Recovery plan concepts

An ordered list of steps that defines how a group of source servers is recovered. A plan belongs to a single AWS account and Region, and its name must be unique within that account and Region.

One entry in the ordered list. A step is either a server step or a wait step. Steps run one at a time, in order.

A step that contains one or more source servers. AWS Elastic Disaster Recovery starts recovery for every server in the step at the same time and considers the step finished only after every server has finished.

A step that pauses the plan for a fixed number of minutes before the next step starts. Use a wait step to give an application time to initialize after its instances launch. A wait step cannot be the first step in a plan.

A per-server setting on a server step that determines whether the failure of that server stops the plan. A Critical server (the default) fails the step and the execution. An Optional server does not. For more information, see Controlling how server failures affect a plan.

A single run of a recovery plan. An execution is a point-in-time copy of the plan: it records the steps and servers as they were when the execution started, so later edits to the plan do not change a run that is already in progress.

Whether the run is a Drill or a Recovery. The mode applies to every server step in the plan and is equivalent to choosing between a drill and a recovery for an individual source server.