1. Upgrades

默认 pod 超时时间为5分钟：kube-controller-manager --pod-eviction-timeout=5m0s

系统升级时更安全的方式是：kubectl drain node-1，节点会被标记为不可调度，会清除 pod，但当 pod 不属于 replicaSet 时，drain 会失败，需要加上 --force，这个 pod 就会丢失，不会再在其他node 上被创建

升级完成之后：kubectl uncordon node-1

kubectl cordon node-2不会清除pod

使用 kubectl get nodes 可以在 VERSION 列查看版本

v1.11.3

• MAJOR

• MINOR : Features, Functionalities

• PATCH : Bug Fixes

ETCD、CoreDNS 的版本独立，kube-apiserver 的版本一般比其他的组件更高，kubectl 的版本可能会高或者低1个版本，其他 Controller-manager、kube-scheduler、kubelet、kube-proxy 版本保持一致。

Kubernetes 只支持最近的3个版本，当使用的 k8s 版本不再支持时，建议一次只升级一个版本。

使用 kubeadm 升级：

• kubeadm upgrade plan 查看最近稳定可用的版本

• kubeadm upgrade apply

先升级 master node，再升级 work node

2. Releases

• 升级master node：

apt-get upgrade -y kubeadm=1.12.0-00

kubeadm upgrade apply v1.12.0

kubectl get nodes 上显示的是apiserver版本

如果master node上有kubelet，则apt-get upgrade -y kubelet=1.12.0-00，然后systemctl restart kubelet

• 升级work node：

kubectl drain node-1 把节点上的pod转移到其他节点

apt-get upgrade -y kubeadm=1.12.0-00

apt-get upgrade -y kubelet=1.12.0-00

kubeadm upgrade node config --kubelet-version v1.12.0

systemctl restart kubelet

kubectl uncordon node-1

查看系统版本：cat /etc/*release*

• Ubuntu 升级 master node 流程：升级 node 用 upgrade apply 版本

On the controlplane node, run the command run the following commands:

apt update : This will update the package lists from the software repository.

apt install kubeadm=1.20.0-00 : This will install the kubeadm version 1.20

kubeadm upgrade apply v1.20.0 : This will upgrade kubernetes controlplane. Note that this can take a few minutes.

apt install kubelet=1.20.0-00 : This will update the kubelet with the version 1.20.

You may need to restart kubelet after it has been upgraded. Run: systemctl daemon-reload, systemctl restart kubelet

• Ubuntu 升级 work node 流程：升级 node 用 upgrade node

If you are on the master node, run ssh node01 to go to node01

apt update : This will update the package lists from the software repository.

apt install kubeadm=1.20.0-00 : This will install the kubeadm version 1.20

kubeadm upgrade node : This will upgrade the node01 configuration. 不用写node-name

apt install kubelet=1.20.0-00 : This will update the kubelet with the version 1.20.

You may need to restart kubelet after it has been upgraded. Run: systemctl daemon-reload, systemctl restart kubelet

Type exit or enter CTL + d to go back to the controlplane node.

uncordon work node 需要切换到 master node

官方文档：https://kubernetes.io/docs/tasks/administer-cluster/kubeadm/kubeadm-upgrade/

3. Rolling Updates and Rollbacks

kubectl rollout status deployment/myapp-deployment：查看状态

kubectl rollout history deployment/myapp-deployment：查看历史

部署策略：

• Recreate：一次销毁所有旧的，新建新版本
• Rolling Update：滚动更新，一次更新一个

更新 image 的方式：

• kubectl apply -f deployment-definition.yml
• kubectl set image deployment/myapp-deployment nginx=nginx:1.9.1：会导致部署定义文件具有不同的配置

回滚：kubectl rollout undo deployment/myapp-deployment，回滚会在原来的replicasets上重新创建