🛠️ Manual Scheduling in Kubernetes

By default, Kubernetes uses the scheduler to assign Pods to nodes. But you can manually schedule a Pod by specifying the nodeName field — bypassing the scheduler entirely.

---

🧾 Use Cases

• Debugging scheduler behavior
• Manually placing a Pod on a specific node
• Simple test setups or demos

---

🧱 YAML Example: Manual Scheduling

apiVersion: v1
kind: Pod
metadata:
  name: manual-pod
spec:
  containers:
  - name: nginx
    image: nginx
  nodeName: worker-node-1

This Pod will only run on worker-node-1, assuming the node exists and is schedulable.

---

🔍 Get Available Node Names

kubectl get nodes

---

🚫 What Happens if Node Is Unavailable?

• Pod stays in Pending state
• Events will show FailedScheduling with no matching node

---

🔄 Switching Back to Scheduler

To let the scheduler choose the node again, simply remove the nodeName field and apply the Pod definition anew.

---

🧪 Alternative: kubectl run with Node Selector

While not full manual scheduling, you can use nodeSelector to influence placement:

kubectl run nginx --image=nginx --overrides='{
  "apiVersion": "v1",
  "spec": {
    "nodeName": "worker-node-1"
  }
}' --dry-run=client -o yaml | kubectl apply -f -

---

❗ Caution

• Manual scheduling does not consider taints, affinities, or resource constraints
• You must know what you're doing — there’s no safety net

---

✅ Summary

Manual scheduling with nodeName: is useful for learning and debugging, but not recommended for production workloads. Use it only when you need full control over placement and understand the cluster state.