From Bugzilla Helper: User-Agent: Mozilla/4.0 (compatible; MSIE 6.0; Windows NT 5.1; SV1; .NET CLR 1.1.4322) Description of problem: I received the following exception when trying to install Fedora Core 4: (unhandled) Exception occurred File "/usr/lib/anaconda/gui/py", line 1137, in handleRender callback self.currentWindow.renderCallback() File "/usr/lib/anaconda/rw/progress_gui.py", line 244, in renderCallback self.intf.icw.nextClicked() ... 8 more lines of this stuff ... System error: pvcreate failed for fsset.partitionDevice instance at 0x2aaaaf137170 I tried it again, and it failed trying to install mozilla. Then I tried it again and it failed trying to install 4Suite-1.0-8.b1.x86.64. Each time I had to restart completely anew; it should not have had to re-partition the hard disk, but it would not continue unless I did. I downloaded FC4-x86_64-disc1 thru 5, and tested all of them, but received the above errors anyway. Version-Release number of selected component (if applicable): How reproducible: Always Steps to Reproduce: 1. Boot into FC4-x86_64-disc1 2. Allow the installer to automatically partition the disk 3. The first time the above error appears, but if that is circumvented the installation still fails trying to install a component. Actual Results: Failed trying to install a specific program, such as mozilla or 4Suite Expected Results: Install and run. Additional info: Fedora is not ready for prime time.
Can you provide the complete traceback you received?
Exception occurred report to http://bugzilla.redhat.com/bugzilla under anaconda File "/usr/lib/anaconda/gui/py", line 1137, in handleRender callback self.currentWindow.renderCallback() File "/usr/lib/anaconda/rw/progress_gui.py", line 244, in renderCallback self.intf.icw.nextClicked() File "/usr/lib/anaconda/dispatch.py", line 242 in moveStep rc=apply(kune, self.bindArgs(args)) // kune may be wrong; can't read own writing File "/usr/lib/anaconda/packages.py", line 582, in turnOnFileSystems thefsset, createLogicalVolumes(instPath) File "/usr/lib/anaconda/fsset.py", line 2066 in setupDevice raise SystemError, pvcreate failed for %S %(volume,) System Error: pvcreate failed for fsset.partitionDevice instance at 0x2aaaaf137170
Almost the same situation in my case. I'm trying to install FC4_x86_64 on Fujistsu-Siemens RX200-S2, select automatic disk partitioning, and have almost the same backtrace, although ending with SystemError: lvmremove failed Backtrace: Traceback (most recent call last): File "/usr/lib/anaconda/gui.py", line 1137, in handleRenderCallback self.currentWindow.renderCallback() File "/usr/lib/anaconda/iw/progress_gui.py", line 244, in renderCallback self.intf.icw.nextClicked() File "/usr/lib/anaconda/gui.py", line 873, in nextClicked self.dispatch.gotoNext() File "/usr/lib/anaconda/dispatch.py", line 174, in gotoNext self.moveStep() File "/usr/lib/anaconda/dispatch.py", line 242, in moveStep rc = apply(func, self.bindArgs(args)) File "/usr/lib/anaconda/packages.py", line 576, in turnOnFilesystems partitions.doMetaDeletes(diskset) File "/usr/lib/anaconda/partitions.py", line 1285, in doMetaDeletes lvm.lvremove(name, vg) File "/usr/lib/anaconda/lvm.py", line 143, in lvremove raise SystemError, "lvremove failed" SystemError: lvremove failed Local variables in innermost frame: lvname: LogVol00 vgname: VolGroup00 args: ['lvm', 'lvremove', '-f', '/dev/VolGroup00/LogVol00'] dev: /dev/VolGroup00/LogVol00 rc: 1280
Find workaround for this problem: Before it goes to partioning, switch to console and using fdisk remove all existing partitions Then, when choose 'Autocreate partitions", but does not allow installer to remove all existing partitions (since there are no partitions existing at the moment), and choose "occupy free or unallocated space"
The suggestion from Nikita Shulga may well work, because if one runs the install twice, it gets by this step on the second try without using any workarounds. But after proceeding for awhile my install failed twice -- once when it was trying install Mozilla and once when installing 4Suite. Has anyone seen the FC4-x86_64 Fedora Core 4 installation go to completion? If so, did they do anything special to get there?
I choose "Server type" of installation, and after this workaround, I had a host running FC4-x86_64. After that I installed all the missing components using 'yum'(like xorg-x11, mozilla, etc)
Same problem exactly. I used Disk Druid to partition as follows and all worked fine afterward: laptop with ~380MB RAM partitions for the single hard drive: /boot ext ~100MB <swap> ~700MB (2x the RAM size) / ext =rest of disk then, I selected "Server type" and went in and changed all the packages the way I wanted them and started the install.
I took Johann's suggestion and the install got further than it ever had before, but still no joy. After Druid, it partitioned the disk and installed a file system. Then it started loading packages. At one point it said it could not read the CD, but offered a retry option, which worked. Then after many minutes and many programs loaded, including the Linux kernel, it tried to load a graphical program to allow one to change the root password; the progress bar said the install was about 1/2 done. The system locked up at that point. The mouse was still functional, but the disk light was off and the CD-ROM drive door would not open. I let it sit like that for about 10 minutes, then hit reset. I tried to restart the install using the acpi=off option; the install said it completed successfully, but it loaded no programs, and upon reboot, the BIOS said there was no operating system on the disk. As stated above, I tested the CD before trying the install, and it said it was OK.
Charles - I believe the initial problem in your bug report is caused by a file descriptor leak described in bug 160693 and fixed in Rawhide. As for the problems with loading packages off the CD, this generally indicates either bad media (though you said the media was tested, which should rule this out), a bad burn (perhaps burned too fast for the media, etc.), or some sort of hardware problem. Can you check the messages on /dev/tty3 and /dev/tty4 as well as the output of the dmesg command to see if there are any errors reported?
I agree that the problem with reading the CD is probably HW. I am planning to copy the ISO files to a different hard disk and try loading from that. What does "fixed in Rawhide" mean? There is not enough O/S on the HD to read the messages on /dev/tty3 or /dev/tty4 or the output of the dmesg command to see if there are any errors reported. I don't have another Linux installation I can use to read the HD. Thanks for your help.
It means I've committed a fix to our development tree in CVS that I believe fixes your problem. You can check the messages on tty3 and tty4 by pressing ctrl-alt-f3 to get away from X and onto tty3, then just alt-f4 to see tty4. You can run dmesg by switching to tty2 via a similar fashion.
I tried the install again, but without success, though it got much further. I verifed the first two *.iso images by downloading a copy from the Fedora cite and comparing them. Then I copied them to a hard disk (HD), and placed that HD and another with 4.2 GB capacity on a motherboard with an AMD 64 installed. Then I started the install from HD using the askmethod method and Disk Druid. I believe the basic install procedure completed because the main progress bar went completely to the right side. Although I was not watching it at the time, what must then have happened was that the install posted a new dialog box saying "Performing Post Install Configuration ... XY%." The mouse pointer had changed to an hour glass and was turning upside down occasionally. XY = 90. The main install program had posted a message: "Installing openldap-clients- 2.2.23-5.x86-64 (332 KB) Client programs for openLDAP." I let it run like that for over one-half hour, but the Post Install Configuratin never terminated. I hit reset and the system rebooted; the BIOS loader must have found a boot record on the HD, because it said it was looking, then OK. But then a message posted saying "missing operating system." By the time Disk Druid had finished there was 3.8 GB on the HD on which I intended to install the Server version of Linux, with almost no extras. Unless that space is not enough I cannot think what could be wrong. I did not think to type crtl- alt-f3 or alt-f4 to look at tty3, tty4, or dmesg.
I have the exact same problem on 2 different machines. I played around with deleting and creating partitions before I saw this post with no luck. Just an FYI.
Update, I ended up using a free tool from seagate that wipes the hard drive. After I did that and reran the installer, fedora installed fine. FYI, I had win2k installled previously, but i never had problems installing over it with previous versions of fedora.
i used fdisk to wipe my partitions then told fc4 to use automatic partitioning, then use existing free space, installation was successful.
I finally got FC4 to install. As far as I know all I did was remove all HD partitions, repartition, and change the network setup from specifying an IP address to use DHCP, and the whole install process went off without a hitch -- EXCEPT it would not install the Grub loader. Essentially, I ran the install over and over again with different options until it copied the loader to the HD. I don't think the installation is any good though. It puts out a message every time I boot it saying the clock is missing interrupts and probably some driver is hogging interrupts. I continued to try to use it though, because it did not seem slow. As I was using the Samba configuration program to get Samba running, the system froze. I have officially given up, permanently. Until I am prepared to come up with the $179 for a version that works, I cannot put any more time into it. All I wanted to do was run the optimized version of Seti@Home; for the few seconds per WU that Linux would save, it just is not worth it.
My soln. was to install FC3 for the first time and then upgrade it through FC4 install DVD. In the HDD of mine Windows XP had been installed. I tried many ways of partitioning on installation of FC4 and got the same error " File "/usr/lib/anaconda/gui.py......" But FC3 had no problem at all with the same HDD. Oh I have to add I did "auto-partitioning" of installer selecting to delete all existing partitions. Once FC3 was installed, upgrading to FC4 was completed. It's easy, all I had to do was to make installation twice, to wait for doubled time. But I could not be so happy about what was successful with FC3 was not with the "upgraded" version.
I've had this same problem with FC4 and an original iMac with 96 MB RAM. I will now be trying wiping the hard drive completely before the installation!
I needed the hard disk I used for this project for a windows installation. I repartitioned the hard disk with Windows, but the grub loader still ran on boot. This was true even after I installed DOS 6.22. When I installed this disk and another in the system, the master boot record on the first disk appears to have been clobbered. The first sector contains a message "Invalid boot record. No operating system." Subsequent sectors contain all zeros. Does anyone know if it is possible to restore the partition table on the first disk, or must I repartition it and start anew?
*** This bug has been marked as a duplicate of 160693 ***