Get best of the week

Lernen, Microservices bereitzustellen. Teil 2. Kubernetes


Hallo Habr!


Dies ist der zweite Teil einer Artikelserie "Lernen, wie man Microservices bereitstellt". Im vorherigen Teil haben wir zwei einfache Microservices geschrieben - ein Backend und ein Gateway - und herausgefunden, wie man sie in Docker-Bilder packt. Im selben Artikel werden wir die Orchestrierung unserer Docker-Container mithilfe von Kubernetes organisieren. Wir werden das System konsequent so konfigurieren, dass es in Minikube gestartet und dann fĂĽr die Bereitstellung in der Google Kubernetes Engine angepasst wird.


Serienplan:


  1. Erstellen Sie Dienste mit Spring Boot und arbeiten Sie mit Docker


    Schlüsselwörter: Java 11, Spring Boot, Docker, Bildoptimierung


  2. Entwicklung des Kubernetes-Konfigurations- und Bereitstellungssystems in Google Kubernetes Engine


    Schlüsselwörter: Kubernetes, GKE, Ressourcenmanagement, Autoscaling, Geheimnisse


  3. Erstellen eines Diagramms mit Helm 3 fĂĽr eine effizientere Clusterverwaltung


    : Helm 3, chart deployment


  4. Jenkins


    : Jenkins configuration, plugins, separate configs repository



Kubernetes:


  • . , . , .
  • . , , .
  • . . .
  • . - , Docker-, Kubernetes , .
  • . ().

Kubernetes — , , . . Kubernetes- .


GitHub .


Kubernetes


Minikube — Kubernetes . . , , GKE. Google Cloud Platform - 300$ . 2+ (n1-standard-1).


:



Kubernetes


Kubernetes . . , .


Kubernetes:


  • Namespace — . , . .


  • Pod — . . , , , , .


  • ReplicaSet — , , . , . ReplicaSet — Deployment.


  • Deployment — , ReplicaSet'. .


  • Service — . , IP-. , . - , Service .


    "" Kubernetes, .


    . , . . , . :


    • ClusterIP — . IP-, .
    • NodePort — IP- ( ClusterIP) , . <NodeIP>:<NodePort>.
    • LoadBalancer — NodePort , . IP- .

    Kubernetes DNS , . .


  • ConfigMap — , , , .


  • Secret — . (â„– SSL-), , base64- , . .


  • HorizontalPodAutoscaler — , .



Minikube configuration


Namespace:

:


kubectl create namespace msvc-ns

:


kubectl config set-context --current --namespace=msvc-ns

'msvc-ns'. , 'default'.


Kubernetes yaml- , CLI. yaml-.


ConfigMap

, . URL- . ConfigMap , ( ).


apiVersion: v1
kind: ConfigMap
metadata:
  name: urls-config
data:
  BACKEND_URL: "http://backend:8080/"

, Kubernetes . , 'backend' 8080 .


Secret

apiVersion: v1
kind: Secret
metadata:
  name: msvc-secret
type: Opaque
stringData:
  secret: secret

Opaque , -. , , Docker-, . stringData. base64, . :


data:
  secret: c2VjcmV0

Deployments

— .


:


apiVersion: apps/v1
kind: Deployment
metadata:
  name: gateway
  labels:
    tier: gateway
    app: microservices
spec:
  replicas: 3
  selector:
    matchLabels:
      tier: gateway
  strategy:
    rollingUpdate:
      maxSurge: 25%
      maxUnavailable: 25%
    type: RollingUpdate
  template:
    metadata:
      labels:
        tier: gateway
    spec:
      containers:
        - name: gateway
          image: anshelen/microservices-gateway:latest
          envFrom:
            - configMapRef:
                name: urls-config
          env:
            - name: SECRET
              valueFrom:
                secretKeyRef:
                  name: msvc-secret
                  key: secret
          readinessProbe:
              httpGet:
                path: /actuator/health
                port: 8080
              initialDelaySeconds: 5
              periodSeconds: 3
          ports:
            - containerPort: 8080
              protocol: TCP
          resources:
            limits:
              memory: "256Mi"
              cpu: "200m"
            requests:
              memory: "128Mi"
              cpu: "50m"

  • metadata.labels
    , . , 'microservices' 'gateway'.
    metadata.annotations — , .


  • spec.replicas
    .


  • spec.selector.matchLabels
    . , tier, 'backend'. spec.template , metadata.labels , .


  • spec.strategy
    spec.strategy . 'rollingUpdate' , ReplicaSet, ReplicaSet', . ( / ) maxSurge maxUnavailable. , . spec.strategy , .


  • spec.templates
    spec.templates .


  • spec.templates.metadata.labels
    , spec.selector.matchLabels, "" .


  • spec.templates.spec.containers.image
    . latest Docker-, , . , — Kubernetes. , . . 'latest' 4 , Jenkins.


  • spec.templates.spec.containers.envFrom.configMapRef
    ConfigMap .


  • spec.templates.spec.containers.env
    'SECRET', - 'secret'.


  • spec.templates.spec.containers.readinessProbe
    . , . Kubernetes , 3 200, .


    initialDelaySeconds — .
    periodSeconds — .


    livenessProbe, ( ).


  • spec.templates.spec.containers.ports
    ports . ( -p 8080:8080).


  • spec.templates.spec.containers.resources
    . limits — , requests — , . 200m — 200 ( ), Mi — .



, , .


Services

:


apiVersion: v1
kind: Service
metadata:
  labels:
    tier: backend
  name: backend
spec:
  ports:
    - port: 8080
      protocol: TCP
      targetPort: 8080
  selector:
    tier: backend

spec.ports.targetPort — , spec.ports.port — . spec.selector , , tier, 'backend'. , ClusterIP, http://backend:8080.


:


apiVersion: v1
kind: Service
metadata:
  labels:
    tier: gateway
  name: gateway
spec:
  ports:
    - nodePort: 30500
      port: 80
      protocol: TCP
      targetPort: 8080
  selector:
    tier: gateway
  type: NodePort

Minikube, , NodePort. spec.ports.nodePort — . , 30000-32767.


Minikube

deploy.yaml
apiVersion: v1
kind: ConfigMap
metadata:
  name: urls-config
data:
  BACKEND_URL: "http://backend:8080/"

---

apiVersion: v1
kind: Secret
metadata:
  name: msvc-secret
type: Opaque
stringData:
  secret: secret

---

apiVersion: v1
kind: Service
metadata:
  labels:
    tier: backend
  name: backend
spec:
  ports:
    - port: 8080
      protocol: TCP
      targetPort: 8080
  selector:
    tier: backend

---

apiVersion: v1
kind: Service
metadata:
  labels:
    tier: gateway
  name: gateway
spec:
  ports:
    - nodePort: 30500
      port: 80
      protocol: TCP
      targetPort: 8080
  selector:
    tier: gateway
  type: NodePort

---

apiVersion: apps/v1
kind: Deployment
metadata:
  name: backend
  labels:
    tier: backend
    app: microservices
spec:
  replicas: 3
  selector:
    matchLabels:
      tier: backend
  strategy:
    rollingUpdate:
      maxSurge: 25%
      maxUnavailable: 25%
    type: RollingUpdate
  template:
    metadata:
      labels:
        tier: backend
    spec:
      containers:
        - name: backend
          image: anshelen/microservices-backend:latest
          envFrom:
            - configMapRef:
                name: urls-config
          ports:
            - containerPort: 8080
              protocol: TCP
          readinessProbe:
            httpGet:
              path: /actuator/health
              port: 8080
            initialDelaySeconds: 5
            periodSeconds: 3
          resources:
            limits:
              memory: "256Mi"
              cpu: "200m"
            requests:
              memory: "128Mi"
              cpu: "50m"

---

apiVersion: apps/v1
kind: Deployment
metadata:
  name: gateway
  labels:
    tier: gateway
    app: microservices
spec:
  replicas: 3
  selector:
    matchLabels:
      tier: gateway
  strategy:
    rollingUpdate:
      maxSurge: 25%
      maxUnavailable: 25%
    type: RollingUpdate
  template:
    metadata:
      labels:
        tier: gateway
    spec:
      containers:
        - name: gateway
          image: anshelen/microservices-gateway:latest
          envFrom:
            - configMapRef:
                name: urls-config
          env:
            - name: SECRET
              valueFrom:
                secretKeyRef:
                  name: msvc-secret
                  key: secret
          readinessProbe:
              httpGet:
                path: /actuator/health
                port: 8080
              initialDelaySeconds: 5
              periodSeconds: 3
          ports:
            - containerPort: 8080
              protocol: TCP
          resources:
            limits:
              memory: "256Mi"
              cpu: "200m"
            requests:
              memory: "128Mi"
              cpu: "50m"


:


kubectl apply -f deploy.yaml

Kubernetes URL :


minikube service gateway --url -n msvc-ns

:


for i in `seq 50`; do curl $(minikube service gateway --url -n msvc-ns) && echo; done

( ):


Number of requests 1 (gateway 544429797, secret secret)
Number of requests 1 (gateway 1543772618, secret secret)
Number of requests 2 (gateway 544429797, secret secret)
Number of requests 3 (gateway 544429797, secret secret)
Number of requests 4 (gateway 544429797, secret secret)
Number of requests 1 (gateway -1940767433, secret secret)
Number of requests 2 (gateway -1940767433, secret secret)
Number of requests 2 (gateway 1543772618, secret secret)
Number of requests 5 (gateway 544429797, secret secret)
...


Number of requests 1 (gateway 544429797, secret secret)
Number of requests 1 (gateway 1543772618, secret secret)
Number of requests 2 (gateway 544429797, secret secret)
Number of requests 3 (gateway 544429797, secret secret)
Number of requests 4 (gateway 544429797, secret secret)
Number of requests 1 (gateway -1940767433, secret secret)
Number of requests 2 (gateway -1940767433, secret secret)
Number of requests 2 (gateway 1543772618, secret secret)
Number of requests 5 (gateway 544429797, secret secret)
Number of requests 3 (gateway 1543772618, secret secret)
Number of requests 6 (gateway 544429797, secret secret)
Number of requests 3 (gateway -1940767433, secret secret)
Number of requests 4 (gateway 1543772618, secret secret)
Number of requests 7 (gateway 544429797, secret secret)
Number of requests 4 (gateway -1940767433, secret secret)
Number of requests 8 (gateway 544429797, secret secret)
Number of requests 9 (gateway 544429797, secret secret)
Number of requests 10 (gateway 544429797, secret secret)
Number of requests 5 (gateway 1543772618, secret secret)
Number of requests 5 (gateway -1940767433, secret secret)
Number of requests 6 (gateway -1940767433, secret secret)
Number of requests 7 (gateway -1940767433, secret secret)
Number of requests 6 (gateway 1543772618, secret secret)
Number of requests 8 (gateway -1940767433, secret secret)
Number of requests 7 (gateway 1543772618, secret secret)
Number of requests 11 (gateway 544429797, secret secret)
Number of requests 12 (gateway 544429797, secret secret)
Number of requests 8 (gateway 1543772618, secret secret)
Number of requests 9 (gateway -1940767433, secret secret)
Number of requests 10 (gateway -1940767433, secret secret)
Number of requests 11 (gateway -1940767433, secret secret)
Number of requests 9 (gateway 1543772618, secret secret)
Number of requests 10 (gateway 1543772618, secret secret)
Number of requests 11 (gateway 1543772618, secret secret)
Number of requests 12 (gateway -1940767433, secret secret)
Number of requests 12 (gateway 1543772618, secret secret)
Number of requests 13 (gateway 544429797, secret secret)
Number of requests 13 (gateway 1543772618, secret secret)
Number of requests 13 (gateway -1940767433, secret secret)
Number of requests 14 (gateway 1543772618, secret secret)
Number of requests 14 (gateway -1940767433, secret secret)
Number of requests 15 (gateway -1940767433, secret secret)
Number of requests 14 (gateway 544429797, secret secret)
Number of requests 15 (gateway 544429797, secret secret)
Number of requests 16 (gateway 544429797, secret secret)
Number of requests 17 (gateway 544429797, secret secret)
Number of requests 15 (gateway 1543772618, secret secret)
Number of requests 16 (gateway 1543772618, secret secret)
Number of requests 16 (gateway -1940767433, secret secret)
Number of requests 17 (gateway 1543772618, secret secret)

, . , , . - , -, , . , , . , 1 , 1000, "" , 1000 . , , , . .



.



kubectl get <object-type> — . 'pod', 'service', 'deployment' . 'all'.
kubectl get <object-type> <object-name> -o yaml — yaml-.
kubectl describe <object-type> <object-name> — .
kubectl cluster-info — .
kubectl top pod/node — /.



kubectl apply -f <file/directory> — .
kubectl delete <object-type> <object-name> — .
kubectl scale deployment <deployment-name> --replicas=n — . : n = 0, n, .
kubectl edit <object-type> <object-name> — .
kubectl rollout undo deployment <deployment-name> — .



kubectl logs <pod-name> — . -f .
kubectl port-forward <pod-name> <host-port>:<container-port> — . .
kubectl exec -it <pod-name> -- /bin/sh — .
kubectl run curl --image=radial/busyboxplus:curl -i --tty — . curl, , , .
kubectl get events --sort-by='.metadata.creationTimestamp' — Kubernetes. , , , .


GKE configuration


, , , Google Kubernetes Engine.


Services

GKE , IP-. , LoadBalancer:


apiVersion: v1
kind: Service
metadata:
  labels:
    tier: gateway
  name: gateway
spec:
  selector:
    tier: gateway
  ports:
    - port: 80
      protocol: TCP
      targetPort: 8080
  type: LoadBalancer

, , . , GCP ( VPC -> IP-) , , spec.loadBalancerIp.


HorizontalPodAutoscalers

HorizontalPodAutoscaler . Minikube (- ), GKE .


apiVersion: autoscaling/v1
kind: HorizontalPodAutoscaler
metadata:
  name: backend
spec:
  scaleTargetRef:
    apiVersion: apps/v1
    kind: Deployment
    name: backend
  minReplicas: 1
  maxReplicas: 3
  targetCPUUtilizationPercentage: 50

spec.scaleTargetRef , backend. , 1 3 50%. , ( ), requests.cpu .


HorizontalPodAutoscaler' .


Quotas

. , (multitenant environment). , :


apiVersion: v1  
kind: ResourceQuota  
metadata:  
  name: msvc-quota  
spec:  
  hard:
    limits.cpu: "2"  
    limits.memory: 4Gi

- , . , , CLI (. kubectl run), LimitRange:


apiVersion: v1  
kind: LimitRange  
metadata:  
  name: msvc-default-resources  
spec:  
  limits: 
    - default:
        memory: "512Mi"  
        cpu: "250m"  
      defaultRequest:  
        memory: "256Mi"  
        cpu: "50m"  
      type: Container

, .


GKE

scripts_gke/.


create_quotas.yaml
apiVersion: v1  
kind: ResourceQuota  
metadata:  
  name: msvc-quota  
spec:  
  hard:
    limits.cpu: "2"  
    limits.memory: 4Gi

---  

apiVersion: v1  
kind: LimitRange  
metadata:  
  name: msvc-default-resources  
spec:  
  limits: 
    - default:
        memory: "512Mi"  
        cpu: "250m"  
      defaultRequest:  
        memory: "256Mi"  
        cpu: "50m"  
      type: Container

deploy.yaml
apiVersion: v1
kind: ConfigMap
metadata:
  name: urls-config
data:
  BACKEND_URL: "http://backend:8080/"

---

apiVersion: v1
kind: Secret
metadata:
  name: msvc-secret
type: Opaque
stringData:
  secret: secret

---

apiVersion: v1
kind: Service
metadata:
  labels:
    tier: backend
  name: backend
spec:
  ports:
    - port: 8080
      protocol: TCP
      targetPort: 8080
  selector:
    tier: backend

---

apiVersion: v1
kind: Service
metadata:
  labels:
    tier: gateway
  name: gateway
spec:
  selector:
    tier: gateway
  ports:
    - port: 80
      protocol: TCP
      targetPort: 8080
  type: LoadBalancer

---

apiVersion: apps/v1
kind: Deployment
metadata:
  name: backend
  labels:
    tier: backend
    app: microservices
spec:
  replicas: 2
  selector:
    matchLabels:
      tier: backend
  strategy:
    rollingUpdate:
      maxSurge: 25%
      maxUnavailable: 25%
    type: RollingUpdate
  template:
    metadata:
      labels:
        tier: backend
    spec:
      containers:
        - name: backend
          image: anshelen/microservices-backend:latest
          envFrom:
            - configMapRef:
                name: urls-config
          ports:
            - containerPort: 8080
              protocol: TCP
          readinessProbe:
            httpGet:
              path: /actuator/health
              port: 8080
            initialDelaySeconds: 5
            periodSeconds: 3
          resources:
            limits:
              memory: "512Mi"
              cpu: "250m"
            requests:
              memory: "256Mi"
              cpu: "50m"

---

apiVersion: autoscaling/v1
kind: HorizontalPodAutoscaler
metadata:
  name: backend
spec:
  scaleTargetRef:
    apiVersion: apps/v1
    kind: Deployment
    name: backend
  minReplicas: 1
  maxReplicas: 3
  targetCPUUtilizationPercentage: 50

---

apiVersion: apps/v1
kind: Deployment
metadata:
  name: gateway
  labels:
    tier: gateway
    app: microservices
spec:
  replicas: 2
  selector:
    matchLabels:
      tier: gateway
  strategy:
    rollingUpdate:
      maxSurge: 25%
      maxUnavailable: 25%
    type: RollingUpdate
  template:
    metadata:
      labels:
        tier: gateway
    spec:
      containers:
        - name: gateway
          image: anshelen/microservices-gateway:latest
          envFrom:
            - configMapRef:
                name: urls-config
          env:
            - name: SECRET
              valueFrom:
                secretKeyRef:
                  name: msvc-secret
                  key: secret
          readinessProbe:
              httpGet:
                path: /actuator/health
                port: 8080
              initialDelaySeconds: 5
              periodSeconds: 3
          ports:
            - containerPort: 8080
              protocol: TCP
          resources:
            limits:
              memory: "512Mi"
              cpu: "250m"
            requests:
              memory: "256Mi"
              cpu: "50m"

---

apiVersion: autoscaling/v1
kind: HorizontalPodAutoscaler
metadata:
  name: gateway
spec:
  scaleTargetRef:
    apiVersion: apps/v1
    kind: Deployment
    name: gateway
  minReplicas: 1
  maxReplicas: 3
  targetCPUUtilizationPercentage: 50


scripts_gke:


kubectl apply -f scripts_gke/

, . URL :


kubectl get svc gateway -o jsonpath='{.status.loadBalancer.ingress[0].ip}'

HorizontalPodAutoscaler' .



Kubernetes- Google Kubernetes Engine.


, , , . , … - , Kubernetes. ( ) Helm.


Im dritten Teil dieser Artikelserie berĂĽhren wir Helm 3, erstellen ein Helmdiagramm fĂĽr unser System und legen es in einem Repository ab, das auf der Grundlage von GitHub-Seiten erstellt wurde.

More articles:


All Articles