New Features

The following new product features and enhancements to existing features are now available in Version 20.0 of CA Fast Load.
cafldb220
The following new product features and enhancements to existing features are now available in Version 20.0 of
CA Fast Load for DB2 for z/OS
(CA Fast Load):
For a list of enhancements for all the
CA Database Management Solutions for DB2 for z/OS
, see New Features on the
CA Database Management Solutions for DB2 for z/OS
product documentation page. To request product enhancements, contact a product manager or post your request in the "Ideas" section of the Broadcom Mainframe Community.
Be sure to download all available maintenance to ensure availability of all features and fixes. Incremental updates cannot be installed selectively. Later increments have dependencies on earlier increments. For a list of available PTFs, access the Maintenance Grid (login required) on
Broadcom Support
Online. We recommend use of the CA SMP/E Internet Service Retrieval service to download maintenance and to keep current with all updates.
DBM 20 IR 8 is the last incremental update for Version 20.0. New features and fixes are no longer bundled in incremental PTFs. Instead, we have transitioned to a continuous delivery release model that provides individual PTFs for new product features and product fixes. For more information, see Transition to Continuous Delivery.
zIIP Support for OUTPUT-CONTROL ALL and REORG (SO16090)
CA Fast Load for DB2 for z/OS
now supports the use of the IBM System z Integrated Information Processor (zIIP) when you specify OUTPUT-CONTROL ALL and OUTPUT-CONTROL REORG. Previously, zIIP support was only available for OUTPUT-FORMAT BUILD. Now, zIIP support is provided for all OUTPUT-CONTROL values.
Offloading zIIP-eligible workloads to zIIP processors can reduce the processing load on your general CPUs. zIIP functionality is enabled by default. Use the zIIP keyword in the SYSIN DD or the PFL-ZIIP parameter in
hlq
.CDBAPARM(UTIL) to control zIIP processing.
More information:
Support for PBR RPN (PBR2) Tablespaces (SO15482)
CA Fast Load now supports partition-by-range relative page numbering (PBR RPN). IBM Db2 12 introduced PBR RPN (known as PBR2) to separate the partition number from the page number for greater flexibility in partition processing.
zIIP Support for OUTPUT-CONTROL BUILD (SO14771)
CA Fast Load for DB2 for z/OS
now supports the use of the IBM System z Integrated Information Processor (zIIP) when you specify OUTPUT-CONTROL BUILD. Offloading zIIP-eligible workloads to zIIP processors can reduce the processing load on your general CPUs.
zIIP functionality is enabled by default and controlled when you specify ZIIP {YES|NO} in the SYSIN DD statement or PFL-ZIIP {YES|NO} in
hlq
.CDBAPARM(UTIL).
More information:
Support for Huffman Compression (SO14029)
CA Fast Load for DB2 for z/OS
supports Huffman data compression. You can now load Huffman encoded objects when the compression dictionary exists and KEEPDICTIONARY is specified. Huffman data compression is introduced with IBM Db2 for z/OS function level V12R1M504.
You can load Huffman encoded objects with CA Fast Load when you use LOAD RESUME, or when you use LOAD REPLACE and specify KEEPDICTIONARY for an object that contains an existing compression dictionary. Otherwise, if SWITCH-TO-IBM YES is specified in
hlq
.CDBAPARM(UTIL), the IBM utility is invoked to load the Huffman encoded object.
More information:
Support for TEMPLATE statements to Allocate Input Data Sets (SO13764)
CA Fast Load now supports the use of the TEMPLATE control statement and INDDN keyword to dynamically allocate SYSULD data sets without static JCL DD statements. You can use variables to define data set names, and set allocation parameters to automate, simplify, and standardize your process. You can take advantage of a z/OS UNIX System Services (USS) pipe for improved performance by defining a USS file PATH name and associated parameters.
More information:
Enhanced Discard Processing with DISCARD-LIMIT Keywords (SO12525)
CA Fast Load now gives you greater flexibility in managing discard processing. The DISCARDS keyword lets you specify the number of invalid records that can be encountered (the DISCARDS limit) before processing stops. The new DISCARD-LIMIT-RESUME keyword lets you specify whether the database object data and associated pageset status will change when you specify RESUME YES and the DISCARDS limit is reached. The new DISCARD-LIMIT-STATUS keyword lets you set a restrictive status on the tablespace and indexspace pagesets when the DISCARDS limit is reached.
Improved Automatic Calculation of ESTIMATED-INPUT (SO11848)
CA Fast Load now provides improved accuracy in the automatic calculation of the size of your input data sets. The sort program uses this value to calculate the size of your sort work data sets. Improved accuracy in the calculation of the size of your data sets can improve sort performance.
You can use the ESTIMATED-INPUT keyword to manually specify the number of rows in the SYSULD input data set. However, we recommend that you omit ESTIMATED-INPUT and let CA Fast Load automatically calculate the size of your input data sets for you.
Db2 Version Support
IBM Db2 12 for z/OS function levels V12R1M100, V12R1M500, and V12R1M
nnn
are now supported.
Version 20.0 of CA Fast Load supports the IBM Db2 12 Continuous Delivery Model, which introduces new features in the maintenance stream for faster adoption. Customers who are moving to Db2 12 should upgrade to Version 20.0. In Version 19.0,
only
V12R1M100 for migration from Db2 11 and function level V12R1M500 are supported for Db2 12.
More information:
Perform an Online Load
CA Fast Load now gives you the ability to perform an online load. This new processing option keeps the table online in read-write (R/W) mode during the load process, loading the data at the end of the table. An online load also does not place the table in copy pending mode, thereby eliminating the need to take a copy of the table after the load. This enhancement provides high availability, decreased CPU cycles, and shorter elapsed processing time. This enhancement also eliminates the need to recover the table if the load fails. Any failed inserts roll back, leaving the table in the same state it was in before the failure. You can then restart the job at the point of failure.
The new online functionality is provided by specifying the new DATA-AVAILABLE CRITICAL keyword option with RESUME YES. You can also specify how many rows to insert at a time and how often to issue a COMMIT with the new COMMIT-FREQUENCY keyword options.
Perform a Fast Switch Load
CA Fast Load now gives you the ability to perform a fast switch load. This new processing option loads the data into a shadow copy of the tablespace, and then switches the shadow copy with the original tablespace. The original tablespace remains available in read-only (RO) mode during the load process. This enhancement enables you to keep a tablespace available while refreshing its data.
The new fast switch functionality is provided by specifying the new DATA-AVAILABLE HIGH keyword option with RESUME NO REPLACE and OUTPUT-CONTROL BUILD.
DELETE-FILES and DELETE-SYSDISC Support
CA Fast Load now supports the forced deletion of output data sets at the end of a successful load. New SYSIN keywords are provided to override the JCL DISP=parameters for the SYSREC, SYSUT1, SYSERR, and SYSDISC data sets. This enhancement lets you use a single keyword to force the deletion of output data sets, which helps free up disk space. Depending on how your JCL is set up, this enhancement can also eliminate the need to manually update multiple JCL jobs if you need to change the data set disposition.
This enhancement applies only to manually allocated output data sets. It does not apply to dynamically allocated data sets.
The new deletion feature is provided by the following new keywords:
  • DELETE-FILES
    Specifies whether to force deletion of the SYSREC, SYSUT1, and SYSERR data sets (or the alternate ddnames specified in the UNLDDN/WORKDDN/ERRDDN keywords).
  • DELETE-SYSDISC
    Specifies whether to force deletion of the SYSDISC data set (or the alternate ddname specified in the DISCARDDN keyword).
UPDATED: Establish Default RECLUSTER Settings for All Jobs
CA Fast Load now gives you the ability to establish the default RECLUSTER keyword settings for all jobs. This keyword specifies whether to sort the data and the clustering index during the load. Previously, specifying RECLUSTER in your SYSIN syntax was the only way to control these sorts.
This functionality is provided by the following new options in
hlq
.CDBAPARM(UTIL):
  • PFL-RECLUSTER-DEFAULT
    Specifies what to sort when you omit RECLUSTER from SYSIN. The following values are valid:
    • SORT-CLINDX
      Sorts the clustering index only. The resultant tablespace rows are in input sequence. This value is the default.
    • SORT-ALL
      Sorts the clustering index and the data. The resultant tablespace rows are in clustering sequence.
    • SORT-NONE
      Does not sort the clustering index or the data. The data must be in the correct order. Rows that are out of order are discarded.
    PFL-RECLUSTER-NO
    Specifies whether to sort the clustering index during a load. The following values are valid:
    • SORT-NONE
      Does not sort the clustering index or the data. The data must be in the correct order. Rows that are out of order are discarded. This value is the default.
    • SORT-CLINDX
      Sorts the clustering index only. The resultant tablespace rows are in input sequence.
  • PFL-RECLUSTER YES
    Specifies what to sort during a load. The following values are valid:
    • SORT-ALL
      Sorts the clustering index and the data. The resultant tablespace rows are in clustering sequence. This value is the default.
    • SORT-CLINDX
      Sorts the clustering index only. The resultant tablespace rows are in input sequence.
The existing RECLUSTER NO and YES keyword options in SYSIN now use the new UTIL parmlib settings:
  • NO uses the PFL-RECLUSTER-NO value.
  • YES uses the PFL-RECLUSTER-YES value.
The default parmlib settings provide the same product behavior that RECLUSTER NO and YES previously provided. However, if you change the parmlib settings, RECLUSTER NO and YES will behave differently. For more information, see the RECLUSTER keyword description.
You can override the parmlib settings by using the following keyword options:
  • The existing NO SORT-CLINDX keyword option
  • The following
    new
    keyword options:
    • NO SORT-NONE
    • YES SORT-CLINDX
    • YES SORT-ALL