How to Fix Kubernetes 0/3 Nodes Are Available (Scheduling Failed)
Quick Fix Summary
TL;DRCheck node taints, resource requests, and node conditions with `kubectl describe node` and `kubectl get events`.
The Kubernetes scheduler cannot find a suitable node to place your pod, resulting in a 'SchedulingFailed' event. This is a critical failure that prevents application deployment or scaling.
Diagnosis & Causes
Recovery Steps
Step 1: Diagnose with kubectl describe and get events
Immediately gather the error details from the pod and cluster events to understand the scheduler's decision.
kubectl describe pod <pod-name> -n <namespace>
kubectl get events --sort-by='.lastTimestamp' -n <namespace>Step 2: Inspect Node Status and Taints
Check if nodes are ready and examine their taints, which can repel pods unless the pod has a matching toleration.
kubectl get nodes
kubectl describe node <node-name> | grep -A 10 -i taintStep 3: Verify Node Resources vs. Pod Requests
Compare the pod's resource requests (CPU/memory) against the allocatable resources on each node.
kubectl describe node <node-name> | grep -A 5 -i allocatable
kubectl describe pod <pod-name> | grep -i requests -A 2Step 4: Check for NodeSelector and Affinity Rules
Ensure your pod's nodeSelector or nodeAffinity rules match labels present on at least one ready node.
kubectl get nodes --show-labels
kubectl describe pod <pod-name> | grep -i node-selector -A 2 -B 2Step 5: Resolve PersistentVolumeClaim Issues
If the pod requires a PersistentVolume, ensure a matching StorageClass exists and the PVC is bound.
kubectl get pvc -n <namespace>
kubectl describe pvc <pvc-name> -n <namespace>Step 6: Add Tolerations or Fix Node State
Based on diagnosis, either add tolerations to your pod spec or cordon/drain/fix the problematic node.
# Example Toleration for a common node taint
tolerations:
- key: "node.kubernetes.io/unreachable"
operator: "Exists"
effect: "NoSchedule"Architect's Pro Tip
"Use `kubectl get pods -o wide --all-namespaces | grep -v Running` to instantly see all non-running pods and their assigned nodes, highlighting cluster-wide scheduling issues."
Frequently Asked Questions
What's the difference between a taint and a nodeSelector?
A taint repels pods from a node unless they tolerate it (node-pushes-pod-away). A nodeSelector attracts a pod to nodes with matching labels (pod-pulls-to-node).
My node shows 'Ready' but pods still won't schedule. Why?
The node may be cordoned (scheduling disabled), have resource pressure (memory/disk), or have unschedulable taints like 'NoSchedule' that your pod doesn't tolerate.
How can I prevent this error during deployment?
Define realistic resource requests/limits in your pod spec, use PodDisruptionBudgets for voluntary disruptions, and ensure node affinity/toleration rules are tested in non-production first.