Kubernetes - DEVOPS 🧊

@(WebDev)[webdev]

Document using languages: portuguese and english

GCP - Google Cloud Platform
Como instalar o Google Cloud SDK

Dinâmica "Superficial"

Source: Introdução ao Kubernetes

Cluster é um conjunto de máquinas (Nodes). Cada máquina possui uma quantidade de vCPU e memória.

Pods: unidade que contém os containers provisionados
O Pod representa os processos rodando no cluster

Services

É uma forma de agregar um conjunto de pods para então implementar políticas de visibilidade.

Type of connections

Sectors

Minikube

Initialization

minikube start

or according the article Minikube on Windows 10 with Hyper-V: minikube start --vm-driver hyperv --hyperv-virtual-switch "Primary Virtual Switch"

Kubectl

Show services kubectl get svc
Show pods kubectl get pods
Show deployments kubectl get deployments

Creating a single pod to test

Source

kubectl apply -f pod.yaml

# kubectl apply -f pod.yaml
apiVersion: v1
kind: Pod
metadata:
  name: pod-exemplo
spec:
  containers:
    - name: pod-exemplo
      image: nginx:1.17-alpine

View log of pod by name kubectl logs pod-exemplo

Creating a deployment

Using command-line

Create a deployment kubectl create deployment hello-nginx --image=nginx:1.17-alpine
- Check it with kubectl get deployments
Create a service kubectl expose deployment hello-nginx --type=LoadBalancer --port=80
- To show it you can run kubectl get services
NAME TYPE CLUSTER-IP EXTERNAL-IP PORT(S)

hello-nginx LoadBalancer 10.97.138.76 pending 80:31365/TCP
To test on browser use minikube service hello-nginx

Using `deployment.yaml`

create a file deployment.yaml with

# kubectl apply -f deployment.yaml
apiVersion: apps/v1
kind: Deployment
metadata:
  name: hello-nginx
  labels:
    app: nginx
spec:
  replicas: 1
  selector:
    matchLabels:
      app: hello-nginx
  template:
    metadata:
      labels:
        app: hello-nginx
    spec:
      containers:
      - name: nginx
        image: nginx:1.17-alpine
        ports:
        - containerPort: 80

Delete previous deployment with kubectl delete deployments --all
kubectl apply -f deployment.yaml

Creating a service

create a file service.yaml with

# kubectl apply -f service.yaml
apiVersion: v1
kind: Service
metadata:
  name: nginx-service
spec:
  type: LoadBalancer
  ports: 
    - port: 80
  selector:
    app: hello-nginx

Run kubectl apply -f service.yaml
Test on command line kubectl get services
Test on browser with minikube service nginx-service

Using ConfigMap to Update nginx settings

create a file configmap.yaml with

# kubectl apply -f configmap.yaml
apiVersion: v1
kind: ConfigMap
metadata:
  name: nginx-conf
  labels:
    app: myapplication
data:
  nginx.conf: |

    server {
      listen 80;
      index index.php index.html;
      root /usr/share/nginx/html;

      rewrite ^/google$ http://google.com permanent;
    }

Add new volume and volumeMount to deployment.yaml

spec:
  containers:
  - name: nginx
    image: nginx:1.17-alpine
    ports:
    - containerPort: 80

    volumeMounts:
    - mountPath:  /etc/nginx/conf.d
      name:  nginx-conf
      readOnly: true

  volumes:
  - name: nginx-conf
    configMap:
      name: nginx-conf
      items:
        - key: nginx.conf
          path: nginx.conf

Test config with kubectl get configmaps
Test on browser minikube service nginx-service

Add Mysql from zero

create a folder mysql and a new file deployment.yaml

# kubectl apply -f deployment.yaml
apiVersion: apps/v1
kind: Deployment
metadata:
  name:  mysql-server
  labels:
    name:  mysql-server
spec:
  replicas: 1

  selector: # Selector using to create a Service
    matchLabels:
      app: mysql-server
      tier: db

  template: # POD 
    metadata:
      labels:
        app: mysql-server
        tier: db
    spec: # spec do POD 
      containers:
      - image:  mysql:5.7
        name:  mysql-server
        args: 
          - "--ignore-db-dir=lost+found"

        env:
        - name: MYSQL_ROOT_PASSWORD
          value: root
        ports:
        - containerPort:  3306
        volumeMounts:
         - name: mysql-persistent-storage
           mountPath: /var/lib/mysql
      
      volumes:
      - name: mysql-persistent-storage
        persistentVolumeClaim:
          claimName: mysql-pv-claim

Creating a persistent volume persistent-volume.yaml
- kubectl apply -f persistent-volume.yaml

apiVersion: v1
kind: PersistentVolumeClaim
metadata:
  name: mysql-pv-claim
spec:
  accessModes:
  - ReadWriteOnce
  resources:
    requests:
      storage: 10Gi

Create Secret

kubectl create secret generic mysql-pass --from-literal=password='a1s2d3f4'
Check it with kubectl get secrets
Change deployment.yaml to read secret:

spec: # spec do POD 
  containers:
  - image:  mysql:5.7
    name:  mysql-server
    args: 
      - "--ignore-db-dir=lost+found"

    env:
    - name: MYSQL_ROOT_PASSWORD
      valueFrom:
        secretKeyRef:
          name: mysql-pass
          key: password

Criando Mysql Service

Create a file service.yaml with

# kubectl apply -f service.yaml
apiVersion: v1
kind: Service
metadata:
  name: mysql-service
spec:
  ports:
    - port: 3306
  selector:
    app: mysql-server
    tier: db
  clusterIP: None

kubectl apply -f service.yaml
Check services with kubectl get svc
kubectl get pods
Executar pod com kubectl exec -it mysql-server-866f9b567b-4bp62 bash
Abrir mysql com: mysql -uroot -p e senha a1s2d3f4

Apagar volume e deployment

kubectl delete PersistentVolumeClaim mysql-pv-claim
kubectl delete deployment mysql

Gerenciando recursos

Testing nginx deployment.yaml:

limitar o uso de cpu de cada Pod com: cpu: "80m"
- unidade de medida de uso de vCPU é chamada milicore mCPU no kubernetes
No caso da memória se o pode bater o limite memory: "100Mi" então ele nem sobe, é bem diferente do caso do cpu que sobe apenas deixando processamento mais lento.
resources:
- requests: significa o requisito mínimo para subir o pod
- limits: significa o máximo de recursos que pode ser consumido pelo pod
- Qualquer coisa além disso o Kubernetes vai tentar duas estratégias:
  1. diminuir a quantidade computacional ou
  2. simplesmente matar o container para não dar um crash

# kubectl apply -f deployment.yaml
apiVersion: apps/v1
kind: Deployment
metadata:
  name: hello-nginx
  labels:
    app: nginx
spec:
  replicas: 2

  selector:
    matchLabels:
      app: hello-nginx
      
  template:
    metadata:
      labels:
        app: hello-nginx
    spec:
      containers:
      - name: nginx
        image: nginx:1.17-alpine
        ports:
        - containerPort: 80

        resources:
          requests:
            memory: "100Mi"
            cpu: "80m"
          limits:
            memory: "150Mi"
            cpu: "100m"

        volumeMounts:
        - mountPath:  /etc/nginx/conf.d
          name:  nginx-conf
          readOnly: true

      volumes:
      - name: nginx-conf
        configMap:
          name: nginx-conf
          items:
            - key: nginx.conf
              path: nginx.conf

HPA - Horizontal Pod Autoscaler

create a folder php-apache
create index.php

<?php

$x = 0.000001;
for ($i = 0; $i <= 100000000; ++$i) {
    $x += sqrt($x);
}
echo  PHP_EOL.'Code.education Rocks! '.PHP_EOL;

create a Dockerfile

# docker build -t lfrichter/php-apache-hpa .
# docker run -it lfrichter/php-apache-hpa bash
# docker push lfrichter/php-apache-hpa
FROM php:7-apache
COPY index.php /var/www/html/index.php
RUN chmod a+rx index.php

create HPA deployment with
- kubectl run php-apache-hpa --image=lfrichter/php-apache-hpa --requests=cpu=200m --expose --port=80
creating command to autoscaling (optional)
- kubectl autoscale deployment php-apache-hpa --cpu-percent=20 --min=1 --max=10
creating a file to autoscaling (better choice)
1. create a file hpa.yaml with:

# kubectl apply -f hpa.yaml
# kubectl get hpa
apiVersion: autoscaling/v1
kind: HorizontalPodAutoscaler
metadata:
  name: php-apache-hpa
spec:
  maxReplicas: 5
  minReplicas: 1
  scaleTargetRef: 
    apiVersion: extensions/v1beta1
    kind: Deployment
    name: php-apache-hpa
  targetCPUUtilizationPercentage: 20

Testing HPA

kubectl run -it loader --image=busybox /bin/sh
- busybox é um linux super pequeno, o mais básico para rodar algo
Criar e executar um looping para essa máquina ficar chamando o php-apache
- default é o namespace padrão do kubernetes
- Acessar página PHP
```
# wget -q -O- http://php-apache-hpa.default.svc.cluster.local;
```
- Criar looping infinito
```
while true; do wget -q -O- http://php-apache-hpa.default.svc.cluster.local; done;
```
Verifique se está funcionando em watch kubectl get hpa

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

readme.md

readme.md

Kubernetes - DEVOPS 🧊

Dinâmica "Superficial"

Services

Type of connections

Sectors

Minikube

Initialization

Kubectl

Creating a single pod to test

Creating a deployment

Using command-line

Using `deployment.yaml`

Creating a service

Using ConfigMap to Update nginx settings

Add Mysql from zero

Create Secret

Criando Mysql Service

Apagar volume e deployment

Gerenciando recursos

HPA - Horizontal Pod Autoscaler

Testing HPA

Files

readme.md

Latest commit

History

readme.md

File metadata and controls

Kubernetes - DEVOPS 🧊

Dinâmica "Superficial"

Services

Type of connections

Sectors

Minikube

Initialization

Kubectl

Creating a single pod to test

Creating a deployment

Using command-line

Using deployment.yaml

Creating a service

Using ConfigMap to Update nginx settings

Add Mysql from zero

Create Secret

Criando Mysql Service

Apagar volume e deployment

Gerenciando recursos

HPA - Horizontal Pod Autoscaler

Testing HPA

Using `deployment.yaml`