Compare Deployment and StatefulSet in OpenShift: when would you choose each?

• A. Use Deployment for stateless apps; StatefulSet for stateful apps requiring stable identities and persistent storage.
• B. Use StatefulSet for stateless; Deployment for stateful.
• C. Use Deployment for long-running jobs; StatefulSet for ephemeral.
• D. Use Deployment for secrets; StatefulSet for config maps.

Answer: (A)

The choice hinges on stateless versus stateful workloads. Deployments are ideal for stateless apps where each pod can be freely replaced, scaled, or updated without needing to preserve identity or local state. They don’t guarantee a stable network identity or persistent storage across restarts, which keeps them simple and highly scalable for front-end services, APIs, and other non-stateful components.

StatefulSets, on the other hand, are designed for stateful workloads that must keep stable identities and durable storage. Each pod gets a predictable, unique name and a stable network identity, and you can attach persistent volumes that remain associated with that pod across rescheduling. This makes StatefulSets suitable for databases, caches, and other systems where preserving state and identity is essential, and where ordered deployment, scaling, and upgrades are important.

So, use a Deployment for stateless applications; use a StatefulSet for stateful applications requiring stable identities and persistent storage. The other options don’t fit because using StatefulSet for stateless workloads adds unnecessary complexity, Deployments aren’t the pattern for long-running jobs (Jobs/CronJobs are), and secrets/config maps aren’t a differentiator between these two controllers.