Setting up replication on vagrant systems with Fatal error on UUID

My work uses Vagrant VMs to quickly create the same developer environment for everyone in the engineering department. I was setting up Master-Slave replication with a developer today on two Percona 5.6 Vagrant VMs and I kept getting an error when starting the slave. This was the first time I had tried to setup replication with one of these Vagrant VMs. The error could be seen in the Last_IO_Error column of "SHOW SLAVE STATUS"

Fatal error: The slave I/O thread stops because master and slave have equal MySQL server UUIDs; these UUIDs must be different for replication to work.

I didn't know until today that MySQL 5.6 has a file called auto.cnf in the data directory with the UUID for that system. Since our Vagrant VMs are all copied from the same system then of course they will all have the same auto.cnf file. Just removing the file and restarting MySQL fixed the issue. MySQL automatically created a new auto.cnf after restart.

This UUID is totally separate from the server-id in the my.cnf file.