This topic describes how to enable storage encryption for an existing table.
Limitations
You cannot enable encryption for the sys tenant.
Background information
The following sections describe how to enable storage encryption for the t1 table in the encrypted sectest_ts1 tablespace.
Enable storage encryption in internal mode
In internal mode, the encryption information of the master key is managed in internal tables, and clogs are not encrypted to avoid circular dependency during log replay.
Log on to a MySQL tenant as an administrator.
Execute the following statement to enable storage encryption in internal mode:
The
tde_methodparameter specifies the encryption method for a transparent tablespace. The default value isnone, which indicates that encryption is disabled for the transparent tablespace.For more information about the
tde_methodparameter, see tde_method.Notice
After the
tde_methodparameter is set, it cannot be modified.obclient> ALTER SYSTEM SET tde_method='internal';Execute the following statement to check whether the value of the
tde_methodparameter isinternalon all OBServer nodes of the tenant:obclient> SHOW PARAMETERS LIKE 'tde_method';If yes, execute the following statement to generate the master key:
Note
The statement takes effect only when the value of the
tde_methodparameter isinternalon all OBServer nodes of the tenant.obclient> ALTER INSTANCE ROTATE INNODB MASTER KEY;Create a tablespace and specify the encryption algorithm.
You can specify one of the following encryption algorithms:
aes-256,aes-128,aes-192, andsm4-cbc. If you set thesectest_ts1 encryptionparameter toy, theaes-256algorithm is used.For example:
obclient> CREATE TABLESPACE sectest_ts1 encryption = 'y';
Move an existing table into an encrypted tablespace
Log on to a MySQL tenant of the database as a regular user.
Log on to the database where the
t1table resides and move the table to thesectest_ts1tablespace.obclient> ALTER TABLE t1 TABLESPACE sectest_ts1;
Perform storage encryption on a table
Log on to a MySQL tenant of the database as a regular user.
Set the
progressive_merge_numparameter to configure a full compaction or progressive compaction task for the table.The
progressive_merge_numparameter specifies the number of rounds of progressive compactions on a table. The default value is0, which indicates incremental compaction. If you set the parameter to1, a full compaction is performed.In general, a full compaction is performed on a table. However, a full compaction of a table with a large amount of data may take a long time. In that case, we recommend that you perform a progressive compaction.
Perform a full compaction on the table
Set the
progressive_merge_numparameter to1.obclient> ALTER TABLE t1 set progressive_merge_num = 1;Log on to the
systenant as therootuser and manually initiate a major compaction.For more information, see Manually initiate a major compaction.
Note
After a full compaction, all table data is encrypted. You can query the
sys.v$encrypted_tablesview to check the encryption status.After the compaction is completed, set the
progressive_merge_numparameter to0.obclient> ALTER TABLE t1 set progressive_merge_num = 0;
Perform a progressive compaction on the table
Set the
progressive_merge_numparameter to a value greater than1, and run theOPTIMIZEcommand.For example:
obclient> ALTER TABLE t1 SET progressive_merge_num = 3; obclient> OPTIMIZE TABLE t1;Log on to the
systenant as therootuser, and manually initiate multiple rounds of progressive compactions to encrypt all the macroblocks of the table and table indexes.Execute the following statement to initiate a progressive compaction:
obclient> ALTER SYSTEM MAJOR FREEZE;The number of rounds of manual incremental compactions is
progressive_merge_num×max_kept_major_version_number.Note
During a progressive compaction, you can query the
sys.v$encrypted_tablesview to check the encryption status in real time.The
max_kept_major_version_numberparameter specifies the number of frozen versions to be retained. It can have a value between 1 and 16, with a default value of2.For more information about the
max_kept_major_version_numberparameter, see max_kept_major_version_number.
After the compaction is completed, you can execute the following statement to check the view and verify whether all macroblocks are encrypted:
For example:
obclient> SELECT * FROM oceanbase.v$encrypted_tables; +-----------+------------------+------------+------------------+---------------+-----------+----------------------------------+------------------+------------------+------------------+--------+ | TENANT_ID | TABLE_ID | TABLE_NAME | TABLESPACE_ID | ENCRYPTIONALG | ENCRYPTED | ENCRYPTEDKEY | MASTERKEYID | BLOCKS_ENCRYPTED | BLOCKS_DECRYPTED | STATUS | +-----------+------------------+------------+------------------+---------------+-----------+----------------------------------+------------------+------------------+------------------+--------+ | 1003 | 1102810162709330 | t1 | 1102810162660329 | aes-256 | YES | 4B0C00AFEFEEF506D4AB804E5071A733 | 1102810162660329 | 2 | 0 | NORMAL | +-----------+------------------+------------+------------------+---------------+-----------+----------------------------------+------------------+------------------+------------------+--------+ 1 row in setCheck the query results. If the value of the
BLOCKS_DECRYPTEDfield is0, all macroblocks are encrypted.For more information about fields in the
v$encrypted_tablesview, see v$encrypted_tables.