Analyzing Parts of the Runtime Environment  

Use

The background processing runtime system consists of the following components (control objects):

Responsible for starting jobs which have a start date and time specified.

If there are jobs with a scheduled start time of After event, At operation mode or After job that cannot be started at their scheduled start time (for example, if there are no free background work processes), then these jobs will be started at the next possible opportunity. The time-based scheduler is also used to start these jobs.

The scheduler is started periodically by the dispatcher in a dialog work process, as long as there is at least one background work process on the application server and the periodic value is greater than zero. The number of background work processes is specified in the profile parameter rdisp/wp_no_btc. The periodic value is set in the profile parameter rdisp/btctime.

(The default value is 60 seconds)

If an event is raised in the R/3 system, a check is made using the event based scheduler to see if there are any jobs waiting for this event. If there are, then the event based scheduler makes sure that these jobs are started. The scheduler always runs in a dialog work process on the application server which the profile parameter rdisp/btcname points at.

A job that is due to be executed will be started by the job starter. This performs all preparatory work such as reading the job data from the database and starting the job steps. The job starter runs in a background work process.

The time-based job scheduler uses a table of times during each run to check whether a switch of operating mode is necessary. If it is, then the scheduler initiates the switch in a dialog process on the application server on which the scheduler is running.

When an R/3 system is started, then a check is made for ‘job corpses’ (zombies). These are jobs which have the status ready or active. Since it is of course impossible for jobs with one of these status values to exist when the R/3 system is started, they are set to have status scheduled or terminated. ‘Job corpses’ are created when, for example, an application server on which a job is running is powered down before the job has finished properly. The background processing runtime system can no longer correct the status contained in the database.

This component allows external programs to be started as part of a job step. An external program is started from the background work process in which the job is running.

Each of these components (control objects) can be analyzed separately.

Procedure

  1. To analyze parts of the runtime environment from the CCMS initial screen, choose Jobs ® Background objects
  2. Choose Maintain. The system displays a list with one entry for each host and control object.
  3. Choose Control object ® Change. The system displays a dialog box in which you can activate a control object for tracing. This object is valid either just for the next run or permanently.
  4. Choose Action log to find out if the control object ran after the trace was switched on.
  5. As long as the control object has run at least once, you can have a look at the trace file on the appropriate application server. To do this, choose Tools ® Administration, Monitor ® System monitoring ® Servers, Goto ® Traces (Transaction SM51)
  6. In the case of control objects that run in dialog work processes, it is not possible to predict in which of the possible work processes the object will be processed. However, it is quite likely to be work process 0 or 1. The best approach is to sort the list of trace files according to date and time The trace entries should be found in one of the most recently written trace files. Look through the files for entries that have an ‘L’ in the first column.

    7. For control objects that run in a background work process, you can find the work process number in the job detail data.

  7. If you cannot evaluate the trace information yourself, you can send the contents of the trace file to SAP for further analysis.