Kubernetes Workloads: ReplicaSets and Deployments
Build a ReplicaSet by hand, observe self-healing, then update its template and end up with Pods running different images. Layer a Deployment on top and confirm the full ownership chain.
Explore tutorials, challenges, courses, and more published by this author.
Build a ReplicaSet by hand, observe self-healing, then update its template and end up with Pods running different images. Layer a Deployment on top and confirm the full ownership chain.
Build a Pod from scratch, exec into it, try to edit an immutable field, explore all three restart policies, and observe graceful termination.
Observe the Delete and Retain reclaim policies in action, recover a Released PV by patching claimRef, and recover data by creating a new PV pointing to the same storage directory.
Explore how the DefaultStorageClass admission controller injects storage classes, explicitly name a StorageClass, switch the cluster default, and inspect where local-path stores volumes on the node.
Learn the three static PV binding patterns - generic match, label selector, and bidirectional claimRef - then use a no-provisioner StorageClass with WaitForFirstConsumer to pin storage to a specific node.
Use emptyDir volumes to share data between containers, understand the volume lifecycle across container restarts and pod deletion, and explore memory-backed ephemeral storage.
Four hands-on challenges covering the full storage stack. Start with emptyDir, move to dynamic provisioning with StorageClass, explore static PV binding patterns including label selectors, bidirectional claimRef, and topology-aware local volumes, then finish with PV reclaim policies and data recovery.
Nine hands-on challenges split into two sections. The resource section covers how requests, QoS classes, priority, and namespace quotas shape scheduling decisions. The placement section covers how affinity, topology spread, and dedicated nodes control where pods land. HPA closes the path by showing how requests drive autoscaling.