Implement burn-in testing

[craiganderson]

Add burn-in (i.e. torture) testing for CPU, memory, disk, and network to validate hardware prior to starting target deployment. Implement basic parsing of kernel & iDrac/iLo logs to identify HW problems.

This should run by default for new hardware, and not run by default on subsequent re-deploy on hardware that it has previous passed on, but should be overrideable as needed (e.g., after replacing failed HW).