On Tue, 24 Apr 2001, Jim Osborn wrote:

initrd is not really a "script" but a boot image, part of the kernel itself.

A clarification on your correction: The initrd is not part of the kernel, but a filesystem image, which is initially loaded into RAM and mounted as /. The kernel can then load modules necessary to enable its SCSI drives, network cards, etc. Finally, with all devices working, it can mount its normal filesystems, even if they be on a module-supported SCSI controller, or NFS. This two-stage init process made possible the all-modular kernel, and made it possible to distribute a 'generic' kernel which would boot on just about any hardware out there. I can't imagine the current YAST2 bootable-CDROM with almost-automagic install and configure without it. Anybody remember the old days, when you began an install by pulling the covers and writing down the ident on all the chips, then spending half an hour reading thru the list of kernels on the CD, hoping to find one that had the set of drivers for your minimum hardware? Ever have to face the choice of buying a different SCSI card, or ask a buddy to compile you a kernel, just so you could get it installed? We've come a long way since Yggdrasil 1.0 'Plug and Play' [sic] Linux! -- Rick Green "I have the heart of a little child, and the brain of a genius. ... and I keep them in a jar under my bed"