distributed: ssh cluster does not close

Hi,

when starting a sshcluster via:

from distributed.deploy.ssh import SSHCluster

def start_cluster():
    hosts = ['localhost']
    nthreads = 1
    nprocs = 2
    c = SSHCluster(hosts * (nprocs+1),
                   connect_options={"known_hosts": None},
                   worker_options={"nthreads": nthreads, },
                   scheduler_options={"port": 0, "dashboard_address": ":8787"},
                   )
    c.scale(nprocs)
    c.close()

if __name__ == '__main__':
    start_cluster()

the scheduler and the worker are still alive even if the calling python process is finished. The dashboard is reachable and the python processes are shown in “top”.

Is there another way to close the ssh cluster?

Using Linux: distributed 2.9.1 dask 2.9.1 python 3.6.9