A systems administrator needs to apply critical security patches to a cluster of load-balanced web servers. The primary requirement is to keep the web application available to users with no perceptible downtime while using only the current server cluster (no duplicate production environment may be provisioned). Which of the following deployment strategies should the administrator use to meet this requirement?
The correct answer is a rolling deployment. By taking one server (or a small subset) out of the load balancer's pool, patching it, verifying its health, and returning it to service before moving to the next, the administrator maintains continuous availability with the existing infrastructure.
A big-bang deployment updates every server at once, producing a full outage.
A recreate deployment shuts down all old instances before starting new ones, which also causes downtime.
A blue-green deployment can achieve zero downtime, but it requires running a duplicate set of servers-disallowed in this scenario. Rolling deployment therefore best satisfies both the no-downtime and no-additional-infrastructure constraints.