RabbitMQ 消息队列

• 普通模式: 默认的模式。消息队列的数据结构存在于每个节点上，但实际消息只存储于某一个节点上。这种模式的优点是性能较高。缺点是，一旦存储数据的节点挂掉，消息就暂时不可用了，需要节点启动后才能恢复。
• 镜像模式: 在普通模式的基础上可配置需要镜像的队列。配置队列后，消息数据会自动同步到集群中的每个节点上。该模式的优点是可用性高，缺点是性能相对较低。

apiVersion: v1
kind: PersistentVolume
metadata:
  name: pv001 
spec:
  storageClassName: standard
  accessModes:
    - ReadWriteOnce
  capacity:
    storage: 20Gi
  hostPath:
    path: /data/pv001/
---
apiVersion: v1
kind: PersistentVolume
metadata:
  name: pv002 
spec:
  storageClassName: standard
  accessModes:
    - ReadWriteOnce
  capacity:
    storage: 20Gi
  hostPath:
    path: /data/pv002/
---
apiVersion: v1
kind: PersistentVolume
metadata:
  name: pv003
spec:
  storageClassName: standard
  accessModes:
    - ReadWriteOnce
  capacity:
    storage: 20Gi
  hostPath:
    path: /data/pv003/

kubectl apply -f pvs.yaml
persistentvolume "pv001" created 
persistentvolume "pv002" created 
persistentvolume "pv003" created

---
apiVersion: v1
kind: ServiceAccount
metadata:
  name: rabbitmq
---
kind: Role
apiVersion: rbac.authorization.k8s.io/v1beta1
metadata:
  name: rabbitmq
rules:
- apiGroups: [""]
  resources: ["endpoints"]
  verbs: ["get"]
---
kind: RoleBinding
apiVersion: rbac.authorization.k8s.io/v1beta1
metadata:
  name: rabbitmq
roleRef:
  apiGroup: rbac.authorization.k8s.io
  kind: Role
  name: rabbitmq
subjects:
- kind: ServiceAccount
  name: rabbitmq

kubectl apply -f ./rabbitmq-rbac.yaml
serviceaccount "rabbitmq" configured
role.rbac.authorization.k8s.io "rabbitmq" configured
rolebinding.rbac.authorization.k8s.io "rabbitmq" configured

kind: Service
apiVersion: v1
metadata:
  name: rabbitmq
  labels:
    app: rabbitmq
spec:
  type: NodePort
  ports:
  - name: http
    protocol: TCP
    port: 15672
    targetPort: 15672
    nodePort: 31672
  selector:
    app: rabbitmq
---
apiVersion: apps/v1
kind: StatefulSet
metadata:
  name: rabbitmq
spec:
  selector:
    matchLabels:
      app: rabbitmq
  serviceName: rabbitmq
  replicas: 3
  template:
    metadata:
      labels:
        app: rabbitmq
    spec:
      serviceAccountName: rabbitmq
      terminationGracePeriodSeconds: 10
      containers:  
      - name: rabbitmq-autocluster
        image: pivotalrabbitmq/rabbitmq-autocluster
        ports:
          - name: http
            protocol: TCP
            containerPort: 15672
          - name: amqp
            protocol: TCP
            containerPort: 5672
        livenessProbe:
          exec:
            command: ["rabbitmqctl", "status"]
          initialDelaySeconds: 30
          timeoutSeconds: 5
        readinessProbe:
          exec:
            command: ["rabbitmqctl", "status"]
          initialDelaySeconds: 10
          timeoutSeconds: 5
        volumeMounts:
        - mountPath: /var/lib/rabbitmq/mnesia
          name: rabbitmq-pvc
        env:
          - name: MY_POD_IP
            valueFrom:
              fieldRef:
                fieldPath: status.podIP
          - name: RABBITMQ_USE_LONGNAME
            value: "true"
          - name: RABBITMQ_NODENAME
            value: "rabbit@$(MY_POD_IP)"
          - name: AUTOCLUSTER_TYPE
            value: "k8s"
          - name: AUTOCLUSTER_DELAY
            value: "10"
          - name: K8S_ADDRESS_TYPE
            value: "ip" 
          - name: AUTOCLUSTER_CLEANUP
            value: "true"
          - name: CLEANUP_WARN_ONLY
            value: "false"
  volumeClaimTemplates:
  - metadata:
      name: rabbitmq-pvc
    spec:
      storageClassName: standard
      accessModes:
        - ReadWriteOnce
      resources:
        requests:
          storage: 20Gi

• 创建了Headless Service用于暴露管理界面的Web端口31672。这里主要是为了演示，在实际应用中，这个可能是没必要的。
• 创建了rabbitmq的StatefulSet，含有3个节点，其中每个节点通过livenessProbe和readinessProbe检查可用性。
• 使用volumeClaimTemplates自动生成volumeClaim，这里的template即模板，会自动给StatefulSet中的每一个节点创建一个Volume Claim，命名为rabbitmq-pvc-0/1/2

kubectl apply -f ./rabbitmq-service.yaml
service "rabbitmq" created
statefulset.apps "rabbitmq" created

• Name: ha_mirror_queue
• Pattern: ^
• Apply to: All exchange and queues
• Defination:
  • ha-mode: all