Skip to content

Encrypted InnoDB tablespace backups

InnoDB supports data encryption for InnoDB tables stored in file-per-table tablespaces. This feature provides an at-rest encryption for physical tablespace data files.

For an authenticated user or an application to access an encrypted F tablespace, InnoDB uses the master encryption key to decrypt the tablespace key. The master encryption key is stored in a keyring.

Percona XtraBackup supports the following keyring components which are stored in the plugin directory:

Percona XtraBackup 8.4 and later versions do not support any plugin versions of the security features.

Note

Enable only one keyring component simultaneously for each server instance. Enabling multiple keyring components is not supported and may result in data loss.

Use the keyring vault component

The keyring_vault component extends the server capabilities. The server uses a manifest to load the component and the component has its own configuration file.

The following example is a global manifest file that does not use local manifests:

{
 "read_local_manifest": false,
 "components": "file://component_keyring_vault"
}

The following example of a global manifest file that points to a local manifest file:

{
 "read_local_manifest": true
}

The following example of a local manifest file:

{
 "components": "file://component_keyring_vault"
}

The configuration settings can be in either a global or a local configuration file.

Example of a configuration file in JSON format
{
 "timeout": 15,
 "vault_url": "https://vault.public.com:8202",
 "secret_mount_point": "secret",
 "secret_mount_point_version": "AUTO",
 "token": "58a20c08-8001-fd5f-5192-7498a48eaf20",
 "vault_ca": "/data/keyring_vault_confs/vault_ca.crt"
}

Find more information on configuring the keyring_vault component in Use the keyring vault component.

The component has no special requirements for backing up a database that contains encrypted InnoDB tablespaces.

The following command creates a backup in the /data/backup directory:

$ xtrabackup --backup --target-dir=/data/backup --user=root

After xtrabackup completes the action, the following message confirms the action:

Confirmation message
xtrabackup: Transaction log of lsn (5696709) to (5696718) was copied.
160401 10:25:51 completed OK!

Prepare the backup with the keyring vault component

To prepare the backup, the xtrabackup binary must access the keyring. The xtrabackup binary does not communicate with the MySQL server or read the default my.cnf configuration file. Specify the keyring settings in the command line:

$ xtrabackup --prepare --target-dir=/data/backup --component-keyring-config==/etc/vault.cnf

The following message confirms that the xtrabackup binary completed the action:

Confirmation message
InnoDB: Shutdown completed; log sequence number 5697064
160401 10:34:28 completed OK!

Restore the backup

As soon as the backup is prepared, you can restore it with the --copy-back option:

$ xtrabackup --copy-back --target-dir=/data/backup --datadir=/data/mysql \
--transition-key=MySecretKey --generate-new-master-key \
--component-keyring-config=/etc/vault.cnf

Use the keyring file component

The keyring_file component is part of the component-based MySQL infrastructure which extends the server capabilities.

The server uses a manifest to load the component, and the component has its own configuration file. See the MySQL documentation on the component installation for more information.

An example of a manifest and a configuration file follows:

An example of ./bin/mysqld.my:

{
   "components": "file://component_keyring_file"
}

An example of /lib/plugin/component_keyring_file.cnf:

{
   "path": "/var/lib/mysql-keyring/keyring_file", "read_only": false
}

For more information, see Keyring Component Installation.

With the appropriate privilege, you can SELECT on the performance_schema.keyring_component_status table to view the attributes and status of the installed keyring component when in use.

The component has no special requirements for backing up a database that contains encrypted InnoDB tablespaces.

$ xtrabackup --backup --target-dir=/data/backup --user=root

The following message confirms that the xtrabackup binary completed the action:

Confirmation message
xtrabackup: Transaction log of lsn (5696709) to (5696718) was copied.
160401 10:25:51 completed OK!

Warning

xtrabackup does not copy the keyring file into the backup directory. To prepare the backup, you must copy the keyring file manually.

Prepare the backup with the keyring_file component

xtrabackup reads the keyring_file component configuration from xtrabackup_component_keyring_file.cnf. You must specify the keyring_file data path if the --component-keyring-config is not located in the attribute PATH from the xtrabackup_component_keyring_file.cnf.

The following is an example of adding the location for the –component-keyring-config:

$ xtrabackup --prepare --target-dir=/data/backup \
--component-keyring-config=/var/lib/mysql-keyring/keyring

xtrabackup attempts to read xtrabackup_component_keyring_file.cnf. You can assign another keyring file component configuration by passing the --component-keyring-config option.

The following message confirms that the xtrabackup binary completed the action:

Confirmation message
InnoDB: Shutdown completed; log sequence number 5697064
160401 10:34:28 completed OK!

The backup is prepared. To restore the backup use the --copy-back option. If the keyring has been rotated, you must restore the specific keyring used to take and prepare the backup.

Incremental encrypted InnoDB tablespace backups with keyring file component

The process of taking incremental backups with InnoDB tablespace encryption is similar to taking incremental backups with unencrypted tablespace.

Create an incremental backup

To make an incremental backup, begin with a full backup. The xtrabackup binary writes xtrabackup_checkpoints into the backup’s target directory. This file contains a line showing the to_lsn, which is the LSN of the database at the end of the backup. First, you need to create a full backup with the following command:

$ xtrabackup --backup --target-dir=/data/backups/base

If you look at the xtrabackup_checkpoints file, you should see the output similar to the following:

Expected output
backup_type = full-backuped
from_lsn = 0
to_lsn = 7666625
last_lsn = 7666634
compact = 0
recover_binlog_info = 1

Now that you have a full backup, you can make an incremental backup based on it. Use a command such as the following:

$ xtrabackup --backup --target-dir=/data/backups/inc1 \
--incremental-basedir=/data/backups/base \

The /data/backups/inc1/ directory should now contain delta files, such as ibdata1.delta and test/table1.ibd.delta. These represent the changes since the LSN 7666625. If you examine the xtrabackup_checkpoints file in this directory, you should see the output similar to the following:

Expected output
backup_type = incremental
from_lsn = 7666625
to_lsn = 8873920
last_lsn = 8873929
compact = 0
recover_binlog_info = 1

Use this directory as the base for yet another incremental backup:

$ xtrabackup --backup --target-dir=/data/backups/inc2 \
--incremental-basedir=/data/backups/inc1 \

Prepare incremental backups

The --prepare step for incremental backups is not the same as for normal backups. In normal backups, two types of operations are performed to make the database consistent:

  • Committed transactions are replayed from the log file against the data files

  • Uncommitted transactions are rolled back

You must skip the rollback of uncommitted transactions when preparing a backup because transactions that were uncommitted at the time of your backup may be in progress, and they will likely be committed in the next incremental backup. Use the --apply-log-only option to prevent the rollback phase. Your incremental backups are useless if you do not use the --apply-log-only option to prevent the rollback phase. After transactions have been rolled back, further incremental backups cannot be applied.

Beginning with the full backup you created, you can prepare it and then apply the incremental differences. Recall that you have the following backups:

/data/backups/base
/data/backups/inc1
/data/backups/inc2

To prepare the backup, you must copy the keyring file manually. The xtrabackup binary does not copy the keyring file into the backup directory.

If the keyring has not been rotated you can use the same one you’ve backed-up with the base backup. If the keyring has been rotated, or you have upgraded the plugin to a component, you must back up the keyring file. Otherwise, you cannot prepare the backup.

To prepare the base backup, you need to run --prepare as usual, but prevent the rollback phase:

$ xtrabackup --prepare --apply-log-only --target-dir=/data/backups/base \
--component-keyring-config=/var/lib/mysql-keyring/keyring
Expected output
InnoDB: Shutdown completed; log sequence number 7666643
InnoDB: Number of pools: 1
160401 12:31:11 completed OK!

To apply the first incremental backup to the full backup, you should use the following command:

$ xtrabackup --prepare --apply-log-only --target-dir=/data/backups/base \
--incremental-dir=/data/backups/inc1 \
--component-keyring-config=/var/lib/mysql-keyring/keyring

The backup should be prepared with the keyring file and type that was used when the backup was being taken. This means that if the keyring has been rotated, or you have upgraded from a plugin to a component between the base and incremental backup, you must use the keyring used when the first incremental backup was taken.

Preparing the second incremental backup is a similar process: apply the deltas to the (modified) base backup, and you will roll its data forward in time to the point of the second incremental backup:

$ xtrabackup --prepare --target-dir=/data/backups/base \
--incremental-dir=/data/backups/inc2 \
--component-keyring-config=/var/lib/mysql-keyring/keyring
Use --apply-log-only when merging all incremental backups except the last one. The previous line does not contain the --apply-log-only option. Even if the --apply-log-only option was used on the last step, the backup would still be consistent, but in that case, the server would perform the rollback phase.

The backup is now prepared and can be restored with --copy-back option. In case the keyring has been rotated you’ll need to restore the keyring which was used to take and prepare the backup.

Restore a backup when the keyring is not available

While this works, the method requires access to the same keyring that the server is using. It may not be possible if the backup is prepared on a different server or at a much later time, when keys in the keyring are purged, or, in the case of a malfunction, when the keyring vault server is unavailable.

The --transition-key=<passphrase> option should be used to make it possible for the xtrabackup binary to process the backup without access to the keyring vault server. In this case, the binary derives the AES encryption key from the specified passphrase and will use it to encrypt tablespace keys of tablespaces being backed up.

Create a backup with a passphrase

The following example illustrates how the backup can be created in this case:

$ xtrabackup --backup --user=root -p --target-dir=/data/backup \
--transition-key=MySecretKey

If --transition-key is specified without a value, xtrabackup will ask for it.

xtrabackup scrapes --transition-key so that its value is not visible in the ps command output.

Prepare a backup with a passphrase

The same passphrase should be specified for the prepare command:

$ xtrabackup --prepare --target-dir=/data/backup \
--transition-key=MySecretKey

There are no --keyring-vault...,``–keyring-file…``, or --component-keyring-config options here, because xtrabackup does not talk to the keyring in this case.

Restore a backup with a generated key

When restoring a backup you need to generate a new master key.

The example for keyring_file component:

$ xtrabackup --copy-back --target-dir=/data/backup --datadir=/data/mysql \
--transition-key=MySecretKey --generate-new-master-key \
--component-keyring-config=/var/lib/mysql-keyring/keyring

The example for keyring_vault component:

$ xtrabackup --copy-back --target-dir=/data/backup --datadir=/data/mysql \
--transition-key=MySecretKey --generate-new-master-key \
--component-keyring-config=/etc/vault.cnf

xtrabackup generates a new master key, stores it in the target keyring vault server, and re-encrypts the tablespace keys using this key.

Make a backup with a stored transition key

Finally, there is an option to store a transition key in the keyring. In this case, xtrabackup must access the same keyring file or vault server during prepare and copy-back steps, but does not depend on whether the server keys have been purged.

In this scenario, the three stages of the backup process look as follows.

  • Back up
$ xtrabackup --backup --user=root -p --target-dir=/data/backup \
--generate-transition-key
  • Prepare

    • keyring_file component variant:
    $ xtrabackup --prepare --target-dir=/data/backup \
    --component-keyring-config=/var/lib/mysql-keyring/keyring
    
    • keyring_vault component variant:
    $ xtrabackup --prepare --target-dir=/data/backup \
    --component-keyring-config=/etc/vault.cnf
    
  • Copy-back

    • keyring_file component variant:
    $ xtrabackup --copy-back --target-dir=/data/backup --datadir=/data/mysql \
    --generate-new-master-key --component-keyring-config=/var/lib/mysql-keyring/keyring
    
    • keyring_vault component variant:
    $ xtrabackup --copy-back --target-dir=/data/backup --datadir=/data/mysql \
    --generate-new-master-key --component-keyring-config=/etc/vault.cnf
    

Get expert help

If you need assistance, visit the community forum for comprehensive and free database knowledge, or contact our Percona Database Experts for professional support and services.


Last update: 2024-09-13