none cleanup: force kubeadm, clear /etc/kubernetes, and use sudo for kubectl kill

hack/jenkins/common.sh

@@ -154,7 +154,7 @@ fi
 if pgrep kubectl; then
   echo "killing hung kubectl processes ..."
   ps -afe | grep kubectl | grep -v grep || true
-  pgrep kubectl | xargs kill || true
+  pgrep kubectl | ${SUDO_PREFIX} xargs kill || true
 fi
 
 mkdir -p "${TEST_HOME}"

hack/jenkins/linux_integration_tests_none.sh

@@ -36,9 +36,11 @@ export KUBECONFIG="/root/.kube/config"
 
 # "none" driver specific cleanup from previous runs.
 # kubeadm
-sudo kubeadm reset || true
+sudo kubeadm reset -f || true
 # Cleanup data directory
 sudo rm -rf /data/*
+# Cleanup old Kubernetes configs
+sudo rm -rf /etc/kubernetes/*
 # Stop any leftover kubelets
 systemctl is-active --quiet kubelet \
   && echo "stopping kubelet" \