Changes in MySQL Enterprise Backup 8.0.18 (2019-10-14, General Availability)

MySQL Enterprise Backup 8.0.18 is the latest release for MySQL Enterprise Backup, and supports MySQL Server 8.0.18 only. For earlier versions of MySQL 8.0, use the MySQL Enterprise Backup version with the same version number as the server. For MySQL Server 5.7, please use MySQL Enterprise Backup 4.1; for MySQL Server 5.6 and 5.5, please use MySQL Enterprise Backup 3.12.

• Functionality Added or Changed

• Bugs Fixed

Functionality Added or Changed

• Close to the end of a backup, mysqlbackup applied the FLUSH TABLES tbl_name [, tbl_name] ... WITH READ LOCK statement on all non-InnoDB tables. With this fix, the statement is not applied to tables that are not included in a partial backup, thus avoiding unncessary locks for some tables. (Bug #29873048)

• The International Components for Unicode (ICU) library for regular expressions, used by MySQL Server 8.0, has now replaced the RE2 library as the library used by MySQL Enterprise Backup for handling regular expressions.

  This change also removes a memory leak associated with the use of the Partial Backup and Restore Options. (Bug #27374240, Bug #29840790, WL #13137)

• The image-to-backup-dir command is now an alias for the extract command. (WL #13278)

• Two new options, --compression-algorithms and --zstd-compression-level, have been introduced for configuring compression for server connections. See Connection Options and Command Options for Connection Compression for details. (WL #1320)

• MySQL Enterprise Backup now supports a faster way to create incremental backups by using the page tracking functionality on MySQL Servers. To use this new feature, set --incremental=page-track. See Incremental Backup Using Page Tracking for details. (WL #10223)

• The --uncompress option is now supported for the extract operation: Files from a compressed single-file backup can now be extracted and uncompressed with a single command. (WL #13250)

Bugs Fixed

• Backups on a Group Replication primary node failed when mysqlbackup tried to upgrade the mysql.backup_history table on the node (see Backup History Table Update for details), and there was a problem causing the node to switch to super_read_only mode. With this fix, mysqlbackup checks if the node is in super_read_only mode during the backup_history table upgrade and, if so, skips the steps in the upgrade process that caused the problem. (Bug #30065637)

• After a partial backup created with the --only-innodb option was restored, the server failed to start due to an assertion error if the clone plugin or thread pool plugin was loaded during the server startup. It was because the backup created with the --only-innodb option did not include the performance schema, and the plugin was looking for the data folder for the performance schema when the restored server was being started. This fix prevents the problem by having mysqlbackup create an empty performance_schema folder under the backup's data directory when creating backups with --only-innodb. (Bug #29999075)

• Wrong value for the --safe-slave-backup-timeout option was printed when the --help option was used with mysqlbackup, if --safe-slave-backup-timeout was set with an unsigned integer. (Bug #29994968)

• A backup failed for a server containing encrypted InnoDB tables if the server was started with --skip-grant-tables. It was because the server only accepted connection through a Unix socket in that case, but mysqlbackup did not pass the value of --keyring-migration-socket to the keyring migration server; this fix makes mysqlbackup pass the option whenever it connects to a server to be backed up with a socket. (Bug #29954367)

• When restoring a full backup of a slave server in a replication setup, mysqlbackup gave warnings for missing relay log information file and master information file even if those files were never used on the backed-up server. This fix removes the unnecessary warnings. (Bug #29941160)

• When the --compress-method option was used at an apply-log operation, the operation failed, and the backup was corrputed. With this fix, mysqlbackup throws an error and quits whenever any compression options are used with apply-log. (Bug #29941117)

• At the end of any apply-log operation, mysqlbackup printed the message INFO: Backup was originally taken with the --include regexp option whenever a partial backup option that made use of regular expressions was used during the backup. With this fix, under the situation, mysqlbackup simply mentions in the log message that the backup is a partial one. (Bug #29872975)

• When a base backup was restored with --skip-relaylog but a subsequent incremental image backup restore did not use the option, the relay log files were copied from the incremental image backup onto the server. With this fix, an incremental image restore, with or without using the --skip-relaylog option, does not restore the relay log files whenever its base backup did not restore the relay log. (Bug #29864964)

• A memory leak that occurred with backups involving encrypted redo log has been removed. (Bug #29841265)

• A memory leak caused by the use of the --optimistic-time option has been removed. (Bug #29841031)

• When a base backup was restored with --skip-binlog but a subsequent incremental image backup restore did not use the option, the binary log files were copied from the incremental image backup onto the server, and the server could not be started because of the incomplete binary log. With this fix, an incremental image restore, with or without using the --skip-binlog option, does not restore the binary log files whenever its base backup did not restore the binary log. (Bug #29802632)

• mysqlbackup hung when performing an incremental or compressed image backup when the --limit-memory option was set to a low value. This fix prevents the problem by adjusting the way data buffers are allocated in the situation. (Bug #29773223)

• mysqlbackup created some duplicate entries in the binary log index file if the binary log was stored outside of the data directory on the backed-up server. (Bug #29564487)

• When restoring an incremental backup created with the --skip-final-rescan option, the external tablespaces on the target server got deleted. It was due to the missing entries for the external tablespaces in the incremental backup's tablespace_tracker file, and this fix corrects the issue. (Bug #29513642)

• Attempts to extract binary log files from a compressed backup failed with a No such file or directory error when the size of the binary log files on the backed up server was greater than 16MB. (Bug #28787312)