Hello satadru,
=== My 2 cents. Not necessarily the opinion of other etherboot
developers ===
> I've been looking at various ways of booting WinPE (Windows
> Preinstallation Environment) off of the network, and after searching the
> net and the etherboot archives, it seems that the two most viable
> methods are setting up a RIS server (ugly) or somehow try to get the RIS
> boot disk tagged for etherboot (ugly, especially considering the likely
> lack of support for my NIC, the sis900).
sis900 not supported by WinPE? Etherboot works fine with it. I even
have one here and happily use it some day or another for netbooting.
Quting:
= WinPE can be booted over the network (using Pre-boot Execution Environment
= [PXE] basic input/output system [BIOS] and a Remote Installation Services
= [RIS] server) or via a CD. Thus, WinPE can be run without the need for
= persistent storage. However, WinPE cannot be booted via Remote Boot Service.
This is what a quick search found out for me. Perhaps this can be done
with etherboot too. However:
= IOS, or by using a remote boot disk for pre-PC98 computers.
MS seems to distribute a bootdisk which supports several NICs (via
packet drivers?) and allows starting RIS from network. If you can get
that disk, try to etherboot an image of that disk.
Perhaps it does not work (if the memory the disc is loaded to is
overwritten by a program on that disc, who knows), but give it a run
and tell us!
> I noticed that IBM has some sort of combined implementation of iscsi and
> something similar to etherboot:
> http://www.haifa.il.ibm.com/projects/storage/iboot/faq.html
Quoting:
= The remote boot offers a seamless solution; changes to the kernel can be made
= permanent. Say, for example, a user wants to add a Linux module to support
= new drivers. With existing methods, the module is lost when the computer is
= shut down and users have to repeatedly add the module every time they boot
= the machine. iBoot is capable of saving these changes, making them permanent.
= iBoot also supports a dual boot option. Users are not limited to a single
= operating system and can select the desired operating system from a menu
= at boot time, offering enhanced flexibility of operation
They seem not to know about NFS, and never to have used etherboot
menuing facilities. This paragraph simply is marketing
blahblah-nonsense.
Their technique seems to be to implement a int13 support routine that
gets harddisk requests and relays them via network. I don't see why it
should not run with DOS/Win9x, but they claim it not do so. Seems not
to be too perfect then.
> How hard would this be to implement inside etherboot? Open Source iscsi
> drivers exist:
> What I would like to be able to do is setup a linux box to serve iscsi,
> and then use a etherboot boot rom to boot off of that network drive.
I don't see sense in here. Why cannot you run a tftp or nfs server?
Serving iSCSI surely is the greater security hole than a chroot'ed
running tftp server for which the source code is available.
Just for loading the kernel and initrd in doubt, as the kernel in turn
could load an iSCSI module and then very well access that remote iSCSI
server.
> Is this too far outside to purview of the etherboot project to be done?
If you like to implement it, please do so. *I* certainly won't, and I
don't expect the experts around here to do so neither. It would
require to implement tcp (at present there only is udp support for
good reasons: Much lighter to implement), and after that one could as
well use ftp or http to retrieve kernel & initrd.
> Etherboot would then have to maintain some process in memory to be able
> to communicate with the iscsi server while the machine was running.
I don't think so. From what I understood (this may be wrong though),
it only needs this stack for "real mode" accesses to iSCSI, like those
done by the BIOS (and Win 2k/XP... first stage), lateron the kernel
loads an own module, which real-mode-drives the NIC and accesses the
iSCSI more directly. Linux kernels are claimed to run without any
real mode stack in memory, just from that kernel driver.
> This would be Really cool,
*not*. Not to much at least. http access in my opinion is much more a
attraction bringer.
> and would also suddenly make all of those
> machines out there with etherboot supported roms iscsi bootable
> machines.
I haven't seen one iSCSI server yet (well, the protocol is quite
fresh), and can't understand why they should be unable to have a tftp
server or a NFS server just for booting purposes.
> If this isn't possible, does anybody know of a iscsi boot
> image that will let me etherboot to a second stage bootloader that would
> support iscsi?
Nope. But you could ask Intel's :-( if they won't open-source their
bootloader some day. OK, bad idea :-/
Hope to help.
Anselm Martin Hoffmeister
Stockholm Projekt Computer-Service
<an...@ho...>
--
Disclaimer - These opiini^H^H damn! ^H^H ^Q ^[ .. :w :q :wq :wq! ^d X^?
exit X Q ^C ^c ^? :quitbye CtrlAltDel ~~q :~q logout save/quit :!QUIT
^[zz ^[ZZZZZZ ^H man vi ^@ ^L ^[c ^# ^E ^X ^I ^T ? help helpquit ^D ^d
man help ^C exit ?Quit ?q CtrlShftDel "Hey, what does this button d..."
|