seth at cql
Apr 8, 2002, 8:03 AM
Post #8 of 8
This is quite true. I made something very similar work recently. It was
with PowerPC hardware, so the boot ROM code was different, but the principle
is exactly the same. This particular boot PROM, in fact, was worse than
anything you should need to deal with in a "normal" situation, because it was
homegrown (not by me) and its DHCP compliance was worse than the broken DHCP
implementations I've dealt with in more mainline hardware.
If anyone wants to get into even more gory details, contact me off list.
Peter's summary is excellent. I hadn't thought of using the chaining
solution; I'm going to give that a try.
On Monday 08 April 2002 03:24 am, you wrote:
> > This is, however, a different and incompatible solution.
> Different, certainly. As you note, pxelinux requires PXE; so when PXE
> isn't available, something else is needed for net booting. The config,
> kernel and initrd must be represented twice on the server.
> But please would you explain "incompatible"? DHCP allows one to use VCIs
> to have multiple configurations for different clients, and ISC DHCP v3
> has a tolerable implementation which allows Etherboot and pxelinux can
> coexist on a network, whether on the same or seprate servers. Of course,
> one should use hpa-tftp to cope with the broken PXE code we are all
> forced to use until a decent open-source PXE comes along.
> Of course, if Mr Geggus finds that Etherboot works nicely on his non-PXE
> nics, he *can* achieve identical configuration for all his workstations
> by chaining Etherboot on his PXE systems (Marty Connor's howto at
> www.ltsp.org has the details).
> SYSLINUX mailing list
> Submissions to SYSLINUX [at] zytor
> Unsubscribe or set options at:
Office: (480) 661-1849
Fax: (480) 614-8909
email: seth [at] cql
pager: 888-605-9296 or email 6059296 [at] skytel