We have recently decided to end our MySQL Enterprise contract. We’re currently testing a Percona cluster, and will likely switch to this architecture across the board if all goes well. My first step is to convert the script I use to populate the dev / test DBs with production data. It’s currently largely using MySQL Enterprise Backup, with mysqldump to preserve some environment-specific schemas. I have some basic questions, as I get started:
First, looking through the documentation and examples, some use innobackupex, while others use xtrabackup. In the current version, is there a difference between these, and if so, what?
The current process backups up the environment-specific data, wipes out the existing data and logs, then creates a backup of productions, streams it across via ssh, restores it on the target, restores the preserved data, and re-deploys the stored procs / functions from SVN. When working with a cluster as the target, what’s the most efficient process for doing this? Assume no other traffic is hitting it at the time, as we shut down access during a sync. If I refresh one machine and leave the others alone, will the changes correctly replicate across, or do I need to wipe them all?