coredns running status but not become ready-CodePudding

I have a k8s cluster with three masters, two workers, and an external haproxy and use flannel as a cni. The coredns have problems, and their status is running, but they don't become ready.

Coredns log

I get the logs of this pod, and I get this message:

[INFO] plugin/ready: Still waiting on: "Kubernetes."

What I do to solve this problem but didn't get any result:
1- check ufw and disable it.
2- check IPtables and flush them.
3- check Kube-proxy logs.
4- check the haproxy, and it is accessible from out and all servers in the cluster.
5- check nodes network.
7- reboot all servers at the end. :))

I get describe po :

describe pod

CodePudding user response：

Lets see your CoreDNS works at all?

You can crete a simple pod and go inside and from there, curl Services via IP:PORT & Service-name:PORT
```
kubectl run -it --rm test-nginx-svc --image=nginx  -- bash
```

IP:PORT
```
curl http://<SERVICE-IP>:8080
```
DNS
```
curl http://nginx-service:8080
```

If you couldn't curl your service via Service-name:PORT then you probably have a DNS Issue....

CoreDNS

Service Name Resolution Problems?

Check CoreDNS Pods are running and accessible?
Check CoreDNS logs

kubectl run -it test-nginx-svc --image=nginx -- bash

Inside the Pod
```
cat /etc/resolv.conf 
```

The result would look like:

nameserver 10.96.0.10 # IP address of CoreDNS
search default.svc.cluster.local svc.cluster.local cluster.local
options ndots:5

If It is NOT working:

I suggest to try re-install it via official docs or helm chart

Try onther CNIs like weave

CodePudding user response：

I solve the problem. It didn't relate to coredns. It was about the IPtables rule. Some rules for Kubernetes svc did not create. Now with the rules which are in the picture, everything is okay:

iptables rules