diff --git a/tools/gate/README.rst b/tools/gate/README.rst
index bdaed3b23e..dfbad083cb 100644
--- a/tools/gate/README.rst
+++ b/tools/gate/README.rst
@@ -6,8 +6,17 @@ locally to aid development and for demonstration purposes. Please note
 that they assume full control of a machine, and may be destructive in
 nature, so should only be run on a dedicated host.
 
-Usage
------
+Supported Platforms
+~~~~~~~~~~~~~~~~~~~
+
+Currently supported host platforms are:
+  * Ubuntu 16.04
+  * CentOS 7
+  * Fedora 25
+
+
+Usage (Single Node)
+-------------------
 
 The Gate scripts use the ``setup_gate.sh`` as an entrypoint and are
 controlled by environment variables, an example of use to run the basic
@@ -20,10 +29,34 @@ integration test is below:
     export PVC_BACKEND=ceph
     ./tools/gate/setup_gate.sh
 
-Supported Platforms
-~~~~~~~~~~~~~~~~~~~
 
-Currently supported host platforms are: \* Ubuntu 16.04 \* CentOS 7
+Usage (Multi Node)
+------------------
 
-With some preparation to docker, and disabling of SELinux operation of
-Fedora 25 is also supported.
+To use for a multinode deployment you simply need to set a few extra environment
+variables:
+
+.. code:: bash
+
+    export INTEGRATION=multi
+    export INTEGRATION_TYPE=basic
+    export PVC_BACKEND=ceph
+    #IP of primary node:
+    export PRIMARY_NODE_IP=1.2.3.4
+    #IP's of subnodes:
+    export SUB_NODE_IPS="1.2.3.5 1.2.3.6 1.2.3.7"
+    #Location of SSH private key to use with subnodes:
+    export SSH_PRIVATE_KEY=/etc/nodepool/id_rsa
+    ./tools/gate/setup_gate.sh
+
+
+Options
+-------
+
+Rather than ceph, you may use a nfs based backend. This option is especially
+useful on old or low spec machines, though is not currently supported with
+Linux Kernels >=4.10:
+
+.. code:: bash
+
+    export PVC_BACKEND=nfs
diff --git a/tools/gate/setup_gate_worker_nodes.sh b/tools/gate/setup_gate_worker_nodes.sh
index 91db2904b7..3d3f43ec1f 100755
--- a/tools/gate/setup_gate_worker_nodes.sh
+++ b/tools/gate/setup_gate_worker_nodes.sh
@@ -14,15 +14,16 @@
 set -ex
 
 : ${SSH_PRIVATE_KEY:="/etc/nodepool/id_rsa"}
+: ${PRIMARY_NODE_IP:="$(cat /etc/nodepool/primary_node_private | tail -1)"}
+: ${SUB_NODE_IPS:="$(cat /etc/nodepool/sub_nodes_private)"}
+
 sudo chown $(whoami) ${SSH_PRIVATE_KEY}
 sudo chmod 600 ${SSH_PRIVATE_KEY}
 
-PRIMARY_NODE_IP=$(cat /etc/nodepool/primary_node_private | tail -1)
 KUBEADM_TOKEN=$(sudo docker exec kubeadm-aio kubeadm token list | tail -n -1 | awk '{ print $1 }')
 
-
 SUB_NODE_PROVISION_SCRIPT=$(mktemp --suffix=.sh)
-cat /etc/nodepool/sub_nodes_private | while read SUB_NODE; do
+for SUB_NODE in $SUB_NODE_IPS ; do
   cat >> ${SUB_NODE_PROVISION_SCRIPT} <<EOS
   ssh-keyscan "${SUB_NODE}" >> ~/.ssh/known_hosts
   ssh -i ${SSH_PRIVATE_KEY} $(whoami)@${SUB_NODE} mkdir -p ${WORK_DIR%/*}