Credits

This documentation is based on the Certified Kubernetes Administrator (CKA) with Practice Tests Udemy course by Mumshad Mannambeth, with supporting labs from KodeKloud.

Working of Scheduler in Kubernetes

• Each pod has a nodeName field, which is empty by default.
• You usually don’t set this field in the manifest—Kubernetes sets it automatically.
• The scheduler looks for pods without a nodeName.
• It runs its algorithm to find a suitable node.
• Once identified, it schedules the pod by setting nodeName through a binding object.

---

Scheduling Without a Scheduler

If there is no scheduler, pods will remain in the Pending state.
You can still assign them manually in two ways:

Case	How it Works	Notes
During Pod Creation	Set the nodeName field in the pod spec before creating the pod. Kubernetes then assigns the pod to that node.	• Only possible at creation time. • Simple and direct method.
After Pod Creation	Create a Binding object and send a POST request to the Pods Binding API, specifying the target node in JSON format.	• nodeName cannot be modified after creation. • Mimics the scheduler’s binding process. • Requires YAML → JSON conversion.

---

Troubleshooting Pods in Pending State

If a pod is stuck in the Pending state, use the following checklist:

Area	What to Check	Command(s)
Pod Events and Description	Review events for scheduling errors or resource issues (CPU, memory, storage).	kubectl describe pod <pod-name>
Node Status	Ensure nodes are Ready and have enough free CPU, memory, and disk space.	kubectl get nodes kubectl describe node <node-name>
Scheduler Availability	Verify kube-scheduler is running. If it’s down, pods will remain Pending.	(Check scheduler pod logs in managed clusters)
Resource Requests and Limits	Confirm pod resources.requests don’t exceed node capacity. Compare with available resources.	(Check pod spec and node resources with kubectl describe)
Taints and Tolerations	Check if nodes have taints; pods need matching tolerations.	kubectl describe node <node-name> \| grep Taints
Node Selectors and Affinity	Validate scheduling rules like nodeSelector, nodeAffinity, or podAffinity/podAntiAffinity.	(Check pod spec YAML)
Persistent Volumes and Claims	Ensure PVCs are bound and a matching PV exists.	kubectl get pvc

---

✅ Quick Reference

1. Check Pod Events → kubectl describe pod.
2. Check Nodes → kubectl get nodes → kubectl describe node.
3. Confirm Scheduler is running.
4. Validate Resource Requests/Limits.
5. Inspect Taints/Tolerations.
6. Review Node Selectors / Affinity rules.
7. Verify PVCs and PVs.