Skip to content

This are some Perl scripts for handle mysql backup and point-in-time-recovery

Notifications You must be signed in to change notification settings

winne27/mysqlBaR

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

7 Commits
 
 
 
 
 
 

Repository files navigation

#Description Some Perl scripts for handle mysql backup and point-in-time-recovery

#Prerequisites

The package "Percona Xtrabackup" must be installed on your server. It is freeware licensed under GPL and you can get it from http://www.percona.com/software/percona-xtrabackup/.

#Installation

Copy myBackup and myRestore to a directory which is in the path. Suggestion: /usr/local/bin Copy my-backup.cnf and my-restore.cnf to /etc or another directory of your joice.

#Create backup user in mySql Database

Create a user (default name: backup) in the mySql database. Give the following rights to him: GRANT RELOAD, LOCK TABLES, REPLICATION SLAVE, REPLICATION CLIENT ON . TO 'backup'@'localhost'

#Configuration

Have a look to the config file my-backup.cnf and change the options for your needs. If point-in-time recovery is requested check if binary logging is enabled in the mysql database. This is done by the following option in my.cnf: log-bin=mysql-bin

#Backup database

The base of this system are rotating backupsets. This means that after a backupset is "full" the next backupset is used. In the case the last backupset is reached all content from the first backupset is erased and then is used as actual backupset. The number of backupsets can be defined in the config file or can be changed by callup parameter.

Each backupset contains directories named "incrementxx" where xx is between 00 and 99. The directory increment00 contains always a full backup. In the config file the number of increments can be defined. If it is set to 0 only a full backup is taken. If this value is grater than 0 incremental backups are taken until the number defined in the config file is reached. In this case the backupset is closed and the next backupset is selected.

For starting the backup put the following to the crontab of root:

sudo -u mysql /usr/local/bin/myBackup >/var/log/mybackup.log 2>&1

#Restore database

For restoring the database a seperate directory is used. The name of this directory is defined in /etc/my-backup.cnf. The default is /var/lib/mysql-restore. This could changed by a calling parameter of myRestore. Use option --help for more information.

There are two methods for restoring the database. In the first method you have to select the backup by identifying the backupset and the increment that should be used for restore. This is done by call parameters of myRestore.

The second method is a point-in-time recovery of the database. Use parameter

--stop_datetime=YYYY-MM-DD-HH-MI-SS

to determine the time of recovery until. The procedure then detects automatically the necessary backupset and the increments and the binary logs needed for the recovery.

In both cases after the successful recovery an alternate database instance with this restored data is started. This instance uses by default the config file /etc/my-restore.cnf. By default this database is listening on port 3307 and on socket /var/lib/mysql-restore/basedir/mysql.sock.

Examples for starting the restore:

myRestore --backupset 2 --increment 5
myRestore --stop_datetime 2012-01-22-18-15-00

Get all possible options by:

myRestore --help

#Stopping the recovered database service

The processid of the database service can be found in /var/lib/mysql-restore/basedir/mysql.pid. With

kill -15 pid

database could be shutdown.

About

This are some Perl scripts for handle mysql backup and point-in-time-recovery

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages