-1

I have to create deployment spec file that will:

  • Launch 3 replicas of the nginx image with label all_label
  • Deployment name star_deploy

Save a copy of this spec file to deployment_spec.yaml or json. After I am done, I need to clean up(delete) any new Kubernetes object which I produced during this task. Based on this task, I do not know how to solve this task. Should I create the deployment and after that, use command

kubectl get deployment star_deploy -o yaml > deployment_spec.yaml

or during creation, use command

kubectl create deployment star_deploy --image=nginx  --dry-run -o yaml > deployment_spec.yaml

What do you think is best way to do it?

Arghya Sadhu
  • 41,002
  • 9
  • 78
  • 107
O.Man
  • 585
  • 2
  • 9
  • 20

2 Answers2

1

I would advocate usage of --dry-run option to create a base yaml file and edit it to make necessary changes and then use kubectl create -f deployment.yaml to actaually create the resources.

Once the task is done you can clean up using kubectl delete -f deployment.yaml

This is following the declarative methodology which makes it easier to define deployments (or any resource) and manage their lifecycle through CI/CD or GitOps.

Arghya Sadhu
  • 41,002
  • 9
  • 78
  • 107
1

Dry run is always good because it give only necessary output without any garbage value like revision number, creationTimestmp etc. And this will give you clean production grade yaml file for deployment. Later you can use this to create daemonsets etc.

kubectl create deployment star_deploy --image=nginx  --dry-run=client -o yaml > deployment_spec.yaml

Note: Always run dry-run as client

Once you generate the yaml then apply this yaml to create deployment by using

kubectl apply -f deployment_spec.yaml

to scale the pods for that deployment you don'even need to edit yaml file.

kubectl scale deployment star_deploy --replicas=3

and to delete the created deployment you can use

kubectl delete deployment star_deploy
Dashrath Mundkar
  • 7,956
  • 2
  • 28
  • 42