-
Notifications
You must be signed in to change notification settings - Fork 9
Description
Related Problem
In the lifecycle of a DSF instance, DSF process plug-ins are uninstalled for various events, for example when updating to a new version at the first or second position of the version (1.0.0.0 -> 1.1.0.0) with new FHIR resources or when installing a bug fix (1.0.0.0 -> 1.0.0.1).
At the moment, neither the running FHIR task resources (status: in-progress) nor the process instances active in BPE are cleaned up. This means that when BPE is restarted, it attempts to load Java classes that no longer exist, for example.
Describe the Solution You’d Like
A cleanup logic would be useful here that deletes both process instances that can no longer be executed in the BPE and changes the corresponding FHIR task resources to the failed state.
Describe Alternatives You’ve Considered
There should be a discussion of how to deal with failed updates, for example (e.g. a process plug-in that has a long-running process instance is to be updated). If, for example, the new process plug-in is not readable due to incorrect file permissions, the process instance could be terminated without being noticed and recovery is not possible even after the file permission problem has been resolved.