Trouble-Shooting a Failed ATC Quality Check

You must distinguish between two types of problems with ATC quality check runs:

• The ATC check run itself failed (status aborted in the ATC monitor).

  In this case, a system problem is usually to blame. The ATC batch job was cancelled, the system was shut down, or some similar problem occurred.

  You can usually restart the ATC check run from the ATC monitor. The check run will continue from the point at which it was interrupted.

• The ATC check run was incomplete and either:

  • A serious error (runtime error or exception) occurred in a check module; or

  • You could not diagnose the problem with objects that could not be checked.

  In this case, you should probably refer the problem to SAP or to some other author of the affected check module.

Here is how to diagnose a check run that has the status aborted:

1. In the system in which the check run was performed, start transaction ATC. In the navigation pane, open Administration > Check Runs > Monitor.

2. Select the aborted run.

3. Click on Display Log. Check the log for messages on the cause of the abnormal ending of the check run.

4. If the log does not show the problem, start transaction SM37 to check the log of the ATC dispatcher job. Search for jobs whose name starts with ATC Project*.

5. Display the Job log for messages that show the cause of the problem.

   A final place to check, if the job log is unexceptional, is the system log (transaction SM21) of the application server on which the ATC job ran.

Here is how to diagnose incomplete results, when a serious error in a check module has occurred or when you cannot determine what the problem with an unchecked object is:

1. In the event of an exception, you can jump from the error display in the result overview to the point in code at which the exception occurred.

2. If a short dump in a check module occurred, then start transaction ST22 and locate the dump. The time at which the short dump occurred are shown in the error display and also in the log of the ATC check run (as described above).

3. If you need to clarify a problem with an object that could not be checked, then do the following:

   • Start by checking the log of the check run, as described above.

   • If there are no useful messages, then try repeating the check run, limiting it to uncheckable objects.

     Before you start the run, increase the ATC log level by choosing Administration > Check Runs > Logging On/Off in transaction ATC.

     You can also start the Object log to record checking at the level of repository objects. Mark Activate Object log when you start the check run.