diff --git a/install/kubernetes/uninstall.sh b/install/kubernetes/uninstall.sh new file mode 100644 index 00000000..fa6ff4b7 --- /dev/null +++ b/install/kubernetes/uninstall.sh @@ -0,0 +1,170 @@ + +#!/bin/bash + +read -p "是否继续执行脚本？(yes/no)" answer +if [[ "$answer" == "yes" ]]; then + # mkdir -p ~/.kube && cp config ~/.kube/config && cp ~/.kube/config /etc/kubernetes/admin.conf + # mkdir -p kubeconfig && echo "" > kubeconfig/dev-kubeconfig + # curl -LO https://dl.k8s.io/release/v1.24.0/bin/linux/amd64/kubectl && chmod +x kubectl && cp kubectl /usr/bin/ && mv kubectl /usr/local/bin/ + node=`kubectl get node -o wide |grep $1 |awk '{print $1}'| head -n 1` + # kubectl label node $node train=true cpu=true notebook=true service=true org=public istio=true kubeflow=true kubeflow-dashboard=true mysql=true redis=true monitoring=true logging=true --overwrite + + # 创建命名空间 + # sh create_ns_secret.sh + + # 部署dashboard + kubectl delete -f dashboard/v2.2.0-cluster.yaml + # 高版本k8s部署2.6.1版本 + kubectl delete -f dashboard/v2.6.1-cluster.yaml + # 部署mysql + kubectl delete -f mysql/pv-pvc-hostpath.yaml + kubectl delete -f mysql/service.yaml + kubectl delete -f mysql/configmap-mysql.yaml + kubectl delete -f mysql/deploy.yaml + # 部署redis + kubectl delete -f redis/pv-hostpath.yaml + kubectl delete -f redis/configmap.yaml + kubectl delete -f redis/service.yaml + # 如果自己需要使用pv来保存redis队列数据，可以修改master.yaml + kubectl delete -f redis/master.yaml + + # 部署prometheus + cd prometheus + kubectl delete -f ./operator/operator-crd.yml + sleep 5 + kubectl delete -f ./operator/operator-crd.yml + kubectl delete -f ./operator/operator-rbac.yml + # kubectl wait crd/podmonitors.monitoring.coreos.com --for condition=established --timeout=60s + kubectl delete -f ./operator/operator-dp.yml + kubectl delete -f ./alertmanater/alertmanager-main-sa.yml + kubectl delete -f ./alertmanater/alertmanager-main-secret.yml + kubectl delete -f ./alertmanater/alertmanager-main-svc.yml + kubectl delete -f ./alertmanater/alertmanager-main.yml + kubectl delete -f ./node-exporter/node-exporter-sa.yml + kubectl delete -f ./node-exporter/node-exporter-rbac.yml + kubectl delete -f ./node-exporter/node-exporter-svc.yml + kubectl delete -f ./node-exporter/node-exporter-ds.yml + kubectl delete -f ./kube-state-metrics/kube-state-metrics-sa.yml + kubectl delete -f ./kube-state-metrics/kube-state-metrics-rbac.yml + kubectl delete -f ./kube-state-metrics/kube-state-metrics-svc.yml + kubectl delete -f ./kube-state-metrics/kube-state-metrics-dp.yml + kubectl delete -f ./grafana/pv-pvc-hostpath.yml + kubectl delete -f ./grafana/grafana-sa.yml + kubectl delete -f ./grafana/grafana-source.yml + kubectl delete -f ./grafana/grafana-datasources.yml + kubectl delete -f ./grafana/grafana-admin-secret.yml + kubectl delete -f ./grafana/grafana-svc.yml + kubectl delete configmap grafana-config all-grafana-dashboards --namespace=monitoring + # kubectl create configmap grafana-config --from-file=./grafana/grafana.ini --namespace=monitoring + # kubectl create configmap all-grafana-dashboards --from-file=./grafana/dashboard --namespace=monitoring + kubectl delete -f ./grafana/grafana-dp.yml + sleep 5 + kubectl delete -f ./grafana/grafana-dp.yml + kubectl delete -f ./service-discovery/kube-controller-manager-svc.yml + kubectl delete -f ./service-discovery/kube-scheduler-svc.yml + kubectl delete -f ./prometheus/prometheus-secret.yml + kubectl delete -f ./prometheus/prometheus-rules.yml + kubectl delete -f ./prometheus/prometheus-rbac.yml + kubectl delete -f ./prometheus/prometheus-svc.yml + # kubectl wait crd/prometheuses.monitoring.coreos.com --for condition=established --timeout=60s + kubectl delete -f ./prometheus/prometheus-main.yml + sleep 5 + kubectl delete -f ./prometheus/pv-pvc-hostpath.yaml + kubectl delete -f ./prometheus/prometheus-main.yml + kubectl delete -f ./servicemonitor/alertmanager-sm.yml + kubectl delete -f ./servicemonitor/coredns-sm.yml + kubectl delete -f ./servicemonitor/kube-apiserver-sm.yml + kubectl delete -f ./servicemonitor/kube-controller-manager-sm.yml + kubectl delete -f ./servicemonitor/kube-scheduler-sm.yml + kubectl delete -f ./servicemonitor/kubelet-sm.yml + kubectl delete -f ./servicemonitor/kubestate-metrics-sm.yml + kubectl delete -f ./servicemonitor/node-exporter-sm.yml + kubectl delete -f ./servicemonitor/prometheus-operator-sm.yml + kubectl delete -f ./servicemonitor/prometheus-sm.yml + kubectl delete -f ./servicemonitor/pushgateway-sm.yml + kubectl delete -f ./prometheus_adapter/metric_rule.yaml + kubectl delete -f ./prometheus_adapter/prometheus_adapter.yaml + cd ../ + + + # 部署gpu的监控 + kubectl delete -f gpu/nvidia-device-plugin.yml + kubectl delete -f gpu/dcgm-exporter.yaml + kubectl delete -f gpu/dcgm-exporter-sm.yaml + + # 部署frameworkcontroller nni超参搜索使用 + kubectl delete serviceaccount frameworkcontroller --namespace kubeflow + kubectl delete clusterrolebinding frameworkcontroller-kubeflow --clusterrole=cluster-admin --user=system:serviceaccount:kubeflow:frameworkcontroller + kubectl delete -f frameworkcontroller/frameworkcontroller-with-default-config.yaml + sleep 5 + # kubectl wait crd/frameworks.frameworkcontroller.microsoft.com --for condition=established --timeout=60s + + kubectl delete serviceaccount frameworkbarrier --namespace pipeline + kubectl delete serviceaccount frameworkbarrier --namespace automl + kubectl delete serviceaccount frameworkbarrier --namespace kubeflow + kubectl delete clusterrole frameworkbarrier --verb=get,list,watch --resource=frameworks + kubectl delete clusterrolebinding frameworkbarrier-pipeline --clusterrole=frameworkbarrier --user=system:serviceaccount:pipeline:frameworkbarrier + kubectl delete clusterrolebinding frameworkbarrier-automl --clusterrole=frameworkbarrier --user=system:serviceaccount:automl:frameworkbarrier + kubectl delete clusterrolebinding frameworkbarrier-kubeflow --clusterrole=frameworkbarrier --user=system:serviceaccount:kubeflow:frameworkbarrier + + # 部署volcano + kubectl delete -f volcano/volcano-development.yaml + kubectl delete secret volcano-admission-secret -n kubeflow + # kubectl apply -f volcano/volcano-development.yaml + # kubectl wait crd/jobs.batch.volcano.sh --for condition=established --timeout=60s + + # 部署istio + kubectl delete -f istio/install-crd.yaml + kubectl delete -f istio/install.yaml + kubectl delete -f istio/install-1.15.0.yaml + + # kubectl wait crd/virtualservices.networking.istio.io --for condition=established --timeout=60s + # kubectl wait crd/gateways.networking.istio.io --for condition=established --timeout=60s + + kubectl delete -f gateway.yaml + kubectl delete -f virtual.yaml + + # 部署argo + kubectl delete -f argo/minio-pv-pvc-hostpath.yaml + kubectl delete -f argo/pipeline-runner-rolebinding.yaml + kubectl delete -f argo/install-3.4.3-all.yaml + + # 部署trainjob:tfjob/pytorchjob/mpijob/mxnetjob/xgboostjobs + kubectl delete -f kubeflow/sa-rbac.yaml + kubectl delete -k kubeflow/train-operator/manifests/overlays/standalone + + # 部署sparkjob + kubectl delete -f spark/install.yaml + + # 部署paddlejob + kubectl apply -f paddle/crd.yaml + kubectl apply -f paddle/operator.yaml + + # 部署管理平台 + kubectl delete configmap kubernetes-config -n infra + + kubectl delete configmap kubernetes-config -n pipeline + + kubectl delete configmap kubernetes-config -n automl + + kubectl delete -f pv-pvc-infra.yaml + kubectl delete -f pv-pvc-jupyter.yaml + kubectl delete -f pv-pvc-automl.yaml + kubectl delete -f pv-pvc-pipeline.yaml + kubectl delete -f pv-pvc-service.yaml + + kubectl delete -k cube/overlays + kubectl delete -f sa-rbac.yaml + sh delete_ns_secret.sh + kubectl label node $node train- cpu- notebook- service- org- istio- kubeflow- kubeflow-dashboard- mysql- redis- monitoring- logging- + + read -p "是否删除所有数据？(yes/no)" answer + if [[ "$answer" == "yes" ]]; then + echo "delete all data !!! " + rm -rf /data/k8s/* + fi + echo "ns未删除，可视情况删除ns" +else + echo "退出脚本！" + exit 0 +fi \ No newline at end of file