Incomplete instructions in Update CoreDNS section on Fargate Getting Started page

[prasadkatti]

This issue is related to the Update CoreDNS section on the Fargate Getting Started page.

I created an EKS cluster with just a single fargate profile to run all of the pods on Fargate. As expected, the coredns pods on my cluster were in PENDING state. The instructions on the page talk about patching the coredns deployment to remove the eks.amazonaws.com~1compute-type annotation. However that annotation was already missing on the pod spec in the coredns deployment.

On inspecting one of the fargate nodes I saw that it had a eks.amazonaws.com/compute-type=fargate:NoSchedule taint on it. So I tried adding a corresponding toleration to the codedns pods and that resolved the issue and my coredns pods got scheduled onto new fargate nodes.