Table of Contents

Standalone Cassandra Backup tool

The StandaloneCassandraBackup.exe tool can be used by an administrator to take a backup of a Cassandra database (either a single node or a cluster).

From DataMiner 10.1.8 onwards, this tool is available on each DMA server in the folder C:\Skyline DataMiner\Tools. As it only affects Cassandra files, it can be used on any DataMiner system regardless of version.

Note

When running one of the below-mentioned commands against a Cassandra cluster or a remote single-node Cassandra, make sure the Cassandra nodetool is accessible remotely on all nodes. For more information, see Making Cassandra nodetool accessible remotely.

Commands

listsnapshot

This command lists all snapshots of the specified Cassandra node.

The listsnapshot command can be run with the following arguments:

  • --host or -h: The IP address of one of the nodes in the cluster of which you want to take a backup.

    Default: localhost

  • --port or -p: The port used to connect to the remote nodetool host.

    Default: 7199

  • --username or -u: The user name used when connecting to the nodetool host.

    Typically, this argument will not be used.

  • --pw: The password used when connecting to the nodetool host.

    Typically, this argument will not be used.

  • --path: The full path to the location of the Cassandra nodetool.

    Default: C:\Program Files\Cassandra\bin\nodetool.bat

Example:

StandaloneCassandraBackup.exe listsnapshot -h 10.11.3.57

clearsnapshot

This command clears the specified snapshot and deletes its data.

Note

This command will automatically be run on every node of a Cassandra cluster.

The clearsnapshot command can be run with the following arguments:

  • --host or -h: The IP address of one of the nodes in the cluster of which you want to take a backup.

    Default: localhost

  • --port or -p: The port used to connect to the remote nodetool host.

    Default: 7199

  • --username or -u: The user name used when connecting to the nodetool host.

    Typically, this argument will not be used.

  • --pw: The password used when connecting to the nodetool host.

    Typically, this argument will not be used.

  • --nameor -n: The name of the snapshot that has to be cleared.

    Default: empty (which clears all snapshots)

  • --path: The full path to the location of the Cassandra nodetool.

    Default: C:\Program Files\Cassandra\bin\nodetool.bat

Example:

StandaloneCassandraBackup.exe clearsnapshot -h 10.11.3.57 -n snapshot1

refresh

This command refreshes the keyspaces of the specified Cassandra cluster as described in the nodetool documentation.

Note

This command will automatically be run on every node of a Cassandra cluster.

The refresh command can be run with the following arguments:

  • --host or -h: The IP address of one of the nodes in the cluster of which you want to take a backup.

    Default: localhost

  • --port or -p: The port used to connect to the remote nodetool host.

    Default: 7199

  • --username or -u: The user name used when connecting to the nodetool host.

    Typically, this argument will not be used.

  • --pw: The password used when connecting to the nodetool host.

    Typically, this argument will not be used.

  • --keyspaces or -k: The name of the keyspaces that have to be refreshed. Multiple keyspaces must be separated by spaces.

    Default: all keyspaces

  • --path: The full path to the location of the Cassandra nodetool.

    Default: C:\Program Files\Cassandra\bin\nodetool.bat

Example:

StandaloneCassandraBackup.exe refresh -h 10.11.3.57 –path C:\Program Files\nodetool.bat

backup

This command creates a snapshot of the specified Cassandra database.

In case of a Cassandra cluster, this command will first retrieve the IP addresses of all cluster nodes using a nodetool status command. Then, a nodetool snapshot command will be sent to each of the nodes. The output of both the status command and the snapshot command will be logged in the C:\Program Files\CassandraBackup\Logs folder, which will automatically be created the first time nodetool is used.

Note

This command will automatically be run on every node of a Cassandra cluster.

The backup command can be run with the following arguments:

  • --host or -h: The IP address of one of the nodes in the cluster of which you want to take a backup.

    Default: localhost

  • --port or -p: The port used to connect to the remote nodetool host.

    Default: 7199

  • --username or -u: The user name used when connecting to the nodetool host.

    Typically, this argument will not be used.

  • --pw: The password used when connecting to the nodetool host.

    Typically, this argument will not be used.

  • --name or -n: The name of the snapshot.

    If no name is provided, Cassandra will automatically generate a name based on the current timestamp.

  • --keyspaces or -k: The name of the keyspaces that have to be included in the snapshot. Multiple keyspaces must be separated by spaces.

    Default: all keyspaces

  • --path: The full path to the location of the Cassandra nodetool.

    Default: C:\Program Files\Cassandra\bin\nodetool.bat

Example:

StandaloneCassandraBackup.exe backup -h 10.11.3.57 -n snapshot21 –path C:\Program Files\nodetool.bat

Restoring a Cassandra backup

As restoring a Cassandra backup requires files to be moved, this cannot be done using the Standalone Cassandra Backup tool.

Proceed as follows to restore a Cassandra backup:

  1. Make sure the table schema exists. If it does not, then recreate it.

  2. Truncate the tables you want to restore. For more information, see the CQL documentation.

  3. Locate the snapshot you want to restore. It will be located in the following folder:

    data_directory/keyspace_name/table_name-UUID/snapshots/

  4. Copy the most recent snapshot folder to the following folder:

    data_directory/keyspace/table_name-UUID

  5. Run nodetool refresh and restart your nodes.

Propose changes