Extractor output data sets


This topic discusses Extractor output data sets, including both primary data sets and alternate data sets.

The starter MVS PAS PROC sends Extractor data to SMF. If data is not written to SMF, you must specify that records be written to Extractor output data sets. In this case, you should have allocated Extractor output data sets during customization. Even if you did specify SMF recording but now want to record to Extractor output data sets, you first need to allocate these data sets.

BMC recommends that at least two data sets be specified to provide for alternate data set support. When two or more data sets are specified, the first one specified is called the primary data set and the others are called alternate data sets.

Alternate data set support

If the Extractor is writing records to an Extractor output data set and that data set fills up, Extractor writing is automatically switched to an alternate data set.

This process is referred to as alternate data set support. This support is not valid when data is being recorded to SMF; SMF provides its own alternate data set support.

The Extractor suspends recording if the current data set becomes full and no other data set is empty. At subsequent recording intervals, if an empty data set is detected, the Extractor resumes recording automatically.

Allocating Extractor output data sets

You can define the same data sets for both CPM and IPM data, or you can define primary and alternate data sets. However, BMC Software recommends that you define at least two data sets for each mode.

If you are using VSAM, the manual customization member CMFJVSAM in hilevel.UBBSAMP contains sample JCL for allocating four data sets: one primary and one alternate data set for CPM mode, and one primary and one alternate data set for IPM mode.

If you are using BSAM, the manual customization member CMFJBSAM in hilevel.UBBSAMP contains sample JCL for allocating four data sets: one primary and one alternate data set for CPM mode, and one primary and one alternate data set for IPM mode. This JCL is defined with block sizes for either a 3380 or a 3390 device type; if you have a different device type, refer to the following table for the correct block size.

BSAM Extractor output data sets use the following attributes:

  • LRECL=32756
  • DSORG=PS (for physical sequential)
  • RECFM=VBS (for variable block sequential)
  • BLKSIZE=variable (depends on the device type; see the following table)

Recommended block sizes for BSAM Extractor output data sets

Device type

Recommended block size

3375

11616

3380

11476

3390

11476

Tape (1600 bpi)

12288

Tape (6250 bpi)

32756

Other

8192

Note

CMF MONITOR increases the block size to 8192 if a smaller block size is defined.

CMF MONITOR adds new records into the primary data set after those that already exist, even if DISP=OLD or DISP=SHR is specified in the JCL. This addition prevents the destruction of any data that was previously collected. CMF MONITOR writes to the beginning of the primary data set and destroys the existing data if the output data set is on tape or if DISP=NEW is coded in the Extractor REPORT control statement.

Note

The DISP=NEW status applies only to the primary data set that the Extractor writes to after initialization. See Report control statement (Analyzer) for more information about the REPORT control statement.

When switching is required, CMF MONITOR writes to an empty data set if one is available. If an empty data set is not available, recording is suspended.

Note

CMF provides two members in hilevel. UBBSAMP that contain JCL to clear your CPM and IPM data sets:

  • CMFJCLRS is a Started Task for clearing data sets.
  • CMFJCLRB is a batch job for clearing data sets.

 

Tip: For faster searching, add an asterisk to the end of your partial query. Example: cert*