Index of files created by Percona XtraBackup¶
Information related to the backup and the server
This file contains information to start the mini instance of InnoDB during the --apply-log. This is NOT a backup of original my.cnf. The InnoDB configuration is read from the file backup-my.cnf created by innobackupex when the backup was made. innobackupex --apply-log uses InnoDB configuration from backup-my.cnf by default, or from innobackupex --defaults-file, if specified. InnoDB configuration in this context means server variables that affect data format, i.e. innodb_page_size, innodb_log_block_size, etc. Location-related variables, like innodb_log_group_home_dir or innodb_data_file_path are always ignored by innobackupex --apply-log, so preparing a backup always works with data files from the backup directory, rather than any external ones.
The type of the backup (e.g. full or incremental), its state (e.g. prepared) and the LSN range contained in it. This information is used for incremental backups. Example of the xtrabackup_checkpoints after taking a full backup:
backup_type = full-backuped from_lsn = 0 to_lsn = 15188961605 last_lsn = 15188961605
Example of the xtrabackup_checkpoints after taking an incremental backup:
backup_type = incremental from_lsn = 15188961605 to_lsn = 15189350111 last_lsn = 15189350111
The binary log file used by the server and its position at the moment of the backup. Result of the SHOW MASTER STATUS.
The xtrabackup binary used in the process.
Contains data needed for running the: --apply-log. The bigger this file is the --apply-log process will take longer to finish.
This file is going to be created when performing the incremental backup. It contains the per-table delta metadata: page size, size of compressed page (if the value is 0 it means the tablespace isn’t compressed) and space id. Example of this file could looks like this:
page_size = 16384 zip_size = 0 space_id = 0
This file contains ranges of skipped secondary index pages. These files are created only when doing the compact backup. The file format is a series of 2-value tuples, with each value being a 4-byte page offset corresponding to the first and the last endpoints of skipped ranges, respectively.
Information related to the replication environment (if using the --slave-info option):
The CHANGE MASTER statement needed for setting up a slave.
Information related to the Galera and Percona XtraDB Cluster (if using the --galera-info option):
Contains the values of status variables wsrep_local_state_uuid and wsrep_last_committed.