On Fri, Sep 08, 2006 at 04:15:04PM -0700, Dan Nicholson wrote: > too difficult to provide this. (I'm ignoring the problem you've > described with the ethX interfaces right now as I'm not exactly > following it.)
I don't know if this will help, but this is how I understand it: The point of the persistance scripts is that it's entirely possible that the order devices get created (and therefore named) may change between boots. What the auto-generating scripts do is make sure that they are always created in the same order as on the first boot. However, we are setting up networking _before_ that first boot. So there's no guarantee that they will be created in the same order when we reboot. For example, suppose we have a tulip card and a realtek card. The tulip should have 192.168.0.1, and the realtek should have 10.0.0.1. Suppose futher that the tulip is eth0 and the realtek is eth1 at the moment. So we set up the networking scripts to assign 192.168.0.1 to eth0 and 10.0.0.1 to eth1. Now suppose they get created the other way round when we reboot. The scripts then generate rules that persistantly assign eth0 to the realtek and eth1 to the tulip. So now the networking scripts will assign 192.168.0.1 to the realtek and 10.0.0.1 to the tulip, and will continue to do so on subsequent reboots. There's no real way to get around this without setting up the network after rebooting. Of course, it doesn't matter if you're using dhcp, but that's not covered in the base LFS book. Alex :-) -- Pippin Computer Monkey to the Pelican www.oxrev.org.uk, www.corpusjcr.org, www.rev.org.uk
pgp7m18L6XWmZ.pgp
Description: PGP signature
-- http://linuxfromscratch.org/mailman/listinfo/lfs-dev FAQ: http://www.linuxfromscratch.org/faq/ Unsubscribe: See the above information page
