On Thu, 2003-08-21 at 17:35, Klingaman, Aaron L wrote: > > Why not do it the other way around -- just bootstrap the base > > group (or if you have plenty of time to kill, the base group > > with things selectively omitted until it starts breaking all > > the time but with modern bandwidth and disk, why bother) and > > then switch over to yum? That was/is how I planned to do it. > > I noted that Seth even put yum in the base group here, which > > makes it pretty easy to do it. > > Well, you still neeed some mechanism to install that base group. Given > the environment my project has to run it: RedHat9 needs glibc 2.3.2, and > the environment its running in (a bootcd) is an older version of glibc. > > Currently, this is working: > > 1. mount the install partition somewhere. > > 2. use cpio to extract the following rpms into that mount point: > bzip2-libs-1.0.2-8.i386.rpm > elfutils-libelf-0.76-3.i386.rpm > popt-1.8-0.69.i386.rpm > rpm-4.2-1.i386.rpm > glibc-2.3.2-11.9.i386.rpm > > This is enough to allow rpm to run > > 3. initialize the rpm db in the mount point using > chroot $SYSDIR /bin/rpm --initdb > > 4. install rpms necessary to run yum, including python and the yum rpm. > turns out this is a fairly big list, like about 50 rpms. > They were installed like: > chroot $SYSDIR /bin/rpm -ivh /tmp/stage2rpms/*.rpm > > 5. configure yum in the mount point > > 6. copy resolv.conf to $SYSDIR/etc/ > > 7. finally: > chroot $SYSDIR yum -y groupinstall 'Base' > > which finishes up the install of the base group. Seems to be working > great. What I don't like is how I had to by hand detiremine which rpms > to install first, to get the yum rpm to install. > > I'm still messing around with this, so I'm open to any suggestions at > this point. > > Aaron I'm assuming that this is a blank system that your installing onto, that has been pre-partitioned? And can I use something like toms boot root disk to get started? -- ***** Not everyone is touched by an Angel.... .... Those that are, never forget the experience *****
