Red Hat Bugzilla – Bug 231494
Anaconda crashes during scan for existing Fedora systems
Last modified: 2007-11-30 17:11:58 EST
Description of problem:
When attempting to do an install, today's and yesterday's boot.iso versions
crash when scanning for existing systems.
Version-Release number of selected component (if applicable):
100% on my system.
Steps to Reproduce:
1. Try to do an install with today's boot.iso file.
Anaconda detects existing Fedora installations.
Due to another problem neither of my hard drives are visible so there are zero
disk partitions to search. This may or may not have any bearing on the above.
I also tried using the updates feature to use update *.py files from cvs, but
saw the same problem.
Created attachment 149601 [details]
Anaconda dump file
I retested this using a boot.iso version from early March 13th and the problem
is still happening.
What kind of hard drives, controller, etc. are you using?
You can look at 227281 to see that saga. This ticket wasn't about getting
anaconda to see the drives, but rather about the failure more.
The controller that has the drives attached is the highpoint 302. It has 2 PATA
WD Caviars attached. (The motherboard has two sets of controllers. The AMD one
has DVD drives attached, and nothing is connected to the promise controller.)
When the problem with the drives not showing up first started happening, I
didn't get tracebacks from anaconda. It just showed an empty list of available
disk drives. Something changed recently with that.
This should be fixed with the booty that I built this afternoon and which will
be in test3
Created attachment 151368 [details]
I just retested this with the Friday (March 30) morning boot.iso and it seems
to still be happening.
However, kudzu changed and my disk drives became visible again, so I had to
change the way I tested this. When I got to the screen after the media check I
switched to tty2 and unloaded the disk controller driver and then continued the
install from X. I then saw a crash at the same point I had previously.
I am attaching the saved traceback.
I retested this with a snapshot from Friday April 13 and things now work as