kubernetes-demo

1. Kubernetes abbrevated k8s
2. Kubernetes Auto-scaling, cloud-agnostic yet integrable technologies.
3. Docker-Swarm, Apache Mesos are another orchesteration.
4. Kubernetes runs anywhere Linux does. a. Your laptop b. Globally distributed data centers. c. Major cloud providers d. Anywhere in between & nearly any combination (not that you'd want to necessarily)
5. @Question: What is "Network Load Balancer", "Application load balancer"
6. minikube is good for Evaluating, Training and Development purpose with kubernetes.
7. @Question: How to use kops for setting up kubernetes on AWS.
8. Kops is a tool for production greade k8s installation, upgrades and management.
9. Kubernetes "Deployments" these are high level constructs that defines an application.
10. A Pod is the instance of docker container, Deployments can have any number of pods required to get the job done. It is a group of one or more containers with shared storage/network and other specification for how to run the container.
11. "Services" are endpoints that expose ports to the outside world.
12. You can create, delete, modify and retereve information about any of these using the kubectl command (as well as the Kubernetes local UI)
13. run kubctl

kubectl run hello-minikub
e --image=gcr.io/google_containers/echoserver:1.10
--port=8080

14. Expose service

kubectl expose deployment hello-minikube --type=NodePort

15. @Question: Type type of other services are available in k8s as --type=NodePort
16. --type=NodePort allow us to have external IP address connected to the port of container that we just created.
17. Command to get running pods

kubectl get pod

That will get you Name, Age, Current Status of the pods. 18. get url of the service with minikube

minikube service hello-minikube --url

19. To delete the deployment hello-minikube

kubectl delete deployment hello-minikube

20. To stop minikube instance.

minikube stop

That will stop virtual machine running in virtual box.

21. Deployments a. "Deployments" are the central metaphor for what we'd consider "app" or "services" b. Deployments are described as a collection of resources and references. c. Deployments take many forms based on the type of services being deployed. d. Typically described in YAML format.

22. Apply deployments

kubectl apply -f deployment.yml

Make sure your VM is running while executing this command. 23. Steps to get deployment working. a. Make sure your minikube instance is working in virtual box. else start this.

minikube start

b. apply deployment through .yml file containing all information of the container being deployed.

kubectl apply -f <deployment>.yml

That will pull the container and make it in running stage c. Expose the port to get the access of choosen port

kubectl expose deployment <deployment name> --type=NodePort

d. Get the end of the service

minikube service <deployment name> --url

You are done 24. Get more detailed information about pod

kubectl describe pods <pod-name>

25. port-forward in kubectl

kubectl port-forward <pod-name> 5000:6000

25. kubectl attach allow you to get attached to a pod to view it's output. Attaches to a process that is already running inside an existing container.

kubectl attach <pod-name>

26. ssh your specific pod/container

kubectl exec -it <container-name-under-pod> bash

27. labeling pod container

kubectl label pods tomcat-deployment-56ff5c79c5-wt9sj healthy=false

28. run more then one replicas / containers in given deployment

kubectl scale --replicas=4 deployment/tomcat-deployment

29. Exposing loadbalancer

kubectl expose deployment tomcat-deployment --type=LoadBalancer --port=8080 --target-port=8080 --name tomcat-load-balancer

You can describe loadbalancer to get more details of loadbalancer

kubectl describe services tomcat-load-balancer

30. Deployments in more details a. Deployments are a high-level object in Kubernete, they define a desired state of an application. b. They Consist of pods. c. Optionally, they can also include ReplicaSets (that are automatically management by the Deployment based on it's replica configurations)

31. What you can do with deployment. a. Create new deployment. b. update created deployment. c. apply rolling updates to pods running on your cluster. d. Rollback to a previous version. e. Pause & Resume a deployment.

32. Get all deployments

kubctl get deployments

33. Get Status of deployment rollout

kubectl rollout status deployment <deployment-name>

34. To update the image of the deployment

kubectl set image deployment/tomcat-deployment tomcat=tomcate:9.0.1

Where tomcat is container name and tomcat:9.0.1 is tomcat container of specific version which need to be upgrade. 35. Some Predefined labels provided by k8s a. hostname, b. operating system type c. architecture

beta.kubernetes.io/arch=amd64
beta.kubernetes.io/os=linux
kubernetes.io/hostname=minikube
node-role.kubernetes.io/master=

36. Label can help you to identify specific resource defined in the deployment. e.g you need to have ssd storage on specific container which need to be access faster. So you can find it with the help of defined labels.
37. Get history of different revisions of contianers done by new rolling out.

kubctl rollout history deployment tomcat-deployment #return all available revisions of rollouts
kubectl rollout history deployment/tomcat-deployment --revision=<available revision>

38. Getting information about the node having all pods installed.

kubectl describe node minikube<name of node>

39. Kubernetes has two type of health checks to ascertain two different things. a. Rediness Probes: To determine when a pod is "ready" (e.g after it has started to see whan it's ready and has loaded what it needs to internally in the image and is ready to take requests from external services) b. Liveness Probes: To determine when a pod is "health" or "unhealthy" after it has become ready.
40. To verify healthness. a. Do succesfully http or tcp request to the pod. b. Do successful command execution on the Pod with the help of exit code. if exit code is 0, it suppose to be successful.
41. There are two different type of checks are available to make the readyness of a pod container. First check is verification of Readness and second check is verification of liveness.
42. kubectl is great command line tool but some time you can also use web ui that can help you to visualise the whole environments.
43. Setting up k8s dashboard ui

kubectl proxy

44. helm is an another tool which handle the most of kubernetes package manager work.

Job 1

1. Deploy and scale mongoDB
2. It should be current version running on your kubernetes cluster with four replica
3. mongodb default port 27017
4. use mongodb latest docker image.

Solution 1

kubectl run mongo-deployment --image=mongo:4.0 --port=27017
kubectl scale --replicas=4 deployment/mongo-deployment

Solution 2

1. Create deployment yaml file

kubectl apply -f mongo-deployment.yaml

And you are done.

DNS & Service Descovery

1. Kuberneties has a built-in DNS service that is launched (and configure) automatically.
2. K8s configures kubelets to tell individual containers to use the DNS service's IP to resolve DNS names.
3. k8s provides preditable dns name.
4. Every service in your kubernetes cluster gets a DNS name
5. Kubernetes has a specific & consistent nomenclature for deciding what this DNS name is ..svc.cluster.local service name is deployment name.
6. But you can use thi swell-known pattern to design loosely coupled services or at least set "Sane-defaults"
7. One could use environment variables to set the DNS name of other services your app may need.
8. Or you could expect service at set DNS names.
9. Or combine the two above and expect service at well-known DNS names and allow environment variables to override ("conventions over configuration"). Like ruby on rails, node are also moving towards conventions over configuration method because names are highly predictable in this case.
10. kube-dns is the DNS which keep track of the given IP address and internal name of the service. and whenever ask it respond with the internal IP of the service.