Date: 2026-01-28
Re-initialized the single-node kubeadm cluster to move the pod CIDR off the LAN subnet. Target pod CIDR: 10.244.0.0/16.
kubeadm reset -f- Removed CNI and kubelet state directories:
/etc/cni/net.d, /var/lib/cni, /var/lib/kubelet, /etc/kubernetes (post-reset cleanup)
kubeadm init --pod-network-cidr=10.244.0.0/16- Applied Calico
v3.31.3
- Applied ingress-nginx
controller-v1.14.2
- Applied local-path-provisioner
v0.0.30
- Updated
configs/kubernetes-network.md with the new pod CIDR
- Control plane components intermittently stabilize (etcd/apiserver running), but the node remains
NotReady.
- CNI not initialized; Calico init container failing to obtain service account token (node authorization/timeout errors).
- Addon pods (CoreDNS, ingress-nginx, local-path) are still
Pending.
- Investigate kubelet node authorization errors for token requests.
- Stabilize apiserver/etcd before re-attempting CNI initialization.
- Re-check kube-proxy and service IP reachability (
10.96.0.0/12).