INTRODUCTION 

The purpose of this blog is to provide you information on :

  1. Deploying kubernetes cluster in private network using kops in h.a mode.
    • Where h.a mode will include
      • Multi – A.Z deployment of master and worker nodes
        • Where each master node apart from hosting its services, will also contain etcd cluster node setup.
      • Auto scaling service for each master and worker node group to maintain their desired count in case of any hardware failure.
  2. Deploy bastion host using kops for accessing private cluster nodes.
  3. Deploy kubernetes dashboard with root access.
  4. Editing master and worker nodes.

Kops is the command line utility provided by AWS to create, destroy and manage kubernetes cluster in a very clean and automated way on AWS infrastructure, It also provide support for other cloud providers like  GCE and VMware which is currently going under testing phase, hence it is not being declared officially by AWS.

Kops make use of s3 bucket in order to store and maintain revisions of kubernetes cluster.

Prerequisite

In order to get started you need to install following tools

Install kops

curl -LO https://github.com/kubernetes/kops/releases/download/$(curl -s https://api.github.com/repos/kubernetes/kops/releases/latest | grep tag_name | cut -d '"' -f 4)/kops-linux-amd64
chmod +x kops-linux-amd64
sudo mv kops-linux-amd64 /usr/local/bin/kops

Install kubectl

curl -Lo kubectl https://storage.googleapis.com/kubernetes-release/release/$(curl -s https://storage.googleapis.com/kubernetes-drelease/release/stable.txt)/bin/linux/amd64/kubectl
chmod +x ./kubectl
sudo mv ./kubectl /usr/local/bin/kubectl

Install aws-cli

# install pip for python3 (recommended)
sudo apt-get install python3-pip 
# install awscli using pip3 command.
pip3 install --upgrade --user awscli
# export the following path in .profile file which present in the user’s home directory
export PATH=/home/ubuntu/.local/bin:$PATH
#Save the file and check the version of aws-cli to verify the installation with following command
aws --version.

1. deploy kubernetes cluster

Export following variables in .profile file.

# Load access and secret key 
export AWS_ACCESS_KEY_ID=
export AWS_SECRET_ACCESS_KEY=
# Dns settings , k8s.local is used when we want use kubernetes dns not our own domain dns.
export NAME=k8s-cluster.k8s.local
# KOPS_STATE_STORE is used to refer bucket name in kops command.
export KOPS_STATE_STORE=s3://k8s-repo;

create s3 bucket with following aws cli command

aws s3api create-bucket --bucket k8s-repo --region us-east-1

Enter the following command to declare kubernetes config.

kops create cluster \
--cloud aws \
--authorization rbac \
--topology private \
--networking kopeio-vxlan \
--master-count 3 \
--master-size t2.micro \
--master-zones us-east-1a,us-east-1b,us-east-1c  \
--master-volume-size=10 \
--node-count 2 \
--node-size t2.micro \
--node-volume-size=10 \
--zones us-east-1a,us-east-1b,us-east-1c \
--ssh-public-key ~/.ssh/id_rsa.pub \
--state ${KOPS_STATE_STORE} \
${NAME}

Where

  • –cloud flag represents the cloud provider.
  • authorization flag represents authentication mode at api server
  • –topology flag represents public/private mode of cluster node networking
  • –network flag represents Container Networking Interface provider.
  • –master-zones flag represents the list of availability zones where master nodes can be deployed in H.A mode.
  • –zones flag represents the list of availability zones where worker nodes can be deployed in H.A mode.
  • –ssh-public-key flag represents public key which is to be imported into the cluster nodes for ssh access.

Enter the following commands to deploy kubernetes cluster on aws infrastructure

# Deploy kubernetes cluster configuration
kops update cluster $NAME --yes
# Validate kubernetes cluster 
kops validate cluster $NAME

2. deploy bastion instance to access cluster instances

# Declare bastion server configuration.
kops create instancegroup bastions --role Bastion --subnet utility-us-east-1a --name ${NAME}
# Deploy bastion server configuration on aws infrastructure.
kops update cluster ${NAME} --yes
# Get endpoint of the bastion's load balancer.
aws elb --output=table describe-load-balancers --region us-east-1 | grep DNSName.\*bastion|awk '{print $4}'

Ssh to bastion server , assuming its endpoint obtained from the last command is bastion-k8s-cluster-k8s-l-vj88hv-150112292.us-east-1.elb.amazonaws.com

# Execute ssh-agent session.
eval `ssh-agent -s`
# Add private key id_rsa to ssh-agent session.
ssh-add ~/.ssh/id_rsa
# Ssh to bastion server
ssh -A admin@bastion-k8s-cluster-k8s-l-vj88hv-150112292.us-east-1.elb.amazonaws.com
# ssh to any cluster node by following command
ssh admin@<node-name>
# Example ssh admin@ip-172-20-57-84.ec2.internal

3 deploy dashboard ui with root access

# Enter the following command to deploy Dashboard-UI on the cluster 
kubectl apply -f https://raw.githubusercontent.com/kubernetes/dashboard/v2.0.0-beta4/aio/deploy/recommended.yaml

Create a file named dashboard-rootuser.yaml and add the following content to it and execute command kubectl apply -f dashboard-rootuser.yaml to create end user’s ServiceAccount named root-user for accessing kubernetes dashboard.

apiVersion: v1
kind: ServiceAccount
metadata:
  name: root-user
  namespace: kubernetes-dashboard

Create a file named rootuser-role.yaml and add the following content to it and execute command kubectl apply -f rootuser-role.yaml to bind a role (i.e giving permission) to ServiceAccount named root-user.

apiVersion: rbac.authorization.k8s.io/v1
kind: ClusterRoleBinding
metadata:
  name: root-user
roleRef:
  apiGroup: rbac.authorization.k8s.io
  kind: ClusterRole
  name: cluster-admin
subjects:
- kind: ServiceAccount
  name: root-user
  namespace: kubernetes-dashboard

Execute the following command to get the token which will be required to login kubernetes dashboard.

kubectl -n kubernetes-dashboard describe secret $(kubectl -n kubernetes-dashboard get secret | grep admin-user | awk '{print $1}')

4 edit master and worker nodes

#Optional
#Set default editor.
export EDITOR=nano

Edit worker nodes

# Edit your node instance group
kops edit ig --name=${NAME} nodes
# Try editing number of worker nodes or updgrading t2.micro to t2.small
# Save your configuration.
# Finally configure your cluster with
kops update cluster ${NAME} --yes
kops rolling-update cluster --yes

Edit you master node

# Edit your master node in availability zone us-east-1a
kops edit ig --name=${NAME} master-us-east-1a
# Try editing number of master node or updgrading t2.micro to t2.small
# Save your configuration.
# Finally configure your cluster with
kops update cluster ${NAME} --yes
kops rolling-update cluster --yes
#Edit this cluster with
kops edit cluster ${NAME}