Update the kubernetes aio guide

The kubernetes aio guide was out of date and not
working.  Kubernetes removed hyperkube from support.

This method is similar to hypercube, but it installs
flannel and dnsmasq.  It is also capable of scaling from
single to multiple nodes using the same repo.

Change-Id: Ife68cfc8e8bb535ccc2a4634522848421a955de6
Partially-implements: blueprint documentation-initialization

doc/source/kubernetes-all-in-one.rst

@@ -7,37 +7,34 @@ Kolla Kubernetes All In One Guide
 Single-node Kubernetes setup
 ============================
 
-http://kubernetes.io/docs/getting-started-guides/docker/
+DNS services
+  - dnsmasq
+  - kube-dns
+  - healthz
 
-The hyperkube container runs the following services:
-  - kube-apiserver (The master)
-  - kubelet (Starts/Stops pods and containers also syncs config)
-  - kube-scheduler (Resource manager)
-  - kube-controller-manager (Manages desired state by monitoring the RC)
-  - kube-proxy (Exposes the services on each node)
-  - etcd (Distributed key-value store)
+Hypercube service
+  - kubernetes-dashboard
+  - kube-addon-manager
+  - controller-manager
+  - apiserver
+  - scheduler
+  - kube-proxy
+  - kubelet
 
-Execute the following commands to create an all-in-one Kubernetes setup with DNS.
+http://kubernetes.io/docs/getting-started-guides/docker-multinode/#setup-the-master-node
 
-::
+Kubernetes manipulates firewall rules so we want it to be the only service on
+the host doing that or some of the containers will fail.  Disable the firewall
+on your host::
 
-   # Get the IP addr of the local host’s docker0 bridge, (use for Hyperkube DNS)
-   export DNS_SERVER_IP=`ip addr show dev docker0 | grep inet | awk '{print $2}'| cut -d'/' -f1`
-   export CLUSTER_DOMAIN="openstack.local"
+  # CentOS
+  systemctl stop firewalld
+  systemctl disable firewalld
 
-   # Start Hyperkube
-   docker run --volume=/:/rootfs:ro --volume=/sys:/sys:rw --volume=/var/lib/docker/:/var/lib/docker:rw --volume=/var/lib/kubelet/:/var/lib/kubelet:rw,shared --volume=/var/run:/var/run:rw --net=host --pid=host --privileged=true --name=kubelet -d gcr.io/google_containers/hyperkube-amd64:v1.3.0 /hyperkube kubelet --resolv-conf="" --containerized --hostname-override="127.0.0.1" --address="0.0.0.0" --api-servers=http://localhost:8080 --config=/etc/kubernetes/manifests --cluster-domain=${CLUSTER_DOMAIN} --allow-privileged=true --v=2 --cluster-dns=${DNS_SERVER_IP}
+Execute the following commands to create an all-in-one Kubernetes setup::
 
-   # Start Kube2Sky on localhost (must start before skydns)
-   docker run -d --net=host --restart=always gcr.io/google_containers/kube2sky:1.12 -v=10 -logtostderr=true -domain=${CLUSTER_DOMAIN} -etcd-server="http://127.0.0.1:4001"
-
-   # Start SkyDNS on localhost (must start after kube2sky)
-   docker run -d --net=host --restart=always -e ETCD_MACHINES="http://127.0.0.1:4001" -e SKYDNS_DOMAIN="${CLUSTER_DOMAIN}." -e SKYDNS_ADDR="0.0.0.0:53" -e SKYDNS_NAMESERVERS="8.8.8.8:53,8.8.4.4:53" gcr.io/google_containers/skydns:2015-10-13-8c72f8c
-
-
-SkyDNS as configured above will any external DNS requests through to the Google
-DNS servers, which should work under most circumstances.  You may change the
-DNS servers 8.8.8.8 and 8.8.4.4 to custom DNS providers if necessary.
+   git clone https://github.com/kubernetes/kube-deploy
+   ./kube-deploy/docker-multinode/master.sh
 
 Download kubectl::
 
@@ -45,55 +42,15 @@ Download kubectl::
    chmod 755 kubectl
    PATH=$PATH:`pwd`
 
-Create a Kubernetes cluster configuration::
-
-  kubectl config set-cluster kolla --server=http://localhost:8080
-  kubectl config set-context kolla --cluster=kolla
-  kubectl config use-context kolla
-
 Try it out::
 
-   kubectl get nodes
+   kubectl get services --all-namespaces
 
+To confirm that DNS services are working, you can start a busybox job which will
+check if ``kubernetes`` is resolvable from inside of it.  If the job completes,
+then DNS is up and running.
 
-Debugging
-=========
+::
 
-kube2sky
---------
-
-kube2sky queries Kubernetes and builds out the necessary etcd records for
-SkyDNS to use.
-
-To check to see if the Kubernetes service has been copied over from Kubernetes
-to SkyDNS, you can check the etcd::
-
-    curl http://127.0.0.1:4001/v2/keys/skydns/local/openstack/svc/default/kubernetes
-
-You should see something like this::
-
-    {"action":"get","node":{"key":"/skydns/local/openstack/svc/default/kubernetes","dir":true,"nodes":[{"key":"/skydns/local/openstack/svc/default/kubernetes/c88f1059","value":"{\"host\":\"10.0.0.1\",\"priority\":10,\"weight\":10,\"ttl\":30,\"targetstrip\":0}","modifiedIndex":137,"createdIndex":137}],"modifiedIndex":92,"createdIndex":92}}
-
-That is the DNS record for the Kubernetes service.
-
-SkyDNS
-------
-
-SkyDNS is a DNS server that serves up data stored in etcd.
-
-After you have verified that kube2sky is creating the necessary records in
-etcd, you can check to see if the SkyDNS server is responding::
-
-    nslookup kubernetes.default.svc.openstack.local 127.0.0.1
-
-You should see something like this::
-
-    Server:   127.0.0.1
-    Address:  127.0.0.1#53
-
-    Name: kubernetes.default.svc.openstack.local
-    Address: 10.0.0.1
-
-From inside a Kubernetes pod, you can use::
-
-    nslookup kubernetes
+  kubectl create -f tools/test-dns.yml
+  kubectl get jobs

tools/test-dns.yml

@@ -0,0 +1,17 @@
+apiVersion: batch/v1
+kind: Job
+metadata:
+  name: test-dns
+spec:
+  template:
+    metadata:
+      name: test-dns
+    spec:
+      restartPolicy: OnFailure
+      containers:
+        - image: busybox
+          imagePullPolicy: IfNotPresent
+          name: busybox
+          command:
+            - nslookup
+            - kubernetes