Using Strobe for Advantage CA-IDMS

This section covers some general information you need to start and configure Strobe for Advantage CA-IDMS. It also explains how to index ADS/O dialogs for the Strobe Performance Profile. Sections describing how to create the indexes using either ISPF panels or the Strobe command language are included.

Invoking Strobefor Advantage CA-IDMS

Depending on whether you are measuring an online region or a batch job, different ways are used to invoke Strobe for Advantage CA-IDMS. This section briefly explains the differences. You can find detailed instructions for submitting measurement requests in the Strobe/ISPF Online Tutorials and in the Using-Strobe-to-measure-online-applications-and-batch-programs.

CA-IDMS-DC/CV Online Regions

If you want measurement data (including ADS/O attribution data) from a CA-IDMS-DC/CV online region, Strobe automatically invokes Strobe for Advantage CA-IDMS when:

• You measure a target program whose name (as given in the PGM= operand of the EXEC statement that invokes the job step) begins with IDMS.
• Your Strobe system programmer specified an appropriate alias to Strobe during installation.

Otherwise, you must specify that Strobe start Strobe for Advantage CA-IDMS when you add a measurement request by doing the following:

• With Strobe/ISPF, enter Y in the DATA COLLECTORS field of the Strobe- Add Active or Strobe- Add Queued panel, then specify Y in the IDMS field on the Strobe- Data Collectors panel (Strobe - Data Collectors Panel).
• With the Strobe command language, include the IDMS keyword parameter when you submit an ADD operation.

If you have no interest in CA-IDMS measurement data, you can stop Strobe from automatically invoking the data collector by doing the following:

• With Strobe/ISPF, enter Y in the DATA COLLECTORS field of the Strobe- Add Active or Strobe- Add Queued panel, then specify N in the IDMS field on the Strobe- Data Collector panel.
• With the Strobe command language, include the NOIDMS keyword parameter when you submit an ADD operation.

If you have no interest in ADS/O attribution data that is available in the Dialog CPU Summary and CPU Usage by ADS/O Statement reports, you can suppress ADS/O attribution data collection by doing one the following:

• With Strobe/ISPF, enter Y in the OTHER PARAMETERS field of the Strobe- Add Active or Strobe- Add Queued panel, then specify on the OTHER PARAMETERS panel:
  IDMS=NOADSO

• With the Strobe command language, submit an ADD operation and include the operand:
  IDMS=NOADSO

Strobe- Data Collectors Panel

CA-IDMS Central Version and Local Mode Batch Regions

Strobedoes not automatically invoke DML attribution when you measure a batch job that issues database DML requests. To obtain DML attribution, you must specify the feature. To specify the DML feature do one the following:

• With Strobe/ISPF, enter Y in the DATA COLLECTOR field of the Strobe- Add Active or Strobe- Add Queued panel, then enter Y in the IDMS BATCH DML field shown previously in Strobe - Data Collectors Panel.
• With Strobe command language, specify DML.

To show wait related to DML sequence numbers, you must compile the program with the DEBUG DML option selected.

Indexing ADS/O Dialogs

When Strobefor Advantage CA-IDMS measures activity within dialogs that were generated with diagnostic tables, it can obtain the statement numbers of the ADS/O commands that are executing. These dialogs can then be indexed, allowing Strobe to associate an ADS/O statement number with its corresponding source statement text. When a dialog has been indexed, Strobe includes both the statement number and all or part of the source statement text in the CPU Usage by ADS/O Statement report. Otherwise, Strobe displays only the statement numbers (if diagnostic tables exist) or the ADS/O FDB CME offsets (if they do not).

You can index your dialogs by executing a separate batch job any time before creating the Strobe Performance Profile. When you request indexing, specify the name of the listing data sets created by the CA-IDMS ADSORPTS utility program, as described below.

Unless you generated diagnostic tables when you compiled your dialogs, Strobe can attribute activity only to an ADS/O FDB CME offset, not to an actual statement number, and it cannot index any of the dialogs. If your Performance Profile displays only the CME offsets, regenerate the dialogs with diagnostic tables and then rerun the measurement session so you can index the dialog.

Producing ADS/O Dialog Listing Data sets

Before you can execute the Strobe ADS/O Indexer STRXADS, you first have to generate the listing data set Strobe uses to match the dialog measurement data it has gathered with the corresponding ADS/O statements.

To generate a listing data set containing the source code for each ADS/O dialog, run the CA-IDMS utility program ADSORPTS with the run-time parameter REPORTS=ALL. (Refer to the Using-Strobe-to-measure-online-applications-and-batch-programs for a list of the DCB required parameters to create the listing data set.) To specify multiple dialogs, use the ADSORPTS DIALOGS parameter, which accepts any of the following values:

• ALL (generate reports for all dialogs).
• Dialog-name (generate a report for a single dialog). To specify multiple dialog names, enter them in a comma-delimited list enclosed by parentheses.
• Dialog-mask-a (generate a report for any dialog that matches the mask criteria).
• Low-dialog-name-high-dialog-name (generate a report for any dialogs within a specified range). The hyphen is required and cannot have surrounding blanks.

The following job control statements will produce reports for all versions of the dialogs named DIAG1, DIAG2, and DIAG3. The ADS/O listing will be placed in the data set named SA.STDM.ADSORPTS.DIAG1.

After you have created the listing data set, supply its name to the Strobe Indexer STRXADS, either through Strobe/ISPF or with the STROX procedure, as described in the next section.

An ADS/O Indexer map data set is identical in format to those produced by the other Strobe Indexers. When you produce a Performance Profile, you can concatenate ADS/O Indexer map data sets with one another and with map data sets produced by the other Strobe Indexers.

Using Strobe/ISPF to Index ADS/O Dialogs

To index your ADS/O dialogs with Strobe/ISPF, select Option 5, INDEX, from the Strobe Options menu. Then enter the following values on the Strobe- Index to Create a Map Dataset panel (Strobe - Index to Create a Map Dataset Panel):

1. Select background or foreground processing by entering the appropriate code in the OPTION field.
2. Enter Y in the ADS/O field.
3. Specify the name of the map data set you want to create in the OUTPUT: Map Dataset field.
4. Specify the name of the listing file produced by the ADSORPTS utility program in the first INPUT: Compiler SYSPRINT Datasets field.

For example, if you want to submit a background processing request to create a map data set called SA.STDM.DIAG1.MAP from the ADSORPTS listing in SA.STDM.ADSORPTS.DIAG1, you would enter the information on the Strobe- Index to Create a Map Dataset panel as shown in the following figure.

Strobe - Index to Create a Map Dataset Panel

Using the STROX Procedure to Index ADS/O Dialogs

Execute the Strobe procedure STROX to index the program and create a map data set, supplying the following parameters:

• MAPPGM — STRXADS (the name of the Indexer program)
• LISTDSN — the output file name produced by the ADSORPTS utility program

• MAPDSN — the name of the output map data set

  //INDEX    EXEC  STROX,
  //               MAPPGM=STRXADS,
  //               LISTDSN='SA.STDM.ADSORPTS.DIAG1',
  //               MAPDSN='SA.STDM.DIAG1.MAP'

To concatenate multiple reports from the ADSORPTS utility and produce only one map data set, override the LISTING DD statement in the STROX procedure with a concatenation of the ADSORPTS listings data set names.

Using Your Own Job Control Statements

Alternatively, you can code the job control statements to execute the CA-IDMS Indexer as shown in the following code block:

Tailoring Strobe Reports

This section explains how to tailor reports in a Performance Profile so only specific measurement data you are seeking is provided. For a complete description of creating Strobe Performance Profile reports, refer to the Using-Strobe-to-measure-online-applications-and-batch-programs.

Suppressing Attribution Reports in the Performance Profile

To limit or suppress the Attribution of CPU Execution Time and Attribution of CPU Wait Time reports in the Performance Profile, you can specify any of the parameters below.

Suppressing Attribution Reports with Low Activity

Suppress Attribution reports for all system service modules in which the total CPU or wait time percentage is less than a specified baseline by doing the following:

• With Strobe/ISPF, enter Y in the Tailor Reports field of the Strobe- Produce a Performance Profile panel, then, on the Strobe- Tailor Reports panel (See the following figure), specify a baseline percentage between 0 and 99.9 in the Compress below % column of the ATTRIBUTION Reports field.
• When you submit a batch job using the STROE or STROXE procedure, include ATTR=nn.n, where nn.n is the baseline percentage (default 2).

Strobe- Tailor Reports Panel (Scrollable)

Suppressing All Attribution Reports

To suppress all the Attribution reports, do one of the following:

• With Strobe/ISPF, enter Y in the Tailor Reports field of the Strobe- Produce a Performance Profile panel, then enter Y in the Suppress column of the ATTRIBUTION Reports field on the Strobe- Tailor Reports panel.
• With the STROE or STROXE procedure, include the NOATTR parameter.

Suppressing CA-IDMS-Specific Attribution Reports

To suppress the Attribution of CPU Execution Time and Wait Time reports for CA-IDMS ADS/O dialogs and batch database application programs only, do one of the following:

• With Strobe/ISPF, enter Y in the Tailor Reports field of the Strobe- Produce a Performance Profile panel, then enter Y in the Suppress reports for IDMS field on the Strobe- Tailor Reports panel.
• With the STROE or STROXE procedure, include the NOATTR=IDMS parameter.