scale down coredns on each master during graceful upgrade (#5344)

This fixes the scenario where masters are upgraded one at a time
and coredns gets improperly scaled back up to 2 replicas.

Change-Id: I7cc9283f40efcfd61b5813c89a5805c95d901567

roles/kubernetes/master/tasks/kubeadm-upgrade.yml

@@ -49,8 +49,8 @@
   retries: 6
   delay: 5
   until: scale_down_coredns is succeeded
+  run_once: yes
   when:
-    - inventory_hostname == groups['kube-master']|first
     - kubeadm_scale_down_coredns_enabled
     - dns_mode not in ['coredns', 'coredns_dual']
   changed_when: false