Red Hat Bugzilla – Bug 63610
Installer shows bug window during package install
Last modified: 2007-04-18 12:42:02 EDT
From Bugzilla Helper:
User-Agent: Mozilla/4.0 (compatible; MSIE 5.5; Windows 98; H010818)
Description of problem:
I was doing a fresh install on a system with the USB keyboard and mouse plugged
in through a USB KVM switch. I got the install going to where it was
installing packages, then I switched the KVM over to another system to do some
web browsing. After 15 minutes or so, I switched back to the installing
system, and it showed a bug window. It gave me the option of writing the error
to a floppy, and here's the contents of that file:
Traceback (innermost last):
File "/var/tmp/anaconda-7.0.1//usr/lib/anaconda/iw/progress_gui.py", line 20,
rc = self.todo.doInstall ()
File "/var/tmp/anaconda-7.0.1//usr/lib/anaconda/todo.py", line 1543, in
File "/var/tmp/anaconda-7.0.1//usr/lib/anaconda/todo.py", line 1298, in
self.rpmFD = os.open(fn, os.O_RDONLY)
OSError: [Errno 2] No such file or directory: '/mnt/source/RedHat/RPMS/libtiff-
Local variables in innermost frame:
self: <todo.ToDo instance at 8444960>
h: <header object at 8b88f30>
intf: <progress_gui.InstallProgressWindow instance at 844ed90>
Version-Release number of selected component (if applicable):
Steps to Reproduce:
Actual Results: See description.
Expected Results: See description.
For the install, I selected all packages. This was the second time I've
installed off this RedHat CDROM. The first installation went easily (no bugs),
and that was to this same system (except with a different hard disk). This
second install stopped with the bug window mentioned above. After saving the
bug, the system rebooted and I ran the install again (entering identical
parameters), and this time I did not touch the KVM switch. This time
everything worked fine. I didn't touch the CDROM between these last two
The system has an Abit BX6r2 motherboard, a GeForce 256 video card,
Soundblaster Live oem audio card, a 3Com 3c905 NIC, an IDE HD & DVD drive, and
a Silitek USB keyboard and 3M USB Renaissance mouse.
This error is because it could not read a package from the CD. Either the CD is
bad or the CD drive is misbehaving.
I think you still have work to do. It's obvious that the CD and drive were
able to read fine since it worked without problem for two other installations.
Your installer should not completely bail just because it had difficulty
reading one file. First, it should do some retries. Second, it should allow
the option of skipping that file if it is not essential. The idea is that
software be user friendly. It shouldn't give a cryptic, non-understandable
message, and force me to start again from scratch without understanding what is
I agree that currently it is not easy to follow why this happens. The kernel
does not give much information about some of the types of read errors that
occur, other than having a system call fail. In our Skipjack beta which just
came out we have tried to use more of the errors which occur to notify the user.
I have seen enough bug reports for anaconda to realize that CD hardware is
finicky, and just because it works once doesn't always guarantee it will always
work. If you could md5sum the CD a few times and see if you get the correct
value it would help identify the source of the problem.
Closing due to inactivity - please reopen if you have additional comments to add
to this bug.
Changed to 'CLOSED' state since 'RESOLVED' has been deprecated.