二进制K8s集群版本升级

一、升级说明

1. 官方建议在小版本之间进行升级，不建议大版本升级，更不建议直接跨大版本去升级
2. 二进制升级时查看官方的更新日志，与当前使用版本进行比对，是否有大的变动，防止升级中出现版本API接口兼容性问题导致升级失败
3. 在升级K8s集群时，二进制的方式只需要拷贝替换对应的二进制执行文件即可
4. 升级过程：先升级ETCD，然后升级K8s的Master节点组件，最后升级K8s集群的工作节点组件(说明：我这里主节点跟工作节点都部署了kubelet和kube-proxy组件，所以进行升级这2个组件时，集群所有节点都需要升级，包括主节点)
5. 升级ETCD和K8s主节点组件对Pod影响较小，升级Kubelet时会对Pod有影响
6. 注意：升级操作必须在业务影响最小的情况下执行，比如晚上

二、升级ETCD

1. 备份ETCD数据
2. 解压缩新版ETCD包
3. 逐个停止ETCD
4. 替换ETCD和ETCDCTL二进制执行文件
5. 启动ETCD

查看当前ETCD集群版本
for node_ip in ${ETCD_IPS[@]}
  do
    echo ">>> ${node_ip}"
    /opt/k8s/bin/etcdctl \
    --endpoints=https://${node_ip}:2379 \
    --cacert=/etc/kubernetes/cert/ca.pem \
    --cert=/etc/etcd/cert/etcd.pem \
    --key=/etc/etcd/cert/etcd-key.pem version
  done

查看当前ETCD集群的leader
/opt/k8s/bin/etcdctl \
  -w table --cacert=/opt/k8s/work/ca.pem \
  --cert=/etc/etcd/cert/etcd.pem \
  --key=/etc/etcd/cert/etcd-key.pem \
  --endpoints=${ETCD_ENDPOINTS} endpoint status

• 备份操作，请查看下面的文章

systemctl stop etcd
tar -zxvf etcd-v3.6.0-linux-amd64.tar.gz
scp etcd-v3.6.0-rc.0-linux-amd64/etcd* [email protected]:/opt/k8s/bin
ssh [email protected] "chmod +x /opt/k8s/bin/*"
ssh [email protected] "systemctl start etcd"
ssh [email protected] "systemctl status etcd"

for node_ip in ${ETCD_IPS[@]}
  do
    echo ">>> ${node_ip}"
    /opt/k8s/bin/etcdctl \
    --endpoints=https://${node_ip}:2379 \
    --cacert=/etc/kubernetes/cert/ca.pem \
    --cert=/etc/etcd/cert/etcd.pem \
    --key=/etc/etcd/cert/etcd-key.pem endpoint health
  done

~]# kubectl get no

三、升级Master节点

[[email protected] v1.24.0]# cd kubernetes/server/bin/

#创建备份目录
[[email protected] bin]# mkdir -p /opt/k8s/work/k8s-v1.24-rc_bak
#备份旧版本二进制执行文件
[[email protected] bin]# cp /opt/k8s/bin/kube-apiserver /opt/k8s/work/k8s-v1.24-rc_bak
[[email protected] bin]# systemctl stop kube-apiserver
[[email protected] bin]# cp -rp kube-apiserver /opt/k8s/bin/


#启动服务
[[email protected] bin]# systemctl start kube-apiserver.service

#检查服务运行状态
[[email protected] bin]# systemctl status kube-apiserver.service

#查看服务启动日志
[[email protected] bin]# journalctl -xe -u kube-apiserver.service -f

#查看版本
[[email protected] bin]# kube-apiserver --version

[[email protected] bin]# mkdir -p /opt/k8s/work/k8s-v1.24-rc_bak
#备份旧版本二进制执行文件
[[email protected] bin]# cp /opt/k8s/bin/kube-controller-manager /opt/k8s/work/k8s-v1.24-rc_bak

[[email protected] bin]# systemctl stop kube-controller-manager

[[email protected] bin]# cp -rp kube-controller-manager /opt/k8s/bin/
#启动服务
[[email protected] bin]# systemctl start kube-controller-manager.service

#检查服务运行状态
[[email protected] bin]# systemctl status kube-controller-manager.service

#查看服务启动日志
[[email protected] bin]# journalctl -xe -u kube-controller-manager.servic
#查看版本
[[email protected] bin]# kube-controller-manager --version

[[email protected] bin]# mkdir -p /opt/k8s/work/k8s-v1.24-rc_bak

#备份旧版本二进制执行文件
[[email protected] bin]# cp /opt/k8s/bin/kube-scheduler /opt/k8s/work/k8s-v1.24-rc_bak

[[email protected] bin]# systemctl stop kube-scheduler

[[email protected] bin]# cp -rp kube-scheduler /opt/k8s/bin/
#启动服务
[[email protected] bin]# systemctl start kube-scheduler.service

#检查服务运行状态
[[email protected] bin]# systemctl status kube-scheduler.service

#查看服务启动日志
[[email protected] bin]# journalctl -xe -u kube-scheduler.service -f
#查看版本
[[email protected] bin]# kube-scheduler --version

四、升级Worker节点

mkdir -p /opt/k8s/work/k8s-v1.24-rc_bak
cd /opt/k8s/work/v1.24.0/

#备份旧版本二进制执行文件
cp /opt/k8s/bin/kube-proxyr /opt/k8s/work/k8s-v1.24-rc_bak

systemctl stop kube-proxy

#替换旧版本二进制执行文件
cp -rp kube-proxy /opt/k8s/bin/


#启动服务
systemctl start kube-proxy.service

#检查服务运行状态
systemctl status kube-proxy.service

#查看版本
kube-proxy --version

[[email protected] ~]# kubectl drain k8s-master1 --delete-emptydir-data --force --ignore-daemonsets

[[email protected] ~]# ssh [email protected] "systemctl stop kubelet"
[[email protected] ~]# ssh [email protected] "systemctl status kubelet"

#备份老版本二进制执行文件
[[email protected] ~]# ssh [email protected] "cp /opt/k8s/bin/kubelet /opt/k8s/work/k8s-v1.24-rc_bak/"

#拷贝新版本二进制执行文件到执行目录下
[[email protected] ~]# ssh [email protected] "cp /opt/k8s/work/v1.24.0/kubelet /opt/k8s/bin/"

[[email protected] ~]# ssh [email protected] "systemctl start kubelet"
[[email protected] ~]# ssh [email protected] "systemctl status kubelet"

[[email protected] ~]# kubectl get no

[[email protected] ~]# kubectl uncordon k8s-master1