Delete container objects when the server is shutting down

Before this, when #stop_worker was called, we would attempt to scale
the deployment. This would work if we had a different number (or zero)
workers configured in the settings, but we don't shut workers down
like that when shutting down the server to preserve the user's settings
for the next startup.

This resulted in worker pods not exiting when the orchestrator pod
stopped and much confusion when the next orchestrator started up.

app/models/miq_server/worker_management/monitor/quiesce.rb

@@ -35,7 +35,10 @@ def quiesce_workers_loop
     @quiesce_loop_timeout = @worker_monitor_settings[:quiesce_loop_timeout] || 5.minutes
     worker_monitor_poll = (@worker_monitor_settings[:poll] || 1.seconds).to_i_with_method
 
-    miq_workers.each { |w| stop_worker(w) }
+    miq_workers.each do |w|
+      MiqEnvironment::Command.is_podified? && w.containerized_worker? ? w.delete_container_objects : stop_worker(w)
+    end
+
     loop do
       reload # Reload from SQL this MiqServer AND its miq_workers association
       break if self.workers_quiesced?