Day-38 - Kubernetes Control Plane Troubleshooting
Day-38
Today I deep-dived into Kubernetes Control Plane Troubleshooting — fixing broken API servers, unhealthy schedulers, kubeconfig failures, and controller crash loops!
1️⃣ “Connection Refused?” → Check kube-apiserver first.
One tiny YAML typo inside /etc/kubernetes/manifests/ can bring your entire control plane down.
Debugging = details.
2️⃣ Using containerd? → crictl ps is the truth.
docker ps won’t show control-plane pods anymore in v1.34 clusters.
Know your runtime = faster troubleshooting.
3️⃣ Kubeconfig mistakes = instant failure.
Fix it quickly by copying:
/etc/kubernetes/admin.conf → ~/.kube/config
4️⃣ Pod stuck in Pending? → Look at kube-scheduler.
Image pull issues or misconfigured scheduler manifests are usually the culprits.
5️⃣ Deployment not recreating pods? → kube-controller-manager is sick.
CrashLoopBackOff often points to cert/hostPath misconfigurations.
6️⃣ kubectl cluster-info + cluster-info dump = underrated superpowers.
Best tools for understanding overall cluster state.
ReadMore: https://lnkd.in/dXm99iqG
Comments
Post a Comment