Day 8/40 — Kubernetes Deployment, Replication Controller and ReplicaSet Explained

The eighth installment of the Tech Tutorials with Piyush 40-day CKA course. This session moves from standalone Pods to the workload controllers that guarantee high availability, self-healing, and zero-downtime updates in production Kubernetes.

Video Summary

This ~35-minute tutorial explains the three core Kubernetes workload controllers that manage Pod replication: ReplicationController (legacy), ReplicaSet (modern), and Deployment (the production standard). Piyush combines whiteboard theory with live terminal demonstrations using kubectl and YAML manifests.

Key Concepts Covered

1. Why Pods Need Controllers A bare Pod is a single point of failure. If it crashes, user traffic receives no response. Kubernetes controllers solve this by continuously monitoring the desired state (number of replicas) and reconciling the actual state — spinning up replacements automatically.

2. ReplicationController (Legacy)

• apiVersion: v1, kind: ReplicationController
• Ensures a specified number of Pod replicas are running at all times
• Uses a template field inside spec to define the Pod blueprint (image, ports, labels)
• If a Pod dies, the controller creates a new one from the template
• Supports manual scaling by updating the replicas field
• Limitation: Cannot adopt existing Pods that were not created by it

3. ReplicaSet (Modern Replacement)

• apiVersion: apps/v1, kind: ReplicaSet
• Functionally similar to ReplicationController but with a critical addition: selectors
• The spec.selector.matchLabels field allows a ReplicaSet to manage existing Pods that match its label criteria, even if they were created independently
• This makes ReplicaSets more flexible and is the reason Deployments use ReplicaSets under the hood
• Scaling methods demonstrated:
  1. Edit YAML → kubectl apply -f
  2. Edit live object → kubectl edit rs <name>
  3. Imperative command → kubectl scale --replicas=10 rs/<name>

4. Deployment (Production Standard)

• apiVersion: apps/v1, kind: Deployment
• A higher-level abstraction that manages ReplicaSets, which in turn manage Pods
• Key superpower: Rolling Updates
  • Updates Pods one-by-one (or in batches) so the application never goes offline
  • Old Pods serve traffic while new Pods are being created and health-checked
  • Ideal for production workloads like banking or trading platforms where downtime is unacceptable
• Rollback capability
  • kubectl rollout history deployment/<name> — view revision history
  • kubectl rollout undo deployment/<name> — revert to the previous revision
• Image updates
  • kubectl set image deployment/<name> container=image:tag — change container image on the live object

Hands-On Commands Demonstrated

Command	Purpose
kubectl apply -f rc.yaml	Create a ReplicationController
kubectl get rc	List ReplicationControllers
kubectl get rs	List ReplicaSets
kubectl get deploy	List Deployments
kubectl get all	Show all resources in the namespace
kubectl describe pod <name>	Inspect Pod details and events
kubectl scale --replicas=5 rs/<name>	Scale a ReplicaSet imperatively
kubectl edit rs <name>	Edit a live ReplicaSet object
kubectl set image deploy/<name> <container>=<image>:<tag>	Update container image
kubectl rollout history deploy/<name>	View deployment revision history
kubectl rollout undo deploy/<name>	Roll back to previous revision
kubectl create deploy <name> --image=<image> --dry-run=client -o yaml > deploy.yaml	Generate a Deployment YAML via dry-run

YAML Structure Highlight

A Deployment manifest contains:

• apiVersion: apps/v1
• kind: Deployment
• metadata (name, labels)
• spec.replicas — desired number of Pods
• spec.selector.matchLabels — which Pods to manage
• spec.template — the Pod template (metadata + spec.containers)

Exam Tips Emphasized

• Time is everything on the CKA exam. Prefer imperative commands (kubectl scale, kubectl create) over manual YAML editing when possible.
• Use --dry-run=client -o yaml to quickly generate manifest skeletons, then modify and apply.
• Know your shortcuts. The exam environment uses vi; practice Shift+A (append at end of line) and other vi navigation commands.
• The cheat sheet is your friend. kubectl <command> --help provides copy-pasteable examples.

Sources

• Tech Tutorials with Piyush — 40-Day CKA Course

See Also

Wiki Concepts

• Deployment, ReplicaSet & Replication Controller — Full technical synthesis of these workload controllers
• Pod Fundamentals — The smallest deployable unit that these controllers manage
• Kubernetes Architecture — Where the Controller Manager fits in the control plane
• CKA Certification — Exam overview and domains
• CKA Study Roadmap — 40-day learning plan
• Why Kubernetes? — Problems that self-healing and replication solve
• Docker Fundamentals — Container basics that underpin Pods
• Client-Server Architecture — Load balancing and high-availability patterns

Related Sources

• CKA Day 7 — Pod Explained — The prerequisite video on standalone Pods
• CKA Day 6 — Kind Cluster Setup — Local cluster used for the hands-on demo

Creator / Entity

• Tech Tutorials with Piyush — Creator of the 40-day CKA challenge series