State inconsistency in the cockpit and in the database tables

I faced a weird issue today. The structure of the workflows is that, there is a main workflow that has a call activity. Once the call activity is done, it again returns back to the main workflow and waits for the next input. When I was trying to update variables for this process instance, I got an error saying
Cannot modify variables for process instance “id”: execution “id” doesn’t exist: execution is null .
I checked the selected process instance id in the process instance table. It had the state as COMPLETED but the end time was still null. The Execution table has no record(which usually is there when the main process instance is still active). When I hit the API to fetch the process instance by id, it gives an error that the process instance does not exist. But in the cockpit view the process instance is still active. The state should not have been COMPLETED in the first place. Why is this happening. What might be the cause?

This topic was automatically closed 90 days after the last reply. New replies are no longer allowed.