Created attachment 381174 [details] Ouput log from console started during installation Description of problem: Problem might be related to bugid 551705 After starting a raid0 array (created in Fedora 11) the volume group does not start under Fedora 12. This does not happen in Fedora 11, and in other distribution installation procedures (e.g., ubuntu). Version-Release number of selected component (if applicable): Fedora 12 How reproducible: see steps to reproduce Steps to Reproduce: 1. create raid0 and and a volume group on top in Fedora 11 2. try to install Fedora 12 on that raid 3. Actual results: volume group can not be made active Expected results: Volume group is active Additional info: # lvm vgchange -a y # (i.e., no output) See attachment for full output
see bug#551705 re the raid problem. thx.
After starting MD array you have to use "vgscan" to rescan new devices _or_ specify VG explicitly (to force scan for that VG). lvm vgscan lvm vgchange -a y <VG_name> (VG name, not PV path! The second vgchange in your log has wrong syntax.) (But according to log, I think there is some misconfiguration on your system, there is existing partition on that MD device - /dev/md0p1 and no lvm2 label is detected there anyway. If you have PV directly on /dev/md0, you have to remove that partition table signature. Kernel MD driver started to recognise partitions on MD in some recent version, maybe that's the reason why F11 works.)
Indeed the existence of the partition (md0p1) caused this problem (especially since the raid was created anyway without a partition). one I deleted that partition (which I have no idea how it was created), the installation went without a problem (or at least the lvm is visible again without doing any vgscan or vgchange). Problems can be closed.