Kubernetes not able to resolve DNS. Container/Pods not able to access Internet.
I have a Kubernetes 2 node cluster on separate AWS EC2 instances (t2.Medium) container networking has been done using: Flannel version: flannel:v0.10.0-amd64 (image) Kubernetes version: 1.15.3
DNS Logs DNS Logs
At times when I delete core-dns pods, the DNS issue gets resolved for some time but it is not consistant. Please suggest what can be done. I flannel mapping may have something to do with this. Please let me know if any other information is also needed.
Errors such you get: nslookup: can't resolve 'kubernetes.default' indicate that you have problem with the coredns/kube-dns add-on or associated Services.
Please check if you did following steps to debug DNS: coredns.
It also seems that like DNS inside busybox does not work properly.
Try to use busybox images <= 1.28.4
Change pod configuration file:
containers:
- name: busybox-image
image: busybox:1.28.3Learn more about most known dns kubernetes issues: kubernetes-dns.