Hi allm my nameservers got changed and rancher containers are not able to ping/curl anything anymore. When updating /etc/resolv.conf on host/node, I am able to ping/curl from host, but still not from rancher containers. Is there any command I need to run to reload dns service to update resolv.conf config?
$ kubectl get pods -o wide -n kube-system | grep coredns coredns-799dffd9c4-jfpxb 1/1 Running 0 166d 10.42.0.2 10.65.5.25 <none> <none> coredns-autoscaler-84766fbb4-xw8bp 1/1 Running 0 166d 10.42.2.2 10.65.5.27 <none> <none> rke-coredns-addon-deploy-job-fsb2l 0/1 Completed 0 166d 10.65.5.25 10.65.5.25 <none> <none>
I tried to
kubectl delete pods -n kube-system coredns-799dffd9c4-jfpxb
and it did re-create a new one, but still the same old resolv.conf config.
I have changed resolv.conf on all nodes, where does it keep picking up the old config from?
Apparently when I try to edit config map of coredns and add
forward . 22.214.171.124 126.96.36.199 save it and then delete pod by
kubectl -n kube-system delete po -l k8s-app=kube-dns it gets deleted, recreated, but still the same old config is in place which confuses me as it should have worked.
I would really appreciate the help, thanks!