First: I have upgraded my purchased product from Linux 6.0 to Red Hat Professional 6.1 and decided to make a new complete installation. So, during the complete SERVER installation from the CD1 I was not asked to create bootable floppy. After completed installation the system attempted to boot from the hard disk but hanged with LI. Strange, because the /boot is created at the beggining of the disk, (and is small: only 17 blocks). Just to help myself I have tried to boot from the bootable floppy which I got during 6.0 installation: My new system booted, but did not activate some loadable modules, because the kernel versions were different and were no suitable module directory. In other words, the default installation of the Server version 6.1 may result in non-bootable system. My hardware: standard IBM Netfinity 7000 server.
A workaround (works if you have 6.0 system, or disk with it, at hand): The following trick worked OK: 1. I have two disks on my system. As reported, the basic "Server" 6.1 installation on one of the disks failed, because lilo hangs with LI and no bootable floppies were created. 2. I had another disk in the system with 6.0 system installed - the 6.0 procedure included creation of bootable floppy, which made the difference. 3. Used the bootable floppy from the 6.0 system to boot from the newly installed 6.1 system. It reported missing /lib/modules/2.2.5-* kernel modules, and as a result I could not use the "mkbootdisk" utility. 4. From another disk (you can do it from another system using floppies) I have copied /lib/modules/2.2.5-* to my newly installed system, which contained own /lib/modules/2.2.12-* modules. 5. Then the system was rebooted (using the same 6.0 bootable floppy). This time no missing modules were reported (obviously). 6. Noted the name of the new 6.1 kernel (ls /lib/modules had shown the full name of the 2.2.12-* kernel, say 2.2.12-22). Executed # mkbootdisk 2.2.12-22 The new created bootable floppy booted the 6.1 version of kernel, as intended. (One could remove now the /lib/modules/2.2.5-*)
As explained above, I was able to create bootable LILO floppy, which boots the 2.2.12-* kernel. However, running lilo with the standard /etc/lilo.conf (which installed lilo on /dev/sda disk), and attemping to boot without bootable floppy, still resulted with handing at "LI" prompt. Now, I experimented with modifications to the /etc/lilo.conf and it happened, that for my HD (which is 9.1 GB SCSI disk) the "linear" option solved the problem. My current lilo.conf file looks like follows: boot=/dev/sda map=/boot/map install=/boot/boot.b prompt linear timeout=50 default=linux image=/boot/vmlinuz-2.2.12-20smp label=linux initrd=/boot/initrd-2.2.12-20smp.img read-only root=/dev/sda8 image=/boot/vmlinuz-2.2.12-20 label=linux-up initrd=/boot/initrd-2.2.12-20.img read-only root=/dev/sda8
Assigned to dledford
In older releases, you were required to select the linear box during the install process to use linear mode, which is required for some scsi controllers and drives. In newer releases, we always use linear mode when installing lilo.
*** Bug 7790 has been marked as a duplicate of this bug. ***