具有calico网络的Crashloopbackoff状态的Coredns

Question

我有一个运行在虚拟机器中的ubuntu 16.04。我使用kubeadm将Kubernetes作为单个节点安装在上面。

但核心pods处于Crashloopbackoff状态。

所有其他pods都在运行。

单接口(Enp0s3)-网桥网络

使用kubectl apply -f https://docs.projectcalico.org/manifests/calico.yaml的应用印花布

output on kubectl describe pod: 
 Type     Reason     Age                  From               Message
  ----     ------     ----                 ----               -------
  Normal   Scheduled  41m                  default-scheduler  Successfully assigned kube-system/coredns-66bff467f8-dxzq7 to kube
  Normal   Pulled     39m (x5 over 41m)    kubelet, kube      Container image "k8s.gcr.io/coredns:1.6.7" already present on machine
  Normal   Created    39m (x5 over 41m)    kubelet, kube      Created container coredns
  Normal   Started    39m (x5 over 41m)    kubelet, kube      Started container coredns
  Warning  BackOff    87s (x194 over 41m)  kubelet, kube      Back-off restarting failed container

Answer 1

我做了一个kubectl logs <coredns-pod>，找到了下面的错误日志，根据建议查看了上面提到的link，在/etc/kubernetes/kubelet/conf.yaml的末尾添加了resolv.conf = /etc/resolv.conf，并重新创建了pod。

kubectl logs coredns-66bff467f8-dxzq7 -n kube-system 
.:53 [INFO] plugin/reload: Running configuration MD5 = 4e235fcc3696966e76816bcd9034ebc7 CoreDNS-1.6.7 linux/amd64, go1.13.6, da7f65b [FATAL] plugin/loop: Loop (127.0.0.1:34536 -> :53) detected for zone ".", see coredns.io/plugins/loop#troubleshooting. Query: "HINFO 8322382447049308542.5528484581440387393." 
root@kube:/home/kube# 

Answer 2

在/etc/resolv.conf (主机)中的行下面注释，并删除kube-system名称空间中的coredns pod。

新的pods以运行状态出现:)

• #nameserver 127.0.1.1