elasticsearch7.x在k8s中的部署

一、说明

二、思路

三、部署

1、建nfs服务器

2、建持久卷

3、部署elasticsearch

四、附件

?pv.yaml内容

elasticsearch.yaml内容

---

一、说明

本文章内容主要的参考来源是https://www.cnblogs.com/javashop-docs/p/12410845.html，但参考文献中的elasticsearch是用的6.x版本，7.x与6.x有些配置上的差异，因此有必要在此基础上记录一下。

二、思路

• 数据的存储

在k8s中的持久化部署不可避免的要用到持久卷，我们采用nfs方式的持久卷来存储es数据。持久卷的详细介绍请见这里：
https://kubernetes.io/docs/concepts/storage/persistent-volumes/

• 节点规划

默认启动5个节点，3主2数据。根据es官方推荐每个节点的智能要分离，因此maseter节点不存储数据，只用来协调。

• 3、多节点的权限问题

es的数据目录默认只允许一个节点访问，但在k8s上采用了持久卷，所有节点的数据都存储在这个卷上，这会导致es的访问权限问题。报错如下：

java.io.IOException: failed to obtain lock on /usr/share/elasticsearch/data/nodes/0

 
 
 
 
 
 
  
* 1

当然可以通过更改es的配置max_local_storage_nodes来允许多个节点访问同一个数据目录，但es官方不推荐这样做。

所以我们的方案是更改每个节点的数据存储目录来解决 _ps:指定es配置项path.data来实现。_举例说明：

节点名

存储目录

es-data-1

/usr/share/elasticsearch/data/es-data-1

es-data-2

/usr/share/elasticsearch/data/es-data-2

三、部署

1、建nfs服务器

对于持久卷的结构规划如下：

/nfs/data/esmaster

es master节点的数据

/nfs/data/esdata

es 数据节点的数据

关于索引的磁盘占用：请根据业务的数据量情况来规划持久卷硬件的情况，根据我们实际测算1000个商品大约需要1MB/每节点

在默认的规划中，我们使用k8s的master节点作为nfs服务器，为上述卷准备了10G的空间，请确保k8s master node 不少于10G的空闲磁盘。请根据您的具体业务情况选择nfs服务器，如果条件允许最好是独立的nfs服务器。根据如上规划建立nfs服务：

#master节点安装nfs
yum -y install nfs-utils

#创建nfs目录
mkdir -p /nfs/data/{mqdata,esmaster,esdata}

#修改权限
chmod -R 777 /nfs/data/

#编辑export文件
vim /etc/exports

粘贴如下内容：
/nfs/data/esmaster *(rw,no_root_squash,sync)
/nfs/data/esdata *(rw,no_root_squash,sync)

#配置生效
exportfs -r

#查看生效
exportfs

#启动rpcbind、nfs服务
systemctl restart rpcbind && systemctl enable rpcbind
systemctl restart nfs && systemctl enable nfs

#查看 RPC 服务的注册状况
rpcinfo -p localhost

#showmount测试，这里的ip输入master节点的局域网ip
showmount -e

如果以看到可被挂载的目录：
# showmount -e 172.17.14.73

Export list for 172.17.14.73:

/nfs/data/esmaster *

/nfs/data/mqdata *

接下来，要在每一个节点上安装nfs服务以便使k8s可以挂载nfs目录

#所有node节点安装客户端
yum -y install nfs-utils
systemctl start nfs && systemctl enable nfs

这样就为k8s的持久卷做好了准备。

2、建持久卷

复制附件中的pv.yaml内容，修改其中的server配置为nfs服务器的ip地址

nfs:

server: 192.168.0.186 #这里请写nfs服务器的ip

在k8s master节点上执行下面的命令创建namespace：
kubectl create namespace ns-elasticsearch

通过下面的命令建立持久卷：
kubectl create -f pv.yaml

通过以下命令查看持久卷是否建立成功：
kubectl get pv

3、部署elasticsearch

由于在elasticsearch.yaml中设置了node信息：nodeSelector🇪🇸 enable，因此需要将node添加标签es:enable
kubectl label nodes es=enable
kubectl label nodes es=enable
kubectl label nodes es=enable

查看node标签
k get node --show-labels

复制附件elasticsearch.yaml的内，并执行下面的命令创建es集群
kubectl create -f elasticsearch.yaml

通过以上部署我们建立了一个ns-elasticsearch的namespace，并在其中创建了相应的pvc、角色账号，有状态副本集以及服务。
kubectl get pods --namespace ns-elasticsearch -o wide

NAME                     READY   STATUS    RESTARTS   AGE   IP            NODE    NOMINATED NODE   READINESS GATES
elasticsearch-data-0     1/1     Running   0          46h   10.244.2.45   vm188   <none>           <none>
elasticsearch-data-1     1/1     Running   2          45h   10.244.1.62   vm187   <none>           <none>
elasticsearch-master-0   1/1     Running   0          46h   10.244.2.44   vm188   <none>           <none>
elasticsearch-master-1   1/1     Running   0          46h   10.244.1.61   vm187   <none>           <none>
elasticsearch-master-2   1/1     Running   0          46h   10.244.0.14   vm186   <none>           <none>

服务

我们默认开启了对外nodeport端口，对应关系：

32000->9200

32100->9300

k8s内部可以通过下面的服务名称访问：

elasticsearch-api-service.ns-elasticsearch:9300

elasticsearch-service.ns-elasticsearch:9200

等待容器都启动成功后验证。

注意：

es的最大内存和最小内存需要保持一致，默认的256m太小，可适当增加，我配置的是1024m。

es7.x参考：

https://blog.csdn.net/chengyuqiang/article/details/89841544

https://www.sohu.com/a/301517999_683048

四、附件

pv.yaml内容

---
#es master节点的持久卷
apiVersion: v1
kind: PersistentVolume
metadata:
  name: pv-es-master
  labels:
    pv: pv-es-master
spec:
  capacity:
    storage: 1Gi
  accessModes:
    - ReadWriteOnce
  persistentVolumeReclaimPolicy: Recycle
  storageClassName: nfs
  nfs:
    server: 192.168.1.190  #这里请写nfs服务器的ip
    path: /nfs/data/esmaster
---
#es数据节点的持久卷
apiVersion: v1
kind: PersistentVolume
metadata:
  name: pv-es-data
  labels:
    pv: pv-es-data
spec:
  capacity:
    storage: 10Gi
  accessModes:
    - ReadWriteOnce
  persistentVolumeReclaimPolicy: Recycle
  storageClassName: nfs
  nfs:
    server: 192.168.1.190  #这里请写nfs服务器的ip
    path: /nfs/data/esdata
---
apiVersion: v1
kind: PersistentVolumeClaim
metadata:
  name: pvc-es-master
  namespace: ns-elasticsearch
spec:
  accessModes:
    - ReadWriteOnce
  resources:
    requests:
      storage: 1Gi
  storageClassName: nfs
  selector:
    matchLabels:
      pv: pv-es-master
---
apiVersion: v1
kind: PersistentVolumeClaim
metadata:
  name: pvc-es-data
  namespace: ns-elasticsearch
spec:
  accessModes:
    - ReadWriteOnce
  resources:
    requests:
      storage: 10Gi
  storageClassName: nfs
  selector:
    matchLabels:
      pv: pv-es-data

elasticsearch.yaml内容

---
apiVersion: v1
kind: ServiceAccount
metadata:
  labels:
    elastic-app: elasticsearch
  name: elasticsearch-admin
  namespace: ns-elasticsearch

---
apiVersion: rbac.authorization.k8s.io/v1
kind: ClusterRoleBinding
metadata:
  name: elasticsearch-admin
  labels:
    elastic-app: elasticsearch
roleRef:
  apiGroup: rbac.authorization.k8s.io
  kind: ClusterRole
  name: cluster-admin
subjects:
  - kind: ServiceAccount
    name: elasticsearch-admin
    namespace: ns-elasticsearch

---
kind: StatefulSet
apiVersion: apps/v1
metadata:
  labels:
    elastic-app: elasticsearch
    role: master
  name: elasticsearch-master
  namespace: ns-elasticsearch
spec:
  serviceName: es-master
  replicas: 1
  revisionHistoryLimit: 10
  selector:
    matchLabels:
      elastic-app: elasticsearch
      role: master
  template:
    metadata:
      labels:
        elastic-app: elasticsearch
        role: master
    spec:
      #将持久卷声明
      volumes:
        - name: pv-storage-elastic-master
          persistentVolumeClaim:
            claimName: pvc-es-master
      nodeSelector:
        es: enable
      containers:
        - name: elasticsearch-master
          image: registry.cn-shanghai.aliyuncs.com/luckyqing/elasticsearch:7.6.2
          lifecycle:
            postStart:
              exec:
                command: ["/bin/bash", "-c", "sysctl -w vm.max_map_count=262144; ulimit -l unlimited;chown -R elasticsearch:elasticsearch /usr/share/elasticsearch/data;"]
          ports:
            - containerPort: 9200
              protocol: TCP
            - containerPort: 9300
              protocol: TCP
          env:
            - name: MY_POD_NAME
              valueFrom:
                fieldRef:
                  apiVersion: v1
                  fieldPath: metadata.name
            #修改es默认的数据存储目录，否则多个节点同时写一个目录es权限不允许
            - name: "path.data"
              value: "/usr/share/elasticsearch/data/$(MY_POD_NAME)"
            - name: "cluster.name"
              value: "elasticsearch-cluster"
            - name: "bootstrap.memory_lock"
              value: "false"
            - name: "discovery.seed_hosts" #7.x的配置方式
              value: "elasticsearch-discovery"
            - name: "node.master"
              value: "true"
            - name: "node.data"
              value: "false"
            - name: "node.ingest"
              value: "false"
            - name: "ES_JAVA_OPTS"
              value: "-Xms1024m -Xmx1024m"
            - name: "cluster.initial_master_nodes"
              value: "elasticsearch-master-0"
          securityContext:
            privileged: true
          #将持久卷映射为数据目录的父目录
          volumeMounts:
            - name: pv-storage-elastic-master
              mountPath: /usr/share/elasticsearch/data/
#      imagePullSecrets:
#        - name: aliyun-secret
      serviceAccountName: elasticsearch-admin
      tolerations:
        - key: node-role.kubernetes.io/master
          effect: NoSchedule

---
kind: Service
apiVersion: v1
metadata:
  labels:
    elastic-app: elasticsearch
  name: elasticsearch-discovery
  namespace: ns-elasticsearch
spec:
  ports:
    - port: 9300
      targetPort: 9300
  selector:
    elastic-app: elasticsearch
    role: master

---
kind: StatefulSet
apiVersion: apps/v1
metadata:
  labels:
    elastic-app: elasticsearch
    role: data
  name: elasticsearch-data
  namespace: ns-elasticsearch
spec:
  serviceName: es-data
  replicas: 2
  revisionHistoryLimit: 10
  selector:
    matchLabels:
      elastic-app: elasticsearch
  template:
    metadata:
      labels:
        elastic-app: elasticsearch
        role: data
    spec:
      #将es-data持久卷声明
      volumes:
        - name: pv-storage-elastic-data
          persistentVolumeClaim:
            claimName: pvc-es-data
      nodeSelector:
        es: enable
      containers:
        - name: elasticsearch-data
          image: registry.cn-shanghai.aliyuncs.com/luckyqing/elasticsearch:7.6.2
          lifecycle:
            postStart:
              exec:
                command: ["/bin/bash", "-c", "sysctl -w vm.max_map_count=262144; ulimit -l unlimited;chown -R elasticsearch:elasticsearch /usr/share/elasticsearch/data;"]
          ports:
            - containerPort: 9200
              protocol: TCP
            - containerPort: 9300
              protocol: TCP
          env:
            - name: MY_POD_NAME
              valueFrom:
                fieldRef:
                  apiVersion: v1
                  fieldPath: metadata.name
            #修改es默认的数据存储目录，否则多个节点同时写一个目录es权限不允许
            - name: "path.data"
              value: "/usr/share/elasticsearch/data/$(MY_POD_NAME)"
            - name: "cluster.name"
              value: "elasticsearch-cluster"
            - name: "bootstrap.memory_lock"
              value: "false"
            - name: "discovery.seed_hosts"
              value: "elasticsearch-discovery"
            - name: "node.master"
              value: "false"
            - name: "node.data"
              value: "true"
            - name: "ES_JAVA_OPTS"
              value: "-Xms1024m -Xmx1024m"
          securityContext:
            privileged: true
          #将持久卷映射到数据目录的父目录
          volumeMounts:
            - name: pv-storage-elastic-data
              mountPath: /usr/share/elasticsearch/data/
      # imagePullSecrets:
      # - name: aliyun-secret
      serviceAccountName: elasticsearch-admin
      tolerations:
        - key: node-role.kubernetes.io/master
          effect: NoSchedule

---
kind: Service
apiVersion: v1
metadata:
  labels:
    elastic-app: elasticsearch-service
  name: elasticsearch-service
  namespace: ns-elasticsearch
spec:
  ports:
    - port: 9200
      targetPort: 9200
      nodePort: 32000
  selector:
    elastic-app: elasticsearch
  type: NodePort

---
kind: Service
apiVersion: v1
metadata:
  labels:
    elastic-app: elasticsearch-service
  name: elasticsearch-api-service
  namespace: ns-elasticsearch
spec:
  ports:
    - port: 9300
      targetPort: 9300
      nodePort: 32100
  selector:
    elastic-app: elasticsearch
  type: NodePort