Skip to content

Latest commit

 

History

History
97 lines (77 loc) · 3.07 KB

getting-stuff-to-cloud.md

File metadata and controls

97 lines (77 loc) · 3.07 KB

Use the steps highlighted here

  • You'll need a digital ocean account prior. Make one!
  • docker-development-YT-series
  • Follow the example until the create our cluster section and continue edit.

Create a cluster

It takes some time for it to start up. Hold up!🤚

doctl kubernetes cluster create ml-zoomcamp-examples --version 1.22.8-do.0 --count 1 --size s-1vcpu-2gb --region lon1

Get the kubeconfig for the cluster

doctl kubernetes cluster kubeconfig save ml-zoomcamp-examples

# you need the config so
cp ~/.kube/config .

Get kubectl

curl -LO https://storage.googleapis.com/kubernetes-release/release/`curl -s https://storage.googleapis.com/kubernetes-release/release/stable.txt`/bin/linux/amd64/kubectl
chmod +x ./kubectl
mv ./kubectl /usr/local/bin/kubectl

# go back to your working directory for the project in your PC
cd ../..

# go back to digital ocean and download the config file: place it the working directory

move your image to digital ocean container registry

doctl registry login 

# follow instructions specified here https://docs.digitalocean.com/products/container-registry/how-to/use-registry-docker-kubernetes/#add-secret-control-panel

build and run images

# build for the trained model
docker build -t ping:v001 . 

# run the file 
docker run -it --rm -p 9696:9696 ping:v001  

# test in another terminal
curl localhost:9696/ping

tag image and push

You'll need to use the API Token route. While Authenticating with the registry

docker tag <my-image> registry.digitalocean.com/<my-registry>/<my-image>
docker push registry.digitalocean.com/<my-registry>/<my-image>

Getting back to kubectl

# confirm if cluster is up
kubectl get nodes

# what about the pods: They'll be some running definitely and they are required 
kubectl get pods -A

Preparation

Get a deployments.yaml file that is generic. You can make this from visual studio code or from the documentation. Make sure you do the following:

  1. ensure that the app, name & ****is consistent e.g ping-deployment and the image has been tagged the right way. The trick is that you should use the registry.digitalocean.com//
  2. Save the file and call it something like deployment.yaml
  3. Run. You may need to create a namespace to handle organisation better.
kubectl create -f deployments.yaml

If you have issues it maybe that the container has been given authorization to run in your cluster. Head to the console and fix that with
Making sure that your cluster and the container registry are in the same region. If they are any passwords you need to add your pods. See https://kubernetes.io/docs/tasks/inject-data-application/define-environment-variable-container/.

kubectl describe nameofpod
kubectl delete deploy ping-deployment
kubectl get pods -w

Expose the app to the internet

Make a service. See example

kubectl apply -f service.yaml