GTIDs and replaying binary logs

[Jon Slusher]

Hi. We recently upgraded to MySQL 5.6 and we are now using GTID-base replication for our topology. We are unclear about a disaster scenario and I would like to get some information about it.

We are trying to decide what it is we need to have backed up in the event we lose the master in such a way that we would have to restore the master and replay binary logs up to a point in time. Let's say there's a DROP statement that found its way to the server unintended.

Currently we have a backup MySQL server and we are using LVM snapshots of its entire data disk, binlogs included. We are also regularly syncing binlogs from this backup server once an hour. If we would have to restore a master to a point in time, before the DROP statement, would the snapshot and the binlogs from the backup server be sufficient? Would there be any situation that we would need backups of the binary logs from the master itself?