From Bugzilla Helper: User-Agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.4) Gecko/20030730 Description of problem: System config: 1. / on LVM1 2. /boot on ext3 mkinitrd does not produce a bootable system. this affects both an upgrade from rh8 to rh9 & a new kernel install. Version-Release number of selected component (if applicable): How reproducible: Always Steps to Reproduce: 1. Install a new RedHat-packaged kernel 2. Reboot on the new kernel 3. Actual Results: Kernel Panic: no root found Expected Results: System should boot Additional info: When creating a new initrd, mkinitrd embeds insmod.static but not insmod.static.old. As a result, no module can be loaded at next boot time & I end-up with a "Kernel Panic" because no root can be found. This bug is of the "How could this ever work?" type. It was already present on RedHat Linux 8. The patch applies to both rh9 & rh8.
Created attachment 95294 [details] Embed both old & new insmod
This is because you installed module-init-tools which move things around poorly on the system. If you want to use 2.4 and 2.[56] kernels, install the modutils from rawhide which also includes the module-init-tools stuff but in a way that doesn't break backwards compatibility.
Uuh? 1. I do not play with 2.[56] kernels 2. I installed my initial rh8 from scratch (rh installed) 3. I upgraded from rh8 to rh9 booting from the rh9 cdrom What is that bad in the above? Shall I assume from now that Rawhide is the only safe branch?
insmod.static.old only ever exists if you install the upstream module-init-tools which is only required for the 2.6 kernel. Otherwise, there is no insmod.static.old to be concerned with. And module-init-tools renaming binaries is *broken* and impossible for distribution scripts to support.