Kubernetes部署zookeeper集群

    其实网络上有很多Kubernetes部署zookeeper集群的文章，搜了好多总觉得每个文章都回拉下一点东西，所以弄好之后把这个yaml文件发到这里，希望有帮到大家。

   这里着重说几点：

1、在containers中如果加command字段可能会造成容器直接认为运行结束退出，所以需要初始化容器的情况下，可以先initcontainers来处理，初始化完成后再正常启动。containers和initcontainers的顺序感觉应该没有影响；

2、如果初始化容器是需要在挂在的pvc中进行操作，那么就需要在initcontainers和containers中都进行volumeMounts操作，我一开始也以为只要在initcontainers中进行了挂在就不需要再containers中再挂载了； 

3、因为集群启动zookeeper的时候需要指定myid，由于kubernetes启动的zookeeper一般是从zookeeper-0开始，但是集群中的myid一般是从1开始，所以在初始化文件时候可能还需要一个简单的运算赋值。

apiVersion: apps/v1
kind: StatefulSet
metadata:
  name: zookeeper
spec:
  selector:
    matchLabels:
      app: zookeeper
  serviceName: "zookeeper"
  replicas: 3
  template:
    metadata:
      labels:
        app: zookeeper
    spec:
      terminationGracePeriodSeconds: 30
      containers:
        - name: zookeeper
          image: <your-image>
          imagePullPolicy: Always
          ports:
            - containerPort: 2181
              name: client
            - containerPort: 2888
              name: server
            - containerPort: 3888
              name: leader-election
          volumeMounts:
          - name: data
            mountPath: /opt/zookeeper/data
          - name: config
            mountPath: /conf
            readOnly: true
          resources: # 设置资源限制
            limits:
              cpu: "1"
              memory: "1Gi"
            requests:
              cpu: "1"
              memory: "1Gi"
      initContainers:
      - name: init-myid
        image: <your-image>
        command: ["sh", "-c", "myid=`echo $(hostname | awk -F'-' '{print $NF}')` && mkdir -pv /opt/zookeeper/logs && if [ ! -f /opt/zookeeper/data/myid ]; then echo $myid && echo $(($myid + 1)) > /opt/zookeeper/data/myid; fi"]
        volumeMounts:
          - name: data
            mountPath: /opt/zookeeper/data
          - name: config
            mountPath: /conf
            readOnly: true
      volumes:
      - name: config
        configMap:
          name: zookeeper-config
          items:
          - key: zoo.cfg
            path: zoo.cfg
          - key: log4j.properties
            path: log4j.properties
  volumeClaimTemplates:
  - metadata:
      name: data
    spec:
      accessModes: [ "ReadWriteOnce" ]
      storageClassName: "nfs-storage"
      resources:
        requests:
          storage: 1Gi