k8s上搭建elasticsearch单节点集群（固定存储）

因为Elasticsearch需要保持数据在固定的磁盘上，属于有状态的应用服务，因此采用k8s的StatefulSet模式部署。

创建PersistentVolumes和Claim

#es-persistent-volume.yaml
apiVersion: v1
kind: PersistentVolume
metadata:
  name: es-nfs-pv
  labels:
    pv: es-nfs-pv
spec:
  capacity:
    storage: 1Gi
  accessModes:
    - ReadWriteOnce
  persistentVolumeReclaimPolicy: Recycle
  storageClassName: nfs
  nfs:
    path: /data/nfs/es/
    server: 192.168.1.1 #要挂载卷的所在机器

---
apiVersion: v1
kind: PersistentVolumeClaim
metadata:
  name: es-nfs-pvc
spec:
  accessModes:
    - ReadWriteOnce
  resources:
    requests:
      storage: 1Gi
  storageClassName: nfs
  selector:
    matchLabels:
      pv: "es-nfs-pv"

deployment-es.yaml文件内容：

kind: StatefulSet
metadata:
  name: elasticsearch-sts
spec:
  selector:
    matchLabels:
      app: es # has to match .spec.template.metadata.labels
  serviceName: "elasticsearch-svc"  #声明它属于哪个Headless Service.
  replicas: 1 # by default is 1
  template:
    metadata:
      labels:
        app: es # has to match .spec.selector.matchLabels
    spec:
      terminationGracePeriodSeconds: 10
      volumes:
       - name: data-storage
         persistentVolumeClaim:
           claimName: es-nfs-pvc #和之前的PersistentVolumeClaim的name一致
      containers:
      - name: elasticsearch
        image: docker.elastic.co/elasticsearch/elasticsearch:7.9.2
        ports:
        - containerPort: 9200
          name: es-cli
        - containerPort: 9300
          name: es-iner
        env:
        - name: discovery.type
          value: single-node  
        - name: http.cors.enabled
          value: "true"
        - name: http.cors.allow-origin
          value: "*"
        volumeMounts:
         - mountPath: "/tmp/data"
           name: data-storage
---
apiVersion: v1
kind: Service
metadata:
  name: elasticsearch-svc
  labels:
    app: elasticsearch-svc
spec:
  type: NodePort
  ports:
  - port: 9200
    nodePort: 30920 #对外暴露的访问端口，需要在30000~32727范围内。
    name: clientport    
 selector: 
    app: es

master节点上执行:

kubectl apply -f es-persistent-volume.yaml

kubectl apply -f deployment-es.yaml

看看执行后的结果：

#查看部署的statefulset
$ kubectl get sts
NAME                READY   AGE
elasticsearch-sts   1/1     1m

#查看service
$ kubectl get svc
NAME                TYPE        CLUSTER-IP       EXTERNAL-IP   PORT(S)          AGE
elasticsearch-svc   NodePort    10.101.52.134            9200:30920/TCP   1m

#查看启动的pod
$ kubectl get pods
NAME                                READY   STATUS    RESTARTS   AGE
elasticsearch-sts-0                 1/1     Running   0          1m

这样部署好之后就可以通过masterIp+nodePort的方式访问Elasticsearch服务了.

• 插入数据:

$ curl -X PUT 127.0.0.1:30920/test/_doc/1?pretty -H 'Content-Type: application/json' -d '{"id":1, "msg":"test"}'
{
  "_index" : "test",
  "_type" : "_doc",
  "_id" : "1",
  "_version" : 4,
  "result" : "updated",
  "_shards" : {
    "total" : 2,
    "successful" : 1,
    "failed" : 0
  },
  "_seq_no" : 3,
  "_primary_term" : 1
}

• 查询：

$ curl -X GET 127.0.0.1:30920/test/_doc/1?pretty
{
  "_index" : "test",
  "_type" : "_doc",
  "_id" : "1",
  "_version" : 4,
  "_seq_no" : 3,
  "_primary_term" : 1,
  "found" : true,
  "_source" : {
    "id" : 1,
    "msg" : "test"
  }
}

---

解决挂载遇到的错误

1. nfs挂载错误“mount: wrong fs type, bad option, bad superblock on ...”，需要在挂载的服务机上启动nfs服务。
   安装nfs-utils：

   yum install nfs-utils

（各个node节点也要安装nfs-utils，不然也会mount失败）。

启动nfs服务：

service nfs start

2. “mount.nfs: access denied by server while mounting”的错误，要export对应的挂载点。
   在/etc/exports添加如下内容：

/home/nfs/data  *(insecure,rw,sync,no_root_squash)
#这里最好不要用async，不然可能会失败

3. 要有对挂载目录的读写权限

4.额外的，挂载命令示例： sudo mount -t nfs 192.168.1.1:/data/nfs /mnt/nfs