within other threads, and these I/Os are not included in the I/O-counts of the API.
To compensate for this limitation GiAPA now calculates the sum of physical I/Os of the thread I/O counts retrieved from the “Open List of Threads” API during HotSpots. This has proven to give a reasonable order of magnitude for the total auxiliary I/Os.
V05M00D released July 19th, 2021
Only very minor changes:
A "File not found" message could be issued when running option 14 (analysis) after manually having changed "Product Libary" to contain other than GIAPALIB.
Explanatory comments added to GiAPA Trial menu.
V0500C released July 3rd, 2021
Correction of problem for the HTML documents generated by GiAPA Menu Option 20 (Fully Automatic Performance Analysis), which showed special some characters used by non-English languages incorrectly (example: ö).
Apple Mac users only:
IBM did not implement the STRPCO command used by GiAPA Menu option 20 to show the results in html format for Apple Mac computers. To see the results in HTML format, Apple Mac users must
- Start IBM’s Access Client connection to the Power i Server.
- Open “Finder”.
- In the Finder window, press Command-K, which opens the “Create connection to Server” window.
- In the input field at the top, run the following statement: Smb://xxx.xxx.x.xxx/GIAPA , where xxx.xxx.x.xxx must be the IP-address of the server
- When connected, a window containing the generated output is displayed.
- Double click on the file name GIAPA20 to see the result in HTML format.
V05M00B released May 28th, 2021
Warning message GIA0070 ("Job exceeded defined limit fpr pages allocated") has been updated and will only be issued for job types B and I. This will avoid the message being sent e.g. for save operations which are handled by tasks (job type V). Field "Pages allocated" can now also show negative values - this change was needed because e.g. reorganization of a file with many deleted records or deletion of a large object could cause "Pages deallocated" to be higher than "Pages allocated" for a job.
V05M00A released May 10th, 2021
Rarely occurring sequence error in program GIAPA149 during data expansion and analysis has been corrected.
V05M00 released April 27th, 2021
A major new version introducing fully automatic performance analysis in the new option 20 of the GiAPA Menu. This is too much to explain here - a picture tells more than a thousand words, so please see
(We have experienced problems, mainly when using Apple (iPad or MAC), when trying to display the results from this new option in HTML format. We are working on finding a solution. In the meantime the results can be seen within the 5250 environment.)
V04M03K released August 10th, 2020
Command GIAPA045 (Where-used for programs) will now automatically display result if running in interactive job.
Automatic analysis of program performance (GiAPA Menu option 19) improved to run more efficiently and to avoid sporadic abnormal termination if scroll down was used after display of last record.
V04M03J released July 20th, 2020
A GiAPA user in Germany wanted to produce management level graphs showing overall resource usage without necessarily having to first run the full expansion and analysis of the collected data. A new command GIAPA055 allows generation of GiAPA standard graphs (Good Morning Report, and Top CPU Jobs) and any user defined graphs based on the raw=unexpanded data.
Error corrected: Jobs opening very many files could cause a sequence error message to be issued in the last step op the data analysis.
V04M03H released June 18th, 2020
The program matching file usage data from different HotSpot intervals has been entirely rewritten, allowing GiAPA’s analysis of the collected data to produce more correct file usage statistics and use less time. The logic of this program is probably the most complex within GiAPA because the data to be matched has no unique keys: A file may be opened several times within a job, and some jobs open hundreds of files. For details on the improvements obtained please follow this link: GiAPA file matching routine.
Also, as a consequence of servers getting more powerful, coping with more jobs running in parallel, GiAPA’s internal tables storing data for all active jobs have been increased to accommodate more jobs.
V04M03G released January 28th, 2020
Correction of command GIAPA050 causing error "Parameters do not match".
V04M03F released January 6th, 2020
Unexpanded GiAPA data members from year 2020 were not shown in correct descending date/time sequence - the sort routine has been rectified.
In connection with the V04M03E change of the default CPU limit for collecting HotSpot information it may be interesting to know exactly how much CPU the GiAPA data collection programs use. A new selection 7 from the GiAPA Menu option 16 will show this in detail.
Error correction: GiAPA Menu option 21, selection 2, might show a slightly too high value for the CPU used in excess of the assigned capacity for uncapped LPARs, if the CPU-assignment contained a decimal.
The text of the warning showed when analyzing GiAPA data from another LPAR was improved/clarified.
V04M03E released December 2nd, 2019
The default CPU limit for collecting HotSpots has been changed from 5.5 % to 4.0 %. The new limit will cause GiAPA to collect call stack and file usage data when a job has used 600 milliseconds within a 15 seconds' collection interval (the old 5.5 % limit coresponded to 825 milliseconds). GiAPA will therefore probably collect around 25 % more HotSpots, thus allowing detection of additional optimization potential.
GiAPA Menu option 19, selection 1 (analyze per user program name across all jobs) will now also show statistics for the most used JAVA classes.
GiAPA Menu option 18, selection 5 is modified to show the full class name for JAVA.
Minor improvements in the editing of call stacks for multithreaded jobs.
V04M03D released November 5th, 2019
Information in file GIAPA192P3 consolidated meaning that new members will occupy less space.
Help texts for “F1=Optimization tips” edited and extended for Automated Program Performance Analysis – help text is now available for > 95 % of active programs.
Duplicates removed from Lock Wait Report (GiAPA Menu option 19, selection 3).
Clear function for File Check data (GiAPA Menu option 85) improved to do a 100 % clean up.
Export function for unexpanded data (GiAPA Menu option 71) will now include record and I/O statistics from object description for most used files.
Improved layout for total I/O statistics on the Automatic File Analysis (GiAPA Menu option 19, selection 2).
V04M03C released August 10th, 2019
Important: Data field sent with message GIA0070 changed. In the previous release we announced a new “Allocation Trap Feature” sending warning message GIA0070 to QSYSOPR MSGQ if a job allocates more pages than a user defined limit. After request from a GiAPA user the first data field in the message was changed to 33 bytes with a layout like this
so that it could be used directly as operand after e.g. the HLDJOB CL command.
The remaining changes were minor justifications, like improved scroll up / scroll down functions and help text clarifications for automated performance analysis, and extension of an internal overflow table used during data collection.
V04M03B released May 7th, 2019
Minor update: New server models added, and layout of file for rejected records updated.
V04M03A released April 22nd, 2019
Error in GiAPA results:
The column reporting “Maximum Pages Used” has not been able to show values higher than 16,777,215 pages work space allocated by a job. The size of any data written to QTEMP is included in this total, and GiAPA has simply shown a blank value if more pages were allocated by a job – the field was not defined large enough to cope with e.g. the very large SQL table builds on today’s larger servers.
To rectify this, the fields sizes in several GiAPA data base files had to be increased. Therefore, updating GiAPA will take somewhat more time than normally – how much depends mainly on the number of records in the members of file GIAPA143P1 used to keep detailed information per collection interval for all jobs.
You might therefore consider using GiAPA option 82 to remove old expanded data before updating GiAPA to V04M03 (= reduce the time needed for file conversion), and to back up GiAPA data before updating.
The update routine has been changed: Instead of calling program GIAPAINST after having restored the new GiAPA update save file, you must use command GIAPAINST.
In addition, if you have stored the results of GiAPA analysis runs in other data libraries (and not only in GIAPALIB), you must also run command GIAPAINST against each such library to ensure that the files are converted to the new layouts.
This option was suggested by an American GiAPA user experiencing performance problems when a job allocated excessive temporary storage, sometimes adding up to several terrabytes – typically caused by SQL code looping while writing to a work file in QTEMP. European GiAPA sites have reported similar cases. Now GiAPA can warn QSYSOPR if excessive work space is allocated.
It works similar to GiAPA’s well-known and popular LoopTrap feature which sends a warning message to QSYSOPR if a job is looping, using CPU without showing any I/Os.
To activate the Allocation Trap Feature please use GiAPA Menu option 78 to modify the new User Installation Parameter defining the limit for when to send a warning. The shipped default value of 999,999,999 pages is so high that a message probably never will be sent.
Message GIA0070 severity code 60 is sent to QSYSOPR if a job allocates more work space than the number of pages (one page = 4K) specified in the Installation Parameters. If a job continues to exceed the limit message GIA0070 is repeated every 20 minutes.
Sort criteria 16 ‘Max pages used’ from the selection panel for the GiAPA Job Performance Summary report (Menu option 15) can be used to find a suitable number of pages to specify as limit for this parameter. Run the report on data for a few busy days and set the limit to e.g. ten times the highest value reported for a single job.
V04M02F released March 19th, 2019
Problem fixed: "Duplicate key" error message could appear when generating graph data from scheduled job.
Additional data needed for AI-analysis (GiAPA Menu option 19) will be included when exporting collected performance data in unexpanded format (GiAPA Menu option 71).
GiAPA Menu option 24 (Statistics by current user) has been improved to include
- new selection 2 combining current user and job user
- new F10 option resulting in summary by user name
Text for command GIAPA021 (display user index attributes) corrected
V04M02E released November 6th, 2018
Improved report of programs/jobs in lock wait status (Menu option 19, selection 3).
Batch jobs using commands to generate several sets of graph input data could end abnormally with duplicate record message due to timing problem caused by the high processor speed of very large servers.
Text for actual program activity in automated performance analysis improved.
Repeated generation of call stack pie charts for a generic job name could result in incorrect data being shown.
V04M02D released September 24th, 2018
Error correction: Problem in data collection program code could cause expansion to terminate abnormally with security code error type F.
Text generated for stacked diagrams improved.
V04M02C released August 23rd, 2018
GiAPA Menu option 16 selection panel was improved to keep any specified job name selection.
Both data collection and analysis runs are modified so that the cause of any problems for a GiAPA run (e.g. locks of an object needed) is shown in a new GIA0003 message appearing in the job log.
Problem fixed: F10=Details from Job Performance Summary Report would cause dump if no detail records were kept for the job.
Run date and from/to time was added to pie charts showing active programs of job call stack.
Problem fixed: Generation of graph data would sometimes fail if F21 was used to limit time interval from Call Stack report.
Explanatory texts for active function in option 19, selection 1 (Program analysis) was improved.
Problem fixed: Use of F4 from option 19, selection 1 (Program analysis) might fail to show previous program, and change of number of HotSpots to show might cause the first record to be skipped.
Code has been added to support the FTP keywords PORT and SECCNN, thus enabling the use of the values *IMPLICIT and *KERBEROS.
Batch job use of command GIAPA110 to start GiAPA data collection while collection was already active would stop the old collection and start a new. This was changed to cancel the new start if collection is running.
V04M02B released May 23rd, 2018
Last GiAPA version that can run under operating system versions 5.4 and 6.1.
New Lock Wait report available as selection 3 under GiAPA Menu option 19.
Processor type is displayed in stead of processor feature, and display of server type and PVU is added to "Statistics from Data Expansion". Furthermore, to support Power 9 servers, auxiliary storage is now shown in GB and memory size in MB (hitherto MB and KB), and main storage size exceeding 2 TB can now be handled.
Due to a change in Op.Sys 7.3 the generation of some GiAPA diagrams could give error message "Input data incomplete" - this problem has been removed.
V04M02A released Febrary 28th, 2018
User total for seconds of wait time on job queues is defined with "only" 9 digits. Data on a GiAPA customer server exceeded that limit and caused GiAPA expansion run to end abnormally. Now the maximum of 999,999,999 (= 277 thousand hours) will be recorded if the user total has a higher value.
V04M02 released February 15th, 2018
New alternative and powerful ways of analyzing program and file performance
The GiAPA Menu comes now with a new option 19 “Program or file performance analysis”. This option provides complete usage statistics for the programs or files based on HotSpots for all jobs included in the expanded data.
Traditionally GiAPA mainly pinpointed jobs that cause many HotSpots. Option 19 analyzes the data across all jobs and will in addition display/identify a program or file that only appears in a few HotSpots per job if:
- the job is running very frequently, or
- the program or file is used by many different jobs.
In addition the new analysis
- combines more information on one screen, allowing for easier identification of optimization potentials, and
- offers tips that reveal which code modifications may result in improved performance.
The overall idea is of course in line with GiAPA’s original objective: to enable the average programmer or operator to become an expert in application performance optimization.
Please see the description of the new options in slides 10 through 21 of Tutorial 8 !
This release also allows selection on lower case job names (i.e., task names are also supported), and gives a more correct I/O count in HotSpots for data base files of long running jobs being active more than one day.