coredns 1.2.2 反復重啟問題

    此環境是有TF(tungstenfabric) 部署的，采用的k8s版本是1.12，coredns是1.2.2

pod就是循環重啟

[root@opcontroller ~]# kubectl get pod -n kube-system -o wide
NAME                                               READY   STATUS                 RESTARTS   AGE    IP                 NODE           NOMINAT NODE
coredns-576cbf47c7-v4hkl                0/1     CrashLoopBackOff   36           161m               10.47.255.251   opcompute      <none>
coredns-576cbf47c7-xtrfb                 0/1     CrashLoopBackOff   36           160m               10.47.255.252   opcompute      <none>

[root@opcontroller ~]# kubectl  describe   pod coredns-576cbf47c7-v4hkl  -n kube-system

解決辦法:

[root@opcontroller ~]# kubectl get cm coredns -o yaml -n kube-system
apiVersion: v1
data:
  Corefile: |
    .:53 {
        errors
        health
        kubernetes cluster.local in-addr.arpa ip6.arpa {
           pods insecure
           upstream
           fallthrough in-addr.arpa ip6.arpa
        }
        prometheus :9153
        proxy . /etc/resolv.conf
        cache 30
        loop  #刪除此處即可
        reload
        loadbalance
    }
kind: ConfigMap
metadata:
  creationTimestamp: 2020-01-03T23:56:52Z
  name: coredns
  namespace: kube-system
  resourceVersion: "221"
  selfLink: /api/v1/namespaces/kube-system/configmaps/coredns
  uid: b6e7d450-2e84-11ea-ae57-000c2939e87

刪除之前的2個pod

[root@opcontroller ~]# kubectl delete pods coredns-576cbf47c7-v4hkl -n kube-system
pod "coredns-576cbf47c7-v4hkl" deleted

過一會k8s會自動拉起