Summary
The user is experiencing an issue with a workflow that contains subworkflows and maptasks, where some arraynodes remain stuck in "running" status, blocking subsequent tasks despite logs indicating that the process has completed. Both the master workflow and subworkflows have a failure policy set to FAIL_AFTER_EXECUTABLE_NODES_COMPLETE. The maptasks are configured with a concurrency of 2 and a minimum success ratio of 0.1. The user is seeking insights into the cause of this problem, mentioning the versions of flytekit (1.12.0) and cluster flyte (1.13.1) being used.