Red Hat Bugzilla – Bug 57047
Installer crashes with python error after generating new hdlists
Last modified: 2007-04-18 12:38:33 EDT
From Bugzilla Helper:
User-Agent: Mozilla/4.78 [en] (X11; U; Linux 2.4.7-10 i686)
Description of problem:
Using all the RPMS directly from the 7.2 CDs for an nfs kickstart install
works fine if I use hdlist and hdlist2 from the CD. If I run
/usr/lib/anaconda-runtime/genhdlist <dir> to create new hdlist and hdlist2,
the kickstart install fails (see error msg below).
Version-Release number of selected component (if applicable):
Steps to Reproduce:
1.prepare for nfs kickstart install
3.begin kickstart install
Actual Results: Install fails and shuts down to "you may safely reboot
Expected Results: a completed kickstart install
The install generates this error:
rpmio_internal.h:199: fdGetFp: Assertion 'fd && fd->magic == 0x04463138'
What release are you running on the machine running genhdlist?
Running Redhat 7.2 on the machine generating the hdlists
Ok.. apparently I lied to you. After deleting my rpm directory and recopying
the files off the CD, I can run genhdlist and an install will still work. I do
note that the hdlist file is a different size from the one on the CD, though
hdlist2 is still the same size.
I will try to replace older rpms with updated versions, run genhdlist and
reinstall. Perhaps I had a bad copy of an rpm. Sorry I didn't try that
Ok. Please reopen the report if that doesn't fix the problem. Thanks for your
Just for the record, I had the same problem with the exact same fd->magic error
message. I traced it down to an RPM of the xforms toolkit that I had added to
my distribution. (I need it to run Geomview, a mathematical application.)
It seems that this RPM actually came from the PLD distribution and something in
kickstart doesn't like it. I can provide the RPM on request, but 700K is a bit
much to attach here. Or search for "xforms-0.89-2.i386.rpm". After hacking the
src.rpm a bit to build, I have a working binary RPM.
So while it would be nice of kickstart failed a bit more gracefully or if one of
the build tools (genhdlist, perhaps) would parse the RPMs in the same manner
that kickstart does, this is certainly not a serious problem.