F8 install failure again & again

[Craig White]

On Sat, 2007-12-15 at 17:47 -0500, Felix Miata wrote:
> I've just made my 5th unsuccessful full attempt to install F8 on a 2 PATA
> disk i815 Dell PIII 1133 MHz 512M RAM system that has the following fully
> functional operating systems:
> Debian Etch
> DOS
> Fedora 6
> Fedora 7
> Knoppix 4.0.2
> Kubuntu 7.10
> OpenSUSE 10.0
> OpenSUSE 10.2
> OpenSUSE Factory
> OS/2
> SUSE 9.3
> WinXP
> Xandros 3
> 
> On each successive attempt I've added more of the "Overriding Automatic
> Hardware Detection" options from docs.fedoraproject.org, the last time using
> about 80% of them.
> 
> On the first aborted attempts, I was trying to include the various IP
> assignments as on "Manually Configuring the Network Settings", but each time
> it wanted manual entry of the IP, using the cmdline parameters only for mask,
> gateway & dns, then aborted immediately upon completing network configuration.
> 
> I gave that up and tried with nothing on cmdline. That got me all the way
> through package selection, but then anaconda aborted as soon as told it to
> proceed to install the selected packages.
> 
> >From then I used as a minimum acpi=off and noapic, and got anaconda to claim
> its first successfully completed installation. It couldn't be booted though.
> First I had to boot it from F7's grub, since the only locations it had
> permitted me to choose from were nowhere and on the mbr. Since MBR is not an
> acceptable location for any additional Linux installation, I was forced to
> choose nowhere. That boot attempt produced the following:
> 
> ...
> Waiting for driver initialization
> Loading ata_generic.ko module
> noresume passed, not resuming...
> creating root device
> 
> That's where it simply stopped.
> 
> I got it to boot using the F7 kernel, whereupon I found no eth0 and corrupted
> rpm DB. modprobe e100 got me network. After finding 97% of the install log's
> entries were rpm errors, rpm --rebuilddb got yum usable for installing the
> crucial mc rpm missing from available packages to select from. Next I did yum
> update. From then on, I couldn't complete a boot with either of the F8
> kernels nor the F7 kernel.
> 
> So I went on to try from scratch again, twice, getting similar results both
> times. More cmdline options made it even worse. Two more times yet again
> anaconda claimed complete, even though it didn't bother installing any
> kernel! Right now on tty2 after 5th full try I'm looking at the output of ls
> -l /mnt/sysimage/boot, and see only:
> 
> drwxr-xr-x 2 root 0 1024 Sep 20 20:21 grub
> 
> Right now, cat /proc/cmdline:
> 
> noapic acpi=off skipddc nodmraid nofirewire noparport nopcmcia nousb vga=788
> 
> Target is Maxtor 20G PATA secondary master 4.8G partition 11 out of 13 total
> shared by DOS, OS/2, SuSE 9.3, OpenSUSE 10 & Knoppix. Seagate PATA primary
> master has 120G with >16  partitions. Primary slave is ZIP100.
> 
> On attempt 4, I tried to install mc and a kernel:
> 
> mount -t ext3 /dev/sdc11 /mnt
> mount -o bind /proc /mnt/proc
> mount -o bind /dev /mnt/dev
> mount -o bind /sys /mnt/sys
> cd mnt
> chroot /mnt
> 
> That resulted in too many errors to remember, but it seems package management
> is unavailable in chroot.
> 
> I'm starting install each time with grub on hda5, specifying
> mirrors.kernel.org /fedora/releases/8/Fedora/i386/os as source.
> 
> Logs from an early attempt:
> http://mrmazda.no-ip.com/tmp/anaconda-ks-f8-gx150.cfg
> http://mrmazda.no-ip.com/tmp/anaconda-f8-gx150.log
> http://mrmazda.no-ip.com/tmp/anaconda-f8-gx150.syslog
> http://mrmazda.no-ip.com/tmp/anaconda-f8-gx150.xlog
> 
> ATM, I'm running a text install using noapic acpi=off, but am not hopeful of
> materially different results.
> 
> Without making hardware changes, I'm at a loss what else to try. Ideas?
----
with hardware that's been around that long, you shouldn't have to add
any of those kernel parameters to get install but you might consider
only installing @base (maybe @editors) and keeping it simple because you
can then boot, do a 'yum groupinstall @whatever' afterwards

Craig