DPRSMF job control statements
This section describes the DPRSMF job stream and the associated job control statements.
Overview of DPRSMF
The DPRSMF job stream extracts data from your SMF files and loads the extracted data into performance data tables for each data type (statistics, accounting, audit, and storage).
SMF data extraction is a multi-step batch execution procedure. You can run this procedure as stand-alone against the SMF data sets, or you can incorporate it into an existing job that also processes this data. You can break the procedure down into its component parts. You can also use generation data groups or other permanent data set types to store and pass the data between runs of the various steps. If required, you can delay some steps, such as the table load, until end of the day processing.
After extracting data from SMF files, DRPSMF submits jobs sequentially to load the extracted data into the performance database. Each job loads a logical set of tables in the performance database. For a sample of the job control statements, see Sample-DPRSMF-JCL.
DPRSMF job stream
The following table describes each step in the job stream:
Step | Description |
---|---|
1 | Extract the Db2 statistics, accounting, and audit records from the input SMF file, creating temporary statistics, accounting, and audit files You can use control statements to select data of only certain specified record types or from specific Db2 subsystems. For more information, see Step 1—SMF extract and file creation. |
2 | Create discrete interval records from the cumulative statistics records To consolidate any statistics intervals that span an SMF tape, this step maintains a spin file which resolves issues caused by processing the SMF tape out of sequence. For more information, see Step 2—Statistics consolidation. |
3 | Build Db2 load control statements, summarize accounting and statistics records, and purge old records You can use control statements to specify the tables in which the extracted data is to be loaded and whether summarization should occur before the load. Other statements allow existing table data to be purged or unloaded. For more information, see Step 3—DB2 performance data tables load procedure. |
4 | Load the performance data tables For more information, see Step 4—Load into DB2 tables. |
This section also contains the following topic:
Related topic