Hello everybody, i have a Question concering MySQL-Backups in a Master/Slave Replication-Setup (MYISAM and InnoDB Tables) and incremental Backups using the Binary-Logs from the Slave.
Well, in case something will go wrong (hopefully not) I'm currently trying to implement a recovery System. Therefor the Slave logs all events it executes in it's own binary-log (log-bin and log_slave_updates activated). Logical Backups are performed with the help of a Perl-Script and mysqldump every 24 hours. The Script stops the Slave (STOP SLAVE), verifies that it's really stopped and performs the Backup for each Database with mysqldump (--opt). After all dumps are created, the script performs FLUSH LOGS to force the start of a new binary-log. Afterwards the Replication is initialized again (START SLAVE). In addition the master.info, the Output of SHOW SLAVE STATUS and finally the relay-log.info and relay-log.index are also backed up after the Slave is stopped. To verify the functionality of the Recovery Process i've copied a Backup from the 03.08.2008 for a specific Database to a Lab-Server. Also the Backup for the same Database from today (04.08.2008) and the binlogs that cover the events starting after the Backup from 03.08 to 04.08. Well and here is my Problem: After Recovering the Backup from 03.08 to my test-Server and replaying the Binary-Logs i compared the results to the Dump from today. In the Dump from today there are more ROWS in specific Tables then in the Dump from yesterday with applied binary logs :-( Why do I miss some ROWS? What is wrong within my recovery procedure? Any hints are greatly appreciated :-) Thanks, Werner -- MySQL General Mailing List For list archives: http://lists.mysql.com/mysql To unsubscribe: http://lists.mysql.com/[EMAIL PROTECTED]