环境准备

我们使用的k8s和ceph环境见：
https://blog.51cto.com/leejia/2495558
https://blog.51cto.com/leejia/2499684

ECK简介

Elastic Cloud on Kubernetes，这是一款基于 Kubernetes Operator 模式的新型编排产品，用户可使用该产品在 Kubernetes 上配置、管理和运行 Elasticsearch 集群。ECK 的愿景是为 Kubernetes 上的 Elastic 产品和解决方案提供 SaaS 般的体验。

ECK使用 Kubernetes Operator模式构建而成，需要安装在Kubernetes集群内，ECK用于部署，且更专注于简化所有后期运行工作：

  • 管理和监测多个集群
  • 轻松升级至新版本
  • 扩大或缩小集群容量
  • 更改集群配置
  • 动态调整本地存储的规模
  • 备份

Kubernetes目前是容器编排领域的领头羊，而Elastic社区发布ECK，使Elasticsearch更容易的跑在云上，也是为云原生技术增砖添瓦，紧跟时代潮流。

部署ECK

部署ECK并查看日志是否正常：

# kubectl apply -f https://download.elastic.co/downloads/eck/1.1.2/all-in-one.yaml

# kubectl -n elastic-system logs -f statefulset.apps/elastic-operator

过几分钟查看elastic-operator是否运行正常，ECK中只有一个elastic-operator pod：

# kubectl get pods -n elastic-system
NAME                 READY   STATUS    RESTARTS   AGE
elastic-operator-0   1/1     Running   1          2m55s

使用ECK部署使用ceph持久化存储的elasticsearch集群

我们测试情况使用1台master节点和1台data节点来部署集群，生产环境建议使用3+台master节点。如下的manifest中，对实例的heap大小，容器的可使用内存，容器的虚拟机内存都进行了配置，可以根据集群需要做调整：

# vim es.yaml
apiVersion: elasticsearch.k8s.elastic.co/v1
kind: Elasticsearch
metadata:
  name: quickstart
spec:
  version: 7.7.1
  nodeSets:
  - name: master-nodes
    count: 1
    config:
      node.master: true
      node.data: false
    podTemplate:
      spec:
        initContainers:
        - name: sysctl
          securityContext:
            privileged: true
          command: ['sh', '-c', 'sysctl -w vm.max_map_count=262144']
        containers:
        - name: elasticsearch
          env:
          - name: ES_JAVA_OPTS
            value: -Xms1g -Xmx1g
          resources:
            requests:
              memory: 2Gi
            limits:
              memory: 2Gi
    volumeClaimTemplates:
    - metadata:
        name: elasticsearch-data
      spec:
        accessModes:
        - ReadWriteOnce
        resources:
          requests:
            storage: 5Gi
        storageClassName: rook-ceph-block
  - name: data-nodes
    count: 1
    config:
      node.master: false
      node.data: true
    podTemplate:
      spec:
        initContainers:
        - name: sysctl
          securityContext:
            privileged: true
          command: ['sh', '-c', 'sysctl -w vm.max_map_count=262144']
        containers:
        - name: elasticsearch
          env:
          - name: ES_JAVA_OPTS
            value: -Xms1g -Xmx1g
          resources:
            requests:
              memory: 2Gi
            limits:
              memory: 2Gi
    volumeClaimTemplates:
    - metadata:
        name: elasticsearch-data
      spec:
        accessModes:
        - ReadWriteOnce
        resources:
          requests:
            storage: 10Gi
        storageClassName: rook-ceph-block

# kubectl apply -f es.yaml

过段时间，查看elasticsearch集群的状态

# kubectl get pods
quickstart-es-data-nodes-0     1/1     Running   0          54s
quickstart-es-master-nodes-0   1/1     Running   0          54s

# kubectl get elasticsearch
NAME         HEALTH   NODES   VERSION   PHASE   AGE
quickstart   green    2       7.7.1     Ready   73s

查看pv的状态，我们可以看到申请的pv已经创建和绑定成功：

# kubectl get pv
pvc-512cc739-3654-41f4-8339-49a44a093ecf   10Gi       RWO            Retain           Bound      default/elasticsearch-data-quickstart-es-data-nodes-0     rook-ceph-block            9m5s
pvc-eff8e0fd-f669-448a-8b9f-05b2d7e06220   5Gi        RWO            Retain           Bound      default/elasticsearch-data-quickstart-es-master-nodes-0   rook-ceph-block            9m5s

默认集群开启了basic认证，用户名为elastic，密码可以通过secret获取。默认集群也开启了自签名证书https访问。我们可以通过service资源来访问elasticsearch：

# kubectl get services
NAME                         TYPE        CLUSTER-IP       EXTERNAL-IP   PORT(S)    AGE
quickstart-es-data-nodes     ClusterIP   None                          4m10s
quickstart-es-http           ClusterIP   10.107.201.126           9200/TCP   4m11s
quickstart-es-master-nodes   ClusterIP   None                          4m10s
quickstart-es-transport      ClusterIP   None                     9300/TCP   4m11s

# kubectl get secret quickstart-es-elastic-user -o=jsonpath='{.data.elastic}' | base64 --decode; echo

# curl https://10.107.201.126:9200 -u 'elastic:J1fO9bu88j8pYK8rIu91a73o' -k
{
  "name" : "quickstart-es-data-nodes-0",
  "cluster_name" : "quickstart",
  "cluster_uuid" : "AQxFX8NiTNa40mOPapzNXQ",
  "version" : {
    "number" : "7.7.1",
    "build_flavor" : "default",
    "build_type" : "docker",
    "build_hash" : "ad56dce891c901a492bb1ee393f12dfff473a423",
    "build_date" : "2020-05-28T16:30:01.040088Z",
    "build_snapshot" : false,
    "lucene_version" : "8.5.1",
    "minimum_wire_compatibility_version" : "6.8.0",
    "minimum_index_compatibility_version" : "6.0.0-beta1"
  },
  "tagline" : "You Know, for Search"
}

不停服，扩容一台data节点：修改es.yaml中data-nodes中count的value为2，然后apply下es.yaml即可。

# kubectl apply -f es.yaml
# kubectl get pods
quickstart-es-data-nodes-0     1/1     Running   0          24m
quickstart-es-data-nodes-1     1/1     Running   0          8m22s
quickstart-es-master-nodes-0   1/1     Running   0          24m

# kubectl get elasticsearch
NAME         HEALTH   NODES   VERSION   PHASE   AGE
quickstart   green    3       7.7.1     Ready   25m

不停服，缩容一台data节点，会自动进行数据同步：修改es.yaml中data-nodes中count的value为1，然后apply下es.yaml即可。

对接kibana

由于默认kibana也开启了自签名证书的https访问，我们可以选择关闭，我们来使用ECK部署kibana:

# vim kibana.yaml
apiVersion: kibana.k8s.elastic.co/v1
kind: Kibana
metadata:
  name: quickstart
spec:
  version: 7.7.1
  count: 1
  elasticsearchRef:
    name: quickstart
  http:
    tls:
      selfSignedCertificate:
        disabled: true
# kubectl apply -f kibana.yaml

# kubectl get pods
NAME                             READY   STATUS    RESTARTS   AGE
quickstart-es-data-nodes-0       1/1     Running   0          31m
quickstart-es-data-nodes-1       1/1     Running   1          15m
quickstart-es-master-nodes-0     1/1     Running   0          31m
quickstart-kb-6558457759-2rd7l   1/1     Running   1          4m3s

# kubectl get kibana
NAME         HEALTH   NODES   VERSION   AGE
quickstart   green    1       7.7.1     4m27s

为kibana在ingress中添加一个四层代理，提供对外访问服务:

# vim tsp-kibana.yaml
apiVersion: k8s.nginx.org/v1alpha1
kind: GlobalConfiguration
metadata:
  name: nginx-configuration
  namespace: nginx-ingress
spec:
  listeners:
  - name: kibana-tcp
    port: 5601
    protocol: TCP

---

apiVersion: k8s.nginx.org/v1alpha1
kind: TransportServer
metadata:
  name: kibana-tcp
spec:
  listener:
    name: kibana-tcp
    protocol: TCP
  upstreams:
  - name: kibana-app
    service: quickstart-kb-http
    port: 5601
  action:
    pass: kibana-app

# kubectl apply -f tsp-kibana.yaml

默认kibana访问elasticsearch的用户名为elastic，密码获取方式如下

# kubectl get secret quickstart-es-elastic-user -o=jsonpath='{.data.elastic}' | base64 --decode; echo

通过浏览器访问kibana：
k8s部署elasticsearch集群_第1张图片

删除ECK相关资源

删除elasticsearch和kibana以及ECK

# kubectl get namespaces --no-headers -o custom-columns=:metadata.name \
  | xargs -n1 kubectl delete elastic --all -n

# kubectl delete -f https://download.elastic.co/downloads/eck/1.1.2/all-in-one.yaml

对接cerebro

先安装Kubernetes应用的包管理工具helm。Helm是用来封装 Kubernetes原生应用程序的YAML文件，可以在你部署应用的时候自定义应用程序的一些metadata，helm依赖chart实现了应用程序的在k8s上的分发。helm和chart主要实现了如下功能：

  • 应用程序封装
  • 版本管理
  • 依赖检查
  • 应用程序分发 
    # wget https://get.helm.sh/helm-v3.2.3-linux-amd64.tar.gz
# tar -zxvf helm-v3.0.0-linux-amd64.tar.gz
# mv linux-amd64/helm /usr/local/bin/helm
# helm repo add stable https://kubernetes-charts.storage.googleapis.com

    通过helm安装cerebro:

    # helm install stable/cerebro --version 1.1.4  --generate-name

    查看cerebro的状态:

    # kubectl get pods|grep cerebro
cerebro-1591777586-7fd87f7d48-hmlp7   1/1     Running   0          11m

    由于默认ECK部署的elasticsearch开启了自签名证书的https服务，故可以在cerebro配置忽略https证书认证(也可以在cerebro中添加自签名证书的ca证书来识别自签名证书)，并重启cerebro:
    1，导出cerebro的configmap：

    # kubectl get configmap cerebro-1591777586 -o yaml > cerebro.yaml

    2，替换configmap中cerebro的hosts相关配置为如下(其中quickstart-es-http为elasticsarch的service资源名字)：

    play.ws.ssl.loose.acceptAnyCertificate = true
hosts = [
  {
    host = "https://quickstart-es-http.default.svc:9200"
    name = "k8s elasticsearch"
  }
]

    3，应用cerebro的configmap并重启cerebro pod：

# kubectl apply -f cerebro.yaml

# kubectl get pods|grep cerebro
cerebro-1591777586-7fd87f7d48-hmlp7   1/1     Running   0          11m
# kubectl get pod cerebro-1591777586-7fd87f7d48-hmlp7 -o yaml | kubectl replace --force -f -

先确认cerebro的service资源，然后配置ingress为cerebro添加7层代理：

# kubectl get services
NAME                         TYPE        CLUSTER-IP       EXTERNAL-IP   PORT(S)    AGE
cerebro-1591777586           ClusterIP   10.111.107.171           80/TCP     19m

# vim cerebro-ingress.yaml
apiVersion: networking.k8s.io/v1beta1
kind: Ingress
metadata:
  name: cerebro-ingress
spec:
  rules:
  - host: cerebro.myk8s.com
    http:
      paths:
      - path: /
        backend:
          serviceName: cerebro-1591777586
          servicePort: 80

# kubectl apply -f cerebro-ingress.yaml

在本地pc的/etc/hosts文件添加host绑定"172.18.2.175 cerebro.myk8s.com"，然后通过览器访问：
k8s部署elasticsearch集群_第2张图片

删除cerebro

# helm list
NAME                NAMESPACE   REVISION    UPDATED                                 STATUS      CHART           APP VERSION
cerebro-1591777586  default     1           2020-06-10 16:26:30.419723417 +0800 CST deployed    cerebro-1.1.4   0.8.4

# heml delete name cerebro-1591777586

参考

https://www.elastic.co/guide/en/cloud-on-k8s/current/k8s-deploy-kibana.html
https://www.elastic.co/guide/en/cloud-on-k8s/current/k8s-kibana-http-configuration.html
https://www.elastic.co/guide/en/cloud-on-k8s/current/k8s-quickstart.html
https://hub.helm.sh/charts/stable/cerebro
https://www.elastic.co/cn/blog/introducing-elastic-cloud-on-kubernetes-the-elasticsearch-operator-and-beyond
https://helm.sh/docs/intro/install/