CentOS - Apr 2014 - Problem with blades that wont kickstart

About 15% of my blades wont kickstart.
I checked to see, and they all seem to be 100% compatible.

I currently do not own/direct access the kickstart server, I will build 
one later.

Is it possible they fat fingered the MAC address of the problem blades 
when they put them into the PXE server? As I can build them from a 
locally mounted virtual CD, but cannot build from PXE

At first I was getting a grub error. When I installed a baseline Centos 
locally, and then pxe booted for the build. I got a
cannot find C0T0   which indicates it cannot find the first drive 
(kickstart uses sda1), yet I look in the drac and I can see C0T0 drive 
as well as C0t1 drive  (this is a raid 1  virtual drive no mirror)

So here is my big question:    can I put my primary drive  in slot sda2 
on the blade... then DD the drive from the working blade, change the MAC 
address in the network interface, change the hostname/IP address in 
/etc/sysconfig/network and /etc/sysconfig/network-scripts/eth0   to 
build the drives?

What are the risks? what are the gotcha's? what problems will I have DD 
ing a drive that is currently the boot drive?

Any suggestions are welcome

-- 
  
Dan Hyatt

Dan Hyatt wrote:
> About 15% of my blades wont kickstart.
> I checked to see, and they all seem to be 100% compatible.
>
> I currently do not own/direct access the kickstart server, I will build
> one later.
>
> Is it possible they fat fingered the MAC address of the problem blades
> when they put them into the PXE server? As I can build them from a
> locally mounted virtual CD, but cannot build from PXE
<snip>
Oh.

How do you *allow* machines to pxeboot? Where I am, the MAC address entry
is put in a special section of the DHCP configuration, and *only* MAC
addresses in that block are allowed to pxeboot.

       mark