Reconciliation of Messages in BPM Contd. – Restart Workflow
This blog is a continuation of blog- Reconciliation of Messages in BPM and it also describes about How to restart the Integration Process in XI after BPM errors.
Once mapping exception occurs, alert will be triggered. In SXMB_MONI- it clearly shows error in BPM.
Technical Details of BPM- To see the complete technical details
Check Alert in the Alert Inbox of RWB-
To check Reconciliation-
Make the required corrections to fix the errors in BPM and activate the component in Integration Repository. Also make sure that Cache is refreshed.
In this example I will change the code, so that it will not affect the file.
To test the Reconciliation, Workflow should be restarted.
Restart Workflow-
Business process can be restarted using the transaction code SWF_XI_SWPR in XI. Or go to SXMB_MONI_BPE->Restart Process After Error
After clicking on Execute button , following screen appears-
After selecting specific,required workflow- clicking on Restart Workflow button takes into following screen.
Now workflow is successfully restarted and Integration process started executing from that point where alert has been triggered and file is written successfully into Receiver system. And here is the original output as required.-
Here you go with the required result. Hence Reconciliation of Messages in BPM.
Note: There will be a performance overhead of transformation step in the BPM. Still ultimate aim with respect to business is to reconcile the messages/data. It will be beneficial, when BPM executes huge amount of Business Logic and runtime exceptions within the BPM after the successful execution of some steps will lead into the message loss.







We too developed a similar king of design for one of our client projects.
the only differene is ----in the place of BPM transformation steps here it is synchronous send step with EH for both system and application errors.
In our design- after the BPM fail the restart is happening from the send step where the exceptions are not handled and thus the user is not getting the alert for a failed restart.
Is it possible to restart the failed BPM from the starting of the BLOCK? ( with some chages in the configurations)
Regards
Sunil.
we have a series of multiple steps in the BPM and if i want to capture the system errros and exceptions for 2 synchronus send steps and 3 Transformation steps then do i need to define for all of them a additional same step as in the 1st part of blog.
any suggestions would be appreciated
Thanks
Srini