The Master-DB instance failed

If your Master-DB instance failed, but you still have a Slave-DB instance operational, you can follow the required actions below to recover from a Master-DB failure.  If you do not have a Slave-DB instance running and your Master-DB failed, see the There are no running DB instances (master or slave)! scenario.

If a Slave-DB is operational, it's better to promote it to master instead of creating a new Master-DB from a previous snapshot.  By promoting the existing slave to master you will minimize data loss and avoid a restoration from an old snapshot that might not have the latest changes which have probably already been copied over to the Slave-DB.