05-kubernetes-best-practices.md

Lab Guide - Kubernetes Best Practices

Exercise 1 - Resource Management

Resource Quota

1. Create a namesapce called "bp-rq"

   kubectl create ns bp-rq

2. Apply the Resource Quota

   kubectl apply -f ./assets/lab-05/resource-quota.yaml -n bp-rq

3. Verify that Used is 0.

   kubectl describe quota -n bp-rq

4. Create a high-priority pod

   kubectl apply -f ./assets/lab-05/high-priority-pod.yml -n bp-rq

5. Verify that the quote has updated

   kubectl describe quota -n bp-rq

Limit Ranges

1. Create a namesapce called "bp-lr"

   kubectl create ns bp-lr

2. Apply the Limit Range

   kubectl apply -f ./assets/lab-05/limit-range.yaml -n bp-lr

3. Describe the Limit Range

   kubectl describe limitrange/limit-mem-cpu-per-container -n bp-lr

4. Create a Pod which define some resource

   kubectl apply -f ./assets/lab-05/resourced-pod.yaml -n bp-lr

5. Review each container in turn

   kubectl describe po/busybox1 -n bp-lr

Exercise 2 - Auto Scaler

Cluster Autoscaler

1. Update your AKS cluster to enable the Cluster Autoscaler.

   az aks update \
   --resource-group <aks-resource-group> \
   --name <aks-cluster-name> \
   --enable-cluster-autoscaler \
   --min-count 2 \
   --max-count 5

HPA

1. Create a namespace called 'wordpress'

   kubectl create ns wordpress

2. Deploy wordpress using Helm:

   helm install wp-release stable/wordpress --namespace wordpress

   We will explore Helm in detail tomorrow.

3. Enable autoscaling:

   kubectl autoscale deployment wp-release-wordpress --cpu-percent=50 --min=1 --max=3 -n wordpress

4. Review HPA:

   kubectl get hpa -n wordpress

Exercise 3 - Backup Your Volumes

Dynamic Provisioning

1. Create new namespace:

   kubectl create ns bp-pvc

2. Create a PVC using default storage class, and a pod:

   kubectl apply -f ./assets/lab-05/pvc.yaml -n bp-pvc

3. Show disk created in Azure portal

4. Create a pod, and mount the disk.

5. Get a shell running, and create a new file in the mounted volume:

   kubectl exec -it mypod2 -n bp-pvc -- /bin/bash
   cd /mnt/mydemo
   hostname >> host.txt
   cat host.txt

6. Exit the shell, and delete the Pod.

   kubectl delete pod mypod2 -n bp-pvc

7. Recreate the pod

   kubectl apply -f ./assets/lab-05/pvc.yaml -n bp-pvc

8. Verify the file still exists:

   kubectl exec -it mypod2 -n bp-pvc -- /bin/bash
   cd /mnt/mydemo
   ls
   cat host.txt

Back up a persistent volume

1. Get the volume

   kubectl get pvc my-data-store2 -n bp-pvc

2. Query for the disk Id

   az disk list --query '[].id | [?contains(@,`pvc-7f700aa1-a3bd-4017-971c-ee054c3cf26a`)]' -o tsv

3. Create a snapshot:

   az snapshot create \
       --resource-group <cluster-mv-resource-group> \
       --name pvcSnapshot \
       --source <disk-id>

Restore and use a snapshot

1. Create a new disk

   az disk create --resource-group <cluster-mv-resource-group> --name pvcRestored --source pvcSnapshot

2. Get the disk ID

   az disk show --resource-group <cluster-mv-resource-group> --name pvcRestored --query id -o tsv

3. Update "restored-disk.yaml" with the disk ID. Create a pod, mounting the restored disk:

   code ./assets/lab-05/restored-disk.yaml
   kubectl apply -f ./assets/lab-05/restored-disk.yaml -n bp-pvc

4. Verify the file still exists:

   kubectl exec -it mypodrestored -n bp-pvc -- /bin/bash
   cd /mnt/mydemo
   ls
   cat host.txt