An external Cassandra database is required when Rhino is configured to provide a key/value store or session ownership store, or when configured to use the pool clustering mode.

These subsystems and functions all require certain keyspaces and tables to be present in the database for them to be able to store and query data. The key/value store and session ownership store each may be configured to automatically create and remove the keyspaces and tables they require at runtime, however this ability is only supported when using the Savanna clustering mode.

When Rhino is configured to use the pool clustering mode, or automatic data definition updates are disabled for the key/value store or session ownership store, then an Administrator must manually create the necessary keyspaces and tables in the database instead.

To do this, follow the steps below:


Start a Rhino node

The Rhino SLEE must first be started such that a management client can connect to the node. It is expected at this point that Rhino will raise an alarm for any keyspace or table that is required to exist in the database but doesn’t. For now, ignore these alarms.


Export persisting resource data definitions from Rhino

The key/value store, session ownership store, and pool maintenance subsystem (used to support the pool clustering mode) are known as persisting resources — internal resources that persist state to an external database. The data definition update statements that these persisting resources use to configure the database schema can be exported from Rhino as described in exporting persisting resource data definitions.


Execute the data definition updates

Use the cqlsh (CQL Shell for Cassandra) tool to connect to the Cassandra database and execute the data definition updates against the database to create the keyspaces and tables.

If the data definitions have been exported from Rhino as a zip file, then these statements can be executed against the database using a single command-line operation such as this:

$ unzip -p | cqlsh

Once all the keyspaces and tables have been created, all corresponding Rhino alarms should clear after a few seconds.

Note It’s possible that subsequent deployment of services or creation of resource adaptor entities in the Rhino SLEE may require new Cassandra keyspaces and tables. If this happens, Rhino will raise new alarms to indicate what is missing. To rectify this, simply export from Rhino a fresh set of data definitions and apply the updated definitions to the database again, as described above. This process can be repeated as often as necessary.
Previous page Next page