Hi,
We are using OKD 4.5 and when one worker node is having network issue, the application getting dns issue when the request is hitting that problematic node. Also we are not able to bring down that problematic node dns pod, its automatically recreate. So how to handle this type of scenario. We are already in the process of upgrading version to OKD 4.9. we dont have any zones in dns.operator/default.
Is there any way to somehow enforce that DNS requests on a node are not going (via round-robin) to any other server, but only to the local coredns service (even if not globally in the cluster then maybe at least in the pod/deployment config)?
We have server in Asia and Europe, as round-robin EU request hit asia worker node coredns. Any thought on this?
Thanks