On Mon, May 24, 2010 at 11:26 AM, Kenneth Gober <[email protected]> wrote:
> are you sure you're solving the right problem?  in a production
environment,
> do you really want production machines to continue to boot to PXE, load an
> installation kernel, and only then discover that a bootable local disk
> exists and reboot from that?  it seems dangerous to allow a situation to
> continue where a misconfiguration of your PXE server could cause all your
> production machines to start re-installing themselves.
> it would seem more sensible to me to set up the machines to only boot to
PXE
> if the local disk isn't bootable.  does your hardware not allow it, forcing
> you to either always boot from PXE first, or never boot from it at all?

I don't think the people who reported the issue were using it in that way - I
believe it was just when they were provisioning machines that they had the
looping problem. I thought if it was doable that it would be more useful for
the
machine to be network available in the installed state rather than running
the
installer ramdisk.

-N

Reply via email to