K8S使用dashboard管理集群
更新:HHH   时间:2023-1-7


K8S使用dashboard管理集群


今年3月份在公司的内部k8s培训会上,开发同事表示使用dashboard的可以满足日常开发需求,例如查看pod的日志,执行exec指令,查看pod的运行状态等,但对basic认证的权限控制表示担忧。
之前介绍过在1.5.2版本上部署dashboard服务,在1.9.1版本离线部署中,也介绍过dashboard服务的RBAC配置和使用技巧。因此本文将在前文基础上完善Heapster的整合与利用token对用户权限进行控制。
dashboard的特点主要如下:
1、能够直观的看到rc、deployment、pod、services等k8s组件的运行情况和日志信息。
2、结合heapster和influxdb后,dashboard的监控图表上可以看到pod的cpu和内存消耗情况。

Heapster介绍

1、Heapster是容器集群监控和性能分析工具,支持Kubernetes和CoreOS。 
2、K8S集群的HPA功能的实现就依赖于这些metric数据,HPA将Heapster作为Resource Metrics API,向其获取metric。
3、Kubernetes有个cAdvisor监控(在1.9版本里面,cAdvisor已经和kubelet整合在一起)。
在每个kubernetes Node上都会运行cAdvisor,它会收集本机以及容器的监控数据(cpu,memory,filesystem,network,uptime)。Heapster是一个收集者,Heapster可以收集Node节点上的cAdvisor数据,将每个Node上的cAdvisor的数据进行汇总,还可以按照kubernetes的资源类型来集合资源,比如Pod、Namespace,可以分别获取它们的CPU、内存、网络和磁盘的metric。默认的metric数据聚合时间间隔是1分钟。还可以把数据导入到第三方工具(如InfluxDB)。

Influxdb数据库介绍

2、Influxdb数据库的相关知识介绍,可参考文档:https://www.jianshu.com/p/d2935e99006e
2、如果对Heapster收集到的metric数据没有持久化的需求,可以不配置Influxdb数据库
3、本文Influxdb数据库的存储采用emptydir的方式实现,实际使用过程中,可以选择吧Influxdb数据库部署在k8s集群外部,或者使用其他存储方案。
4、如果有需要的话,还可以集成一个grafana做web展示。Grafana配置可参考文档:https://blog.51cto.com/ylw6006/2084403

一、获取相关镜像

需要科学上网方式获取到dashboard相关的镜像文件,仓库可纳入本地仓库统一管理

# cat /etc/systemd/system/docker.service.d/http-proxy.conf [Service]
Environment="HTTP_PROXY=http://192.168.115.2:1080"# systemctl  daemon-reload# systemctl restart docker# docker pull k8s.gcr.io/kubernetes-dashboard-amd64:v1.8.3# docker pull k8s.gcr.io/heapster-influxdb-amd64:v1.3.3# docker pull k8s.gcr.io/heapster-amd64:v1.4.2

二、准备配置文件

1、k8s-dashborad-sa.yaml文件,secrct和serviceaccount配置

# cat k8s-dashborad-sa.yaml # ------------------- Dashboard Secret ------------------- #apiVersion: v1kind: Secretmetadata:
  labels:
    k8s-app: kubernetes-dashboard
  name: kubernetes-dashboard-certs
  namespace: kube-systemtype: Opaque---# ------------------- Dashboard Service Account ------------------- #apiVersion: v1kind: ServiceAccountmetadata:
  labels:
    k8s-app: kubernetes-dashboard
  name: kubernetes-dashboard
  namespace: kube-system

2、k8s-dashborad-rbac.yaml文件,配置 Role和Role Binding

# cat k8s-dashborad-rbac.yaml # ------------------- Dashboard Role & Role Binding ------------------- #kind: RoleapiVersion: rbac.authorization.k8s.io/v1metadata:
  name: kubernetes-dashboard-minimal
  namespace: kube-systemrules:
  # Allow Dashboard to create 'kubernetes-dashboard-key-holder' secret.- apiGroups: [""]
  resources: ["secrets"]
  verbs: ["create"]  # Allow Dashboard to create 'kubernetes-dashboard-settings' config map.- apiGroups: [""]
  resources: ["configmaps"]
  verbs: ["create"]  # Allow Dashboard to get, update and delete Dashboard exclusive secrets.- apiGroups: [""]
  resources: ["secrets"]
  resourceNames: ["kubernetes-dashboard-key-holder", "kubernetes-dashboard-certs"]
  verbs: ["get", "update", "delete"]  # Allow Dashboard to get and update 'kubernetes-dashboard-settings' config map.- apiGroups: [""]
  resources: ["configmaps"]
  resourceNames: ["kubernetes-dashboard-settings"]
  verbs: ["get", "update"]  # Allow Dashboard to get metrics from heapster.- apiGroups: [""]
  resources: ["services"]
  resourceNames: ["heapster"]
  verbs: ["proxy"]
- apiGroups: [""]
  resources: ["services/proxy"]
  resourceNames: ["heapster", "http:heapster:", "https:heapster:"]
  verbs: ["get"]

---apiVersion: rbac.authorization.k8s.io/v1kind: RoleBindingmetadata:
  name: kubernetes-dashboard-minimal
  namespace: kube-systemroleRef:
  apiGroup: rbac.authorization.k8s.io
  kind: Role
  name: kubernetes-dashboard-minimalsubjects:- kind: ServiceAccount
  name: kubernetes-dashboard
  namespace: kube-system

3、k8s-dashborad-deployment.yaml配置文件,定义创建pod的模板和副本数

# cat k8s-dashborad-deployment.yaml # ------------------- Dashboard Deployment ------------------- #kind: DeploymentapiVersion: apps/v1beta2metadata:
  labels:
    k8s-app: kubernetes-dashboard
  name: kubernetes-dashboard
  namespace: kube-systemspec:
  replicas: 1
  revisionHistoryLimit: 10
  selector:
    matchLabels:
      k8s-app: kubernetes-dashboard
  template:
    metadata:
      labels:
        k8s-app: kubernetes-dashboard
    spec:
      containers:
      - name: kubernetes-dashboard
        image: k8s.gcr.io/kubernetes-dashboard-amd64:v1.8.3
        ports:
        - containerPort: 8443
          protocol: TCP
        args:
          - --auto-generate-certificates          # Uncomment the following line to manually specify Kubernetes API server Host
          # If not specified, Dashboard will attempt to auto discover the API server and connect
          # to it. Uncomment only if the default does not work.
          # - --apiserver-host=http://my-address:port
        volumeMounts:
        - name: kubernetes-dashboard-certs
          mountPath: /certs          # Create on-disk volume to store exec logs
        - mountPath: /tmp
          name: tmp-volume
        livenessProbe:
          httpGet:
            scheme: HTTPS
            path: /
            port: 8443
          initialDelaySeconds: 30
          timeoutSeconds: 30
      volumes:
      - name: kubernetes-dashboard-certs
        secret:
          secretName: kubernetes-dashboard-certs
      - name: tmp-volume
        emptyDir: {}
      serviceAccountName: kubernetes-dashboard      # Comment the following tolerations if Dashboard must not be deployed on master
      tolerations:
      - key: node-role.kubernetes.io/master
        effect: NoSchedule

4、 k8s-dashborad-service.yaml配置文件,定义service

# cat k8s-dashborad-service.yaml   # ------------------- Dashboard Service ------------------- #kind: ServiceapiVersion: v1metadata:
  labels:
    k8s-app: kubernetes-dashboard
  name: kubernetes-dashboard
  namespace: kube-systemspec:
  ports:
    - port: 443
      targetPort: 8443
      nodePort: 8490
  type: NodePort
  selector:
    k8s-app: kubernetes-dashboard

三、通过配置文件创建dashboard

# kubectl create -f .# kubectl get pod,deployment,svc -n kube-system

四、配置使用basic认证方式

默认情况下只支持kubeconfig和令牌认证

# echo 'admin,admin,1' > /etc/kubernetes/basic_auth_file # grep 'auth' /usr/lib/systemd/system/kube-apiserver.service   
  --authorization-mode=Node,RBAC \
  --runtime-config=rbac.authorization.k8s.io/v1alpha1 \
  --enable-bootstrap-token-auth=true \
  --token-auth-file=/etc/kubernetes/token.csv \
  --basic-auth-file=/etc/kubernetes/basic_auth_file \# grep  ‘basic’  k8s-dashborad-deployment.yaml   (配置在args下面)
     - --authentication-mode=basic# systemctl daemon-reload# systemctl restart kube-apiserver # kubectl apply -f k8s-dashborad-deployment.yaml

将admin用户和cluter-admin role进行角色绑定

# curl --insecure https://vm1:6443 -basic -u admin:admin  # kubectl create clusterrolebinding  \login-on-dashboard-with-cluster-admin  \
--clusterrole=cluster-admin --user=admin# curl --insecure https://vm1:6443 -basic -u admin:admin

五、访问测试


六、整合heapster和influxdb

在没有配置heapster和influxdb的情况下,pod的metric信息是无法获取到的,而早前版本K8S的HPA特性依赖的metric数据来源恰巧就是heapster和influxdb。

1、准备yaml配置文件

# cat heapster-sa.yaml apiVersion: v1kind: ServiceAccountmetadata:
  name: heapster
  namespace: kube-system
# cat heapster-rbac.yaml kind: ClusterRoleBindingapiVersion: rbac.authorization.k8s.io/v1beta1metadata:
  name: heapsterroleRef:
  apiGroup: rbac.authorization.k8s.io
  kind: ClusterRole
  name: system:heapstersubjects:- kind: ServiceAccount
  name: heapster
  namespace: kube-system
# cat heapster-deployment.yaml apiVersion: extensions/v1beta1
kind: Deployment
metadata:
  name: heapster  namespace: kube-system
spec:
  replicas: 1
  template:
    metadata:
      labels:
        task: monitoring
        k8s-app: heapster
    spec:
      serviceAccountName: heapster
      containers:
      - name: heapster
        image: k8s.gcr.io/heapster-amd64:v1.4.2
        imagePullPolicy: IfNotPresent
        command:
        - /heapster
        - --source=kubernetes:https://kubernetes.default
        - --sink=influxdb:http://monitoring-influxdb.kube-system.svc:8086
# cat heapster-service.yaml   apiVersion: v1kind: Servicemetadata:
  labels:
    task: monitoring
    kubernetes.io/cluster-service: 'true'
    kubernetes.io/name: Heapster
  name: heapster
  namespace: kube-systemspec:
  ports:
  - port: 80
    targetPort: 8082
  selector:k8s-app: heapster
# cat influxdb-deployment.yaml apiVersion: extensions/v1beta1kind: Deploymentmetadata:
  name: monitoring-influxdb
  namespace: kube-systemspec:
  replicas: 1
  template:
    metadata:
      labels:
        task: monitoring
        k8s-app: influxdb
    spec:
      containers:
      - name: influxdb
        image: k8s.gcr.io/heapster-influxdb-amd64:v1.3.3
        volumeMounts:
        - mountPath: /data
          name: influxdb-storage
      volumes:
      - name: influxdb-storage
        emptyDir: {}
# cat influxdb-service.yaml   apiVersion: v1kind: Servicemetadata:
  labels:
    task: monitoring
    kubernetes.io/cluster-service: 'true'
    kubernetes.io/name: monitoring-influxdb
  name: monitoring-influxdb
  namespace: kube-systemspec:
  ports:
  - port: 8086
    targetPort: 8086
  selector:
    k8s-app: influxdb



获取heapster中的获取支持的metrics

# kubectl run -i --tty curl --namespace=kube-system  \--image=registry.59iedu.com/webwurst/curl-utils /bin/sh 
# curl http://heapster/api/v1/model/metrics# curl http://heapster/api/v1/model/debug/allkeys

# kubectl get node # kubectl top node


当heapster和influxdb pod都正常运行的时候,在dashboard里面就可以看到CPU和内存的监控数据了。

七、配置用户权限

1、删除apiserver里面basic认证相关的配置后重启apiserver
--basic-auth-file=/etc/kubernetes/basic_auth_file

# systemctl daemon-reload# systemctl  restart kube-apiserver

2、删除clusterrolebinding

# kubectl delete  clusterrolebinding  login-on-dashboard-with-cluster-admin

3、修改k8s-dashborad-deployment.yaml文件
去掉- --authentication-mode=basic参数

4、创建普通用户,赋予所有namespace下资源的get、watch和list权限。
这里通过clusterrole和culsterrolebinding赋予所有namespace相关资源的get、watch、list权限,实际应用环境建议使用创建role和rolebinding指定特定的namespace相关资源权限,各资源权限的赋予规则遵循最小权限原则。

# cat rbac-yang.yaml kind: ClusterRoleapiVersion: rbac.authorization.k8s.io/v1metadata:
  name: role-yangrules:- apiGroups: [""]
  resources: ["*"]
  verbs: ["get","watch","list" ]
- apiGroups: ["storage.k8s.io"]
  resources: ["*"]
  verbs: ["get","watch","list" ]
- apiGroups: ["rbac.authorization.k8s.io"]
  resources: ["*"]
  verbs: ["get","watch","list" ]
- apiGroups: ["batch"]
  resources: ["*"]
  verbs: ["get","watch","list" ]
- apiGroups: ["apps"]
  resources: ["*"]
  verbs: ["get","watch","list" ]
- apiGroups: ["extensions"]
  resources: ["*"]
  verbs: ["get","watch","list" ]
---kind: ClusterRoleBindingapiVersion: rbac.authorization.k8s.io/v1metadata:
  name: role-bind-yangsubjects:- kind: ServiceAccount
  name: yang
  namespace: kube-systemroleRef:
  kind: ClusterRole
  name: role-yang
  apiGroup: rbac.authorization.k8s.io
# kubectl create sa yang -n kube-system# kubectl create -f rbac-yang.yaml # kubectl -n kube-system describe secret $(kubectl -n kube-system get secret | grep yang | awk '{print $1}')

5、测试普通用户的权限




6、创建super用户admin

# kubectl create sa admin -n kube-system# cat rbac-admin.yaml apiVersion: rbac.authorization.k8s.io/v1beta1
kind: ClusterRoleBinding
metadata:
  name: admin
roleRef:
  apiGroup: rbac.authorization.k8s.io
  kind: ClusterRole
  name: cluster-admin
subjects:
- kind: ServiceAccount
  name: admin
  namespace: kube-system# kubectl create -f rbac-admin.yaml # kubectl -n kube-system describe secret $(kubectl -n kube-system get secret | grep admin | awk '{print $1}')


使用admin用户的token登陆后继承cluster-admin的权限

参考:
https://github.com/kubernetes/dashboard/wiki/Creating-sample-user
https://github.com/kubernetes/dashboard/wiki/Access-control
https://github.com/kubernetes/heapster/blob/master/docs/model.md


返回云计算教程...