Skip to content

Latest commit

 

History

History
152 lines (132 loc) · 3.9 KB

INSTALL.K8S.md

File metadata and controls

152 lines (132 loc) · 3.9 KB
Raw

Installing Kubernetes

On the master node, start etcd

docker run \
  --net=host \
  --detach \
  gcr.io/google_containers/etcd:2.0.12 \
  /usr/local/bin/etcd \
    --addr=127.0.0.1:4001 \
    --bind-addr=0.0.0.0:4001 \
    --data-dir=/var/etcd/data

Download the latest stable kubernetes.tar.gz from: https://github.com/kubernetes/kubernetes/releases

Untar the file and look for kubernetes-server-linux-amd64.tar.gz. Untar that file too. Copy kube-apiserver, kube-controller-manager, kube-scheduler, kubelet and kubectl to a directory.

On the master node, start the following daemons: kube-apiserver, kube-controller-manager, kube-scheduler. You can start first two of them in two modes: plain text HTTP or HTTPS. For plain text use the following commands:

  • kube-apiserver
nohup ./kube-apiserver \
  --service-cluster-ip-range=192.168.200.0/24 \
  --address=0.0.0.0 \
  --etcd-servers=http://127.0.0.1:4001 \
  --v=2 \
  2>&1 > /dev/null &
  • kube-controller-manager
nohup ./kube-controller-manager \
  --master=127.0.0.1:8080 \
  --v=2 \
  2>&1 > /dev/null &

For HTTPS use, the certificates and token should be generated beforehand. Please refer to this document on how to generate necessary certificates. And to this document on how to create static token files.

After the preparations the following commands should be run:

  • kube-apiserver
nohup ./kube-apiserver \
  --service-cluster-ip-range=192.168.200.0/24 \
  --address=0.0.0.0 \
  --etcd-servers=http://127.0.0.1:4001 \
  --v=2 \
  --secure-port=443 \
  --tls-cert-file=/etc/kubernetes/ssl/apiserver.pem \
  --tls-private-key-file=/etc/kubernetes/ssl/apiserver-key.pem \
  --client-ca-file=/etc/kubernetes/ssl/ca.pem \
  --service-account-key-file=/etc/kubernetes/ssl/apiserver-key.pem \
  --token-auth-file=/etc/kubernetes/auth/token.csv \
  2>&1 > /dev/null &
  • kube-controller-manager
nohup ./kube-controller-manager \
  --master=http://127.0.0.1:8080 \
  --v=2 \
  --service-account-private-key-file=/etc/kubernetes/ssl/apiserver-key.pem
  --root-ca-file=/etc/kubernetes/ssl/ca.pem
  2>&1 > /dev/null &
  • kube-scheduler
nohup ./kube-scheduler \
  --master=127.0.0.1:8080 \
  --v=2 \
  2>&1 > /dev/null &

On the minions, you need to download a few upstream CNI plugins (as root or using sudo)

mkdir -p /opt/cni/bin && cd /opt/cni/bin
wget https://github.com/containernetworking/cni/releases/download/v0.2.0/cni-v0.2.0.tgz
tar xfz cni-v0.2.0.tgz

On minions, start the kubelet specifying that the network plugin is of type CNI and the network plugin directory to be /etc/cni/net.d. e.g:

nohup ./kubelet \
  --api-servers=http://10.33.74.22:8080 \
  --v=2 \
  --address=0.0.0.0 \
  --enable-server=true \
  --network-plugin=cni \
  --network-plugin-dir=/etc/cni/net.d \
  2>&1 > /dev/null &

If kube-apiserver and kube-controller-manager were started in HTTPS mode run the following commands:

echo "apiVersion: v1
kind: Config
clusters:
- name: local
  cluster:
    certificate-authority: /etc/kubernetes/ssl/ca.pem
users:
- name: kubelet
  user:
    client-certificate: /etc/kubernetes/ssl/worker.pem
    client-key: /etc/kubernetes/ssl/worker-key.pem
contexts:
- context:
    cluster: local
    user: kubelet
  name: kubelet-context
current-context: kubelet-context" > /etc/kubernetes/worker-kubeconfig.yaml

nohup ./kubelet \
  --api-servers=https://10.33.74.22 \
  --v=2 \
  --address=0.0.0.0 \
  --enable-server=true \
  --network-plugin=cni \
  --network-plugin-dir=/etc/cni/net.d \
  --kubeconfig=/etc/kubernetes/worker-kubeconfig.yaml \
  --tls-cert-file=/etc/kubernetes/ssl/worker.pem \
  --tls-private-key-file=/etc/kubernetes/ssl/worker-key.pem \
  2>&1 > /dev/null &

You can then verify that all your nodes are registered by running the following on the master node.

./kubectl get nodes