(disclaimer: i'll be a bit vague about some of this since i don't have a test system to try to reproduce it until later today or this weekend.)

during a RHEL SA course i was teaching this week, i was using both centos 5.5 and RHEL 6.0 beta 2 and, as a fun exercise, i was showing how to "git" checkout the kernel source tree, configure it, build, install and boot to a new kernel. sadly, at no time did that exercise actually work, so i just want to ask a general question -- should i, with a standard install of centos 5.5 with all of the required development packages, be able to checkout the kernel source, and build and boot a new kernel?

upon reflection, the issues might have to do with the fact that LVM was in use and perhaps the initrd didn't have LVM support built in but, again, i can't check that until later today at the earliest.

so, here's the hypothetical scenario i'm going to test later:

* gateway, 64-bit laptop, with 64-bit centos 5.5, fully updated * /boot is primary ext4 partition * / and /home are logical volumes in a single volume group * "git clone" latest kernel source * "make defconfig" * "make" * "make modules_install" * "make install"

is there any reason why the above shouldn't work? (and instead of "make defconfig", i can copy the /boot/config file for the stock kernel and "make oldconfig". i'll almost certainly try both just to cover all the possibilities.)

so ... should this theoretically work? is there a web page that actually walks through the process? because we were having all kinds of grief just installing the kernel or booting and failing to find the root device (which is why i suspect something LVM-related.)

i'm open to suggestions. thanks.

rday

p.s. obviuosly, i'd love to hear that someone else has already done this and it worked just fine.