You'll have to test that theory out for how you intend to get all the data back to the master eventually.
The instructions for doing this are NOT something you want to rely on. You want to test this. Possibly in a sandbox scenario to be sure.
Note that a backup/restore set is marvelous. But if you have *never* done a restore from your backup set as a test, many old-school system administrators would say the backup set is not valid. This stems from "if it hasn't been tested, you have to assume it's not valid because it very well may not be!" after experiencing old tape backup failures. Still applies.
The same applies to recover from a prime failure followed by use of the slave and eventual recovery back to the primary. If you haven't tested it all the way through, it may fail and you may lose all your data. So ... full backup and restore to a different database (to be certain your DB backup/restore works), then test your primary fail and everything else all the way through to getting all the new data back to the primary. Make changes to the data on the slave and be sure those changes eventually end up back in the primary.
At least until Kumba automates all of this for you.
(And remember to thank Kumba and Matt for all this crap working in the first place!)
Nefariousparity wrote:Ahh "The Great" William Conely, thank you very much.
Conehead, what? Ouch. And I prefer "The Nefarious" rather than "The Great", but with your username I can see why you'd avoid that. Or maybe there should be a captain in there somewhere.