From Bugzilla Helper: User-Agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.6) Gecko/20040510 Description of problem: When running an upgrade (graphical mode) from a FC1 system the installer complains about two partitions having the same disk label and refuses to continue. The two partitions DO have the same disk label since they are part of a RAID-1 volume but the label is not used in /etc/fstab (or anywhere else I know about) on the original FC1 installation. Work around: Boot the old system and run '/sbin/tune2fs -L "" /dev/mdX' where X is the raid-1 volume containing the disk label. Version-Release number of selected component (if applicable): How reproducible: Always Steps to Reproduce: 1.Create a RAID-1 volume with an ext2/3 file system 2.Label the raid volume 3.Try to run an upgrade to FC2. Actual Results: Told that the installer could not continue with two partitions having the same label. Asked to reboot the system (and abort the upgrade). Expected Results: Be able to run the upgrade without having to erase the disk label. Additional info:
I'm investigating this currently - I have setup a Fedora Core 1 install using RAID-1 kickstarted with the following partitioning setup: clearpart --linux part /boot --fstype ext3 --size=150 part raid.15 --size=4096 part raid.13 --size=4096 part / --fstype ext3 --size=2048 part swap --size=1024 raid /usr --fstype ext3 --level=RAID1 raid.15 raid.13 I boot into FC1 and use e2label /dev/md0 /usr to label the RAID device. I then manually perform a Fedora Core 2 upgrade - that all seems fine here - no error messages from anaconda about disk labels. I've also tried with / on RAID-1 too and I sill don't see an error. Could you expand on your steps to reproduce and also tell me at what stage you see the error.
Since my original system is upgraded I created a new FC1 installation in a VMWare virtual machine with two "physical" disks, sda and sdb. FC1 installation (graphical mode) Custom installation, minimal /etc/fstab: /dev/sda1 / ext3 defaults 1 1 /dev/sdb1 /usr ext3 defaults 1 2 ... /dev/sda2 swap swap defaults 0 0 /dev/sdb2 swap swap defaults 0 0 /dev/md0 /mnt/foo ext3 defaults 1 2 /etc/raidtab: raiddev /dev/md0 raid-level 1 nr-raid-disks 2 chunk-size 64 persistent-superblocks 1 nr-spare-disks 0 device /dev/sda5 raid-disk 0 device /dev/sdb5 raid-disk 1 /sbin/mkraid /dev/md0 /sbin/mkfs.ext3 /dev/md0 /sbin/e2label /dev/md0 "MISC" FC2 upgrade (graphical mode) ... Language Selection: English (English) Keyboard Configuration: Swedish Monitor Configuration: Unprobed Monitor Select "Upgrade an existing installation" and klick "Next". A notice pops up with the message: Duplicate Labels Multiple devices on your system are labelled MISC. Labels across devices must be unique for your system to function propery. Please fix this problem and restart the installation process. [Reboot] Boot into FC1 and run umount /mnt/foo /sbin/tune2fs /dev/md0 -L "" /sbin/reboot FC2 upgrade gives an error message at the same step as above: Error Error mounting device md0 as /mnt/foo: Invalid argument This most likely means this partition has not been formatted. Press OK to reboot your system. [OK] Boot FC1 to verify that /mnt/foo is mounted and ready. Place a # before /dev/md0 in /etc/fstab to run the FC2 upgrade process without any problem. Boot the upgraded FC2, uncomment /dev/md0 in fstab and reboot. /mnt/foo is mounted and ready. Add a disk label with /sbin/e2label /dev/md0 "MISC" and reboot to verify that /mnt/foo is mounted without problem.
Thanks for the detail - I'm investigating further.
Does this still happen with newer releases?
I just did an upgrade from FC3 to FC4 test 3 as described in comment #2 without any problems.
As per comment #5, resolving this as closed:rawhide.
I just ran into this problem with the upgrade from FC3 to FC4. I am using three raid1 arrays. e2label suggests that the partitions and arrays don't have labels, since it comes back with a blank line for each. I even tried echo \"`e2label device`\" for each device and just got "" each time. The log from anaconda that I had it save to a floppy suggests that there are extended ascii characters for labels of two of the devices. I will attach the error log.
Created attachment 115516 [details] Dump of the Anaconda log