kubernetes不重置集羣更新節點ip

kubernetes節點ip變更後通信就會出現問題，我們只需要通過kubeadm init phase命令，重新生成config文件和簽名文件就可以了。

 

1、備份當前k8s集羣配置文件

cp -r /etc/kubernetes /etc/kubernetes_bak$(date +"%Y%m%d%H%M%S")

 

2、批量替換k8s配置文件中ip地址（如果配置了hosts、kubelet，同樣需要一起替換）

sed -i "s/${OLD_IP}/${NEW_IP}/g" `grep -rl "${OLD_IP}" /etc/kubernetes`

sed -i "s/${OLD_IP}/${NEW_IP}/g" /etc/hosts

if [ -f "/etc/default/kubelet" ]; then
  sed -i "s/${OLD_IP}/${NEW_IP}/g" /etc/default/kubelet
elif [ -f "/etc/sysconfig/kubelet" ]; then
  sed -i "s/${OLD_IP}/${NEW_IP}/g" /etc/sysconfig/kubelet
fi

 

3、重新生成apiserver證書

rm -rf /etc/kubernetes/pki/apiserver.*
kubeadm init phase certs apiserver --apiserver-advertise-address ${NEW_IP}

 

4、重新生成admin配置（並更新k8sconfig）

rm -rf /etc/kubernetes/admin.conf
kubeadm init phase kubeconfig admin --apiserver-advertise-address ${NEW_IP}
\cp /etc/kubernetes/admin.conf ~/.kube/config

 

5、重啓docker和kubelet

systemctl restart docker && systemctl restart kubelet
systemctl status docker && systemctl status kubelet

 

6、更新kube-proxy配置，並重啓kube-proxy服務

kubectl -n kube-system get cm kube-proxy -oyaml > /etc/kubernetes/kube-proxy.conf
sed -i "s/${OLD_IP}/${NEW_IP}/g" /etc/kubernetes/kube-proxy.conf
kubectl -n kube-system apply -f /etc/kubernetes/kube-proxy.conf
kubectl -n kube-system delete pod -l k8s-app=kube-proxy

注意：

1、如果不更新kube-proxy配置，連接將會失敗（因爲default.svc.kubernetes對應地址依然是舊的ip）

2、如果是多節點，刪除kube-proxy服務時會卡主，這是由於工作節點kubelet配置沒有更新，導致節點NotReady，刪除節點上pod自然會卡主（強制終止即可，或者參考步驟8提前更新工作節點）

 

 

 

7、檢查節點和系統狀態

kubectl get node | grep -vE 'STATUS|Ready'
Ready kubectl get pod -n kube-system |grep -vE 'STATUS|Running'

 

8、如果是多節點，需要修改工作節點kubelet.conf配置，並重啓kubelet

sed -i "s/${OLD_IP}/${NEW_IP}/g" `grep -rl "${OLD_IP}" /etc/kubernetes`
systemctl restart kubelet

 

 

>>>如果你還是覺得很麻煩，我這邊已經提供現成的自動化腳本：

      
#!/bin/bash

if [ "$UID" -ne 0 ]; then
   echo "[ERROR]: require root user"
   exit 1
fi
echo "If there is a worker node, first execute the following command to update:
sed -i 's/<OLD_IP>/<NEW_IP>/g' `grep -rl '<OLD_IP>' /etc/kubernetes`
systemctl restart kubelet
"

NEW_IP=$(ip route get 8.8.4.4 | head -1 | awk '{print $7}')
IP_CONF=$(cat /etc/kubernetes/manifests/kube-apiserver.yaml | grep "advertise-address")
OLD_IP=${IP_CONF#*=}
while [ "$1" != "" ]; do
   case $1 in
       --new-ip)
       NEW_IP=$2
       shift 2
       ;;
       --old-ip)
       OLD_IP=$2
       shift 2
       ;;
       *)
       echo "[ERROR] invalid argument '$1'"
       usage
       exit 1
    esac
done

echo "Update node ip: $OLD_IP > $NEW_IP"
echo "Backup k8s config: /etc/kubernetes_bak$(date +"%Y%m%d%H%M%S")"
cp -r /etc/kubernetes /etc/kubernetes_bak$(date +"%Y%m%d%H%M%S")

echo "======================= Update k8s config ==============================="
grep -rl "${OLD_IP}" /etc/kubernetes
sed -i "s/${OLD_IP}/${NEW_IP}/g" `grep -rl "${OLD_IP}" /etc/kubernetes`
sed -i "s/${OLD_IP}/${NEW_IP}/g" /etc/hosts

if [ -f "/etc/default/kubelet" ]; then
  sed -i "s/${OLD_IP}/${NEW_IP}/g" /etc/default/kubelet
elif [ -f "/etc/sysconfig/kubelet" ]; then
  sed -i "s/${OLD_IP}/${NEW_IP}/g" /etc/sysconfig/kubelet
fi
echo

echo "======================== Generate new certificate =========================="
rm -rf /etc/kubernetes/pki/apiserver.*
kubeadm init phase certs apiserver --apiserver-advertise-address ${NEW_IP}
echo

echo "======================= Generate new config ================================"
rm -rf /etc/kubernetes/admin.conf
kubeadm init phase kubeconfig admin --apiserver-advertise-address ${NEW_IP}
\cp /etc/kubernetes/admin.conf ~/.kube/config
echo

echo "======================= Restart docker and kubelet ========================="
systemctl restart docker && systemctl status docker | head -n10
systemctl restart kubelet && systemctl status kubelet | head -n10
echo

echo "=========================== Check node status =============================="
while kubectl get nodes| grep master | grep NotReady; do sleep 5 && echo "waiting node ready..."; done
echo

echo "======================== Update kube-proxy config ==========================="
kubectl -n kube-system get cm kube-proxy -oyaml > /etc/kubernetes/kube-proxy.conf
sed -i "s/${OLD_IP}/${NEW_IP}/g" /etc/kubernetes/kube-proxy.conf
kubectl -n kube-system apply -f /etc/kubernetes/kube-proxy.conf
kubectl -n kube-system delete pod -l k8s-app=kube-proxy
echo

echo "=========================== Check pod status ==============================="
while kubectl get pod -n kube-system| grep -vE "STATUS|Running"; do sleep 5 && echo "waiting pod ready..."; done
echo "IP updated successfully."