Hi,
I will soon be using a PXE server to install CentOS onto 240 SuperMicro servers, via a kickstart file. The servers are configured to attempt a network boot (PXE) first and my kickstart file is configured to reboot each host after the installation completes, so they are accessible via SSH. My issue is that that will create a closed loop of repeatedly installing to each host. Is there a way to prevent the PXE boot option after the install is completed or is there a way to ignore the second request for an IP from a host?

Configure the installation targets so that the first boot option is hard disk, second option network. Ensure there is nothing bootable on the hard disk. Switch them on.

A PXE server is normally a combination of DHCP and TFTP server. I doubt that you can configure them so that they only accept one request from a given address. Besides, if a PXE-based installation fails for whatever reason (if 240 servers boot at the same time, overload and timeouts are not totally out of the question), it would fail for good in this scenario.

It was recommended (from another source) that I try ipmiutil. I have not had a chance to try it out, but I'll mark this thread as solved, since the tool does claim to change the boot order for the second reboot.