On Sun, 3 Feb 2002, Steven J. Yellin wrote:
> On Sat, 2 Feb 2002, Ian Firla wrote: > > > > There must be a way of doing that, but I don't know it. Instead you > > > could copy the kernel rpm's from Redhat's site or from a mirror, then > > > install them yourself, after finding out how such a thing is done -- see > > > > > > http://www.redhat.com/support/resources/howto/kernel-upgrade/kernel-upgrade.html > > > > > > One of the effects of the procedure will be to update the rpm database. > > > > but on the effect of doing that will be to "upgrade" my 2.4.16 kernel to > > 2.4.9... > > > Ok. So maybe you can't get as recent a version rpm as you'd like from > RedHat, and you have to either use an older version or get a more recent > one from elsewhere. For example, from http://rpmfind.net/linux/rpm2html/ > one may search for "kernel" and find rpm's for kernels up to 2.4.17-1 on > ftp://rpmfind.net/linux/PLD/current/dists/nest/PLD/i686/PLD/RPMS/ 2.4.17-1 at rpmfind is for PLD linux, not Redhat. The RPM with the latest kernel for Redhat is 2.4.9-21 The only options, it seems to me, are to roll your own rpm, wait for Redhat to do so, or, as someone on the Oxlug mailing list has suggested to me: >Recently Alan Cox has added an rpm target to the main kernel Makefile, so >you can do 'make xconfig rpm', and have an rpm built for you. Then you can >either install, or use --justdb to tell rpm about it, but not mess with >the filesystem. Or, as has also been suggested: >My, not terribly nice, solution I've used in some places is to have an >empty kernel RPM (and corresponding kernel headers RPM), with a version >corresponding to my real kernel. Saves the faff of actually trying to >RPM it whenever you want to change the configuration. The trouble with using a generic RPM for a kernel is that, often, there's a lot of overhead built in by distros like Redhat to allow for maximum portability. That usually means a large and somewhat bloated kernel. My machine has been much more spritely since I rolled my own kernel. What surprised me was the effect it'd have on the rpm database. I must say, the Cox-based solution sounds sensible. All the best, Ian _______________________________________________ Seawolf-list mailing list [EMAIL PROTECTED] https://listman.redhat.com/mailman/listinfo/seawolf-list
