Using the Checkpoint Checker › Functional Overview
Functional Overview
Many batch jobs use DLI or DB2® as a database. If a batch job abends, usually backout processing is started automatically by the database system. The backout process works until a synchpoint is found. The DLI checkpoint command provides such a synchpoint. If checkpoint writing is not implemented in the application program, or if it is done in an improper way, it takes a long time to perform the backout process. This process can sometimes take several hours, exceeding the time limit for the job or job step, which can cause an abend or require an operator to cancel the job.
The following points are major considerations to implement checkpoint writing in a program:
- The program requires an IMS environment, even if it is only using DB2 SQL.
- To provide a program with checkpoint writing requires a detailed understanding of and intervention into the logic of the application program.
- The benefits of checkpoint writing are as follows:
- High resource consumption and elapsed time for backout processing is avoided.
- Jobs that implement checkpoint writing can be canceled by the operator on demand without long delays and can be restarted.
- The Checkpoint Checker helps answer the following essential questions for MVS systems personnel:
- Which job steps are using DLI or DB2 databases?
- How high is the checkpoint writing frequency? For example, if 30 minutes elapse between each checkpoint written, checkpoint writing does not have a significant impact.
- Which job steps run without checkpoint writing?
- How important is the consumption habit of each single job step, as compared with all other job steps?
