Hello,

I am reading for weeks several documents about building a kernel.
I am running fedora 8. So I read also about "building a kernel on the Fedora" way.

When you build a kernel on the Fedora way, you are using rpm, an the result is a kernelpackage .rpm.

So this package can only be used to be installed on an "already" working system.

I want to build a new (small) system that can be run (and started) from a pxe-server.
I want to run it on a thin client.
I think to build a monolithic kernel. (Start from Vanilla ?) (how can I start from my Fedora kernel source?)
Is there a way to "check-out" out of this rpm : the kernel and initrd. ?
Who can guide me how I can build a small rootfs? (in ram)

I have found an interesting article : (on kernel 2.4)

http://www.linuxfocus.org/English/Ju...ticle252.shtml

Does someone know such an article for 2.6?

Who can help me to understand/build a initrd? and a rootfs ?

thanks,

I can understand what you're trying to do, but you're asking too much questions that could be answered by reading the "right" docs. I could suggest you start technology-wise from the top down with PXE (something like http://www.ltsp.org/, http://www.kegel.com/linux/pxe.html (which mentions LTSP as well, also see: http://wiki.ltsp.org/twiki/bin/view/Ltsp/LBE) or http://pxe.dev.aboveaverageurl.com/index.php/Main_Page) to get the general idea for what's included slash necessary. Once you dig PXE then knowing you can build a kernel from your Fedora manually without it resulting in an RPM should do it. BTW the http://www.linuxfocus.org/English/Ju...ticle252.shtml you posted *was* updated to use kernel 2.6 AFAIK.