When an administrator updates an enterprise application that is installed on a running server, the product manages the replacement itself. The update replaces the application files in the configuration repository, the changes are synchronised to the nodes, and each affected server then applies them. Where the server has to restart the application to load the new version, it stops the application, replaces the files, and starts it again without administrator action. Option C describes that sequence.
The other options misdescribe the behaviour. Option A, no interruption, applies only when the change can be handled through fine-grained operations and the application does not have to be restarted, and it does not describe the general case; users can still experience a brief outage or a lost session when classes are reloaded. Option B suggests that a manual redeploy is required, which is not the case, since the update is itself the deployment step. Option D is incorrect because an automatic restart takes place, and the administrator is not required to start the application by hand.
Administrators concerned with availability use a clustered topology and a rolling update, taking members out of service in sequence, so that the application remains available while each member is updated. Session persistence or replication keeps in-flight sessions from being lost during the restarts.
[Reference: Updating enterprise applications; Fine-grained updates (Replace, add, or delete files); Application restart after update.]
Submit