Kubernetes · Orchestration under pressure
What is a ReplicaSet and how does it reconcile replicas?
Short answer
A ReplicaSet ensures a specified number of matching pods exist. It creates or deletes pods to match its selector and replica count, and Deployments use it to manage revisions.
A ReplicaSet has a label selector, a desired replica count and a pod template. Its controller lists pods matching the selector and compares the count with the desired count. Shortfalls cause creations; excess matching pods are deleted.
- The selector is immutable after creation on most API versions
- Adopting existing pods is possible if labels match, which can be dangerous
- A ReplicaSet has no rollout strategy; a Deployment adds that
- ReplicaSets do not guarantee pod identity or ordering
- Deleting a Deployment cascades to its ReplicaSets and pods
kubectl get rs -l app=api
kubectl describe rs api-7c9d8f6b5
kubectl get pods -l app=api --show-labelsIn day-to-day work you operate Deployments, not ReplicaSets. The ReplicaSet is an implementation detail that becomes visible during rollouts, rollbacks and debugging when you need to see which template revision is producing pods.
Follow-ups you should expect
Interviewers use these to check whether the first answer was understanding or memorisation.
- 1Why is the ReplicaSet selector immutable?
- 2How can a ReplicaSet adopt an existing pod?
- 3What is the difference between a Deployment and a ReplicaSet?
- 4How do you find which ReplicaSet owns a pod?
Say it in the interview
A ReplicaSet is a controller that keeps a fixed number of pods matching a selector. It creates or deletes pods to close the gap, but it has no rollout strategy or identity guarantees. In practice Deployments own ReplicaSets and I only inspect them during rollout debugging. I keep selectors specific because accidental adoption of pods is a real operational hazard.
Use this as a model, not a script. Saying it in your own words is what makes it survive follow-up pressure.
Related questions
Reading is step one
Now say it out loud.
Understanding an answer and delivering it under pressure are different skills. Rehearse this topic in a live Kubernetes mock interview and get scored feedback on depth, clarity and confidence.