Hey all! New to the list, and fairly new to Grub2.
I've got about 400+ machines as part of a farm. They are re-imaged
routinely on demand, each image a different preset for the role it needs
to play (even swapping OS at times.)
Traditionally, we used pxelinux to image this stuff remotely. I really,
really would prefer to use Grub2 for this, but I've run into a small
problem that I'm hoping the list here can lend a hand with.
The machines are set to network boot as a default. I'd like to keep
this behavior, and let the timeout fallthrough to the primary hard
drive. (This way, I can use tftp-hpa to rewrite read requests for
'grubenv' to get per-machine environmental vars, and remote-control the
imaging. That part of this process is working great.)
It looks like I can accomplish this via:
set root=(hd0,1)
chainloader +1
boot
The problem is that there aren't any drives loaded by default from the
(pxe) environment.
sh:grub> ls
(pxe)
I discovered that I could 'insmod biosdisk', and lo-and-behold get my
hd* nodes (woo!):
sh:grub> insmod ext2
sh:grub> insmod biosdisk
sh:grub> ls
(hd0) (fd0) (pxe)
However, I can't actually -use- hd0 for anything. I'm not sure why.
I'm guessing it has something to do with what filesystem it thinks hd0
is:
sh:grub> ls -lh
Device hd0: Filesystem type pxefs <---- uh oh
Device fd0: Unknown filesystem
Device pxe: Filesystem type pxefs
sh:grub> chainloader (hd0,1)+1
error: no such partition
... and here's where my diagnostic abilities stop. :)
Can anyone suggest other things to try, reasons why this might be
happening, or other ways to achieve my goal? Could this just be a bug?
All input welcome and appreciated.
--
Mahlon E. Smith
http://www.martini.nu/contact.html
_______________________________________________
Grub-devel mailing list
Grub-devel@gnu.org
http://lists.gnu.org/mailman/listinfo/grub-devel
