Kubernetes 部署 RabbitMQ 集群

一、安装步骤

ConfigMap

configmap.yaml

apiVersion: v1
kind: ConfigMap
metadata:
  name: rabbitmq-config
  namespace: default
data:
  enabled_plugins: |
      [rabbitmq_management,rabbitmq_mqtt,rabbitmq_peer_discovery_k8s].
  rabbitmq.conf: |
      ## Cluster formation. See https://www.rabbitmq.com/cluster-formation.html to learn more.
      cluster_formation.peer_discovery_backend  = rabbit_peer_discovery_k8s
      cluster_formation.k8s.host = kubernetes.default.svc.cluster.local
      ## Should RabbitMQ node name be computed from the pod's hostname or IP address?
      ## IP addresses are not stable, so using [stable] hostnames is recommended when possible.
      ## Set to "hostname" to use pod hostnames.
      ## When this value is changed, so should the variable used to set the RABBITMQ_NODENAME
      ## environment variable.
      cluster_formation.k8s.address_type = hostname
      ## How often should node cleanup checks run?
      cluster_formation.node_cleanup.interval = 30
      ## Set to false if automatic removal of unknown/absent nodes
      ## is desired. This can be dangerous, see
      ##  * https://www.rabbitmq.com/cluster-formation.html#node-health-checks-and-cleanup
      ##  * https://groups.google.com/forum/#!msg/rabbitmq-users/wuOfzEywHXo/k8z_HWIkBgAJ
      cluster_formation.node_cleanup.only_log_warning = true
      cluster_partition_handling = autoheal
      ## See https://www.rabbitmq.com/ha.html#master-migration-data-locality
      queue_master_locator=min-masters
      ## This is just an example.
      ## This enables remote access for the default user with well known credentials.
      ## Consider deleting the default user and creating a separate user with a set of generated
      ## credentials instead.
      ## Learn more at https://www.rabbitmq.com/access-control.html#loopback-users
      loopback_users.guest = false

RBAC

rbac.yaml

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

Service

service.yaml

apiVersion: v1
kind: Service
metadata:
  labels:
    app: rabbitmq
  name: rabbitmq
  namespace: default
spec:
  ports:
  - name: http
    port: 15672
    protocol: TCP
    targetPort: 15672
  - name: amqp
    port: 5672
    protocol: TCP
    targetPort: 5672
  - name: mqtt
    port: 1883
    protocol: TCP
    targetPort: 1883
  selector:
    app: rabbitmq
  type: ClusterIP

StatefulSet

statefulset.yaml

apiVersion: apps/v1
# See the Prerequisites section of https://www.rabbitmq.com/cluster-formation.html#peer-discovery-k8s.
kind: StatefulSet
metadata:
  name: rabbitmq
  namespace: default
spec:
  serviceName: rabbitmq
  # Three nodes is the recommended minimum. Some features may require a majority of nodes
  # to be available.
  replicas: 3
  selector:
    matchLabels:
      app: rabbitmq
  template:
    metadata:
      labels:
        app: rabbitmq
    spec:
      serviceAccountName: rabbitmq
      terminationGracePeriodSeconds: 10
      nodeSelector:
        # Use Linux nodes in a mixed OS kubernetes cluster.
        # Learn more at https://kubernetes.io/docs/reference/kubernetes-api/labels-annotations-taints/#kubernetes-io-os
        kubernetes.io/os: linux
      containers:
      - name: rabbitmq-k8s
        image: rabbitmq:3.8.2-management
        volumeMounts:
          - name: config-volume
            mountPath: /etc/rabbitmq
        # Learn more about what ports various protocols use
        # at https://www.rabbitmq.com/networking.html#ports
        ports:
          - name: http
            protocol: TCP
            containerPort: 15672
          - name: amqp
            protocol: TCP
            containerPort: 5672
          - name: mqtt
            protocol: TCP
            containerPort: 1883
        livenessProbe:
          exec:
            # This is just an example. There is no "one true health check" but rather
            # several rabbitmq-diagnostics commands that can be combined to form increasingly comprehensive
            # and intrusive health checks.
            # Learn more at https://www.rabbitmq.com/monitoring.html#health-checks.
            #
            # Stage 2 check:
            command: ["rabbitmq-diagnostics", "status"]
          initialDelaySeconds: 60
          # See https://www.rabbitmq.com/monitoring.html for monitoring frequency recommendations.
          periodSeconds: 60
          timeoutSeconds: 15
        readinessProbe:
          exec:
            # This is just an example. There is no "one true health check" but rather
            # several rabbitmq-diagnostics commands that can be combined to form increasingly comprehensive
            # and intrusive health checks.
            # Learn more at https://www.rabbitmq.com/monitoring.html#health-checks.
            #
            # Stage 1 check:
            command: ["rabbitmq-diagnostics", "ping"]
          initialDelaySeconds: 20
          periodSeconds: 60
          timeoutSeconds: 10
        imagePullPolicy: IfNotPresent
        env:
          - name: MY_POD_NAME
            valueFrom:
              fieldRef:
                apiVersion: v1
                fieldPath: metadata.name
          - name: MY_POD_NAMESPACE
            valueFrom:
              fieldRef:
                fieldPath: metadata.namespace
          - name: RABBITMQ_USE_LONGNAME
            value: "true"
          # See a note on cluster_formation.k8s.address_type in the config file section
          - name: K8S_SERVICE_NAME
            value: rabbitmq
          - name: RABBITMQ_NODENAME
            value: [email protected]$(MY_POD_NAME).$(K8S_SERVICE_NAME).$(MY_POD_NAMESPACE).svc.cluster.local
          - name: K8S_HOSTNAME_SUFFIX
            value: .$(K8S_SERVICE_NAME).$(MY_POD_NAMESPACE).svc.cluster.local
          - name: RABBITMQ_ERLANG_COOKIE
            value: "your_cookie"
      volumes:
        - name: config-volume
          configMap:
            name: rabbitmq-config
            items:
            - key: rabbitmq.conf
              path: rabbitmq.conf
            - key: enabled_plugins
              path: enabled_plugins

注意：RABBITMQ_ERLANG_COOKIE

# 生成RABBITMQ_ERLANG_COOKIE
echo $(openssl rand -base64 32)

参考文档：GitHub - rabbitmq/rabbitmq-peer-discovery-k8s: Kubernetes-based peer discovery mechanism for RabbitMQ

二、镜像模式

设置全部为镜像模式

# 列出策略
rabbitmqctl list_policies
# 设置镜像模式
rabbitmqctl set_policy H "^" '{"ha-mode":"all" , "ha-sync-mode":"automatic"}'
# 再次列出策略
rabbitmqctl list_policies

参考文档：Classic Queue Mirroring — RabbitMQ

三、压测结果

压测工具：GitHub - rabbitmq/rabbitmq-perf-test: A load testing tool

./runjava com.rabbitmq.perf.PerfTest -h amqp://guest:[email protected]:5672 -x 1 -y 2 -u "throughput-test-1" -a --id "test 1"

1. 单节点方案压测

压力模型	发送速率（msg/s）	接收速率（msg/s）
1生产者、2消费者	5266	3673
2生产者、4消费者	2972	2372
2生产者、8消费者	2794	1445

2. 集群方案压测

压力模型	发送速率（msg/s）	接收速率（msg/s）
1生产者、2消费者	15740	14041
2生产者、4消费者	29676	26524
2生产者、8消费者	23039	22822

3. 集群方案（镜像模式）压测

压力模型	发送速率（msg/s）	接收速率（msg/s）
1生产者、2消费者	10756	9314
2生产者、4消费者	14760	11592
2生产者、8消费者	17934	13135