Cassandra Restore Operations

You can restore the following data:

You can restore the Cassandra commit log in addition to the data.

For Cassandra configurations, you can perform an in-place (to the same destination), or an out-of-place (to a new destination) restore operation.

For Scylla configurations, you can perform an in-place (to the same destination) restore operation.

The Commvault software also has the following options:

Have the Commvault software invoke the Cassandra bulk loader or sstableloader to load the data into the database so that users do not have to manually run the sstableloader
Perform a restore with or without the sstableloader
Replace a dead node
Perform a disaster recovery
Do not recover the database. After the restore is complete, manually run the sstableloader command to recover the database
Perform a stage-free recovery. In this method, you can eliminate the need to restore the backup data to a staging location. The stage-free recovery process exports the backup data as 3DFS network share, and mounts the share on the client nodes. The sstableloader is then invoked from the mounted directory. You cannot perform a stage-free restore operation for a docker deployment.

Selecting the Data to Restore

The Commvault software provides a granular browse option.

The Browse page displays all the cluster nodes in the left pane.

Expand the nodes to view the keyspaces, and the ColumFamily entities that keyspace contains.

To include data from all the nodes in the restore, perform the following operations:

Entity	Steps
Keyspace	Right-click the keyspace, and then click Select Keyspace from all nodes.
ColumnFamily	Right-click the ColumnFamily, and then click Select ColumnFamily from all nodes.

You can restore to the most recent commit log when you restore the data

If you want to restore the commit log to a point in time, then you must browse to that point-in-time.

You can restore the commit log to different clusters that have the same Cassandra version and also to a new Cassandra instance.