Continual Replication Sync

Danil Zburivsky (The Pythian Group) speaks about "Continual Replication Sync" at the 2010 O'Reilly MySQL User Conference & Expo.

Slides - http://assets.en.oreilly.com/1/event/36/Continual%20Replication%20Sync%20Pres...
From the official conference site at http://en.oreilly.com/mysql2010/public/schedule/detail/13428

Replication can get out of sync without warning. If you are using a slave for backup, disaster recovery, or even read-scaling, you want to make sure the slave has the same data as the master.

We will show the procedure we developed to use mk-table-checksum to minimize locking and slave lag that would be caused by locking large tables. We will discuss the problems we have run into with mk-table-checksum and how we have gotten around or fixed the issues—including when a table cannot be split into pieces and false positives generated in the MySQL error log.

Whether you use statement-based replication (the default), row-based replication or the mixed method, you can use mk-table-checksum to ensure your slave is in sync with your master. You can also use it even if your slave is lagging behind your master.

We will also show you how to fix the problems you encounter.

Category:

Science & Technology

Tags:

• replication
• mysql
• mysqlconf
• danil zburivsky
• pythian
• sync
• replication sync
• checksum
• consistency
• mk-table-checksum
• maatkit