you can use a SingleProc as failure genarator. Add a failure control method. In this method, you can define which of your machines will be failed and you can disrupt the machine or remove the disruption.
So at any time only one machine will be failed.
Hope this helps