I've been encountering few major issues with my K8s cluster:
- Windows pods running on my Windows nodes are unable to communicate (internally within the cluster) with my linux pods or services. BUT my Linux pods are able to communicate with my Windows pods.
- Windows pods are also unable to resolve dns names for any service within K8s
- Windows pods are unable to resolve domain names for anything on the internet. Though after opening a shell in the windows pod and manually adding 8.8.8.8 as a DNS server, it would start working
- Windows pods are unable to communicate with other pods using the FQDN or ClusterIPs of services created for windows pods.
Things I've verified:
- CoreDNS is running
- I see no warnings or errors in logs or k8s description of either CoreDNS or Calico pods
I suspect that the issue might be with calico, but I'm quite lost on where to even begin now to troubleshoot these issues. Googling turned up this similar issue, but I've been unsuccessful so far
I've also tried running the following command, but no luck:
kubectl -n kube-system rollout restart deployment coredns
Any ideas any what I can try to take a look at? below is a diagram that I created to try and help visually explain my issue:
