Hi everyone,

I am trying to set up an automated CentOS installation system. My idea is to use PXE to boot the kernel/initrd and pass the kickstart options to the kernel from there.

This works correctly until it boots up Linux. As soon as it does that it makes a DHCP request and, unsurprisingly, it gets the same one (same MAC) but it also gets the filename and next-server details. Then Anaconda tries to use those details instead of the ones I specified on the command line.

You can see on one of the other VTs that it is trying to load /pxelinux.0 instead of using http://server/1.ks.

I can see three possible solutions for this: -Forcing (somehow) Anaconda to use a fixed IP address (possibly via a %pre script?), or, at least, making it use the kickstart file that gets passed from pxelinux. -Configuring the DHCP server so it knows how to differentiate between requests (not too sure on how to do it yet). -Doing a complete hack. When it gets the second request, it tries to access nfs:/server:/pxeboot.0. I could create the NFS share and add the file there, maybe symlinked to the right one, but it would stop me from being able to have different installations selectable from pxelinux.

My ideal solution would be the first, as the other two seem to be workarounds, more than solutions.

Any help will be greatly appreciated.

Gabriel

___________________________________________________________ What kind of emailer are you? Find out today - get a free analysis of your email personality. Take the quiz at the Yahoo! Mail Championship. http://uk.rd.yahoo.com/evt=44106/*http://mail.yahoo.net/uk