Restore Database to a Single Node & Replicate it to Other Nodes
The common approach to restoring a cluster is to restore the backed-up database to a single server, and then expand the database group to additional nodes, allowing normal replication.
Restore Database to Multiple Nodes Simultaneously
You can create the cluster in advance, and restore the database to multiple nodes simultaneously.
This procedure is advisable only when restoring a Snapshot.
- Simultaneously restoring a logical-backup by multiple nodes, triggers each node to send change-vector updates to all other nodes.
- Simultaneously restoring a snapshot does not initiate this behavior, because the databases kept by all nodes are considered identical.
- On the first node, restore the database using its original name.
- On other nodes, restore the database using different names.
- Wait for the restoration to complete on all nodes.
Soft-delete the additional databases (those with altered names).
- Delete the databases from the cluster with
HardDelete set to
false, to retain the data files on disk.
- Rename the database folder on all nodes to the original database name.
- Expand the database group to all relevant nodes.