Create a log file group, and assign one or more undo log files to it (an undo log file is also sometimes referred to as an undofile).

Create a tablespace; assign the log file group, as well as one or more data files, to the tablespace.

Create a Disk Data table that uses this tablespace for data storage.

We create a log file group named lg_1 using CREATE LOGFILE GROUP. This log file group is to be made up of two undo log files, which we name undo_1.log and undo_2.log, whose initial sizes are 16 MB and 12 MB, respectively. (The default initial size for an undo log file is 128 MB.) Optionally, you can also specify a size for the log file group's undo buffer, or permit it to assume the default value of 8 MB. In this example, we set the UNDO buffer's size at 2 MB. A log file group must be created with an undo log file; so we add undo_1.log to lg_1 in this CREATE LOGFILE GROUP statement:

CREATE LOGFILE GROUP lg_1
    ADD UNDOFILE 'undo_1.log'
    INITIAL_SIZE 16M
    UNDO_BUFFER_SIZE 2M
    ENGINE NDBCLUSTER;

To add undo_2.log to the log file group, use the following ALTER LOGFILE GROUP statement:

ALTER LOGFILE GROUP lg_1
    ADD UNDOFILE 'undo_2.log'
    INITIAL_SIZE 12M
    ENGINE NDBCLUSTER;

Some items of note:

Now we can create a tablespace, which contains files to be used by NDB Cluster Disk Data tables for storing their data. A tablespace is also associated with a particular log file group. When creating a new tablespace, you must specify the log file group which it is to use for undo logging; you must also specify a data file. You can add more data files to the tablespace after the tablespace is created; it is also possible to drop data files from a tablespace (an example of dropping data files is provided later in this section).

Assume that we wish to create a tablespace named ts_1 which uses lg_1 as its log file group. This tablespace is to contain two data files named data_1.dat and data_2.dat, whose initial sizes are 32 MB and 48 MB, respectively. (The default value for INITIAL_SIZE is 128 MB.) We can do this using two SQL statements, as shown here:

CREATE TABLESPACE ts_1
    ADD DATAFILE 'data_1.dat'
    USE LOGFILE GROUP lg_1
    INITIAL_SIZE 32M
    ENGINE NDBCLUSTER;

ALTER TABLESPACE ts_1
    ADD DATAFILE 'data_2.dat'
    INITIAL_SIZE 48M
    ENGINE NDBCLUSTER;

The CREATE TABLESPACE statement creates a tablespace ts_1 with the data file data_1.dat, and associates ts_1 with log file group lg_1. The ALTER TABLESPACE adds the second data file (data_2.dat).

Some items of note:

Now it is possible to create a table whose nonindexed columns are stored on disk in the tablespace ts_1:

CREATE TABLE dt_1 (
    member_id INT UNSIGNED NOT NULL AUTO_INCREMENT PRIMARY KEY,
    last_name VARCHAR(50) NOT NULL,
    first_name VARCHAR(50) NOT NULL,
    dob DATE NOT NULL,
    joined DATE NOT NULL,
    INDEX(last_name, first_name)
    )
    TABLESPACE ts_1 STORAGE DISK
    ENGINE NDBCLUSTER;

The TABLESPACE ... STORAGE DISK option tells the NDBCLUSTER storage engine to use tablespace ts_1 for disk data storage.

Once table ts_1 has been created as shown, you can perform INSERT, SELECT, UPDATE, and DELETE statements on it just as you would with any other MySQL table.

For table dt_1 as it has been defined here, only the dob and joined columns are stored on disk. This is because there are indexes on the id, last_name, and first_name columns, and so data belonging to these columns is stored in RAM. In NDB Cluster 7.3 and NDB Cluster 7.4, only nonindexed columns can be held on disk; indexes and indexed column data continue to be stored in memory. This tradeoff between the use of indexes and conservation of RAM is something you must keep in mind as you design Disk Data tables.