On Monday 13 February 2006 02:58, Ross Boylan wrote: > I'm tempted to file a bug against the installation manual, but am not > sure if that's appropriate. Could anyone advise me about that? Well, the appendix has not been really written for newfangled setups like that, nor has it been adapted for possible extra demands that recent kernels and use of udev may put on it. The permutations are just to many. > Here are some theories about what the problem is: > 1. I've missed something. > 2. The manual is incomplete. > 3. debootstrap is incomplete. > 4. udev is missing something. > 5. evms needs to cater to this situation. I have no idea which of those is most appropriate, but filing a bug against the manual saying "it does not work in my setup" is not going to solve anything (except maybe the addition of a note saying that the appendix wont work for setting up an evms system). > There's a technology overload going on: evms requires that /dev/evms > be available. I'm guessing this depends on udev with recent kernels. > udev requires procfs, but that is not mounted in the chroot (though it > can be mounted, and would be mounted if the mount -a worked). > > I notice that the debootsrap created system does not have the udev > package installed. I wonder if it should for newer kernels (running > 2.6.14). If you're going to install a system using debootstrap, you're supposed to be a little more savvy than people using the installer. debootstrap is just not smart enough (and not intended to be) to solve such riddles. My suggestion would be that you try to find out on other lists what the best solution is, document that as a patch (probably a short extra para) for the appendix and submit that as a bug report against the manual. That is, of course, if you're willing to make the effort. Cheers, FJP
Attachment:
pgpvQnnc_H760.pgp
Description: PGP signature