kubectl
1. 安装 kubectl
1.1. mac 安装 kubectl
brew info kubernetes-cli
brew search kubernetes-cli
brew install kubernetes-cli
brew install bash-completion
kubectl completion bash > $(brew --prefix)/etc/bash_completion.d/kubectl
pip3 install kube-shell
# 卸载
brew uninstall kubernetes-cli
# 指定版本安装(这个安装报错失败)
brew install https://github.com/Homebrew/homebrew-core/blob/d09d97241b17a5e02a25fc51fc56e2a5de74501c/Formula/kubernetes-cli.rb
# 采用二进制方式安装指定版本(需翻墙)
curl -LO https://storage.googleapis.com/kubernetes-release/release/v1.9.6/bin/darwin/amd64/kubectl
chmod +x /Users/yanglei/Downloads/kubectl
mv /Users/yanglei/Downloads/kubectl /usr/local/bin/kubectl
将集群的 /root/.kube/config 拷贝到 mac，即可管理集群了。
kubectl label nodes 10.0.43.33 k8s.wonhigh.cn/namespace=bst-petrel-st --overwrite kubectl label nodes 10.0.43.9 k8s.wonhigh.cn/namespace=bst-petrel-st --overwrite
2. 常用操作
# 查看节点资源耗用情况
kubectl top node
# 按 label 查看节点资源耗用情况
kubectl top node -l k8s.wonhigh.cn/namespace=wonhigh-petrel-dev
# 禁止节点调度
kubectl cordon 10.240.116.53
# 驱逐节点 pod
kubectl drain 10.240.116.53 --ignore-daemonsets --delete-local-data
# 强制驱逐节点 pod
kubectl drain 10.240.116.53 --ignore-daemonsets --delete-local-data --force
# 关闭服务、重启机器...
# 检查机器...
# 删除已退出的所有容器
docker rm `docker ps -a | grep Exited | awk '{print $1}'`
# 恢复节点调度
kubectl uncordon 10.240.116.53
# 使用 busybox 容器测试集群，比如网络、dns 等是否正常
kubectl run -it --rm busybox2 --image=busybox /bin/sh
kubectl run -it --rm busybox2 --image=busybox --overrides='{ "apiVersion": "apps/v1", "kind": "Deployment", "spec": { "template": { "spec": { "nodeSelector": { "kubernetes.io/hostname": "10.0.42.176" } } } } }' /bin/sh
# 也可以获取官方的编排文件进行改造
wget https://k8s.io/examples/admin/dns/busybox.yaml
kubectl create -f busybox.yaml
kubectl exec -it busybox2 nslookup kubernetes.default
kubectl exec -it busybox3 nslookup kubernetes.default
kubectl exec -it busybox2 nslookup test-dop-server.belle.net.cn
kubectl exec -it busybox3 nslookup test-dop-server.belle.net.cn
kubectl run -it --rm alpine --image=alpine --overrides='{ "apiVersion": "apps/v1", "kind": "Deployment", "spec": { "template": { "spec": { "nodeSelector": { "kubernetes.io/hostname": "10.0.42.176" } } } } }' /bin/sh
apk add curl
apk add tcpdump
# 扩容
kubectl scale deploy kubernetes-dashboard --replicas=1 -n kube-system
kubectl scale rc oms-e-api.1.0.1.rc17 --replicas=3 -n belle-petrel-prod
重启所有pods
(删除所有pods，然后k8s根据deployment中的设置重建)
kubectl delete pod $(kubectl get pods | grep -v NAME | awk '{print $1}')
更新deployment数量
kubectl scale deployment depName --replicas 1
获取异常容器
kubectl get pods | grep -v Running
查看pod状态信息
kubectl describe pod podName
查看pod日志信息
kubectl logs podName
将k8s node变为不可用状态
kubectl patch node nodeIP -p '{"spec":{"unschedulable":true}}'
将k8s node变为可用状态
kubectl patch node nodeIP -p '{"spec":{"unschedulable":false}}'
3. 不常用操作
# dns 组件检查
kubectl get endpoints kube-dns -n kube-system
kubectl exec -it kube-dns-84777cd667-84hzf -c kubedns -n kube-system nslookup kubernetes.default 127.0.0.1
kubectl exec -it kube-dns-84777cd667-84hzf -c kubedns -n kube-system nslookup test-dop-server.belle.net.cn 127.0.0.1
kubectl exec -it kube-dns-84777cd667-jg5f9 -c kubedns -n kube-system nslookup test-dop-server.belle.net.cn 127.0.0.1
kubectl exec -it kube-dns-84777cd667-p8bvd -c kubedns -n kube-system nslookup test-dop-server.belle.net.cn 127.0.0.1
kubectl exec -it kube-dns-84777cd667-pnsnm -c kubedns -n kube-system nslookup test-dop-server.belle.net.cn 127.0.0.1
kubectl exec -it kube-dns-84777cd667-sbk5k -c kubedns -n kube-system nslookup test-dop-server.belle.net.cn 127.0.0.1