Failover

Full recovery (failover) recovers protected workloads as native instances on the recovery site after a disaster — or as a planned promotion. Run it from the recovery site's management console: Configuration → Protection Plan → (plan) → Actions → Recover.

Run a full recovery

1. Select virtual machines

   Choose the VMs to recover and the recovery source:

   • Latest — recovers from the most recent replicated copy.
   • Point In Time — recovers from an available checkpoint; pick a common checkpoint for all VMs or per-VM checkpoints.

   If pre/post scripts are configured, provide guest credentials (username/password or a credentials file). Credentials are used only for the workflow and never stored; without scripts they are optional.

2. Tools and scripts

   • Install System Agents — installs EC2Launch2 + SSM Agent (AWS), the google-compute-engine agent set (GCP), the Azure VM agent (Azure), or VMware Tools (VMware) into recovered guests. Required when assigning static IPs on VMware targets.
   • Install Cloud Packages — installs the AWS or Google Cloud SDK where applicable.
3. Summary

   Review and click Finish. Recovery jobs start for the selected VMs; monitor each job's current step under Jobs → Recovery.

After failover

• Verify each recovered VM: recovery jobs expose the instance IP (RDP file download for Windows, SSH for Linux). The Windows-only validation step flags guests needing attention; Refresh recovery status re-checks status against the platform.
• VMs boot in the configured boot order with the configured delay.
• When the original site is back, return workloads with failback (reverse replication).
• Recovered copies that are no longer needed can be removed with Actions → Cleanup Recoveries.

Related docs

• Failback

• Recovery testing

• Status check