3782084 - Custom Mass Data Run Updates Instances After Run Has Failed.

You observe that your custom Mass Data Run has failed with errors but some instances are still updated. The change log indicates these updates happened after the run failed. 

• SAP Business ByDesign
• SAP Cloud for Customer 
• SAP Cloud Application Studio 

1. Go to Custom Mass Data Run Work Center and click on the application log. 
2. You will find error with job in finished status.
3. You can see the run ended after 10 hours.
4. Go to the work center relevant to the records to be processed in the failed run.
5. Go to the Changes tab.
6. You can see that the attribute/s were updated by the MDR user and the time and date are after the run failed. 

- When an MDRO runs, the main process creates work packages that execute the application logic.
- After triggering the work packages, the main process waits up to 10 hours for results. If the results do not arrive within that timeframe, the main process sets the work package status to not valid and ends its execution. It does not control the sub processes, so those continue to run to completion even if the main process has already failed.
- The MDR splits the selected instances into batches of 50. With parallel processing, these batches run simultaneously; without this option, they run sequentially.
- As a result, once the main process fails after the ten-hour wait, no additional batches are scheduled. The last batch that was already triggered will continue independently.

- Optimize the number of instances processed within a package so it completes within 10 hours; this will improve MDRO performance overall.
- Build smaller packages of requested data and make sure memory is reduced between packages.
- Check and fine tune the PDI coding to fetch a lesser amount of data according to your requirement so that the MDR will not fail.
  Consider using parallel processing; instances will be handled in smaller packages (50 per package), which helps avoid the dump in the UI and ensures a failure in one package does not  impact others. 

• Refer to: Cloud Application Studio: Performance Check
• Refer to: Performance Improvements with ABSL
• Refer to: Performance ABSL Trace – New functionality
• Refer to: Cloud Application Studio Performance Best Practices

mdr, mass data run, rfc exception system_failure, short dump, work package status not valid, c4c batch job, service_request bo, accesscontrollist.entry, memory limit, package size, pdi, absl, performance tuning, application log, day band calculation , KBA , AP-RC-BDS-MDR , ByDesign Studio Mass Data Run , Problem

SAP Business ByDesign all versions ; SAP Cloud Applications Studio all versions ; SAP Cloud for Customer add-ins all versions ; SAP Cloud for Customer core applications all versions