Hi all,
I complete the lab and all item is fix but lab mark incomplete, is it i have something not done ?
Thank you
If you are studying CKAD, know that the master node and node01 problems require CKA knowledge to solve.
Use the answers from the solution for those, then try to solve the rest yourself.
Hello,
This is regarding the Kubernetes Challenge #2.
The CoreDNS pods have an error related to the image and are not running.
The Kubernetes cluster is at v1.30.
The answer in the github link for this issue is to update the coredns deployment image version to registry.k8s.io/coredns/coredns:v1.8.6.
For Kubernetes v1.30, is the correct coredns version 1.8.6?
Is there a Kubernetes documentation page, where we can find the right version for all Kubernetes components with regards to a particular Kubernetes version?
I’m not sure why the challenge has you update CoreDNS to 1.8.6. The standard install versions of CoreDNS for different K8s versions is in github, and for 1.30.0, that would be v1.11.1. But apparently, using 1.8.6 does work, which I’ve verified in the lab. So 1.8.6 is similar enough to 1.11.1 to work, even if it isn’t the version generally used. v1.3.1, which is what you find installed in the lab initially, either that’s too old, or just doesn’t actually exist:
Warning Failed 8m15s (x4 over 9m35s) kubelet Failed to pull image “registry.k8s.io/kubedns:1.3.1”: rpc error: code = NotFound desc = failed to pull and unpack image “registry.k8s.io/kubedns:1.3.1”: failed to resolve reference “registry.k8s.io/kubedns:1.3.1”: registry.k8s.io/kubedns:1.3.1: not found
1 Like
It’s not that older versions of CoreDNS won’t work in newer clusters - they will. CoreDNS is just a DNS server and it doesn’t depend on cluster APIs that may become deprecated.
Newer versions of CoreDNS are not released to support newer kubernetes versions. It will be for things like small bug fixes, performance improvements and any enhancements to the DNS specification.
Kubernetes will recommend certain versions of it for different cluster versions based on the fact that they have tested it, although it’s not unlikely that CoreDNS 1.11 would run fine on a much older cluster, although the manifest/helm chart for installing it might need some modifications.
1 Like
per the original poster @mengkiat0211…
if all checks are green but the task does not complete, i found the service task instructions do not ask for a nodeport. once that is done all is well.