mkinitrd creation while Fedora installation with anaconda

[Olivier BONHOMME <obonhomme nerim net>]

Hello everybody,

I am trying to resolve an issue about anaconda but I need some clues in 
order to do that. I am building a custom Fedora Distribution. The 
installation is made using anaconda. For the kernel, I made my own RPM 
with my specific spec file.

My RPM seems to be OK : when I install it, mkinitrd is created and it is 
possible for me to boot on the kernel installed with this RPM.

But when I put this RPM on my CD and when I try to make a fresh 
installation, the initrd file is not generated and as a consequence, it 
is impossible for me to boot on the kernel.

I tried to look at the anaconda code but I am not a python specialist. 
That's why I would want to know what is the necessary condition for 
anaconda to create the initrd file while the installation.

When I searched on the code, I found that the initrd is created when 
some modules are present (foundModule variable set to 1 which launchs 
the recreateInitrd function) but I can't find precisely what is needed 
in order to modify my spec file.

Does anybody can give me more precisions about this issue ?

Thank your for your answers

Best Regards,
Olivier BONHOMME