This How to guide can be used for trouble shooting BPC data manager package (DMP) issues.
When you are troubleshooting BPC data manager issues you have to look in several areas for the diagnostic information and determining the root cause.
Some Background Information:
BPC Data Manager Packages map to process chains in BW through the BPC Data Manager Organize Package menu options. When a BPC package is executed from BPC Data Manager it releases the process chain it is mapped to on the server. Users who are familiar with BW and process chains know that process chains are executed through transaction RSPC. RSPC releases a job for each process chain which is then executed in the background. Whenever a user runs a Data Manager package from BPC client, it executes RSPC on the BW server which releases a job to execute the relevant process chain.
Data manager issues can be due to errors on system or application level. Both types of issues can trigger system dumps. Depending on when and why a system dump occurs the error may or may not be captured in a data manager log.
Where to start troubleshooting - Initial Steps information gathering:
Always check all of the following areas to identify the root cause and the scope:
- Using SM37 check the status of the job and the detail job log
To find the job log ID you must have either the time stamp or check UJD_STATUS based on BPC application set (environment), application (model), user and package IDs. See 1645625 for details
- Check if there is any system dump associated with the problem. To find the dump:
Open the detailed job log in transaction code SM37 and review the failed step.
Using ST22 transaction, search for any system dumps by the user ID and timestamp.
Using UJFS find the data manager logs. See Note 1645625 - How to Reproduce and Analyze BPC Data Manager Issues for details.
Using transaction RSPC log view review the variant that failed
- For additional information:
- Check application log using SLG1 for sub-component UJD.
- Check the system log using transaction code SM21.
What to do if the information gathered by the initial steps is not enough
Debug the data manager package by using the program UJD_TEST_PACKAGE.
Various problems and root causes and how to find them
When a data manager package fails all of the above steps except for step 5 should be performed to find out precisely what happened. The root cause of an error can be at the system level, memory and data base issues, or application level incorrect mapping and conversion. Typically system level errors generate a dump but even application level errors can trigger a data base error and cause a system dump. If a dump occurs, depending on what triggered the dump and at what stage of the process it can end the data manager process immediately and therefore the failed status may not get recorded either in job log or the data manager log.
After reviewing all the available information, gathered as described in initial steps, if the root cause is still unclear then run UJD_TEST_PACKAGE for further debugging.