Description of problem: The installation program traps itself in an interactive loop before collecting all information, and never actually starts (much less finishes) installing the system onto the hard drive. Version-Release number of selected component (if applicable): anaconda from squashfs.img on Fedora-8-Test-1-Live-i686.iso [no /var/log/rpmpkgs on running LiveCD system, so how should one extract version info for this package?] How reproducible: Solidly. Has failed every time in several attempts. Steps to Reproduce: 1. Download Fedora-8-Test-1-Live-i686.iso 2 # sha1sum -c SHA1SUM Fedora-8-Test-1-Live-i686.iso: OK. 3. Boot a system from the Fedora-8-Test-1-Live-i686.iso, select "Verify and run from image", media check PASS, 4. Select "Install to Hard Drive" iocn on desktop. 5. Make choices and select Next on each screen, but no further progress can be made after Network device screen. Actual results: A dialog titled "Error!" is displayed containing: [! icon] An error occurred when attempting to load an installer interface component. className = TimezoneWindow Exit Retry Selecting "Retry" just leads to a repetition (even with many attempts) of the same dialog, with no progress, Selecting "Exit" leads to a dialog titled "Exiting" containing: [! icon] Your system will now be rebooted. Reboot But no restart actually happens (and not clear that one should happen). Expected results: The timezone installation dialog and the remainder of the installer should be displayed, so that the installation completes. Additional info: When a Nautilus terminal window (from Applicatgions, System Tools, Terminal) is open, the following lines will be displayed in it: Loading /lib/kdb/keymaps/i386/qwerty/us.map.gz No module named zonetab [this line repeated for each Retry attempt ...] The terminal lurking behind Ctrl+Alt+F3 shows: : INFO : moving (1) to step autopartitionexecute INFO : Detected 512M of memory INFO : Swap attempt of 512M to 1024M INFO : moving (1) to step partition WARNING : /usr/lib/anaconda/gui.py:1158: Warning: IA__g_object_remove_toggle_ref: couldn't find toggle ref 0x73ecfc0(0x964798c) self.currentWindow = None INFO : moving (1) to step partitiondone INFO : moving (1) to step bootloadersetup INFO : moving (1) to step bootloader INFO : moving (1) to step networkdevicecheck INFO : moving (1) to step network INFO : moving (1) to step timezone
Are there any error messages on tty4? This sounds like bad media, kernel driver troubles, etc.
Created attachment 161837 [details] SQUASHFS errors found via /var/log/messages Once I finally invested the time to attempt to repeat the situation, after several tries I was able to verify my memory that tty4 was still at its login prompt, with no information logged from this "Install to Hard Drive" attempt. As originally reported, this CD passed both the sha1sum and media check tests. But this time I noticed that /var/log/messages has some evidence to support your expectation. See the attached file for aquashfs errors from the CD, which are obviously intermittent since this process sometimes runs to completion without encountering these errors from this CD on the given CD drive. Would it be appropriate to share your insights in a more detailed error message available in this situation? E.g. a dialog titled "Error!" would be displayed containing: [! icon] An error occurred when attempting to load an installer interface component. className = TimezoneWindow [twisty] Details This likely arises from a media or device error: Consider: - Exiting and running the media check test, if this particular disc has not already passed that test. - Viewing the recent contents of /var/log/messages to see if the device reading this disc encountered errors. If so, Exit, and restart this procedure in case this may have been a transient error. Exit Retry Include mention of tty4 if you know that sometimes it reveals useful messages in this situation. Is there a simple way to distinguish between different data being returned from the same pits in the optical media on different runs, and squashfs driver failures?