Skip to content

Booting Hypervisors

Brian Akins edited this page Jan 9, 2015 · 2 revisions

The initial hypervisor will need to be booted from a thumb drive or a CD-ROM or another "external" method. (Note: this is probably true for at least one hypervisor on each broadcast domain for DHCP, etc to work)

Hypervisors net boot from this initial one:

  • Hypervisor sends a DHCP request
  • DHCP server running on another hypervisor responds with address, etc, including a bootfile.
  • Hypervisor requests the boot file - which is the first stage of iPXE - via TFTP.
  • Hypervisor the does a second DHCP request with the user-class set to iPXE.
  • DHCP server checks the flag and returns a URL to the ipxe configuration service.
  • Hypervisor does a GET of the URL.
  • the ipxe configuration service looks at the query string and uses the IP there to lookup the Hypervisor. It returns a custom iPXE config that include url's for the correct versioned kernel and initrd for the Hypervisor.
  • Hypervisor, using iPXE, fetches this kernel and initrd and boots.
Clone this wiki locally