CA Database Management Solutions for IMS for z/os

Transcription

1 CA Database Management Solutions for IMS for z/os Best Practices Guide Version

2 This Documentation, which includes embedded help systems and electronically distributed materials, (hereinafter referred to as the Documentation ) is for your informational purposes only and is subject to change or withdrawal by CA at any time. This Documentation may not be copied, transferred, reproduced, disclosed, modified or duplicated, in whole or in part, without the prior written consent of CA. This Documentation is confidential and proprietary information of CA and may not be disclosed by you or used for any purpose other than as may be permitted in (i) a separate agreement between you and CA governing your use of the CA software to which the Documentation relates; or (ii) a separate confidentiality agreement between you and CA. Notwithstanding the foregoing, if you are a licensed user of the software product(s) addressed in the Documentation, you may print or otherwise make available a reasonable number of copies of the Documentation for internal use by you and your employees in connection with that software, provided that all CA copyright notices and legends are affixed to each reproduced copy. The right to print or otherwise make available copies of the Documentation is limited to the period during which the applicable license for such software remains in full force and effect. Should the license terminate for any reason, it is your responsibility to certify in writing to CA that all copies and partial copies of the Documentation have been returned to CA or destroyed. TO THE EXTENT PERMITTED BY APPLICABLE LAW, CA PROVIDES THIS DOCUMENTATION AS IS WITHOUT WARRANTY OF ANY KIND, INCLUDING WITHOUT LIMITATION, ANY IMPLIED WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE, OR NONINFRINGEMENT. IN NO EVENT WILL CA BE LIABLE TO YOU OR ANY THIRD PARTY FOR ANY LOSS OR DAMAGE, DIRECT OR INDIRECT, FROM THE USE OF THIS DOCUMENTATION, INCLUDING WITHOUT LIMITATION, LOST PROFITS, LOST INVESTMENT, BUSINESS INTERRUPTION, GOODWILL, OR LOST DATA, EVEN IF CA IS EXPRESSLY ADVISED IN ADVANCE OF THE POSSIBILITY OF SUCH LOSS OR DAMAGE. The use of any software product referenced in the Documentation is governed by the applicable license agreement and such license agreement is not modified in any way by the terms of this notice. The manufacturer of this Documentation is CA. Provided with Restricted Rights. Use, duplication or disclosure by the United States Government is subject to the restrictions set forth in FAR Sections , , and (c)(1) - (2) and DFARS Section (b)(3), as applicable, or their successors. Copyright 2013 CA. All rights reserved. All trademarks, trade names, service marks, and logos referenced herein belong to their respective companies.

3 CA Technologies Product References This document references the following CA Technologies products: CA Chorus Software Manager (CA CSM) CA Chorus (CA Chorus) CA Database Analyzer for IMS for z/os (CA Database Analyzer) CA Database Copier for IMS for z/os (CA Database Copier) CA Database Organizer for IMS for z/os (CA Database Organizer) CA Datacom /AD (CA Datacom/AD) CA Datacom /DB (CA Datacom/DB) CA Dataquery for CA Datacom (CA Dataquery for CA Datacom) CA High Performance Recovery for IMS for z/os (CA High Performance Recovery) CA Secondary Index for IMS for z/os (CA Secondary Index) CA Secondary Index Builder for IMS for z/os (CA Secondary Index Builder)

4 Contact CA Technologies Contact CA Support For your convenience, CA Technologies provides one site where you can access the information that you need for your Home Office, Small Business, and Enterprise CA Technologies products. At you can access the following resources: Online and telephone contact information for technical assistance and customer services Information about user communities and forums Product and documentation downloads CA Support policies and guidelines Other helpful resources appropriate for your product Providing Feedback About Product Documentation If you have comments or questions about CA Technologies product documentation, you can send a message to techpubs@ca.com. To provide feedback about CA Technologies product documentation, complete our short customer survey which is available on the CA Support website at Best Practices Guide Process These best practices represent years of product experience, much of which is based on customer experience reported through interviews with development, technical support, and technical services. Therefore, many of these best practices are truly a collaborative effort stemming from customer feedback. To continue and build on this process, we encourage users to share common themes of product use that might benefit other users. Please consider sharing your best practices with us. To share your best practices, contact us at techpubs@ca.com and preface your subject line with "Best Practices for product name" so that we can easily identify and categorize them.

5 Documentation Changes The following documentation updates have been made since the last release of this documentation: Introduction (see page 9) This chapter has been updated to comply with Mainframe 2.0 standards. Implement a Proactive Preventive Maintenance Strategy (see page 13) This is a new topic that describes recommended maintenance strategy. The Staying Current on Maintenance and Releases topic has been removed.

6

7 Contents Chapter 1: Introduction 9 Chapter 2: Installation Best Practices 11 Using CA CSM and Electronic Software Delivery Implement a Proactive Preventive Maintenance Strategy Review ITK STC Health Check Processing Signing Up for Hyper Notifications Registering on CA Support Website Applying IBM Maintenance Chapter 3: Configuration Best Practices 17 Use Sample JCL to Enable Processing Set Defaults Using Global Parameters Customize and Start the ITK STC Use Warm Start Options to Shut Down and Restart ITK STC and Pause Enabled IMS Control Regions Generate a Dump when An Abend Occurs Use EXCP and MMGR for Reading Databases Turn Off Largest Records Data Collection Use Expand and Compress Defaults During Unload and Reload Evaluate Use of DBDCHANGE Keyword in CA Database Organizer Use LBI for an Output Tape Data Set Optimize Database Unload Allocate Buffered DSGs Rebuild Indexes During Reorganization Increase Sort Work File Size Use IIR Reporting Functions Chapter 4: Integration Points 35 Perform Pointer Checking During Backup Perform Backup and Pointer Checking After Recovery Evaluate Performance Profiles for Randomizers Invoke HDSORT Dynamically Rebuild Indexes Automatically Contents 7

8 Appendix A: Performance Best Practices 41 CA Database Analyzer Performance Improvements CA Database Copier Performance Improvements CA Database Organizer Performance Improvements CA High Performance Recovery Performance Improvements CA Secondary Index Builder Performance Improvements Index 43 8 Best Practices Guide

9 Chapter 1: Introduction The guide provides a brief introduction to the CA Technologies mainframe management strategy and features, and describes the best practices for installing and configuring your product. The intended audience of this guide is systems programmers and administrators who install, maintain, deploy, and configure your product. Chapter 1: Introduction 9

10

11 Chapter 2: Installation Best Practices The following topics provide best practices for implementing Mainframe 2.0 strategies, and installing and configuring CA Database Management Solutions for IMS for z/os. This release includes the following Mainframe 2.0 features: CA CSM and Electronic Software Delivery Standardized installation using SMP/E Integration with z/os Health Checker This section contains the following topics: Using CA CSM and Electronic Software Delivery (see page 11) Implement a Proactive Preventive Maintenance Strategy (see page 13) Review ITK STC Health Check Processing (see page 14) Signing Up for Hyper Notifications (see page 15) Registering on CA Support Website (see page 15) Applying IBM Maintenance (see page 16) Using CA CSM and Electronic Software Delivery CA CSM provides a web interface, which works with Electronic Software Delivery (ESD) and standardized installation, to provide a common way to manage CA mainframe products. Use CA CSM and ESD to acquire, install, and maintain the CA Database Management Solutions for IMS for z/os. CA CSM lets you download product and maintenance releases over the Internet directly to your system from the CA Support website. The same interface is used to install the downloaded software packages using SMP/E. When you order the CA Database Management Solutions for IMS for z/os, you receive the authorizations and instructions to access, download, and prepare the installation files without the need for a physical tape. Maintenance can be applied as needed without waiting to obtain a special tape or cartridge. Chapter 2: Installation Best Practices 11

12 Using CA CSM and Electronic Software Delivery Additional Considerations: Downloaded files are delivered in a new pax.z format that eliminates the need to create an installation cartridge from an ESD image. You can download the pax.z file from the CA Support website. To download the files and create the installation data sets, perform the following steps: 1. Review the instructions for installing from DASD provided in the CA Enhanced Electronic Software Delivery Pax-Enhanced ESD User Guide before proceeding with the remaining steps. This guide is available from the following link: 2. Prepare your mainframe environment to receive the pax.z file. Important! This step requires a UNIX System Services (USS) directory to receive the pax file and to perform the unpack steps. Verify that all users working with ESD pax files have writer authority to the directory and that enough space has been allocated to receive the file before you start the download. Note: For additional ESD information, go to the mainframe web page on the CA website. Under Events, we offer an ESD webcast to explain the enhanced ESD process. 3. Download the pax.z file (include your USS directory and TCP/IP profile data set in the JCL). 4. Unpack the ESD pax file and copy the ESD installation files to z/os data sets. 5. Proceed with the product installation as described in the CA Database Management Solutions for IMS for z/os Installation Guide. To install a new release of the products or apply maintenance (application of all or selective fixes within the same release), perform SMP/E RECEIVE, APPLY, and ACCEPT processing through native SMP/E or the Software Delivery Manager (SDM). Note: If you used ESD to install a new or maintenance release of the CA Database Management Solutions for IMS for z/os, specify also the DASD high-level qualifier to use the files obtained from ESD in the installation process. Detailed information is provided in the CA Database Management Solutions for IMS for z/os Installation Guide. More Information: For more information about CA CSM, see the CA CSM Product Guide or CA CSM Readme. 12 Best Practices Guide

13 Implement a Proactive Preventive Maintenance Strategy Implement a Proactive Preventive Maintenance Strategy CA Technologies formerly delivered product maintenance using Service Packs. We have replaced this model with CA Recommended Service (CA RS) for z/os, which provides more flexibility and granular application intervals. CA RS is patterned after the IBM preventive maintenance model, Recommended Service Upgrade (RSU). With CA RS, you can install preventive maintenance for most CA Technologies z/os-based products in a consistent way on a schedule that you select (for example, monthly, quarterly, annually, and so on). We recommend that you develop and implement a proactive preventive maintenance strategy whereby you regularly apply maintenance. You could follow the same schedule that you use to apply IBM maintenance, or you could implement a schedule for CA Technologies products only. Keeping your products current with maintenance helps your team remain productive and minimize errors while safely protecting your systems. If you do not install preventive maintenance regularly, you risk encountering known problems for which we have published and tested fixes. Our mainframe maintenance philosophy is predicated upon granting you the flexibility to maintain your sites and systems in a manner that is consistent with industry best practices and your site-specific requirements. Our philosophy focuses on two maintenance types. Understanding each of these types can help you maintain your systems in the most efficient manner. Note: This philosophy applies to the CA Next-Generation Mainframe Management stack products. For legacy products, contact CA Support for maintenance details. Corrective Maintenance Helps you address a specific and immediate issue. This type of maintenance is necessary after you encounter a problem. CA Technologies may provide a test APAR if a new problem is uncovered, or a confirmed PTF if the problem has already been resolved. Your primary goal is to return your system to the same functional state that it was before you experienced the issue. This type of maintenance is applied on an as-needed basis. Chapter 2: Installation Best Practices 13

14 Review ITK STC Health Check Processing Preventive Maintenance Lets you apply PTFs that CA Technologies has created and made public. You may or may not have experienced the issues that each PTF addresses. CA RS provides a way to easily identify published maintenance that has been successfully integration-tested with other CA Technologies products, current z/os releases, and IBM subsystems, such as CICS and DB2. Major CA RS service levels are published quarterly, with updates for HIPER and PE-resolving PTFs that are published monthly. After you test the CA RS level, we recommend that you accept that level before you apply a new CA RS level. You can initiate a maintenance installation activity at any time, and install the current CA RS level of maintenance (recommended) or an earlier level. In addition, you may want to install maintenance to support a new hardware device, software upgrade, or function using our FIXCAT method. For all maintenance, before you initiate any maintenance action, obtain the current SMP/E HOLDDATA. Important! CA Chorus Software Manager (CA CSM) - formerly known as CA Mainframe Software Manager (CA MSM) - is an intuitive web-based tool that can automate and simplify many CA Technologies product installation and maintenance activities. We strongly recommend that you use CA CSM to maintain your CA Technologies z/os-based products. More Information: For the steps to apply preventive maintenance using CA CSM or from CA Support Online on see the Installation Guide for your product and the CA CSM online help. Review ITK STC Health Check Processing The health checks continually monitor the system and software to provide feedback on whether the software continues to be configured optimally. Review the messages generated from the following health check that is provided for the ITK STC to verify that all of the required subtasks are executing: Contains the jobname of the ITK started task. Early notification on the potential need to fix the ITK STC helps avoid unscheduled downtime and assist you in pro-actively maintaining your software. 14 Best Practices Guide

15 Signing Up for Hyper Notifications Additional Considerations: The ITK started task attaches many tasks on which work can be dispatched. It is critical that all tasks are available and running. This health check verifies that each subtask is running and that there are no abending tasks. When all subtasks are operating, the health check generates a message that indicates the ITK STC is operating and that there are no abending tasks. When one or more subtasks contain errors, the health check generates an exception error that assigns a severity level to the exception and lists the abending tasks. This message provides an explanation, system action, operator response, system programmer response, problem determination, source, and reference documentation. Additional messages are also generated to help diagnose and correct the condition. More Information: For more information about the messages generated from the health check, see the CA Database Management Solutions for IMS for z/os Message Reference Guide. Signing Up for Hyper Notifications CA Support is always alert to any known problems that could affect customers. We recommend that you sign up for hyper notifications for all CA Technologies products on the CA Support website so that you receive notification of potential problems as soon as possible. Select Hyper notifications and follow that dialog to receive notifications for your products. Early notification about the potential need to fix the CA Database Management Solutions for IMS for z/os helps avoid unscheduled downtime and assist you in proactively maintaining your software. Registering on CA Support Website Register your database team members on the CA Support website. The CA Support web pages contain not only fixes, but also valuable information, technical articles, and recommendations regarding CA Technologies products. Go to and register your address and the CA Database Management Solutions for IMS for z/os that you have. Chapter 2: Installation Best Practices 15

16 Applying IBM Maintenance The CA Support website has guides available for download and searchable knowledge forums: Knowledge Base Search Support By Product or Solution Download Center Licensing CA Tech Insider Subscription Compatibilities CA Support users are able to review tickets (both open and closed), search the knowledge database, and download copies of guides. This access gives them a significant productivity boost, without having to disturb other employees with questions. Programmers and DBAs can be more productive by being able to get questions answered quickly by searching knowledge bases and communicating with CA Support on any open issues. Learning how to navigate the CA Support website can help you become more knowledgeable about CA Technologies products. Greater product knowledge helps you get more value out of them. You can benefit from shorter time to resolution when other users report problems on the CA Support website, and better quality assurance when upgrading to new software. Reviewing technical documents for your products can unearth previously unused functionality or can provide enhancements to the usage of known functionality. The technical documents can also describe ways of using the products that had not been considered and could be beneficial. Regular monitoring of the CA Support website verifies that you are aware of the latest software changes and information about upgrade functionality that meets future requirements. Applying IBM Maintenance We recommend that you review the current list of IBM APARs and apply the APARs that are appropriate to your environment. To obtain a current list of IBM APARs, contact CA Support at Missing APARs can affect the operation or performance of the CA Database Management Solutions for IMS for z/os. 16 Best Practices Guide

17 Chapter 3: Configuration Best Practices This section contains the following topics: Use Sample JCL to Enable Processing (see page 17) Set Defaults Using Global Parameters (see page 18) Customize and Start the ITK STC (see page 19) Use Warm Start Options to Shut Down and Restart ITK STC and Pause Enabled IMS Control Regions (see page 20) Generate a Dump when An Abend Occurs (see page 21) Use EXCP and MMGR for Reading Databases (see page 22) Turn Off Largest Records Data Collection (see page 23) Use Expand and Compress Defaults During Unload and Reload (see page 23) Evaluate Use of DBDCHANGE Keyword in CA Database Organizer (see page 24) Use LBI for an Output Tape Data Set (see page 25) Optimize Database Unload (see page 25) Allocate Buffered DSGs (see page 29) Rebuild Indexes During Reorganization (see page 30) Increase Sort Work File Size (see page 32) Use IIR Reporting Functions (see page 33) Use Sample JCL to Enable Processing Sample JCL is provided in high-level.cimtsamp for you to edit and customize as needed. This provides an easy method to get started using the CA Database Management Solutions for IMS for z/os for specific functions. By editing and customizing the sample JCL, you are able to quickly get started using the CA Database Management Solutions for IMS for z/os at your site. More Information: See the product reference guides for detailed information about the EXEC, DD, and control statements that can be specified. Chapter 3: Configuration Best Practices 17

18 Set Defaults Using Global Parameters Set Defaults Using Global Parameters The CA Database Management Solutions for IMS for z/os support a function known as global parameter processing. This function lets you establish function-specific defaults, in a global parameter data set, that supplement any JCL-coded control statements. We recommend that you use global parameters to set installation-wide defaults and simplify JCL preparation. To set installation defaults using global parameters, perform the following steps: 1. Customize the sample product members in the parmlib (high-level.cimtparm) to contain the desired parameters. 2. Copy the modified members to the production global parameter library. 3. Do either of the following to activate the global parameters: Assemble and link a DFSMDA member for the parmlib data set. Code a JCL statement for it under the IDIPARM DD. After this feature is activated, the CA Database Management Solutions for IMS for z/os will always attempt to allocate the global parameter data set, and to read the control statements for the function being performed. Using global parameter defaults simplifies the preparation of JCL for individual jobs and helps eliminate duplicate statements. This saves you time setting up new jobs in your environment, and lets you establish installation-specific standards that let you reuse common data as needed. Any parameters that are not provided in the JCL will be found in the global parameter data set for the function being performed. Additional Considerations: This feature is optional. If you do not update these members, the provided global defaults match the documented defaults for each product. No user JCL changes are required to implement this function, but if desired, a DD statement may be added to override the standard data set name. More information: The CA Database Management Solutions for IMS for z/os Installation Guide contains more information about modifying these members and additional information about syntax rules. Do not change the provided settings unless you are familiar with the affected function and the impact the change you make will have on its processing. 18 Best Practices Guide

19 Customize and Start the ITK STC Customize and Start the ITK STC The ITK started task (STC) enables Full Function and HALDB online and one-step reorg in CA Database Organizer, snapshot image copy processing (FLASH=YES) in CA Database Copier, and Recovery Analyzer in CA High Performance Recovery. The ITK started task is needed to issue IMS commands and to use BMP pause and resume functionality. It can also be used to enable the option MMGRIO=YES (Media Manager) with an unauthorized CA Database Management Solutions for IMS for z/os products load library. The ITK started task must be running on the same system as the target IMS system and on every LPAR containing a member of the IMSplex in a data-sharing environment. One ITK started task can service multiple IMS subsystems, even when each is running a different version of IMS. However, you can execute multiple ITK started tasks on a single LPAR, for example, to run multiple online reorgs on two distinct IMS subsystems that reside on this LPAR, using two different versions of CA Database Management Solutions for IMS for z/os. Defining the parameters first reduces the time it takes to get the ITK STC started and running. Additional Considerations: When you first execute the ITK STC, it will interactively request values for parameters that have not yet been coded. You respond through the operator console. Any values you supply are rewritten into the ITK STC parameter member. If you do not supply a value for a specific parameter, you can give a null reply. You can include the startup parameters in the parmlib member with values or null values. More Information: For more information about how to customize and start the ITK started task, the ITK started task control statements and related commands, see the Installation Guide. Chapter 3: Configuration Best Practices 19

20 Use Warm Start Options to Shut Down and Restart ITK STC and Pause Enabled IMS Control Regions Use Warm Start Options to Shut Down and Restart ITK STC and Pause Enabled IMS Control Regions The ITK STC is shut down using the MVS STOP command (/P jobname). By default, a START of the ITK STC uses the warm start options for both the STC and PAUSE features (TYPE=W,PAUSE=W). There are no restart options associated with the starting of a PAUSE enabled IMS control region. We recommend that you use the default options of TYPE=W, PAUSE=W. When a shutdown is requested, active work in STC is permitted to quiesce or complete before shutdown. However, a shutdown of IMS with active FFOR or PAUSE activity causes an active FFOR job to fail. To determine if there is work being processed before requesting a shutdown, complete the following steps: 1. Issue the MVS Modify command to determine if there is active work being processed: /F stc_jobname,display ACTIVE 2. Issue the DISABLE command to prevent any new FFOR job from starting: /F stc_jobname,disable If the STC is not terminated, issue an /F stc_jobname,enable to allow new FFOR jobs. Using the default options, you can shut down and restart the STC and IMS regions in any order. This allows you to minimize the impact on your current operational procedures and select different technologies for starting these regions. For example, you can use the IBM Automatic Restart Manager (ARM) to start the ITK STC after an IPL. Additional Considerations: Situations can arise that require a cold start of the STC or PAUSE (TYPE=C,PAUSE=C). For example: Applying CA Technologies maintenance with ++HOLD information indicating that a cold start of the STC or PAUSE feature is required. Errors in Subsystem communications or XCF that are not resolved by a warm start of the STC. 20 Best Practices Guide

21 Generate a Dump when An Abend Occurs Do the following to completely reload and reinitialize the STC PAUSE environment without an IPL, including the Subsystem Communications Facilities of the STC: 1. Shut down all PAUSE enabled IMS control regions. Wait for completion. 2. STOP the ITK STC. 3. Restart ITK STC with option TYPE=C,PAUSE=C. Wait for completion. 4. Restart the IMS control regions. Notice the following when considering cold start: PAUSE=C allows the refresh of PAUSE-specific modules in CSA without requiring a restart of the associated IMS (as required for a TYPE=C restart). If the ITK STC is cold started (TYPE=C,PAUSE=any), it will no longer communicate with PAUSE enabled IMS control regions until they are restarted. TYPE=C refreshes the ITK STC modules and storage resident in CSA. This causes the STC to lose the ability to support PAUSE enabled IMS control regions that were started before the COLD start (TYPE=C) of the ITK STC. These IMS regions must be restarted before an FFOR job can be executed. A cold start of the PAUSE feature (TYPE=W,PAUSE=C) can be performed without restarting the IMS control regions, but the Subsystem Communications Facilities of the STC will not be refreshed. More Information: For more information about restarting ITK STC and pause enabled IMS, see the CA Database Organizer Reference Guide. Generate a Dump when An Abend Occurs When a user exit fails or you debug a problem working with CA Support, you generate a dump. To generate a dump, specify ABENDDUMP=Y in your JCL. This is the default. A dump will be generated to SYSABEND or SYSUDUMP when an operating system ABEND occurs. The dump provides information about your system when the ABEND occurred. This helps save time in determining the cause of the ABEND and resolving the issue. Chapter 3: Configuration Best Practices 21

22 Use EXCP and MMGR for Reading Databases Additional Considerations: Typically, if an ABEND occurs, the Extended Specific Task Abnormal Exit (ESTAE) recovery routine of the product formats the information in the Program Status Word (PSW) and various registers, and also writes it to the messages file. Therefore, a dump is not necessary. However, occasionally, you may need to generate a dump when a user exit fails or to debug a problem if requested by CA Support. Use EXCP and MMGR for Reading Databases Verify that EXCP (OSAM) and MMGR (VSAM) are used in CA Database Analyzer, CA Database Copier, and CA Database Organizer for reading databases through EXCPIO=Y and MMGRIO=Y. The latter requires APF authorization or the presence of the ITK STC. Also, unless I/O reports are required, setting MMSMFREPORT=N helps save a small amount of CPU. Verifying that the database is read as efficiently as possible helps optimize the performance of CA Database Analyzer, CA Database Copier, and CA Database Organizer. Additional Considerations: The EXCPIO control statement uses EXCP code for reading OSAM databases when set to YES. Note: If the database is allocated in cylinders, the I/O counts reported to SMF reflects the number of cylinders read. If the database was allocated in tracks or blocks, the I/O counts reported to SMF reflects the number of tracks read. The MMGRIO control statement uses the IBM Media Manager interface to read VSAM databases. Note: APF authorization is required to use the Media Manager interface. If APF authorization is not provided, you can still use the Media Manager interface if the ITK STC is running. The MMSMFREPORT control statement reports I/O counts for Media Manager access. The utility ignores the control statement, unless you specify MMGRIO=YES. The I/O counts are for job accounting purposes. These counts will not directly relate to the number of I/Os performed because they reflect the number of CIs read. More Information: For more information about using the EXCPIO, MMGRIO, and MMSMFREPORT control statements, see the CA Database Analyzer, CA Database Copier, and CA Database Organizer reference guides. 22 Best Practices Guide

23 Turn Off Largest Records Data Collection Turn Off Largest Records Data Collection If the largest records data is not needed, we recommend that you turn off the largest records data collection to save CPU resources. Specify LARGESTRECORDS=0 to turn off data collection. CPU resources will be saved during data collection. Additional Considerations: The LARGESTRECORDS control statement defines the number of the largest records to track and report on using the Largest Database Records report during the processing of the database. More Information: For more information about the LARGESTRECORDS control statement and report information, see the CA Database Organizer for IMS for z/os Reference Guide. Use Expand and Compress Defaults During Unload and Reload If you have a database whose data is compressed and you have no requirement for the unload file to contain uncompressed data, the unload and reload process runs faster if the data is not expanded during unload and compressed during reload. This can be achieved by using the default settings of EXPAND=N for unload and COMPRESS=N for reload. The unload and reload process runs faster if compressed data is not expanded, which saves time and resources. Additional Considerations: To use an IBM compatible unload file, specify EXPAND=YES on the unload to verify that the unload file can be reloaded by any utility that accepts IBM DFSURGU0 unload files as input. During a CA Database Organizer reload, the EXPAND control statement is not meaningful. If CA Database Organizer finds a compressed segment in the input file during a reload, the segment is inserted into the database without calling the exit. If the segment in the input file is in expanded form but it is defined as compressed in the DBD, the segment edit/compression routine is called. Chapter 3: Configuration Best Practices 23

24 Evaluate Use of DBDCHANGE Keyword in CA Database Organizer The COMPRESS control statement is useful when the reload input file already contains compressed segments but CA Database Organizer cannot determine the same. This occurs if the unload file was created by HSSR or some other non-ca Database Organizer utility that did not mark each record as expanded or compressed. If CA Database Organizer unload created the unload file, reload compresses them, regardless of the value specified for COMPRESS. More Information: For more information about using the EXPAND and COMPRESS control statements, see the CA Database Organizer for IMS for z/os Reference Guide. Evaluate Use of DBDCHANGE Keyword in CA Database Organizer Only use DBDCHANGE=Y when you know the DBD has changed. The keyword imposes significant overhead on the reload process. Optimizing the database reload processing helps improve system performance for CA Database Organizer. Additional Considerations: The DBDCHANGE control statement indicates that it is okay to use a different DBD name during the reload than the DBD name that was used during the unload. Identify the changed DBD name using the DBDNAME control statement and specify FILEFORMAT=LONG or DFSURGU0. With the IMS-compatible DFSURGU0 file format, DBDCHANGE is not needed and is ignored if specified because unrestricted changes to the DBD are allowed. DBD changes of any type are allowed and processed correctly with a return code of 0. If a compressed segment is unloaded in compressed format (EXPAND=NO) and reloaded with a DBD that defines the segment as not compressed, an error occurs and the reload terminates. To unload the segment in compressed format and reload the compressed data using a DBD that says the segment is not compressed, specify DBDCHANGE=YES in the reload. 24 Best Practices Guide

25 Use LBI for an Output Tape Data Set When you change the DBD, the records on the unload file can be from an unload, but must be FILEFORMAT=LONG or DFSURGU0, or IMS-compatible DFSURGU0 file format. DBDCHANGE=YES cannot be used with FILEFORMAT=XSHORT during reload. More Information: For more information about the DBDCHANGE keyword, see the CA Database Organizer for IMS for z/os Reference Guide. Use LBI for an Output Tape Data Set Specify LARGEBLKS=Y if any output file is directed to a 3490 in CA Database Copier, CA Database Organizer, and CA Secondary Index Builder. Maximizing the database block size helps optimize the performance of CA Database Copier, CA Database Organizer, and CA Secondary Index Builder. Additional Considerations: The LARGEBLKS control statement uses the IBM Large Block Interface (LBI) for an output tape data set. This is useful with cartridge devices because the data set block size can be larger than the standard z/os default maximum block size of 32 KB. A larger block size helps downsize the number of I/Os and elapsed times. Note: For 3480s and 3490s, the maximum block size is 64 KB. For 3590s, the maximum block size is 256 KB. Input tape data sets created using the LBI are automatically detected and no keyword is required. Optimize Database Unload Unload is a logical process that writes out segments by following pointers from each root segment down through the database hierarchy. When the database is well organized, these segments are adjacent to each other. This allows a sequential read of the database to retrieve most of the segments, and sequential reads are efficient. When the database becomes disorganized, however, perhaps because update-type activities have caused segments to become distant from each other, optimizing the I/O can be a challenge. A sequential read will still probably retrieve most of the segments but the remainder will have to be retrieved using random reads, which are inefficient. Chapter 3: Configuration Best Practices 25

26 Optimize Database Unload To optimize the database unload processing, specify the following control statements in your JCL as applicable: BUF4SEQINIT Specifies the number of buffers to allocate to the sequential thread (processing) of the database. BUF4RANDSEQ Specifies the number of buffers to allocate to the random thread (processing) of the database. To achieve optimal performance, most of the database I/O should be done on the sequential thread. In most cases, the default values for BUF4SEQINIT and BUF4RANDSEQ work well; however, they may need adjustment if the database is highly disorganized. Typically, increasing the number of buffers causes more segments to be read from a block cached in storage rather than having to read the block again from the database. The cost of this will be greater CPU in managing the increased number of buffers and for CA Database Organizer, more storage will be used (the storage is above the line (31 bit)). Conversely, decreasing the number of buffers causes more blocks to be read from the database again, but less CPU will be used in managing the buffers and for CA Database Organizer, less storage will be used. Because there is a main storage association with a buffer of roughly the database block size +160 bytes, do not set the BUF4SEQINIT value too high; otherwise, the region may run out of storage. Setting the BUF4RANDSEQ specification to a high value is not recommended because the likelihood of that setting reaping dividends is low. In addition, you can use the RANDOMTRACKIO control statement to force unload to always read in a full track of data when a block is to be read using the random thread. The original idea behind this was that doing consecutive READs for the number of blocks per track was an efficient way of reading data from a CKD device. Further, if those data blocks contained a significant number of related segments, then you would see a benefit compared to reading each block individually, perhaps in a random sequence. However, with the newer DASD architectures, this argument no longer holds true, so we recommend that you always set RANDOMTRACKIO=N. If you have a highly disorganized data set group (DSG) that causes performance problems during unload, consider using the BUFFEREDDSG control statement with CA Database Organizer. When specified, CA Database Organizer uses a highly tuned, sequential I/O to read the DSG into storage (hiperspace). This enables all pointers to be resolved from storage rather than DASD and helps avoid the destructive random I/O that tends to slow down the unload process. 26 Best Practices Guide

27 Optimize Database Unload Optimizing the database unload processing helps improve system performance for CA Database Organizer (UNLOAD, EXTRACT, and API mode) and CA Secondary Index Builder (INDEXCREATE and INDEXEXTRACT). Additional Considerations: The Database I/O Statistics section from the DBORPTS data set is especially useful in optimizing the I/O side of unload. Here is an example of this data: DDNAME: ddname DATA SET NAME: database.dataset.name SEQUENTIAL RANDOM BUFFER HITS 57,850 0 PHYSICAL I/O 3,574 0 READS 3,574 0 WRITES 0 0 ERASES 0 0 CHECKS 0 0 POSITIONS 0 0 ENDREQS 0 0 BH BUFFERS DBO A/M BUFFERS 30 0 ACTUAL A/M BUFFERS 30 0 LOAD MODE I/O 0 FORMAT WRITES 0 FORMAT I/O 0 Remembering that the unload process unloads segments in hierarchical sequence but must read database blocks to do so, here is an explanation of the salient terms: SEQUENTIAL Shows the values related to the sequential processing thread. RANDOM Shows the values related to the random processing thread. BUFFER HITS Specifies the number of times a segment has been found in a database block that was cached as a result of a prior read. PHYSICAL I/O READS Specifies the number of EXCPs issued against the data set. Specifies the number of GETs or READs issued against the data set. Chapter 3: Configuration Best Practices 27

28 Optimize Database Unload POSITIONS Specifies the number of points issued against the data set. On the sequential thread, this number should always be zero. On the random thread, if RANDOMTRACKIO=N is set, this number should be the same as the number of READS. If RANDOMTRACKIO=Y is set, the number of POSITIONS is a fraction of the number of READs. (If you have 15 blocks per track, the number of POSITIONs is 1/15th the number of READs). BH BUFFERS Specifies the BUF4SEQINIT and BUF4RANDSEQ specifications. If the report shows a significant number of READs on the random thread, an increase in the number of buffers allocated to the sequential thread might improve the situation. This can be achieved by setting BUF4SEQINIT to a value higher than the value shown under BH BUFFERS. If, however, the number of BUFFER HITS on the sequential thread does not increase, you will most likely have incurred more CPU for no result. If so, consider further increasing the BUF4SEQINIT specification. However, if again, no improvement is seen, we recommend that you revert to the original value. You can specify the number of BH buffers in CA Database Organizer as follows: BUFSPACE Specifies a numerical value denoting a number of MB. If BUFSPACE=n is set, the number of sequential buffers is as follows: (n* )/block size BUF4SEQINIT Specifies a numerical value explicitly stating the number of sequential buffers. SEQBFRS Specifies a numerical value that, when multiplied by the database blocks per track, determines the number of sequential buffers. If none of these values is specified, the following defaults are used: For single-dsg databases: (database blocks per track)*15*4 For multi-dsg databases: (database blocks per track)*15*2 28 Best Practices Guide

29 Allocate Buffered DSGs Example: Determine Number of Sequential Buffers and Above-the-Line Storage Requirements For a single-dsg database with a block size of 4096 bytes, the number of sequential buffers is 12 * 15 * 4, or 720. The corresponding (above-the-line) storage requirement in bytes is ( ) * 720 bytes, or 3,064,320 bytes. For a multi-dsg database with a block size of 4096 bytes, the number of sequential buffers is 12 * 15 * 2, or 360. The corresponding (above-the-line) storage requirement in bytes is ( ) * 360 bytes, or 1,532,160 bytes for every DSG. Allocate Buffered DSGs If you have a highly disorganized data set group (DSG) that causes normal processing performance to increase during unload, use the BUFFEREDDSG control statement with CA Database Organizer to allocate buffered DSGs. A database can become disorganized over time when its logical and physical storage allocations contain many scattered areas of storage that are too small, not physically contiguous, or too disorganized to be used productively. When BUFFEREDDSG is specified, CA Database Organizer uses a highly tuned, sequential I/O to read the DSG into storage (hiperspace). This enables all pointers to be resolved from storage rather than DASD and helps avoid the destructive random I/O that tends to slow down the unload process. Optimizing the database unload processing helps improve system performance for CA Database Organizer (unload, extract and API mode) and CA Secondary Index Builder (index create and extract). Additional Considerations: Valid for CA Database Organizer. The BUFFEREDDSG control statement specifies whether to allocate and copy into hiperspaces a specific data set group (DSG). When the primary DSG of an HDAM database is specified, only the overflow areas are buffered. Use this statement when performance from normal processing is not providing expected results and many buffer hits are registering under Random in the Database I/O Statistics report. When specified during unload, CA Database Organizer will copy the DSG into one or more hiperspaces and unload it from there instead of dropping into random I/O when the database segments get too far apart to remain in the sequential buffers. Chapter 3: Configuration Best Practices 29

30 Rebuild Indexes During Reorganization For database DSGs that are disorganized, this processing can significantly improve performance. Note: The size of the DSG determines the amount of storage needed for buffering. If it is too large, auxiliary storage can be impacted to the point that the system can no longer function properly until the job is canceled. The DSG cannot be larger than 4 GB (VSAM and HALDB) or 8 GB (OSAM non-haldb only). This control statement is valid only during unload processing. This control statement has the following format: BUFFEREDDSG=dd1 dd1 Specifies the ddname of the data set to be buffered from the DD1 statement in the DBD. The problematic DSGs are usually HDAM overflow and secondary DSGs that are large. If specified, RANDOMTRACKIO is turned off for the DSG being buffered. Note: If DD1 is the HDAM DSG#1, only the overflow is buffered. Example: Allocate Buffered Secondary DSGs For example, specify BUFFEREDDSG=DSG2 where DSG2 is the ddname of the secondary DSG of a given DBD. Rebuild Indexes During Reorganization It is more efficient to rebuild secondary indexes of a database during a reorganization instead of after. To do this, specify AUTOINDEXBUILD=YES in the CA Database Organizer and CA Secondary Index Builder reload control statements. Creating index work files during the reload process and using those work files to rebuild the indexes helps eliminate the need to completely reread the database in a subsequent CA Secondary Index Builder step. Additional Considerations: For full function database reorganizations with AUTOINDEXBUILD=YES, you can write the index records to DFSURWF1 (SPLITOUTPUT=NO) or to separate work files (SPLITOUTPUT=YES, the default) to make the database reload faster. Allocate these work files through the JCL using ddnames DDNAM@nn, where DDNAM represents the first 5 characters of the ddname of the primary database, and nn is a numerical suffix from 01 to Best Practices Guide

31 Rebuild Indexes During Reorganization For example, if the DBD has a DATASET macro with a specification of DD1=ABCDEFGH, specify the work files as follows: //ABCDE@01 DD SPACE=(CYL,500) //ABCDE@02 DD SPACE=(CYL,500) //ABCDE@03 DD SPACE=(CYL,500) //ABCDE@04 DD SPACE=(CYL,500) For HALDB database reorganizations with AUTOINDEXBUILD=YES, note the following: SPLITOUTPUT=YES is defaulted and required. ALLPARTS=YES must be specified. Allocate the HALDB work files through the JCL using ddnames PSNAM@nn, where PSNAM represents the first five characters of the PSINDEX DBNAME and nn is a numerical suffix from 01 to 99. For example, if HALDB=MASTERH has three PSINDEXes DBD=PSIN1DX, DBD=PSIN2DX and DBD=PSIN3DX, with two, three, and four partitions respectively, specify the work files as follows: //PSIN1@01 DD SPACE=(CYL,500) * DBD=PSIN1DX, partition 1 //PSIN1@02 DD SPACE=(CYL,500) * DBD=PSIN1DX, partition 2 //PSIN2@01 DD SPACE=(CYL,500) * DBD=PSIN2DX, partition 1 //PSIN2@02 DD SPACE=(CYL,500) * DBD=PSIN2DX, partition 2 //PSIN2@03 DD SPACE=(CYL,500) * DBD=PSIN2DX, partition 3 //PSIN3@01 DD SPACE=(CYL,500) * DBD=PSIN3DX, partition 1 //PSIN3@02 DD SPACE=(CYL,500) * DBD=PSIN3DX, partition 2 //PSIN3@03 DD SPACE=(CYL,500) * DBD=PSIN3DX, partition 3 //PSIN3@04 DD SPACE=(CYL,500) * DBD=PSIN3DX, partition 4 More Information: For more information about the AUTOINDEXBUILD control statement, see the CA Database Organizer for IMS for z/os Reference Guide. Chapter 3: Configuration Best Practices 31

32 Increase Sort Work File Size Increase Sort Work File Size To improve sort initialization and run-time performance, override the estimated number of sort records and default maximum record length of 1000 bytes as follows (cards must begin in column 2 or after): CA-SORT: //SORTCNTL DD * RECORD TYPE=V,LENGTH=(090) OPTION FILSZ= E DFSORT: //DFSPARM DD * RECORD TYPE=V,LENGTH=(090) OPTION FILSZ= E SYNCSORT: //$ORTPARM DD * (Note: //DFSPARM can be used instead of //$ORTPARM) RECORD TYPE=V,LENGTH=(090) OPTION FILSZ= E In the previous examples, the sort record length is variable; no record to be sorted is larger than 90 bytes, and the estimated number of records to sort is 75,000,000. SORT operates more efficiently and helps optimize available resources. Additional Considerations: To calculate the sort record maximum length, use the following formula: RECORD TYPE=V,LENGTH=(nn) nn Specifies the length as the highest value of all "LARGEST RECORD (BYTES) = " for indexes shown in DBORPTS plus the following, rounded up to an even number: For full function primary and secondary indexes, add 18. If there is only one SI (INDEXDBDNAME=) or primary index (PRIMIND=), add 12. If the database is logically related through paired segments and more than one secondary index is created using a single FUNCTION=INDEXCREATE, add 22. For HALDB PSINDEXes, PIs, and ILDSes, add Best Practices Guide

33 Use IIR Reporting Functions For HALDB ILDSes only (BUILDILDS=ONLY and no PRIMIND=), nn is 32. CA Secondary Index Builder generates this automatically. Do not override the CA Secondary Index Builder-generated value. More Information: For more information about optimizing CA Secondary Index Builder performance, see the CA Secondary Index Builder for IMS for z/os Reference Guide. Use IIR Reporting Functions We recommend that you use the IMS Information Repository (IIR) reporting and utility functions to improve your ability to recognize trends in your environment based on historical data collected by CA Database Analyzer, CA Database Copier, CA Database Organizer, CA High Performance Recovery, and CA Secondary Index Builder. To enable IIR reporting functions, perform the following steps: 1. Install the CA Datacom/AD environment. 2. Install the IIR database. After you have installed the CA Datacom/AD environment and IIR database, you can begin collection of the execution data in the IIR tables by specifying IIRRECORD=YES in the utility JCL. This information can be collected over a period to provide detailed information for trend analysis, performance monitoring, and debugging. IIR helps you recognize trends based on historical data collected by these products and does not affect the overall performance of the monitored utility jobs. The information collected provides significant benefit to the IMS user in performance tuning, problem determination, and resource planning. Additional Considerations: The following product functions can optionally store information in the IIR repository: ALL in CA Database Analyzer COPY and IICOPY in CA Database Copier RELOAD and UNLOAD in CA Database Organizer RECOVERY in CA High Performance Recovery INDEXBUILD, INDEXCREATE, and INDEXEXTRACT in CA Secondary Index Builder Chapter 3: Configuration Best Practices 33