Kubernetes : Basics
- Virtual machines comes with unwanted apps and services Eg: Browser
- All process / Threads were not is user control .
- Memory allocation was not in user control.
- Syncing Virtual machines is not easy
- Lot of configuration was needed
- To delete and reconfigure was a nightmare.
- configure, automate, and manage applications using containers Image
- Open Source
- Facility for scheduling of containers
- Creation, deletion, and movement of containers
- Easy scaling of containers
- Monitoring
- online : katacoda , https://labs.play-with-k8s.com/
- K3 - https://k3s.io/
- microK8
- miniKube (InBuilt Docker)
- K3
- Kind etc.,
- Cluster: A collection of nodes(Computer) represent a cluster.
- Pod:
- Runs Container Image
- unique ip address
- Pods are the smallest unit in kubernetes
- private, isolated network.
- Container Image : code
- kubectl: Command creates a proxy ,forwarding communications into the cluster using API.
minikube start
minikube stop
kubectl create namespaces somename
kubectl config set-context --current --namespace=somename
kubectl run -it alpine --image=alpine # Create a Pod
kubectl get pod
kubectl exec -it <pod-id> bash #exit
kubectl logs <podname>
# Create a Deployment with 2 pods
kubectl create deployment mydep --image=nginx --replicas=2
#configmap(env) , secret , cronJob
kubectl get deployment
kubectl describe deployment mydep
# Create a service to expose deployment
kubectl expose deployment mydep --port=80 --type=LoadBalancer
kubectl get services
kubectl describe service mydep
#Port forward to outside world
kubectl port-forward svc/mydep 8080:80&
#open browser > localhost:8080
kubectl get deployment mydep -o yaml >> mydep.yml
kubectl get service mydep -o yaml >> mydepsvc.yml
kubectl delete deployment mydep
kubectl delete service mydep
kubectl apply -f mydep.yml
kubectl apply -f mydepsvc.yml
kubectl port-forward svc/mydepsvc 8080:80&
Env Variables
kubectl create configmap env --from-literal=key=1
kubectl get configmap env -o yaml
kubectl get configmap env -o=jsonpath='{.data.key}'
Secret
kubectl create secret my-secret literal-token --from-literal pass="123"
kubectl get secret my-secret -o=jsonpath='{.data.pass}' | base64 --decode
CronJob /Job
kubectl create cronjob hello --image=busybox:1.28 --schedule="*/1 * * * *" -- echo "Hello World"
kubectl create job hello --image=busybox:1.28 -- echo "Hello World"
Export Yaml
kubectl create deployment html --image=nginx --dry-run=client -o yaml > deployment.yaml # Volume Mount
Stateful
- All Databases ie., Relational and Non-Relational
- All replicas are synced (maintain Data Consistency)
- Mandatory Persistant Volumes ( Fault Tolerance)
- Have Serial ID alloted from 0 to n
- Complex to create
- Not Recommened to use Databases in Kubernetes*
- Uses Master and Slave :
- Master is id-0
- only one allowed to change data
- Data gets transfered from id-0 > id-1> id-2 ....
Stateless:
- Eg: Rest Api Server
- Not Synced
- Not mandatory Persistant Volumes
apiVersion: v1
kind: Pod
metadata:
name: myapp
spec:
containers:
- name: myapp
image: dockerimage
command: [ "env" ] # print env var
- SERVICE Git:
- Service /portForward (Testing only)
- ClusterIp (Within Cluster)
- Loadbalancer (Commonly Used to expose Application)
- Ingress
Simple Service/port-forward (Used for Testing Only - )
apiVersion: v1
kind: Service
metadata:
name: newservice
spec:
selector:
app: targetappname
ports:
- port: 8085
targetPort: 8085
name: httpTest
#kubectl get service
#kubectl port-forward svc/newservice 8085:80
#kubectl delete service newservice
ClusterIP
- Within Cluster Only http://localhost:8080/api/v1/proxy/namespaces/<NAMESPACE>/services/<SERVICE-NAME>:<PORT-NAME>/
apiVersion: v1
kind: Service
metadata:
name: clusterservice
spec:
selector:
app: targetappname
type: ClusterIP
ports:
- name: http
port: 80
targetPort: 80
protocol: TCP
#kubectl proxy --port=8080
LoadBalancer - load balancer functionality - Only Used in Cloud Services ( AWS, GCP ,Azure ..) - Each service needs new LoadBalancer and new IP address.
apiVersion: v1
kind: Service
metadata:
name: myloadbalancer
spec:
type: LoadBalancer
selector:
app: targetappname
ports:
- port: 8000
targetPort: app_port
Ingress (Similar to port-forward / Simple Service ) - Ingress is an object that allows access to your Kubernetes services from outside - Needs ingress controller running. - Ingress controller ( will not run by default ) - Ingress Resource
kubectl get pods --all-namespaces | grep ingress
kubectl get service --all-namespaces | grep ingress
kubectl get Ingress ingress-name
kubectl delete ingress ingress-name
apiVersion: v1
kind: ConfigMap
metadata:
name: myconfig
data:
TEST_ENV: test
---
apiVersion: v1
kind: Pod
metadata:
name: myapp
spec:
containers:
- name: busybox
image: k8s.gcr.io/busybox
command: [ "env" ] # print env var
envFrom:
- configMapRef:
name: myconfig
#kubectl apply -f configmap.yaml
#kubectl logs myapp
#kubectl delete -f configmap.yml
apiVersion: v1
kind: Secret
metadata:
name: my-secret
type: Opaque
#$echo "test" | base64
data:
testpassword: dGVzdAo=
---
apiVersion: v1
kind: Pod
metadata:
name: env-pod
spec:
containers:
- name: test
image: alpine
command: ['env']
env:
- name: USER
valueFrom:
secretKeyRef:
name: my-secret
key: testpassword
#kubectl describe secrets
#kubectl get secret my-secret -o jsonpath='{.data}'
#echo '[encoded-value]' | base64 --decode
#kubectl delete secret my-secret
apiVersion: v1
kind: ConfigMap
metadata:
name: test-file
data:
data.csv: |
name,age
sam,1,
tom,2
---
apiVersion: v1
kind: Pod
metadata:
name: test-pod
spec:
restartPolicy: Never
containers:
- name: busybox
image: k8s.gcr.io/busybox
command: [ "sleep", "3600" ]
volumeMounts:
- name: config
mountPath: "/datadir"
readOnly: true
volumes:
- name: config
configMap:
name: test-file
#kubectl apply -f configmap-file.yaml
#kubectl get pod
#kubectl exec -it test-pod sh
# $cat datadir/data.csv
#kubectl delete -f configmap-file.yaml
Create new Only (Cannot give pod in command) :
- kubectl create -f manifest.yml #create new Pod
Create new / Overwrite existing x 3 :
- kubectl apply -f manifest.yaml
- kubectl apply -f https://git.io/vPieo
- cat <<EOF | kubectl apply -f -
apiVersion: v1
kind: Pod
metadata:
name: busybox-sleep
spec:
containers:
- name: busybox
image: busybox:1.28
args:
- sleep
- "1000000"
EOF
Pre-Req:
- $minikube
- make sure Docker application is running
- run "eval $(minikube docker-env)" #make use of docker in minikube Docker Tutorial
-
Remove Taint :
kubectl get nodes //copy node name kubectl describe node node1| grep -i taint //copy name kubectl taint node nodename pasteabove- -
Create namespace:
kubectl create namespaces somename kubectl config set-context --current --namespace=somename
apiVersion: v1
kind: Pod
# Pod / Job
metadata:
name: hi
spec:
containers:
- name: hi
image: ubuntu
command: ['sh', '-c', 'echo "Hello, Kubernetes!" && sleep 3600']
# imagePullPolicy: Never
kubectl create -f hi.yml
kubectl exec -it hi -- /bin/bash
Activity : Pull local Docker image into minikube
Create a file with name Dockerfile
Add below lines :
FROM alpine:3.4
RUN apk update
RUN apk add vim
RUN apk add curl
open new terminal
minikube start
eval $(minikube docker-env)
docker build -t foo:1.0 .
docker images #Check if foo is created
kubectl run foo -it --image=foo:1.0
- $cat /proc/version
- $exit
kubectl get pods
kubectl get deployments
kubectl delete deployment foo
Activity :Create "Hello world" python program push and pull from remote Docker Pre-Req:
- Linux machine
- Goto Docker-hub and create a account
- Login to Docker-hub and create a repository
- Restart your system and Open Bios
- Enable vt-x / virtualization
- Turn on Docker application
- Make sure
- Create a folder called apps
- Create 3 files inside it .
Create deployment.yaml
apiVersion: v1
kind: Pod
metadata:
name: foo
spec:
containers:
- name: whatever
image: index.docker.io/usn/repo_name:latest
imagePullPolicy: Always
imagePullSecrets:
- name: my_registry
main.py
print('hi')
Dockerfile" (no extension)
FROM python:3.7
RUN mkdir /app
WORKDIR /app
ADD . /app/
EXPOSE 5000
CMD ["python","-u", "/app/main.py"]
Steps:
- cd into apps
- sudo docker images #empty table
- sudo docker build -t any_name:v1 . # note there is '.' at the end
- sudo docker images
- sudo docker run -p 4000:80 any_name
- sudo docker images #note down the id /name, usually it is latest
- sudo docker login
- sudo docker tag TAG_id usn/repo_name:TAG_NAME_of_image
#docker tag 3a4677d31cde usn/test_repo:latest
- sudo docker push usn/repo_name:TAG_NAME_of_image
#docker push usn/repo:latest
- kubectl apply -f deployment.yaml #pull image from docker hub & create pod
- kubectl logs -f foo #hi
- kubectl get pods #shows all pods
- kubectl delete pod pod_name #to delete pod
#Status =CrashLoopBackOff. Because we just have 1 print statement , so whenever the application was closing after "hi".The pod try to restart service and had done it mutiple times
Activity : Send arguments from different terminal
kubectl attach redis_container -i
Activity :Forward ports from Pods to your local machine
kubectl port-forward redis-izl09 6379
telnet localhost 6379 #nc -l -p 6379
Ref: https://kubernetesbyexample.com/ https://training.play-with-docker.com/ops-s1-hello/
