Common Node/compute problems
This section lists common issues you might see with worker nodes in your EKS cluster, along with potential solutions.
Node/Nodes can’t join the cluster
For this problem, what has changed, for example, a new node being added or an IP whitelist being changed, will determine what needs fixing. Often, the error is simply nodes being in a NotReady
/Unknown
state, shown next, which has a number of different root causes:
$ kubectl get node NAME  STATUS   ROLES    AGE    VERSION ip-172-31-10-50.x.internal    NotReady    <none>   7d4h   v1.24.13-eks-0a21954 ip-172-31-11-89.x.internal    Unknown    <none>   7d4h   v1.24.13-eks-0a21954
If you’re running self-managed nodes, did EC2 run the bootstrap.sh
script? With a managed node...