[syslinux] chainbooting pxe servers stalls

[Shantanu Gadgil]

--- On Mon, 9/10/12, Sean Mitchuson <smitchuson at murraystate.edu> wrote:

> From: Sean Mitchuson <smitchuson at murraystate.edu>
> Subject: Re: [syslinux] chainbooting pxe servers stalls
> To: "For discussion of Syslinux and tftp-hpa" <syslinux at zytor.com>
> Date: Monday, September 10, 2012, 10:04 PM
> That's exactly what I think my
> problem is and I'm currently working on
> fixing it.
> 
> I will reply again with my findings.  Thank you.
> 
> On Mon, Sep 10, 2012 at 11:21 AM, Gene Cumm <gene.cumm at gmail.com>
> wrote:
> 
> > On Sep 10, 2012 11:09 AM, "Sean Mitchuson" <smitchuson at murraystate.edu>
> > wrote:
> > >
> > > I've started using pxe servers for OS deployment
> at my organization and
> > I'm
> > > trying to get chainbooting working but I seem to
> run into an issue that
> > > causes a hard lock up of a system.
> > >
> > > MENU BEGIN ^1. Boot Satellite PXE
> > >         LABEL
> Satellite PXE
> > >         KERNEL
> pxechain.com
> > >         APPEND
> (Server IP Address)::pxelinux.0
> > > Is part of my config file for my main pxe server
> to boot over to the
> > > satellite based pxe server but when I choose this
> option I see
> > > TFTP boot: (server ip address) pxelinux.0 and it
> locks up the machine and
> > I
> > > have to hard reset it.
> > >
> > > So I'm not sure where you start troubleshooting.
> >
> > For each server, what version of PXELINUX and where did
> you obtain the
> > binaries?  Did you build them yourself?
> >
> > Could
> >
> > http://www.syslinux.org/wiki/index.php/Common_Problems#PXE_Calls_in_4.00-4.03apply
> > (for the source PXELINUX)?
> >
> > If not, could you describe the machines you've tested
> with?  Make, model,
> > BIOS revision/version?  If not using an onboard
> NIC with the BIOS's PXE
> > OROM, please describe the NIC and the PXE stack you're
> using.

Hello Sean,

I think I have a similar setup except I just use satellite DHCP (PXE) servers which only need the subnet specific info (IP, router) and the IP of the TFTP server.
This reduces the overhead of having too many files per satellite server.
I use dnsmasq as the satellite DHCP/PXE server, which just redirects any PXE booting machine to the main (only) TFTP server.

Further installation continues from the TFTP server only.

I had to do this as DHCP forwarding has been disabled on the intermediate router (I presume yours could be a similar case?!?)

Anyway, be wary of the 4.10preXXX builds as they tend to show symptoms like you have described.
The 4.06pre11 work fine on my setup (which I am currently using).

Regards,
Shantanu