How Orchestrator Manages Storage Failover
Storage failover is a process of switching your virtual infrastructure from the source storage system in the production site to the destination storage system in the DR site. During storage failover, fully functional VMs are recovered to the required snapshots on the specified target storage system — as a result, you can access services and applications you need with minimum disruption.
Orchestrator orchestrates storage failover in the following way when running a NetApp storage plan:
- Before Orchestrator starts processing VM groups included in the plan, it performs a number of pre-plan steps to prepare the failover environment:
- Orchestrator runs the VM Power Actions step to shut down source VMs running in the production vCenter Server. To answer VM questions that appear while shutting the VMs down, Orchestrator applies default answers specified in the vCenter Server settings.
The VM Power Actions step has a preconfigured timeout parameter that defines the exact amount of time for the step to execute. If step execution time exceeds the defined parameter value, Orchestrator powers the source VMs off.
- Orchestrator runs the Storage Failover step. It breaks the SnapMirror relationship between the source and destination storage volumes, mounts the destination volumes to the target vCenter Server, and then mounts the recovered datastores to all hosts in the required storage recovery location.
For more information on the way Orchestrator identifies storage recovery locations required to recover datastores, see How Orchestrator Processes Resources.
- When processing VM groups, Orchestrator registers target VMs on hosts in the DR site, and then powers the VMs on.
For more information on the way Orchestrator defines hosts where recovered VMs will be registered, see How Orchestrator Processes Resources.
- After Orchestrator finishes processing VM groups, it performs a number of post-plan steps to finalize the storage failover process:
- Orchestrator runs the Unregister VMs step to unregister source VMs from hosts in the production site.
- Orchestrator runs the Unmount Datastore step to unmount the source volumes from the source vCenter Server.
- [This step applies only if you have selected the Reprotect storage volumes after failover check box at the Reprotect Volumes step of the Run Plan wizard]
Orchestrator runs the Protect Storage Volumes step to reprotect volumes included in the plan by resynchronizing the data protection relationship in the reverse direction.
When Orchestrator orchestrates storage failover, it handles specific internal elements under the hood — protection groups and storage items:
Protection groups and storage items were introduced into Orchestrator to exclusively support the storage failover process. That is why the Orchestrator UI does not show these elements, but you may come across them in some reports and log entries.