Red Hat Bugzilla – Bug 231426
RAID installation not detected for upgrade
Last modified: 2007-11-30 17:11:58 EST
Description of problem: Fedora installation on RAID 1 volume was not detected by the installer. The configuration used 3 sata hard drives with 1 hot spare. There was also a second raid 5 partition which was used for LVM. This was also not detected. The next screen showed the partition layout as having one RAID 1 device of the smaller size with 4 hot spares. The correct layout should have been one RAID1 device of the smaller size, one RAID5 device of the larger size, and LVM over the RAID5. Version-Release number of selected component (if applicable): How reproducible: Consistently. Steps to Reproduce: 1. Install FC6 with 3 drives. Use RAID1 with 1 hot spare for / and RAID5 and LVM for /home. 2. Attempt to upgrade using FC7 test 2. Actual results: Fedora installation is not detected. Partition layout in next screen is incorrect. Expected results: Fedora Core 6 installation should be detected on RAID 1 partition. Upgrade should complete successfully on this partition. Additional info:
Fedora 7 test bugs should be filed against "devel", not against test1/2/3. This isn't obvious, I know. Moving this report so it isn't lost. This is a bulk message -- I apologize if this was actually meant to be targeted against a different release. If so, please fix or let me know. Thanks.
A couple of questions: Are all the drives SATA drives? Are they all on the same controller? Do they get different device names in F7 than they had in FC6? Is this regular Linux software RAID (mdraid) or some other type of BIOS/hardware RAID?
Answers: Yes. All are SATA drives. All are on the same controller, built into the ASUS P5LD2-VM motherboard. I was using regular software RAID (mdraid). No BIOS/hardware stuff. I really don't know about the device names any more. It is possible that they may have changed. I have since reinstalled, so the old layout is gone. It is possible to reproduce the above configuration, though. Also, I have not tested this against F7 test4. Any reason to believe it would be different?
Were there other drives in the system (USB, IDE, etc.) that might have caused the devices to change numbering?
No drives were added or removed. Nothing was plugged into USB. I see no obvious reason for devices to have changed numbering. If they did somehow change numbering, should I not have still expected this to work given that all three drives were still present?
rawhide as of tomorrow should be better here with switching to mdadm across the board which should be a little bit more resilient to things moving around. Can you try with tomorrow's tree and see if it does any better?
This works for me today. Was able to upgrade both a 2-drive mdraid system (mix of raid0 and raid1 arrays) and a 4-drive mdraid system (raid1 and raid6).
All our testing indicates that this is fixed in rawhide and F7rc2. If possible, could you retest and confirm the fix?