Quick Guide to VCF Automation for VCD Administrators

 Most Common Azure Kubernetes Errors & How to Fix Them:

1️⃣ Pods Stuck in "Pending" State

🔹 Issue: Insufficient resources or missing node pools.

✅ Solution:

Run kubectl describe pod <pod-name> to check for resource constraints.

Scale up the node pool (az aks scale --resource-group <rg> --name <cluster> --node-count <N>).

Check node taints & tolerations preventing scheduling.

2️⃣ ImagePullBackOff / ErrImagePull

🔹 Issue: The container image cannot be pulled.

✅ Solution:

Verify image availability (kubectl describe pod <pod-name>).

Ensure credentials are correct for private registries (kubectl create secret docker-registry).

Check network/firewall rules blocking the registry.

3️⃣ CrashLoopBackOff

🔹 Issue: The container repeatedly crashes.

✅ Solution:

Check logs (kubectl logs <pod-name>) and events (kubectl describe pod).

Debug locally using docker run before deploying.

Ensure readiness/liveness probes are correctly configured.

4️⃣ Node Not Ready

🔹 Issue: Node is unavailable in the cluster.

✅ Solution:

Run kubectl get nodes & kubectl describe node <node-name>.

Check if the VM is healthy in the Azure portal.

Restart or scale up the node pool.

5️⃣ OOMKilled (Out of Memory)

🔹 Issue: Container exceeds allocated memory.

✅ Solution:

Adjust resource requests/limits in your YAML file:

resources:

 requests:

 memory: "512Mi"

 limits:

 memory: "1Gi"

Use kubectl top pods to identify high memory usage.

6️⃣ RBAC Authorization Error (Forbidden)

🔹 Issue: Insufficient permissions for a user or service account.

✅ Solution:

Assign the correct role using:

kubectl create rolebinding <binding-name> --clusterrole=<role> --user=<user> --namespace=<namespace>

Verify RBAC rules with kubectl auth can-i.

7️⃣ Service Not Accessible (Pending or No External IP)

🔹 Issue: LoadBalancer or Ingress controller misconfiguration.

✅ Solution:

Ensure the service type is correct (kubectl get svc).

Use kubectl describe svc <service-name> to check external IP allocation.

For Ingress, verify Azure Application Gateway/NGINX ingress settings.

8️⃣ PersistentVolumeClaim (PVC) Stuck in Pending

🔹 Issue: Storage class or capacity issues.

✅ Solution:

Ensure the correct storage class is used (kubectl get sc).

Check if the requested storage size is available in Azure Disk/File Share.

Verify kubectl describe pvc for binding errors.

9️⃣ DNS Resolution Failure in Pods

🔹 Issue: Pods cannot resolve internal/external domains.

✅ Solution:

Restart CoreDNS (kubectl rollout restart deployment coredns -n kube-system).

Check kubectl get svc -n kube-system for the DNS service.

Validate resolv.conf inside the pod (kubectl exec -it <pod-name> -- cat /etc/resolv.conf).