Resolving Insufficient Storage Issues for RDS for MariaDB Instances

In most cases, a table contains primary key indexes and secondary indexes. More secondary indexes mean that the table takes up more space.

Optimize the data structure of the table to reduce the number of secondary indexes.

If large fields of the binary large object (BLOB), TEXT, or VARCHAR data type are defined in the schema of a table, the table takes up a lot of space.

Compress data before you insert the data into the table.

If the fragmentation ratio of an InnoDB table is high, there will be an excessive number of idle tablespaces. InnoDB manages tablespaces by page. If some records of a full page are deleted and no new records are inserted into the positions these records were deleted from, a large number of tablespaces will be idle.

Run the show table status like '<Name of the table>'; command to query idle tablespaces. If there are too many tablespaces, run the optimize table '<Name of the table>'; command to manage the tablespaces.

When you perform a semi-join, distinct, or sort operation without using an index on a table, a temporary table is created. If the temporary table contains an excessive amount of data, the storage usage for the temporary table may be excessively high.

When you execute data definition language (DDL) statements to rebuild tablespaces that are used to store the data of a large table, the temporary table that is generated from an index-based sort operation will also be large.

View the plans that the DDL statements were based on. Check whether the Using Temporary field was specified.

Before you execute DDL statements on large tables, check whether your DB instance provides sufficient storage space. If the available storage space is insufficient, scale up the storage space of your DB instance before executing the statements.