Configuration Best Practices › Tuning for Improved Performance and Lower TCO › Tailor System Generation Parameters

Tailor System Generation Parameters

Set CA IDMS system generation parameters to the recommended value in the tables shown in Additional Considerations, and monitor your system performance to fine tune the parameter setting. It is usually best to over-allocate resources such as storage pools, program and reentrant pools, RCEs, and so on. As a general rule, it is better to be slightly over-allocated at the beginning and allow the system to run through a peak load period. You can then use the peak load statistics to reconfigure the system to optimum levels.

Business Value:

Adjusting the CA IDMS system generation parameters appropriately for your site will provide optimal runtime performance of a CA IDMS system. This will help to meet Service Level Agreements for response time and effectively use available computing resources.

Additional Considerations:

The following tables contain recommended values for CA IDMS system generation parameters. The tables contain recommendations for the following system generation statements:

• SYSTEM
• ADSO
• PROGRAM
• TASK
• STORAGE POOL and XA STORAGE POOL

The parameter values fall into these three categories and are listed under the column Recommendation in the tables that follow.

• Best Setting-If there is a value in the Recommendation column, then this value has been found to be the best setting for that parameter.
• Set then Tune-If there is a value in the Recommendation column, followed by <site specific>, then the value listed is a reasonable starting point but will need to be monitored and tuned using available CA IDMS Performance Monitor functions and commands, as well as various DCMT and OPER commands.
• Site specific-If <site specific> is listed in the Recommendation column, then you need to determine the appropriate values for your site. Guidelines for determining the appropriate value are provided in the notes following the table. When upgrading to a new release of CA IDMS, the site specific parameters related to Storage and Program Pools should be set at a minimum to the values specified in the previous release; however, we recommend that you set the values 10 to 20 percent higher, and then tune down after running through your busiest processing period.

  System Statement

The following table represents recommended values for parameters set using the SYSTEM statement.

Notes:

LOADLIST

LOADLIST tailoring can provide significant CPU savings and response time reduction but incorrect specification can and will result in exactly the opposite. Initially try using the default SYSLOAD loadlist with the NODYNAMIC option of the Program Definition statement.

MAX ERUS/MAX TASK

There are no magic numbers for these two parameters. They need to be determined by each site. We recommend that you specify high values for both on the system statement, and then use the following command to vary max tasks up or down until the optimum number is found.

DCMT VARY ACIVE TASK MAX TASK nnn

You could initially set both parameters to 100. If you are using CA IDMS/DC, then you set the MAX ERUS parameter to accommodate any batch to CV jobs you may need to run. Sites that only run CICS COBOL or PL/I DML transactions can set the MAX TASK value to a very low number; 5 is recommended. A low value allows for any online work that may need to be processed (for example, signing on through the system console).

CICS TS sites should also note that if the value of the CA IDMSMAX ERUS parameter is set below the CICS MAXTASK parameter, they could experience 1473 abends. If MAX ERUS is reached, meaning all the allocated ERE control blocks that maintain the link between the external session and the central version are in use, the next CICS task that needs to interact with the central version will not be able to and will abend with a 1473 error code. You can prevent this by specifying a value for MAX ERUS that is higher than the CICS MAXTASK parameter value. In this way, CICS tasks will be able to acquire an ERE. You can then control the number of CICS tasks that are allowed to run within the central version using the DCMT VARY ACTIVE TASK MAX TASK command.

PROGRAM POOLs

Select values for your PROGRAM POOL parameters and then tune them. You can determine the appropriateness of your current settings by using various statistics such as the output of the DCMT DISPLAY ACTIVE PROGRAMS and DCMT DISPLAY ACTIVE REENTRANT PROGRAM commands.

When upgrading to a new release of CA IDMS, the size of all below the line PROGRAM and REENTRANT pools should be allocated at the size they were in the prior release.

STORAGE POOL

The STORAGE POOL parameter specifies the size of storage pool 0 that is used for system type storage. This pool could and would be used for other types of storage, user and shared, if there were no other pools defined in the system; however, you will always define other pools. The 1200 K allocation is probably on the high side, and can be reduced after determining how much storage is used during your peak processing period. Use statistics displays from DCMT, OPER, and CA IDMS Performance Monitor to help determine the appropriateness of your current setting. For example, the OPER WATCH STORAGE command displays the number of times a short-on-storage (SOS) condition occurred and the number of waits for storage.

SYSTRACE

Internal system trace data is sometimes needed for debugging. When a problem exists, CA Technical Support may ask you to set the SYSTRACE parameter to ON and set the number of entries to 9999 using SYSGEN or higher using DCMT vary commands.

XA STORAGE POOL

This parameter controls storage pool 255 that is the XA equivalent of storage pool 0. This pool must be defined. All system type storage is allocated out of this pool. If this pool fills, the storage pool 0 will be used; however, using the storage pool 0 should be avoided. An allocation of 12000 is likely higher than what you will actually need, but it is better to over-allocate than under-allocate, as mentioned earlier.

ADSO Statement

The following table represents recommended values for parameters set using the ADSO statement.

Program Definition Statement

The following table represents recommended parameter values for each PROGRAM statement.

Task Definition Statement

The following table represents a recommended parameter value for each system generation TASK statement.

STORAGE POOL and XA STORAGE POOL Statements

The following table represents recommended parameter values for the STORAGE POOL and XA STORAGE POOL statements. These statements are used to create user-defined secondary storage pools.

Notes:

STORAGE POOLs

The size of USER storage pools, those numbered 1 thru 127 for below the line pools and numbered 128 thru 254 for above the line storage pools, should initially be allocated to at least the size of the USER storage pools used in the prior release. As mentioned previously, it is best to allocate these pools larger than they were in the prior release. You may want to increase the size by 25 percent initially, and then cut them back after finding the true HWM after processing through your peak period.

The below the line and above the line storage pools should be defined as shown in the following lists.

Below the line storage pools:

• One storage pool that contains types (User, User Kept)
• One storage pool that contains types (Shared, Shared Kept)
• One storage pool that contains types (Terminal, Database)

The sizes of these below the line storage pools can be very small as the amount of below the line storage allocated is minimal.

Above the line (XA) storage pools:

• One storage pool that contains types (User, User Kept)
• One storage pool that contains types (Shared, Shared Kept)
• One storage pool that contains types (Terminal, Database)

As a size recommendation, we suggest that you overallocate. If you had one XA storage pool that contained all types in the prior release and its size was 10 M, allocate each pool at 10 M. You will not need all of this storage. After running through your peak processing time ,you can determine the HWMs and adjust as needed.

The High Performance Storage Protect Option requires the storage pools be defined in this manner. These definitions are also valid for non-protect, as well as full storage protect, usage.

More Information:

For detailed descriptions of all system generation (SYSGEN) options and parameters, see the CA IDMS System Generation Guide.

For descriptions and examples of all DCMT and OPER commands that can be used to monitor resources within the CA IDMS system, see the CA IDMS System Tasks and Operator Commands Guide.

For detailed online and batch reporting facilities available in the CA IDMS Performance Monitor product, see the CA IDMS Performance Monitor User Guide and the CA IDMS Performance Monitor System Administration Guide.