Core 4 Anaconda Problems

[Rob Wright]

On Fri, 15 Jul 2005 23:41:12 -0400
  Jim Cornette <fc-cornette at insight.rr.com> wrote:
> Rob Wright wrote:
> 
>> On Thu, 14 Jul 2005 17:30:52 -0400
>>  <fedorarob at poncacity.net> wrote:
>>
>>> This happened to me with core 3, and now with core 4. 
>>>I'm installing 
>>> along, fresh install completely. As soon as all the 
>>>packages are 
>>> transferred from the CD, anaconda bails out with a list 
>>>of gui.py 
>>> errors and tells me to reboot. Is anaconda damaged, or 
>>>am I just 
>>> gonna have to stick with FC2? I realize the Fedora Core 
>>>is a test 
>>> distro, but I can't test it if I can't even install it. 
>>>I've burned a 
>>> fresh set of discs, twice, isos downloaded from two 
>>>different 
>>> mirrors. My hardware is ok. If I install FC2 I get the 
>>>exact same 
>>> problem with updating from the discs. I can use yum to 
>>>update to Core 
>>> 4, but that is such a pain in the ass.
>>>
>>
>> Just a little more information as I'm still working on 
>>this problem. 
>> I'm going through the suggestions that have been offered 
>>so far, but 
>> no luck.
>>
>> Going through a graphical install, the packages aren't 
>>actually 
>> "installed" it seems. The last line of the error that 
>>I'm getting is:
>>
>> RuntimError:/usr/sbin/kudzu cannot be run
>>
>> I booted into rescue off the cd and I was able to run 
>>kudzu-probe from 
>> there. I decided I'd try to install once more, using the 
>>noprobe and 
>> text options. It's copying files now, but there's a 
>>continously 
>> scrolling message on the screen:
>>
>> rpmdb: PANIC: fatal region error detected; run recovery
>>
>> Which also happens in the background in a graphic 
>>install, but I don't 
>> see it until the aborted install forces the reboot and X 
>>closes.
>>
> 
> I just did a search for "fatal region error detected" 
>and most responses referred to the rpm database being 
>corrupted. If you are trying to install, this sounds like 
>it is not reading what was written to disks. By your 
>description, this seems to be the case.
> 
> WAG1:
> Try linux selinux=0 as a boot option. (I have no idea if 
>this is preventing writing, but I disabled selinux on 
>most development upgrades for safety during package 
>upgrading with development.
> #2
> Try linux ide=nodma
> 
> To get away from the guessing, thee is some documentatio 
>somewhere forkernel options and what they do. Have you 
>checked your hardware, cabling, ran memtest at boot  ... 
>?
> 
> Sorry not much help on this problem.
> 
> Jim
> 
> I'm stumped with your problem
> 

Jim,

Thanks for the suggestions, but nothing has worked. To 
summarize:

All cabling and hardware has been checked out.
memtest reports no probems in RAM.
ide=nodma and selinux=0 made no difference.
Same thing happened with Core 3 and was resolved only by 
going back to Core 2.

So, I'm going to try and install a different distro, just 
to see if it makes any difference.