Fix pods unable to send traffic to ClusterIP

Flannel with VXLAN suffers from a bug[1] where pods on the same node are
unable to send traffic to a service's ClusterIP when the endpoint is on
the same node.

This is due to improper NATTing of the return traffic.

The fix is to load the br_netfilter module as specified in the
kubernetes doc.[2]

[1] flannel-io/flannel#1702
[2] https://kubernetes.io/docs/setup/production-environment/container-runtimes/#forwarding-ipv4-and-letting-iptables-see-bridged-traffic

Change-Id: Ic182bba9d480421c2cb581558ebde8dfb20421c8

magnum/drivers/common/templates/kubernetes/fragments/configure-kubernetes-master.sh

@@ -41,8 +41,11 @@ EOF
         systemctl restart NetworkManager
     fi
 elif [ "$NETWORK_DRIVER" = "flannel" ]; then
-    $ssh_cmd modprobe vxlan
-    echo "vxlan" > /etc/modules-load.d/vxlan.conf
+    $ssh_cmd modprobe -a vxlan br_netfilter
+    cat <<EOF > /etc/modules-load.d/flannel.conf
+vxlan
+br_netfilter
+EOF
 fi
 
 

magnum/drivers/common/templates/kubernetes/fragments/configure-kubernetes-minion.sh

@@ -47,8 +47,12 @@ EOF
         $ssh_cmd systemctl restart NetworkManager
     fi
 elif [ "$NETWORK_DRIVER" = "flannel" ]; then
-    $ssh_cmd modprobe vxlan
-    echo "vxlan" > /etc/modules-load.d/vxlan.conf
+    $ssh_cmd modprobe -a vxlan br_netfilter
+    cat <<EOF > /etc/modules-load.d/flannel.conf
+vxlan
+br_netfilter
+EOF
+
 fi
 
 mkdir -p /srv/magnum/kubernetes/