Cluster upgrade strategies in Kubernetes - Practice Problems & Coding Challenges

Challenge - 5 Problems

🧠 Conceptual

intermediate

Understanding Rolling Upgrade Strategy

Which statement best describes the rolling upgrade strategy in Kubernetes cluster upgrades?

A Upgrade nodes one at a time, maintaining cluster availability throughout.

B Upgrade all nodes simultaneously, causing a brief downtime.

C Upgrade only the master nodes and leave worker nodes unchanged.

D Upgrade nodes randomly without any order or plan.

💡 Hint

Think about how to keep the cluster available during upgrades.

The rolling upgrade strategy upgrades nodes one by one, ensuring the cluster remains available without downtime.

💻 Command Output

intermediate

Output of draining a node before upgrade

What is the expected output when running kubectl drain node-1 --ignore-daemonsets --delete-local-data?

Kubernetes

kubectl drain node-1 --ignore-daemonsets --delete-local-data

A Warning: ignoring DaemonSet-managed pods: kube-proxy, fluentd

B Error from server (NotFound): nodes "node-1" not found

C node/node-1 cordoned and drained

D node/node-1 drained

💡 Hint

Draining a node marks it unschedulable and evicts pods.

The command cordons the node (marks unschedulable) and drains pods, so the output confirms both actions.

🔀 Workflow

advanced

Correct sequence for a blue-green cluster upgrade

What is the correct order of steps in a blue-green upgrade strategy for Kubernetes clusters?

A 1,3,2,4

B 1,2,3,4

C 2,1,3,4

D 3,1,2,4

💡 Hint

Think about deploying, testing, switching traffic, then removing old cluster.

First deploy green cluster, then test it, switch traffic, and finally remove blue cluster.

❓ Troubleshoot

advanced

Troubleshooting node upgrade failure due to pod eviction

During a node upgrade, kubectl drain fails with the error: error when evicting pod: pods "my-pod" is not managed by a controller. What is the most likely cause?

A The pod has completed successfully and is terminating.

B The node is already drained and cordoned.

C The pod is managed by a ReplicaSet controller.

D The pod is a static pod and cannot be evicted automatically.

💡 Hint

Pods not managed by controllers are harder to evict.

Static pods are created directly by kubelet and are not managed by controllers, so they cannot be evicted by drain.

✅ Best Practice

expert

Best practice for minimizing downtime during cluster upgrade

Which practice best minimizes downtime during a Kubernetes cluster upgrade?

A Upgrade all nodes at once during a maintenance window.

B Use a rolling upgrade with pod disruption budgets configured.

C Delete all pods before upgrading nodes to avoid conflicts.

D Upgrade only worker nodes and skip control plane nodes.

💡 Hint

Consider how to keep pods available during node upgrades.

Using rolling upgrades with pod disruption budgets ensures pods remain available and downtime is minimized.