Jeremy Katz wrote:
On Mon, 2008-01-28 at 10:59 -0600, Jerry Vonau wrote:
Cain, Brian (GE Healthcare) wrote:
I have a PXE boot environment set up for installing Fedora/RHEL. Has anyone looked at making the LiveCD loadable from the network? If not, are there any pointers that might help?
Mohammed Khan did some work a while back. I wasn't able to get it to work (failed at the pivot root), but I dig very deep on it. It works for him, so it's worth looking at.
https://www.redhat.com/archives/fedora-livecd-list/2007-June/msg00021.ht ml
I'm happy someone else thinks like me. I posted to the anaconda list, a hack using nsf as the storage point for the live image, but that only did an install.(July/Aug 07). If that is what you want, it works. I'm currently playing around with anaconda's loader to see if I can get that to run the livecd, seems to me like a good starting point, with a gui for dmc's persistent work. With anaconda's hardware detection, I'm hoping to make a portable persistent live-usb disk. I've been (very) slowly porting the livecd's initrd parts to C code for anaconda's loader. I've got a mix of shell and C that currently can mount the livecd, and use that in place of the stage2.img file. (boy is it hard to keep up with some of the changes going on) Next is to get the run-init part to work without freaking out anaconda. Any interested?
Rather than hijacking anaconda pieces (and the doom that is the anaconda initrd), the better approach is probably to get to where we can use the standard mkinitrd pieces to build the live initrd. This likely is going to involve switching the initrd over to using bash as opposed to nash as its interpreter. I started some work in this area, but realistically, it's going to take some large-ish changes.
But I like anaconda's initrd, ;-) Cool, that is a whole pile less convoluted that playing in C, looking forward to your progress. Got an svn sandbox somewhere?
Jerry