每个k8s节点修改时区
timedatectl set-timezone Asia/Shanghai
date -s 正确时间
hwclock -w
每个k8s master节点执行:
cd deploy/scripts/
./fix-k8s-master-nodes.sh
cd ../../
每个k8s worker节点执行:
cd deploy/scripts/
./fix-k8s-worker-nodes.sh
cd ../../
部署机安装docker,保持网络在线,执行:
cd prepare/
./bundle.sh
以上命令会下载prometheus-operator的源码包并自动将项目依赖的全部docker镜像下载保存
拷贝部署机整个目录至一个k8s worker节点,修改一下deploy/scripts/generate-offline-package.sh和push-offline-images.sh脚本里的变量:
将MyImageRepositoryIP=192.168.9.20修改成实际的私有镜像仓库地址,并执行:
cd deploy/scripts/
./generate-offline-package.sh
./push-offline-images.sh
./deploy.sh
补充模块启用:
cd deploy/addon-exporter/
kubectl -n monitoring apply -f etcd-monitor.yaml
cd deploy/addon-exporter/exporter-kube-dns/
kubectl apply -f kube-dns.yaml
kubectl -n monitoring apply -f prometheus-k8s-service-monitor-kube-dns.yaml
cd deploy/addon-exporter/
kubectl apply -f k8s/
通过Grafana UI界面导入以下目录内的etcd、kube-dns的json模板即可使用相应的dashboard了。
deploy/grafana-dashboard