---

一、创建命名空间elk

kubectl create ns elk

二、安装NFS

ubuntu16.04.4 部署NFS共享

mkdir /nfs_data/es
chmod -R 777 /nfs_data/es
vim /etc/exports
#添加如下内容
/nfs_data/es    192.168.0.0/16(rw,sync,no_subtree_check,no_root_squash)

#重启nfs
systemctl restart nfs

#查看共享目录
showmount -e 192.168.1.2

返回值如下，表示创建成功

三、创建StorageClass

参考文章：K8S之持久卷（PV和PVC）

四、创建es有状态服务es-cluster.yaml

会自动创建PV和PVC

apiVersion: apps/v1
kind: StatefulSet
metadata:
  name: es-cluster
  namespace: elk
spec:
  # 必须设置
  serviceName: es-cluster-svc
  replicas: 3
  selector:
    # 设置标签
    matchLabels:
      app: es-cluster
  template:
    metadata:
      # 此处必须要与上面的matchLabels相同
      labels:
        app: es-cluster
    spec:
      # 初始化容器
      # 初始化容器的作用是在应用容器启动之前做准备工作，每个init容器都必须在下一个启动之前成功完成
      initContainers:
        - name: increase-vm-max-map
          image: busybox:1.32
          command: ["sysctl", "-w", "vm.max_map_count=262144"]
          securityContext:
            privileged: true
        - name: increase-fd-ulimit
          image: busybox:1.32
          command: ["sh", "-c", "ulimit -n 65536"]
          securityContext:
            privileged: true

      # 初始化容器结束后，才能继续创建下面的容器
      containers:
        - name: es-container
          image: elasticsearch:7.9.1
          ports:
            # 容器内端口
            - name: rest
              containerPort: 9200
              protocol: TCP

          # 限制CPU数量
          resources:
            limits:
              memory: 1536Mi
              cpu: 1000m
            requests:
              cpu: 100m
              memory: 1536Mi
          # 设置挂载目录
          volumeMounts:
            - name: es-data
              mountPath: /usr/share/elasticsearch/data
            - name: es-plugins
              mountPath: /usr/share/elasticsearch/plugins

          # 设置环境变量
          env:
            # 自定义集群名
            - name: cluster.name
              value: k8s-es
            # 定义节点名，使用metadata.name名称
            - name: node.name
              valueFrom:
                fieldRef:
                  fieldPath: metadata.name
            # 初始化集群时，ES从中选出master节点
            - name: cluster.initial_master_nodes
              # 对应metadata.name名称加编号，编号从0开始
              value: "es-cluster-0,es-cluster-1,es-cluster-2"
            - name: discovery.zen.minimum_master_nodes
              value: "2"

            # 发现节点的地址，discovery.seed_hosts的值应包括所有master候选节点
            # 如果discovery.seed_hosts的值是一个域名，且该域名解析到多个IP地址，那么es将处理其所有解析的IP地址。
            - name: discovery.seed_hosts
              value: "es-cluster-0.es-cluster-svc,es-cluster-1.es-cluster-svc,es-cluster-2.es-cluster-svc"
            # 配置内存
            - name: ES_JAVA_OPTS
              value: "-Xms512m -Xmx512m"
            - name: network.host
              value: "0.0.0.0"

  volumeClaimTemplates:
    - metadata:
        # 对应容器中volumeMounts.name
        name: es-data
        labels:
          app: es-volume
      spec:
        # 存储卷可以被单个节点读写
        accessModes: [ "ReadWriteOnce" ]

        # 对应es-nfs-storage-class.yaml中的metadata.name
        storageClassName: managed-nfs-storage

        # 申请资源的大小
        resources:
          requests:
            storage: 10Gi
    - metadata:
        # 对应容器中volumeMounts.name
        name: es-plugins
      spec:
        # 存储卷可以被单个节点读写
        accessModes: [ "ReadWriteOnce" ]

        # 对应es-nfs-storage-class.yaml中的metadata.name
        storageClassName: managed-nfs-storage
        # 申请资源的大小
        resources:
          requests:
            storage: 2Gi

五、创建es服务es-service.yaml

apiVersion: v1
kind: Service
metadata:
  name: es-cluster-svc
  namespace: elk
spec:
  selector:
    # 注意一定要与"es-cluster.yaml"中spec.selector.matchLabels相同
    app: es-cluster

  # 设置服务类型
  type: NodePort
  ports:
    - name: rest
      # 服务端口
      port: 9200
      # 应用端口（Pod端口）
      targetPort: 9200
      # 映射到主机的端口，端口范围是30000~32767
      nodePort: 29200

六、检查es集群

查看集群节点：http://192.168.1.2:29200/_cat/nodes?v

查看集群健康状态：http://192.168.1.2:29200/_cat/health?v

七、部署kibana

apiVersion: apps/v1
kind: Deployment
metadata:
  name: kibana
  namespace: elk
spec:
  selector:
    matchLabels:
      app: kibana
  replicas: 1
  template:
    metadata:
      labels:
        app: kibana
    spec:
      restartPolicy: Always
      containers:
        - name: kibana
          image: kibana:7.9.1
          imagePullPolicy: Always
          ports:
            - containerPort: 5601
          env:
            - name: ELASTICSEARCH_HOSTS
              value: http://es-cluster-svc:9200
            - name: I18N.LOCALE
              value: zh-CN    #汉化
          resources:
            requests:
              memory: 1024Mi
              cpu: 50m
            limits:
              memory: 1024Mi
              cpu: 1000m
---
apiVersion: v1
kind: Service
metadata:
  name: kibana
  namespace: elk
spec:
  type: NodePort
  ports:
    - name: kibana
      port: 5601
      targetPort: 5601
      nodePort: 25601
  selector:
    app: kibana

kibana的启动时间大概2分钟

八、访问kibana

http://192.168.1.2:25601

九、参考文章

在k8s上部署ES集群