Mailing List Archive

Maes Rich wrote:
>
> I'm using rocks cluster software to startup a small blade cluster in my
> lab. The blades I have redboot installed on them. I am able to boot
> any client that has pxeboot capability using my setup, except for these
> blades. My thoughs were using redboot commands to do the following:
>
> 1. load -v -r -b 0x8000000 -h 10.1.1.1 pxelinux.0
> 2. start
>
> This however caused my blade to reboot spontaneously. Should have known
> it couldn't be that easy.
>
> Has anyone attempted this before? Thanks in advance.
>

Have you asked on the Redboot mailing lists?

FWIW, pxelinux.0 would have to be loaded at 0x7c00, and a PXE API needs
to be presented. Otherwise you have to load Etherboot/gPXE as an
intermediate stage in order to provide the PXE stack.

-hpa

_______________________________________________
SYSLINUX mailing list
Submissions to SYSLINUX@zytor.com
Unsubscribe or set options at:
http://www.zytor.com/mailman/listinfo/syslinux
Please do not send private replies to mailing list traffic.

Op 18-07-2007 om 11:15 schreef Maes Rich:
> The subject says it all.
>
> I'm using rocks cluster software to startup a small blade cluster in my
> lab. The blades I have redboot installed on them. I am able to boot
> any client that has pxeboot capability using my setup, except for these
> blades. My thoughs were using redboot commands to do the following:
>
> 1. load -v -r -b 0x8000000 -h 10.1.1.1 pxelinux.0
> 2. start
>
> This however caused my blade to reboot spontaneously.
> Should have known it couldn't be that easy.

:-) nobody said it would be easy


> Has anyone attempted this before?


My guess is that redboot binaries have another header format,
then the pxelinux.0 has.

An simple check would be doing

file pxelinux.0

and

file a_redboot_executable


After prepending pxelinux.0 with the redboot header info,
you will mostly likely will be missing PXE code in the redboot boot ROM.


My advice is to contact the people from the Etherboot project,
if you want to continue with "pxelinux.0 on redboot"



Cheers
Geert Stappers


P.S.

I wonder why to load pxelinux.0 to load a kernel,
where redboot can load a kernel at once ...

_______________________________________________
SYSLINUX mailing list
Submissions to SYSLINUX@zytor.com
Unsubscribe or set options at:
http://www.zytor.com/mailman/listinfo/syslinux
Please do not send private replies to mailing list traffic.

Geert, to answer your PS.
I'm trying to bring up Rocks on my little cluster here. Rocks, depends
on a PXE type mechanism for node installation. Unfortunatley, my nodes
don't support a CD_ROM for normal kickstart installation, so I am
resorting to using a PXE boot method. The other unfortunate is that
they have have redboot on them as opposed to .... almost anything else
that supports PXE. :(

So I am trying this and that to get this little cluster to play with OTS
cluster software. It's a little frustrating.

I'm going to look into the Etherboot project as you recommend. In fact
looking at my dhcpd.conf file I see that if the client reports as a
etherboot client, it still gets handed the pxelinux.0 file so there must
be some high synergy there.

Thanks for your help.

Rich Maes
Lead Systems Engineer
Isothermal Systems Research, Inc.
office: 509-232-3458
mobile: 509-701-1356
www.spraycool.com


-----Original Message-----
From: syslinux-bounces@zytor.com [mailto:syslinux-bounces@zytor.com] On
Behalf Of Geert Stappers
Sent: Wednesday, July 18, 2007 12:06 PM
To: syslinux@zytor.com
Subject: Re: [syslinux] Is there a howto for using redboot to kickoff
apxelinux boot?

Op 18-07-2007 om 11:15 schreef Maes Rich:
> The subject says it all.
>
> I'm using rocks cluster software to startup a small blade cluster in
my
> lab. The blades I have redboot installed on them. I am able to boot
> any client that has pxeboot capability using my setup, except for
these
> blades. My thoughs were using redboot commands to do the following:
>
> 1. load -v -r -b 0x8000000 -h 10.1.1.1 pxelinux.0
> 2. start
>
> This however caused my blade to reboot spontaneously.
> Should have known it couldn't be that easy.

:-) nobody said it would be easy


> Has anyone attempted this before?


My guess is that redboot binaries have another header format,
then the pxelinux.0 has.

An simple check would be doing

file pxelinux.0

and

file a_redboot_executable


After prepending pxelinux.0 with the redboot header info,
you will mostly likely will be missing PXE code in the redboot boot ROM.


My advice is to contact the people from the Etherboot project,
if you want to continue with "pxelinux.0 on redboot"



Cheers
Geert Stappers


P.S.

I wonder why to load pxelinux.0 to load a kernel,
where redboot can load a kernel at once ...

_______________________________________________
SYSLINUX mailing list
Submissions to SYSLINUX@zytor.com
Unsubscribe or set options at:
http://www.zytor.com/mailman/listinfo/syslinux
Please do not send private replies to mailing list traffic.


_______________________________________________
SYSLINUX mailing list
Submissions to SYSLINUX@zytor.com
Unsubscribe or set options at:
http://www.zytor.com/mailman/listinfo/syslinux
Please do not send private replies to mailing list traffic.

Op 18-07-2007 om 14:11 schreef Maes Rich:
> From: [mailto:syslinux-bounces@zytor.com] On Behalf Of Geert Stappers
> Sent: Wednesday, July 18, 2007 12:06 PM
>
> >
> > P.S.
> >
> > I wonder why to load pxelinux.0 to load a kernel,
> > where redboot can load a kernel at once ...
>
> Geert, to answer your PS.
> I'm trying to bring up Rocks on my little cluster here. Rocks, depends
> on a PXE type mechanism for node installation. Unfortunatley, my nodes
> don't support a CD_ROM for normal kickstart installation, so I am
> resorting to using a PXE boot method. The other unfortunate is that
> they have have redboot on them as opposed to .... almost anything else
> that supports PXE. :(
>
> So I am trying this and that to get this little cluster to play with OTS
> cluster software. It's a little frustrating.

I think you try to say that you have (cluster) software that depends
on the (Intel only) PXE software. And that the only networkboot
protocol stack you known is the Pre eXecution Environment.


> I'm going to look into the Etherboot project as you recommend. In fact
> looking at my dhcpd.conf file I see that if the client reports as a
> etherboot client, it still gets handed the pxelinux.0 file so there must
> be some high synergy there.

I foresee that you will make the same error as the error which started
this E-mail thread. Starting a non-redboot binary in the redboot
environment. I assume that you will culpritt redboot falsely.

Okay, hopefully I have now your attention.

Consider the redboot environment a sane start, and actually it is.
It runs ELF binaries.

Etherboot provides binaries for IBM-PC, either a BIOS ROM or a "bootable
DOS floppy".

You need the Etherboot binary because it provides PXE for the OTS Cluster software,
where 'PXE' is literally Pre eXecution Environment.


My advice:

Build a ELF binary of Etherboot (or gPXE).
That will get you further as your original post.


> Thanks for your help.

My pay-pal account is available on request.
Also a shipping address for sending a redboot system.


Cheers
Geert Stappers

_______________________________________________
SYSLINUX mailing list
Submissions to SYSLINUX@zytor.com
Unsubscribe or set options at:
http://www.zytor.com/mailman/listinfo/syslinux
Please do not send private replies to mailing list traffic.