Skip to content

Latest commit

 

History

History
67 lines (49 loc) · 3.33 KB

hello-world.md

File metadata and controls

67 lines (49 loc) · 3.33 KB
Raw

hello, world

Whatever Kubernetes Installation you use, here is a first example how to use it, inspired by Kubernetes' tutorial or GKE's Quickstart.

echoserver

k create deployment echoserver --image=k8s.gcr.io/echoserver:1.10
k port-forward deployment/echoserver 8080

The echoserver web page is now available at http://localhost:8080. We can also create a service to expose it externally, like this:

$ k expose deployment echoserver --type=LoadBalancer --port=8080
$ k get service echoserver
NAME         TYPE           CLUSTER-IP   EXTERNAL-IP     PORT(S)          AGE
echoserver   LoadBalancer   10.x.y.z     a.b.c.d         8080:32151/TCP   1m

The echoserver web page is now available at http://a.b.c.d:8080. If the external IP address is still shown as "pending", then wait for a minute and retry. This assumes that your Kubernets Cluster has a Load Balancer, which actually will assign such an external IP. That's usually the case in a "managed hosted Kubernetes", such as GKE or EKS etc. but otherwise:

  • When using minikube, then you can just use minikube service instead:

    minikube service --url=true echoserver

  • When using kind you could e.g. use extra port mappings (such as these).

  • When using a "bare metal" kind of installation on a real physical machine (e.g. with Talos), or e.g. the minikube None driver which would run the echoserver container directly on the respective machine, you could also expose it as a NodePort service for it to be available on http://localhost:30xxx, where 30xxx is the port shown:

    $ k expose deployment echoserver --type=NodePort --port=8080
$ k get service echoserver
NAME         TYPE       CLUSTER-IP       EXTERNAL-IP   PORT(S)          AGE
echoserver   NodePort   10.109.191.180   <none>        8080:30xxx/TCP   55s

Now execute the following commands to have a look at what Kubernetes created based on the commands above:

$ k describe service echoserver
$ k get -o yaml service echoserver
$ k describe deployment echoserver
$ k get -o yaml deployment echoserver
$ k describe replicaset echoserver
$ k get -o yaml replicaset
$ k get -o yaml pod
$ k describe pod echoserver
(...)
Limits:
  cpu:                500m
  ephemeral-storage:  1Gi
  memory:             2Gi

Note how the CPU, Memory and Storage Limits were automatically chosen. In a next step, you'll want to set more appropriate limits yourself. To remove this first example from your cluster, clean up both the service and deployment, which will also remove its replicaset and pod:

k delete service echoserver
k delete deployment echoserver

Dashboard for Minikube

minikube dashboard

If you run minikube inside a Virtual Machine (VM) instead of on your Laptop/Desktop workstation, then you'll need to set up some port forwarding to be able to access services in your local web browser.