Appendix A Frequently Asked Questions for MySQL Enterprise Backup

This section lists some common questions about MySQL Enterprise Backup, with answers and pointers to further information.

Questions

A.1: What versions of the MySQL server does MySQL Enterprise Backup 4.1 support?
A.2: What is the big ibdata file that is in all the backups?
A.3: Can I back up non-InnoDB data with MySQL Enterprise Backup?
A.4: What happens if the apply-log or apply-incremental-backup step is interrupted?
A.5: Why is the option --defaults-file not recognized?
A.6: Can I back up a database on one OS platform and restore it on another one using MySQL Enterprise Backup?
A.7: What if I have included the binary log or relay log in my backup but do not want to restore it?
A.8: What would happen if I start a server directly using a raw directory backup, without running either the copy-back or the apply-log operation?

Questions and Answers

A.1: What versions of the MySQL server does MySQL Enterprise Backup 4.1 support?

See Section C.3, “Compatibility with MySQL Versions” for details of compatibility between different releases of MySQL Enterprise Backup and MySQL Server.

A.2: What is the big ibdata file that is in all the backups?

You might find your backup data taking more space than expected because of a large file with a name such as ibdata1. This file represents the InnoDB system tablespace, which grows but never shrinks as a database operates, and is included in every full and incremental backup. To reduce the space taken up by this file in your backup data:

After doing a full backup, do a succession of incremental backups, which take up less space. The ibdata1 file in the incremental backups is typically much smaller, containing only the portions of the system tablespace that changed since the full backup.
Set the configuration option innodb_file_per_table=1 before creating your biggest or most active InnoDB tables. Those tables are split off from the system tablespaces into separate .ibd files; the tables can then be individually included or excluded from backups, and disk space is freed when the tables are dropped or truncated.
If your system tablespace is very large because you created a high volume of InnoDB data before turning on the innodb_file_per_table setting, you might use mysqldump to create a dump of your entire server instance, then turn on innodb_file_per_table before re-creating the database, so that all the table data is kept outside the system tablespace.

A.3: Can I back up non-InnoDB data with MySQL Enterprise Backup?

While MySQL Enterprise Backup can back up non-InnoDB data (like MYISAM tables), the MySQL server to be backed up must support InnoDB (i.e., the backup process will fail if the server was started up with the --innodb=OFF or --skip-innodb option), and the server must contain at least one InnoDB table.

A.4: What happens if the apply-log or apply-incremental-backup step is interrupted?

If mysqlbackup is interrupted during the apply-log or apply-incremental-backup stage, the backup data is OK. The file operations performed by those options can be performed multiple times without harming the consistency of the backup data. Just run the same mysqlbackup command again, and when it completes successfully, all the necessary changes are present in the backup data.

A.5: Why is the option --defaults-file not recognized?

When you specify the --defaults-file option, it must be the first option going after mysqlbackup. Otherwise, the error message makes it look as if the option name is not recognized.

A.6: Can I back up a database on one OS platform and restore it on another one using MySQL Enterprise Backup?

See Section C.2, “Cross-Platform Compatibility” for details.

A.7: What if I have included the binary log or relay log in my backup but do not want to restore it?

If you want to skip the restore of the binary log, relay log, or both during a restore, use the --skip-binlog option, the --skip-relaylog option, or both with your copy-back or copy-back-and-apply-log command.

A.8: What would happen if I start a server directly using a raw directory backup, without running either the copy-back or the apply-log operation?

This should never be attempted. Not only would the server crash, but the backup would likely get corrupted and become unusable. This is because the directory backup contains metadata created by mysqlbackup that the MySQL server would not understand; also, the raw backup might be inconsistent and need to be brought up-to-date by an apply-log operation, so that changes made to the database during the backup process can be applied.