Kubernetes

Basics

in the Cluster:

Master

- manage all nodes
- start app/schedule app container

Node

- kublet -  talk to ma master
- container runtime - like docker, to pull the container image from a registry, unpacking the container and running the app

pods

• private, isolated network, visible from other pods and services within same k8s cluster

• containerized apps, all containers run in a pod. If you have containers that depends on each other, they should be in a single pod

  • share resources/namespace - The containers in a Pod share one IP Address and port space, are always co-located and co-scheduled, and run in a shared context on the same Node. And can communicate with each other
    • storages, as Volumes
    • networking, as unique cluster IP address
    • container images versions or ports
  • deployment create pods, make sure the containers are running even when the nodes failed
  • use proxy to forward communications to cluster-wide, private network
    • kubectl proxy
    • curl http://localhost:8001/version
  • the pod endpoint will be created automatically based on the pod name

  Services

• an abstraction which defines a logical set of Pods and a policy by which to access them. e.g routes traffic across a set of Pods. Persistent Endpoint for Pods
  • use YAML/JSON
  • user labels to select pods
  • ​Services can be exposed in different ways by specifying a type in the ServiceSpec, a.k.a different levels of access to a set of pods
    • Cluster IP, internal only
    • NodePort, give each node a accessible IP (from external nodeport to internal target port)
    • Load Balancer, add a load balancer from cloud provider which force the traffic from the service to nodes within it

kubectl

• kubectl get nodes
• deploy app kubectl run APP_NAME --image=DOCKER_IMAGE_URL --port=PORT_NUMBER
• check deployments kubectl getdeployments
• kubectl get pods get existing pods
• kubectl describe pods details about the Pod’s container: IP address, the ports used and a list of events related to the lifecycle of the Pod
• kubectl logs $POD_NAME
• to execute commands kubectl exec -ti $POD_NAME bash
  • kubectl exec -ti $POD_NAME curl localhost:8080 access from cluster
• service
  • kubectl expose deployment/kubernetes-bootcamp --type="NodePort" --port 8080 expose the nodeport to external traffic, the NodePort that it is assigned will allow access from external
  • kubectl label pod $POD_NAME app=v1 could have multiple labels

Monitor and Health Checks

• readiness check, if fails will remove load balancer
• live probe, restart container
• cat pods/healthy-xxx.yaml

Secrets and Configmaps

• dont put secrets in dockerfile

  #configmaps

  • can be used as env vars
  • can tell the downstream pods that configuration is changed along a pod or restart if necessary
  • kubectl create configmap MAP_NAME --from-file xx/xxx.conf

    secrets

  • kubectl create secret
  • kubectl create -f pods/xxx.yaml mount the secret as the volume of the pod

Deployments

deployments help to manage the replicas, to make sure it reaches the desired state, handle the creation, update and deletion

deployments/xxx.yaml

• specify the replicas
• label
• images

#create a deployment
kubectl create -f deployments/xxx.yaml
#check
kubectl describe deployments DEPLOYMENT_NAME
#create service
kubectl create -f services/xxx.yaml

kubectl get services xxx

kubectl get replicasets

#change the replica number in deployment yaml 
vim deployments/xxx.yaml
kubectl apply -f deployments/xxx.yaml

#then we can see the update from replicasets and pods

Update

pod with app v2 will be deployed in one pod, and at this moment, there will be both v1 and v2. Then, route the traffic to v2, and remove v1 in this pod, and then cycle it to the next pods

1. update the deployment yaml file to use v2 container image
2. and then run kubectl apply
3. kubectl describe deployments xxx check the RollingUpdateStrategy
4. kubectl get pods to check the period of the update
5. kubectl describe pods POD_NAME to check the version