helm 类似于Linux系统下的包管理器，如yum/apt等，可以方便快捷的将之前打包好的yaml文件快速部署进kubernetes内，方便管理维护。

安装

  1. yum install -y ntpdate
  2. ntpdate ntp.aliyun.com
  5. cd /opt 
  6. wget https://get.helm.sh/helm-v3.7.1-linux-amd64.tar.gz
  7. tar -zxvf helm-v3.7.1-linux-amd64.tar.gz
  8. cp linux-amd64/helm /usr/local/bin
  10. # 查看是否安装完成
  11. helm version
  13. # 安装repo
  14. helm repo add apphub https://apphub.aliyuncs.com
  16. # 查找安装包
  17. helm search repo redis
  19. # 安装应用 比如redis
  20. helm install redis apphub/redis
  22. #查看状态
  23. helm status redis
  25. helm get values redis

一键安装 Prometheus 和 Grafana

  1. # 添加仓库
  2. helm repo add prometheus-community https://prometheus-community.github.io/helm-charts
  4. # 列出有哪些仓库
  5. helm repo list
  7. # 更新仓库
  8. helm repo update
  11. kubectl create ns monitor
  12. helm install prometheus-stack prometheus-community/kube-prometheus-stack -n monitor
  14. # 查看安装状态
  15. kubectl --namespace monitor get pods -l "release=prometheus-stack"
  18. #查看命名空间的组件
  19. kubectl get all -n monitor
  21. [root@control-plane daemonset]# kubectl get all -n monitor
  22. NAME                                                       READY   STATUS              RESTARTS   AGE
  23. pod/prometheus-stack-grafana-7d5754bbdd-pn25l              0/2     PodInitializing     0          113s
  24. pod/prometheus-stack-kube-prom-operator-7d9cf4f57c-z9t8q   0/1     ContainerCreating   0          113s
  25. pod/prometheus-stack-kube-state-metrics-8476b7f76f-5hts2   0/1     ContainerCreating   0          113s
  26. pod/prometheus-stack-prometheus-node-exporter-7c7xx        1/1     Running             0          114s
  28. NAME                                                TYPE        CLUSTER-IP       EXTERNAL-IP   PORT(S)    AGE
  29. service/prometheus-stack-grafana                    ClusterIP   10.102.99.8      <none>        80/TCP     114s
  30. service/prometheus-stack-kube-prom-alertmanager     ClusterIP   10.102.38.65     <none>        9093/TCP   114s
  31. service/prometheus-stack-kube-prom-operator         ClusterIP   10.101.143.49    <none>        443/TCP    114s
  32. service/prometheus-stack-kube-prom-prometheus       ClusterIP   10.104.232.98    <none>        9090/TCP   114s
  33. service/prometheus-stack-kube-state-metrics         ClusterIP   10.111.141.9     <none>        8080/TCP   114s
  34. service/prometheus-stack-prometheus-node-exporter   ClusterIP   10.103.135.203   <none>        9100/TCP   114s
  36. NAME                                                       DESIRED   CURRENT   READY   UP-TO-DATE   AVAILABLE   NODE SELECTOR   AGE
  37. daemonset.apps/prometheus-stack-prometheus-node-exporter   1         1         1       1            1           <none>          114s
  39. NAME                                                  READY   UP-TO-DATE   AVAILABLE   AGE
  40. deployment.apps/prometheus-stack-grafana              0/1     1            0           114s
  41. deployment.apps/prometheus-stack-kube-prom-operator   0/1     1            0           114s
  42. deployment.apps/prometheus-stack-kube-state-metrics   0/1     1            0           114s
  44. NAME                                                             DESIRED   CURRENT   READY   AGE
  45. replicaset.apps/prometheus-stack-grafana-7d5754bbdd              1         1         0       113s
  46. replicaset.apps/prometheus-stack-kube-prom-operator-7d9cf4f57c   1         1         0       113s
  47. replicaset.apps/prometheus-stack-kube-state-metrics-8476b7f76f   1         1         0       113s
  49. # 转发端口  本地来访问 Prometheus 
  50. kubectl port-forward -n monitor prometheus-prometheus-stack-kube-prom-prometheus-0 9090
  52. # 转发端口  本地来访问 Grafana  默认用户名是 admin，密码是 prom-operator
  53. kubectl port-forward -n monitor prometheus-stack-grafana-5b6dd6b5fb-rtp6z 3000
  55. 破译密码过程
  56. kubectl get secret prometheus-stack-grafana -n monitor -o jsonpath='{.data}'
  58. kubectl get secret prometheus-stack-grafana -n monitor -o jsonpath='{.data.admin-user}' | base64 --decode
  59. kubectl get secret prometheus-stack-grafana -n monitor -o jsonpath='{.data.admin-password}' | base64 --decode
  61. # 转发端口  本地来访问  kube-state-metrics 
  62. kubectl port-forward -n monitor prometheus-stack-kube-state-metrics-c7c69c8c9-bhgjv 8080
  65. 提示错误：
  66. E1124 03:46:48.075668    9302 portforward.go:400] an error occurred forwarding 3000 -> 3000: error forwarding port 3000 to pod d1a7b62b484532c715cc932a4935645f01085cc2f5f229561aac7fa883bb430d, uid : unable to do port forwarding: socat not found
  68. yum install -y socat

HPV弹性扩容

  1. wget https://github.com/kubernetes-sigs/metrics-server/releases/download/v0.5.2/components.yaml
  3. vi components.yaml
  5. 在此下面增加两个配置
  6.     - --metric-resolution=15s
  8.     - --kubelet-insecure-tls=true
  9.     - --kubelet-preferred-address-types=InternalIP
  11. kubectl create -f components.yaml
  14. vi nginx-deploy-hpa.yaml
  16. apiVersion: apps/v1
  17. kind: Deployment
  18. metadata:
  19.   name: nginx-deployment
  20.   namespace: demo
  21. spec:
  22.   selector:
  23.     matchLabels:
  24.       app: nginx
  25.       usage: hpa
  26.   replicas: 1
  27.   template:
  28.     metadata:
  29.       labels:
  30.         app: nginx
  31.         usage: hpa
  32.     spec:
  33.       containers:
  34.       - name: nginx
  35.         image: nginx:1.19.2
  36.         ports:
  37.         - containerPort: 80
  38.         resources:
  39.           requests:
  40.             memory: "64Mi"
  41.             cpu: "250m"
  42.           limits:
  43.             cpu: "500m"
  44.             memory: "128Mi"
  46. # 创建
  47. kubectl create -f nginx-deploy-hpa.yaml
  49. kubectl get deploy -n demo
  50. kubectl get pod -n demo
  52. # 导出service
  53. kubectl expose deployment/nginx-deployment -n demo
  55. kubectl get svc -n demo
  57. kubectl get endpoints -n demo
  59. #创建HPA  cpu大于50%得进行弹性扩容
  60. #通过kubectl autoscale的命令，我们就可以将 Deployment 的副本数控制在 1~10 之间，CPU 利用率保持在 50% 以下，
  61. 即当该 Deployment 所关联的 Pod 的平均 CPU 利用率超过 50% 时，就增加副本数，直到小于该阈值。
  62. 当平均 CPU 利用率低于 50% 时，就减少副本数：
  63. kubectl autoscale deploy nginx-deployment -n demo --cpu-percent=50 --min=1 --max=10
  65. # 查看hpa
  66. kubectl get hpa -n demo
  69. # 测试
  70. kubectl run demo-benchmark --image httpd:2.4.46-alpine -n demo -it sh
  71. ab -n 50000 -c 500 -s 60 http://10.109.93.199/   ---->  kubectl get svc -n demo 拿到服务IP
  73. # 查看hpa扩容详情 
  74. kubectl get hpa -n demo -w
  75. kubectl get deploy -n demo -w
  77. # 查看节点监控信息
  78. kubectl get podmetrics -n demo
  79. kubectl get podmetrics -n demo nginx-deployment-6d4b885966-zngnd -o yaml

常用仓库

  1. helm repo add stable https://kubernetes.oss-cn-hangzhou.aliyuncs.com/charts
  2. helm repo add aliyun https://apphub.aliyuncs.com/
  3. helm repo add bitnami https://charts.bitnami.com/bitnami/
  4. helm repo add azure https://mirror.azure.cn/kubernetes/charts/

资料