On 19 Jun 2002, seth vidal wrote: > 1. stop using lilo ;) Oh, I did, and am. In fact, getting the system to boot again I learned how to drive grub at the command line (determining along the way that here is yet another package that has documentation from hell). I just have a small string of systems in various states of legacy because two or three of them are/were being used as specialized servers (requiring custom built stuff that were a LOT of work to get working, originally). They were inside a firewall, safe, and unused except for totally boring beowulfish stuff and print serving, so why invest energy in upgrading? Now the HP officejet stuff is in 7.3 -- no incentive at all to keeping them behind just to preserve a setup (however much of a bitch it was to hack ghostscript and the new printer database system the first time). I'm crossing my fingers that I can reinstall printer and scanner drivers in five notes. The system in question was sufficiently munged by all my hacking that I had to do a full reinstall anyway. Both yup and yum barfed "instantly" with a python error. Not a big deal. > 2. the lilo stuff is stolen from up2date - chances are your lilo.conf > was munged or odd, or something was not available in it that should be. > (did you rpm -e a kernel that you didn't remove from the lilo.conf?) Sigh. Of course -- the mkinitrd kernel 2.4.3 conflict. I actually did have a clue and installed grub before rebooting, but I didn't get the methodology right the first time -- install on /dev/hda (to replace lilo in the MBR) plus the (hd0,[0,1,2...]) notation. I basically installed it on the wrong partition and got stuck in lilo. This isn't a bug in yum, I know, EXCEPT insofar as you try (if you try) to write an upgrade option that manages e.g. the kernel/mkinitrd conflict. It will either need a bit of smarts to correctly patch lilo.conf (risky, obviously) or it will need a helpful little message like: "Remove kernel XXX from lilo.conf and rerun lilo or install and configure grub" or it will need to steal the grub/lilo sequence from the regular RH upgrade/install, if possible. So it is a line on a possible feature. rgb Robert G. Brown http://www.phy.duke.edu/~rgb/ Duke University Dept. of Physics, Box 90305 Durham, N.C. 27708-0305 Phone: 1-919-660-2567 Fax: 919-660-2525 email:rgb@xxxxxxxxxxxx