Hey Guys, You've been a great help in the past and I'm hoping you can point me in the right direction this time with some more good advice. Background. Working on upgrading our old RH7.x based Kiosk OS to Fedora Core 2. Installation script handles installing FC2 rpms from /usr/src/fedora/ and custom packages from /usr/src/fedora/extras. At the moment this all works fine installing to a HDD on the build host (i.e put a drive in as /dev/hdc and choose it as a target using the dialog based interface to the script). Previously we made an automated version of this process that ran from a CD-ROM. Moving this CD-ROM setup to install the FC2 rpms has been a right pain. Just as I thought I'd solved all the problems I starting running into issues with exec-shield enabled programs running in %post scripts after the rpm installs. The kernel was an older 2.4.18 kernel and didn't have the extensions for exec-shield. So I took a chance and replaced the vmlinuz with the one from the FC2 CD, made sure my initrd had insmod in it and the modules from stage2.img's modules dir in /modules (with the cpio extracted out). Now when I try to insert the ext3 module I get a -1 error with an unknown symbol in module. Now I'm thinking that I might just put what was in my initrd into the stage2.img and have a /usr/bin/anaconda symlink to my own install script. i.e. Let the FC2 cd boot as it normally would, have the loader program insert all the modules and call the anaconda script, which is a link to my own. Would this work ? I've seen mention of genhdlist commands mentioned before and wonder if that's only used by anaconda proper and not needed by the earlier part of the boot process. TIA, -- Glen Gray <glen@xxxxxxxxxx> 17 Dame Court Senior Software Engineer Dublin 2, Ireland Lincor Solutions Ltd. Ph: +353 (0) 1 6746413
