coredns CrashLoopBackOff 报错

1.kubectl logs -f coredns-99b9bb8bd-47mvf -n kube-system

.:53
2018/09/22 07:39:37 [INFO] CoreDNS-1.2.2
2018/09/22 07:39:37 [INFO] linux/amd64, go1.11, eb51e8b
CoreDNS-1.2.2
linux/amd64, go1.11, eb51e8b
2018/09/22 07:39:37 [INFO] plugin/reload: Running configuration MD5 = f65c4821c8a9b7b5eb30fa4fbc167769
2018/09/22 07:39:38 [FATAL] plugin/loop: Seen "HINFO IN 8205794187887631643.5216586587165434789." more than twice, loop detected

参考这篇文章： https://github.com/coredns/coredns/issues/1986

从这篇文章得到提示，coredns pod会取宿主机的/etc/resolv.conf里面定义的nameserver作为自己的upstream server。而ubuntu的这个文件定义的nameserver是127.0.0.1，那么问题来了，coredns向127.0.0.1查询，而127.0.0.1代表本机，就这样形成死循环，也解释了为什么上面的日志提示"loop detected"

• 修改/etc/resolv.conf
  这个文件是自动生成的，无法手工修改。上网找了一大堆办法

修改 /etc/network/interfaces
auto enp2s0
iface enp2s0 inet dhcp
dns-nameservers 114.114.114.114

systemctl restart networking

修改/etc/dhcp/dhclient.conf
prepend domain-name-servers 114.114.114.114;

systemctl restart networking
统统没用，/etc/resolv.conf里面只有nameserver 127.0.0.1

苍天不负有心人，终于被我找到这篇文章

https://askubuntu.com/questions/627899/nameserver-127-0-1-1-in-resolv-conf-wont-go-away/627900#627900
原来只要简单粗暴地，把原来的/etc/resolv.conf删掉，重新创建一个新文件，把nameserver写进去，就行了

 

参考：http://blog.51cto.com/355665/2178181