kubernetes-deployment

Helps explain how Kubernetes is deployed in Google Cloud.

Initial Steps

• Create a Google Cloud Account at https://cloud.google.com/
• Install the GCP SDK from here https://cloud.google.com/sdk/
• Make sure gcloud works on your system. If it does not, add this to your .bash_profile

export GPATH=/Users/<yourlocation>/google-cloud-sdk/bin
export PATH=$PATH:$GPATH

• Enable the following APIs
  

Google Container Engine
Google Compute Engine
Google Container Registry

• Create a Project and note down the project_id
  
• Perform gcloud init and follow through the instructions

Get started with it.

• Get your project ID and export it to your system: export PROJECT_ID=<your_project_id>
• Create a Dockerfile for your project. Or, you can clone mine.
• Build the docker image : docker build -t <gcr.io/tag_name> .
• Push your image to Google Cloud Registry : gcloud docker --push gcr.io/tag_name
• Create a Cluster : gcloud container cluster create <cluster_name>
• Make sure your kubectl has cmd access :

gcloud container clusters get-credentials <cluster_name> --zone <your-zone> --project <project_id>

• To get a UI for what's goin on with your cluster : kubectl proxy and browse to 127.0.0.1/ui

Set up your access credentials for kubectl with your context,cluster and username/token.

This is going to go in your .kube/config file.

clusters:
- cluster:
    certificate-authority: <pem file that you want>
    server: https://<IP of your kube-api-haproxy>
  name: <name of the cluster>

contexts:
- context:
    cluster: <name of the cluster>
    user: <name of the user>
  name: <name of the context>

users:
- name: <name of the user>
  user:
    token: <kubectl token>

Deployments, Pod and Scaling

• Deploy it :

kubectl run <cluster_name> --image=<image_path> --port=<your-choice-of-port>

• Check if it's deployed and other details

kubectl get deployments
kubectl get pods
kubectl cluster-info
kubectl describe node <name of the node>

• Expose your deployment as a service

kubectl expose deployment <cluster_name> --type=<whatever_type_you_want>

• Find out where your service is

kubectl get services --show-labels=true <cluster_name>

• Browse to the external-IP:port to see if your webservice is running.

• Scale it : kubectl scale deployment <cluster_name> --replicas=<# of replicas ya want>

Other commands

• To view the config kubectl config view
• Set Cluster kubectl config set-cluster <name_of_cluster> --server=https://<haproxy_ip>
• Set Context kubectl config set-context <name_of_context> —cluster=<name_of_cluster>
• Use Context kubectl config use-context <name_of_context>
• Exec a command on a given container in a pod kubectl exec -it <podname> -c <container-name>
• Get logs for a given container in a pod kubectl logs <deployment-name> <podname>
• Drain or Cordon

kubectl cordon <node_name>
kubectl uncordon <node_name>
kubectl drain <node_name>

• Get all endpoints kubectl get ep --show-labels=true

My build

Dockerhub Image : https://hub.docker.com/r/shreyasgune/kubernetes-app-174114/tags/

GCR Image : https://gcr.io/kubernetes-app-174114/shreyasgune/kubernetes-app-174114