We seem to have a ton of phantom/orphan pip_con_scan_com threads. We have stopped everything on the cluster and confirmed there are no task_runs, statistic queries, or async mv refreshes, and no queries. However we have roughly 7500

pip_con_scan_com

threads that alternate consuming all the cpus on the entire cluster. Because there are no queries or task runs or anything we have no ability to kill the threads. Additionally, when we attempt to kill the threads by id they are not found and restarting the nodes also does not clear the these threads. Any idea how we can identify the source of these threads or kill them? StarRocks Version: 4.0.0-dccb665