If you are redeploying a process, and, during design time, the Keep Running Instances option is selected, then redeployment fails in the following scenarios:
Parallel or inclusive gateway pair is removed Inclusive-complex gateway pair is changed to inclusive-inclusive gateway pair Subprocess is removed or its loop characteristic is changed User task is moved to another branch in gateway pair or outside the gateway pair Activity level is changed., for example, the activity is moved inside a subprocess or gateway structure Subprocess or event subprocess or a call activity is added Event subprocess is changed from interrupting to non-interrupting Boundary event is changed from interrupting to no- interrupting Boundary event is added User task implementation is changed to use another human task
Referance:
https://docs.oracle.com/cd/E29542_01/user.1111/e15175/bpmug_alt_flw_mig.htm#BPMUG87383
Hi,
From the usage point of view, has anyone done/do such a redeployment in a production environment? Do you have any data? From my understanding this brings a lot of risks to running instances.
Hi,
Yes, deployment with keep running instances option brings lot of risks, so my recommendation is DO NOT TRY this option if your new development matches any scenario in the post.