Isolate scenario args between iterations

We have several scenarios which modifies input data. Iterations of such
scenarios are not isolated between each other.

Also, this patch returns NeutronNetworks.create_and_update_ports scenario
from unstable to regular job.

Change-Id: Ifa5d5ecd04507edb00e9befa1c7add075010258c

rally-jobs/rally-neutron.yaml

@@ -566,6 +566,34 @@
         failure_rate:
           max: 20
 
+  NeutronNetworks.create_and_update_ports:
+    -
+      args:
+        network_create_args: {}
+        port_create_args: {}
+        ports_per_network: 5
+        port_update_args:
+            admin_state_up: False
+            device_id: "dummy_id"
+            device_owner: "dummy_owner"
+            name: "_port_updated"
+      runner:
+        type: "constant"
+        times: {{smoke or 20}}
+        concurrency: {{smoke or 10}}
+      context:
+        network: {}
+        users:
+          tenants: {{smoke or 3}}
+          users_per_tenant: {{smoke or 2}}
+        quotas:
+          neutron:
+            network: -1
+            port: -1
+      sla:
+        failure_rate:
+          max: 20
+
   Quotas.neutron_update:
     -
       args: