First modify run() in kickstart.py file found in stage2.img -- have it check for the existence of a temp file, if that file doesn't exist then write the file and execv so that anaconda re-reads the ks.cfg file. In the %pre section of the ks.cfg check if the temp file exists and if it doesn't, then use sed to modify the partition lines in /tmp/ks.cfg depending on dtype. Good Luck! "Bodle, Donald E" <donald_bodle@reyrey To: "'kickstart-list@xxxxxxxxxx'" <kickstart-list@xxxxxxxxxx> .com> cc: Sent by: Subject: selecting disk for onpart in Pre8.0 kickstart kickstart-list-admin @redhat.com No Phone Info Available 04/04/2003 12:56 PM Please respond to kickstart-list I have a kickstart config file which allows me to do a fresh o/s install without blowing away user data. It is predicated on the idea that we always partition our systems the same way. So, I don't do a clearpart, and I use onpart as follows: part /boot --onpart sda1 part swap --onpart sda7 part / --onpart sda10 part /tmp --onpart sda6 part /usr --onpart sda5 part /var --onpart sda8 part /opt --noformat --onpart sda11 part /rs1 --onpart sda9 I don't even mention the partition that has the user data, so it is left totally unaffected. My problem is this. I'm trying to do this in such a way as to support both ide and scsi disk systems. Right now I have to use 2 different CD versions, one for each. Is there a way to set a variable in the %pre section that can be used in place of the h or s in as in: part /boot -onpart ${dtype}da1 There are different ways to determine what dtype should be in %pre. My problem is whether or not I can make use of that information in the -onpart option. Note I'm currently using RH 7.1 with latest errata, so I don't think there is a %include option is this version. If so, or when I go to RH 8/9, I could probably use the %pre to build the target of the include then (I hope). Donald E. Bodle, Jr. Sr. Systems Integration Engineer Platform Development The Reynolds and Reynolds Co. (937) 485-1954