README.md

Kubernetes Basics (Pods, Deployments, Services, Namespaces, Health Checks)

Expected outcome

In this lab you will deploy some basic kubernetes resources for a sample application.

How to

1. Create a Pod manifest file (pod.yaml) that has the following parameters

   • name: my-pod
   • Uses 2 Labels
     • zone = prod
     • version = v1
   • image = evillgenius/kuar:1
   • Exposes port 8080

   Example

2. Deploy the Pod manifest

   Execute the manifest file: kubectl create -f config\pod.yaml

   Expose the Pod by port forwarding: kubectl port-forward my-pod 8080:8080 (will run in the background)

   Verify the the endpoint locally: http://localhost:8080

   Stop the port forwarding command.

3. Create a Deploy manifest file (deploy.yaml) using the same parameters as above but add

   • Create 3 replicas of the app
   • Use a RollingUpdate strategy maxUnavailable = 1 and maxSurge = 1

   Example

4. Deploy the Deploy manifest

   Execute the manifest file: kubectl create -f config\deploy.yaml

   Expose the Pod by port forwarding: kubectl port-forward my-pod 8080:8080 (will run in the background)

   Verify the the endpoint locally: http://localhost:8080

   Stop the port forwarding command (Control + c)

5. Create a Service manifest file (service.yaml) that exposes the container using a LoadBalancer

   Example

6. Deploy the Service manifest

   Execute the manifest file: kubectl create -f config\service.yaml

   Note: Creating a public endpoint can take up to a few minutes. You can run kubectl get service -w until an external IP is shown. You will also be able to view the public IP from the portal.

   Verify the the endpoint publicly: http://:8080/

7. Clean Up

   kubectl delete pods,deployments,services,namespaces --all
   

   Note: some namespaces can not be deleted, and the kubernetes service should restart automatically.

8. Create an App manifest file (`myhealthyapp.yaml') that combines the Deploy and Service manifests and adds Health Checks

   Add:

   • The kuar app has an http /healthy path listening on port 8080 for Liveness
   • The kuar app has an http /ready path listening on port 8080 for Readiness

   Example

9. Deploy the App manifest

   Execute the manifest file: kubectl create -f config\myhealthyapp.yaml

10. Change your context

    The myhealthyapp manifest put everything in the 'my-ns' namespace. The current context is the default k8s namespace. In order to see the new deployment/services/pods, the namespace context must be my-ns.

    Use the kubectl config view command to view the current cluster configuration. This should look something like:

[vagrant@default-centos-74 hackfest]$ kubectl config view
apiVersion: v1
clusters:
- cluster:
    certificate-authority-data: REDACTED
    server: https://hackfestk8-kubernetes-hackf-ea55d6-fa0dd490.hcp.eastus.azmk8s.io:443
  name: hackfestK8sCluster
- cluster:
    certificate-authority-data: REDACTED
    server: https://hackfestk8-kubernetes-hackf-ea55d6mgmt.eastus.cloudapp.azure.com
  name: hackfestk8-kubernetes-hackf-ea55d6mgmt
contexts:
- context:
    cluster: hackfestK8sCluster
    user: clusterUser_Kubernetes-Hackfest_hackfestK8sCluster
  name: hackfestK8sCluster
- context:
    cluster: hackfestk8-kubernetes-hackf-ea55d6mgmt
    user: hackfestk8-kubernetes-hackf-ea55d6mgmt-admin
  name: hackfestk8-kubernetes-hackf-ea55d6mgmt
current-context: hackfestk8-kubernetes-hackf-ea55d6mgmt
kind: Config
preferences: {}
users:
- name: clusterUser_Kubernetes-Hackfest_hackfestK8sCluster
  user:
    client-certificate-data: REDACTED
    client-key-data: REDACTED
    token: dac003cceb03ceb3977f83a7bf9b28f7
- name: hackfestk8-kubernetes-hackf-ea55d6mgmt-admin
  user:
    client-certificate-data: REDACTED
    client-key-data: REDACTED
[vagrant@default-centos-74 hackfest]$ 

The "current-context" needs to bee changed. Before changing the context, a new (human readable) context must be created. This is done, as follows, using the information that was displayed on your terminal (similar to above):

kubectl config set-context my-ns --namespace=my-ns \
  --cluster=hackfestk8-kubernetes-hackf-ea55d6mgmt \
  --user=hackfestk8-kubernetes-hackf-ea55d6mgmt-admin

With this, the namespace will show up in the kubectl config view similar to the following:

- context:
    cluster: hackfestk8-kubernetes-hackf-ea55d6mgmt
    namespace: my-ns
    user: hackfestk8-kubernetes-hackf-ea55d6mgmt-admin
  name: my-ns

With the new context created, simply switch to that context and then all actions occur in that context:

[vagrant@default-centos-74 hackfest]$ kubectl config use-context my-ns
Switched to context "my-ns".
[vagrant@default-centos-74 hackfest]$ kubectl get pods
NAME                                READY     STATUS    RESTARTS   AGE
my-healthydeploy-2191437352-8jxn6   1/1       Running   0          1h
my-healthydeploy-2191437352-k7dph   1/1       Running   0          1h
my-healthydeploy-2191437352-lcr09   1/1       Running   0          1h
[vagrant@default-centos-74 hackfest]$ 

Once in the right context, you can check the status of the service to get the endpoint IP address: kubectl get service my-healthysvc, eventually it will show an external IP.

Verify the the endpoint publicly: http://:8080/

Navigate to the liveness and readiness pages: http://:8080/-/liveness & http://:8080/-/readiness

Open the K8S cluster and monitor failures as you experiment with sending failures

10. Clean Up again

Advanced areas to explore

1. Deploy a Pod that has multiple containers in a single Pod.
2. Change the image version of your deployment and cause a RollingUpdate.
3. Rollback an update