CoreDNS 的pod为CrashLoopBackOff 状态问题的解决办法

测试环境发现运行在Master节点上的CoreDNS 的pod为CrashLoopBackOff 异常状态状态，导致业务应用大量pod均出现异常，查看日志显示大量如下报错：

# 查看详细信息
kubectl -n kube-system describe pod coredns-6dc69b487c-svxxx

# 查看CoreDNS 的pod日志
kubectl -n kube-system logs -f –tail=200 coredns-6dc69b487c-svxxx

# 输出日志
Failed to list *v1.Service: Get https://10.96.0.1:443/api/v1/services?limit=500&resourceVersion=0: dial tcp 10.96.0.1:443: connect: no route to host
Failed to list *v1.Service: Get https://10.96.0.1:443/api/v1/services?limit=500&resourceVersion=0: dial tcp 10.96.0.1:443: connect: no route to host
Failed to list *v1.Service: Get https://10.96.0.1:443/api/v1/services?limit=500&resourceVersion=0: dial tcp 10.96.0.1:443: connect: no route to host

根据日志分析，该pod无法和10.96.0.1之间建立路由连接，请求被拒绝，所有要允许容器pod 的 IP段通过防火墙，那么容器IP段如何查询呢？

通常是查看该地址对应的svc，因为k8s集群有两个段，一个k8s-service服务IP段，容器IP段，两者之间无法找到路由。

kubectl get svc -A |grep 10.96.0.1
default kubernetes ClusterIP 10.96.0.1 <none> 443/TCP

解决

猜测可能是iptables防火墙未放行，由于部署之前已经关闭掉了firewalld，因此尝试清空iptables，放开连接。

#执行如下命令，清空iptables

systemctl stop kubelet
systemctl stop docker
iptables –flush
iptables –tnat –flush
systemctl start kubelet
systemctl start docker

# 通过deployment的方式去重启Pod

kubectl scale deployment coredns –replicas=0 -n kube-system
deployment.extensions/coredns scaled
kubectl scale deployment coredns –replicas=1 -n kube-system
deployment.extensions/coredns scaled

#再次查看pod运行状态，此时已经正常

kubectl get pod -A |grep core
kube-system coredns-7dc79b4f7c-clvz4 1/1 Running