[syslinux] Chain loading hard disk with gPXELinux

Tim Franssen mail at timendus.com
Tue May 12 12:11:47 PDT 2009


On Tue, May 12, 2009 at 5:59 PM, H. Peter Anvin <hpa at zytor.com> wrote:

> I don't know of "the" localboot bug.  There is "the" chainloading bug
> (gPXELINUX can't load another NBP), but that doesn't affect localboot.
> There were problems with localboot a long time ago, but they should have
> long since been resolved.
>
> What version are you running?


Oh, I'm sorry, I suppose I should have mentioned that :)

We're currently using 0.9.5 but I also tried 0.9.7 today, which is the
current download. I did something like this:

default local
> label local
>     localboot 0
>

0.9.5 gave me an error (not sure what it was, something like "can't find
another NBP"?)
0.9.7 said "Press <something> to enter gPXE command line" or something
similar and then halted.

The above configuration was aquired over HTTP, the gpxelinux.0 binary over
TFTP. I'm sorry I can't be more accurate, I'll have to look things up
tomorrow. Let me Google this a bit, I know I read something about this... I
know for sure that I read I should use chain.c32 somewhere, I didn't come up
with that myself :)

These seem to be about my issue:
http://syslinux.zytor.com/archives/2008-September/010767.html
http://www.etherboot.org/support/issue18

If you look at the e-mail that I've sent before, you will see a setup
> where PXELINUX uses gPXE-provided HTTP instead of TFTP, which I think
> you would find convenient.  If you use the RFC options mentioned
> therein, you will also avoid the default "countdown" before of PXELINUX,
> and thus avoid any additional delays.  - Shao
>

I don't seem to be able to find it. I'll give it another try tomorrow.

Again, thanks for your responses!

Tim



More information about the Syslinux mailing list