Jakub Gutkowski PRO
Dad, husband, blogger, developer, geek. Loves to play with novelties, learn new languages and libraries, avid conference attender, passionate about meeting new people, always willing to help.
Organizational stuff
If you can’t feed a team with two pizzas, it’s too large. That limits a task force to five to seven people, depending on their appetites
Jeff Bezos
this is a twist
pets vs cattle (another twist)
what exactly happens when execute create?
ClusterIP
ClusterIP (how it works)
NodePort
LoadBalancer
LoadBalancer
k8s in action, manning
Ingress - special type, needs to be installed
kubectl
# what is node
kubectl explain node
# what is pod
kubectl explain pod
kubectl explain pod.spec
kubectl explain pod.spec.containers
kubectl explain pod.spec.containers.command
# what is service
kubectl explain service
# what is RESOURCE_NAME
kubectl explain RESOURCE_NAME
# get status information about all nodes
$ kubectl get node
NAME STATUS ROLES AGE VERSION
NODE_NAME-0 Ready agent 25d v1.11.3
NODE_NAME-1 Ready agent 25d v1.11.3
NODE_NAME-3 Ready agent 25d v1.11.3
# get status information about single node
$ kubectl get node NODE_NAME
NAME STATUS ROLES AGE VERSION
NODE_NAME Ready agent 25d v1.11.3
# get status information about single RESOURCE (pod, service...)
$ kubectl get RESOURCE RESOURCE_NAME
# get details about node
$ kubectl describe node NAME
# get node yaml representation
$ kubectl get node NAME -o yaml
deployment
# create pod from file
kubectl create -f pod.yaml
# create pod and save base configuration so it can be used later by apply
kubectl create -f pod.yaml --save-config=true
# create pod or update if exists
kubectl apply -f pod.yaml
# replace pod
kubectl replace -f pod.yaml
# get the current pod definition
kubectl get pod NAME -o yaml|json
delete
# deletes pod described in file
$ kubectl delete -f pod.yaml
# delete pod of name
$ kubectl delete pod POD_NAME
how to access pod from the command line
# allowing access to the pod by url http://127.0.0.1:8888
kubectl port-forward pod/NAME 8888:pod_container_port
# accessing container in pod/running app in container
kubectl exec -it POD_NAME bash
kubectl exec -it POD_NAME -c container_name bash
kubectl exec POD_NAME -- cmd -p1 -p2
kubectl exec POD_NAME cmd
# and
kubectl proxy
# adding
kubectl label pod NAME key1=value1 key2=value2
# changing
kubectl label pod NAME key1=newValue --overwrite
kubectl
# get services
kubectl get service
kubectl get services
kubectl get svc
NAMESPACE NAME TYPE CLUSTER-IP EXTERNAL-IP PORT(S) AGE
default kubernetes ClusterIP 10.0.0.1 <none> 443/TCP 97d
default my-svc NodePort 10.0.18.147 <none> 8080:31238/TCP 97d
default my-second-svc NodePort 10.0.45.201 <none> 8080:32245/TCP 97d
# removes service
kubectl delete svc NAME
# create/update service
kubectl apply -f service.yaml
kubectl
# create/update deployment
kubectl apply -f deployment.yaml
# create/update deployment and keep track of operations
kubectl apply -f deployment.yaml --record=true
# get all deployments
kubectl get deployments
kubectl get deployment
kubectl get deploy
# get replica sets
kubectl get rs
managing deployments from kubectl
# return status of current deployment
# this waits if the status is not finished
$ kubectl rollout status deployment/NAME
# get a history of rollout's
$ kubectl rollout history deployment/NAME
# get history details of rollout N
$ kubectl rollout history deployment/NAME --revision=N
# undo rollout
$ kubectl rollout undo deployment/NAME
# create deployment with pod and one replicaset
kubectl run NAME --image gutek/dumpster:v1
# create 5 replicas
kubectl run NAME --image gutek/dumpster:v2 --replicas=5
# run pod that will be removed once done
kubectl run -it NAME --image=some/img --rm --restart=Never -- curl http://onet
# we can use different generators
# only pod
kubectl run NAME --image=some/img --generator=run-pod/v1
# deployment
kubectl run NAME --image=some/img --generator=extensions/v1beta1
# List the environment variables defined on all pods
kubectl set env pods --all --list
# Update all containers in all replication controllers in the project to have ENV=prod
kubectl set env deploy --all ENV_VAR=VALUE_1
# Set a deployment's container image to 'nginx:1.9.1'
kubectl set image deployment/DEPLOYMENT_NAME CONTAINER_NAME=nginx:1.9.1
# scale to 3 replicas deployment of name DEPLOYMENT_NAME
$ kubectl scale --replicas=3 deployment/DEPLOYMENT_NAME
# scale to 3 replicas ReplicaSet of name RS_NAME
$ kubectl scale --replicas=3 rs/RS_NAME
# create service of type NodePort
# for deployment hello-minikube
# and name it front-minikube
kubectl expose deployment hello-minikube --type=NodePort --name=front-minikube
# create service of type NodePort
# for pod POD_NAME
# and name it POD_NAME_svc
kubectl expose pod POD_NAME --type=NodePort --name=POD_NAME_svc
# check
$ helm version
# linux
$ sudo snap install helm --classic
# mac
brew install kubernetes-helm
# win
choco install kubernetes-helm
# or by visual studio code kubernetes plugin
basic commands
# initialize helm usage
$ helm init
# refresh repo
$ helm repo update
# list all installed packages
$ helm list
# list all even deleted pacakges
$ helm list --all
# uninstall package but keep it locally
$ help delete NAME
# remove deleted pacakges completly
$ helm del --purge NAME
# install package
$ help install channel/package-name
multi container systems
Do domu
By Jakub Gutkowski
Dad, husband, blogger, developer, geek. Loves to play with novelties, learn new languages and libraries, avid conference attender, passionate about meeting new people, always willing to help.