qafro1 / ee

hello-app app to k8s

app

Objective:

This approach to create a Python application within a Docker container and then load balance the application within Kubernetes (minikube).

Prerequisites

1. Docker machine
2. Docker Hub account for docker repo
3. Minikube for Kubernetes
4. kubectl

Documentation for the above are provided below:
Docker: https://docs.docker.com/v17.12/install/

Dockerhub: https://hub.docker.com/

minikube: https://kubernetes.io/docs/tasks/tools/install-minikube/

kubectl: https://kubernetes.io/docs/tasks/tools/install-kubectl/

Implementation
Kubernetes coordinates a highly available cluster of computers that are connected to work as a single unit. A Service in Kubernetes is an abstraction which defines a logical set of Pods and a policy by which to access them. LoadBalancer - This provides an externally-accessible IP address that sends traffic to the correct port on your cluster nodes.

Diagram:

                            Kubernetes cluster 

| Loadbalancer|--------->| Services ------------>Pod1 & pod2

Installing
As discussed in the prerequisites the following needs to be checked or installed from the documentations provided. However - This project is using a sample DockerHub account for repo to get started and recommends to create your own account then follow procedures for future useage.

Configuration
Check if docker is installed: docker --version,if not install docker: https://docs.docker.com/v17.12/install/

Dockerhub: Docker Hub can automatically build images from source code in an external repository and automatically push the built image to your Docker repositories. This project has been configured and setup it will be using docker image: qafro/hello-python When you push code to a source code branch in GitHub for one of those listed image tags, the push uses a webhook to trigger a new build, which produces a Docker image. The built image is then pushed to the Docker Hub registry. Documentation to Set up automated builds for docker: https://docs.docker.com/docker-hub/builds

Minikube & kubectl should be configured according to the Documentations.

Testing
In the dev environment - steps to run test locally.

1. Check if docker is installed: docker --version.
2. Run a clone of the project: git clone https://github.com/qafro1/app.git and change into that directory : cd app
3. Build the app which build the image of app with tag hello-python: docker build -t hello-python .
4. Run the app locally: docker run --name hello-python -p 5000:5000 hello-python
5. Test the app if it's running successfully with port 5000: curl http://localhost:5000
6. Result display: Hello World...*

Running
This project will deploy hello-python python app to kubernetes cluster using minikube. This section will get Minikube up running and deploy the kubernetes manifest files. Assumming the Minikube & kubectl are configured. Also the docker image used as stated: qafro/hello-python

To deploy it using minikube, we are using deploy.yaml and service.yaml configuration file, which will tell the Kubernetes cluster the configuration.

1. minikube server needs to run: minikube start
2. check the Minikube nodes to show minikube Ready master: kubectl get node
3. Create a deployment: kubectl apply -f deploy.yaml
4. Create a create a service: kubectl apply -f service.yaml
5. Check both kubernates deployment & service are running: kubectl get deploy,svc
6. To display the result in a browser: minikube service hello-python
7. Result display : Hello World...*

Troubleshooting.

Debugging Pods
Check the current state of the Pod and recent events with the following command: kubectl describe pods ${POD_NAME}
View the state of the containers in the pod.

Pod is crashing or otherwise unhealthy
Take a look at the logs of the current container: kubectl logs ${POD_NAME} ${CONTAINER_NAME}