How Instant Recovery Works

To perform instant recovery, Veeam Explorer for Microsoft SQL Server uses the Veeam Explorers Recovery Service. The service is installed on the backup server and every mount server deployed in the backup infrastructure, and runs in the background. During instant recovery, Veeam Explorer for Microsoft SQL Server uses the service running on the mount server associated with the backup repository. This removes the necessity to keep Veeam Explorer for Microsoft SQL Server running during an entire instant recovery session. Once you have started an instant recovery session, you can close Veeam Explorer for Microsoft SQL Server and then launch it later when necessary.

When you start an instant recovery session in Veeam Explorer for Microsoft SQL Server, the following happens:

  1. Veeam Explorer for Microsoft SQL Server uses backup files to start 2 iSCSI mount sessions in parallel:
    • The first mount is used to publish the database on the target production server and attach it to the target SQL Server instance.
    • The second mount is used to copy database files to the target server in the background.
  1. While the published database is online, all changes to the published database are saved in the write cache on the mount server.

You can view and change the write cache location in the mount server settings. For more information, see the Specify Mount Server Settings section of the Veeam Backup & Replication User Guide.

  1. After all database files are copied to the target server, Veeam Explorer for Microsoft SQL Server uses the write cache to synchronize changes in database files.
  2. After synchronization, you can switch over to the up-to-date copy of the database on the production server. Depending on the option you choose, switchover starts in one of the following ways:
  • Automatically, immediately after synchronization
  • Automatically, according to a specified schedule
  • Manually

During switchover, Veeam Explorer for Microsoft SQL Server does the following:

  1. Stops the published database.
  2. Uses the write cache to synchronize differences between the published database and the copied database files.
  3. Drops the published database.
  4. Starts the recovered database.

Instant recovery session is resilient to network disruption, backup server or mount server crashes. If anything disrupts the restore process, the restore process stays in the waiting mode and performs 10 automatic retries every 5 minutes. If the retries fail, you can launch retry after the server or network is up.