From Bugzilla Helper: User-Agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.6) Gecko/20040211 Firefox/0.8 Description of problem: I'm using a Fedora Core 1 system. My /usr is a raid 5 software array. Rebooting and running the Fedora Core 2 installer (I have the DVD version, downloaded & recorded), after I choose to upgrade my FC1 system, installer says it's unable to mount /dev/md1 to /usr because of "No such file or directory". Then all I can do is click on "ok" to reboot the system. How reproducible: Always Steps to Reproduce: Upgrade a Fedora Core 1 system with /usr (or anything else, I guess) as a raid 5 software array.
Are there any errors with starting the RAID array on tty4 from the kernel? Can you switch to tty2 and see if raidstart /dev/md1 works?
Ok, it seems you've found the problem :) # raidstart /dev/md1 * missing components of raid device md1. The raid device needs 3 drive(s) and only 2 (was/were) found. This raid device will not be started. could not find devices associated with raid device md1 In fact my md1 array misses a disk. I have 2 disks out of 3 (the third one is damaged and I can't buy another one right now). Shouldn't it mount it anyway? It should be able to use it even if there's one (and just one) disk missing... As I'm using it right now. The tty4 looks fine. I think it's not interesting anymore, so I won't attach the output. Just tell me if you need it :) What should I do now? Is there a way to make it use that md1 array anyway? Thank you :)
anaconda intentionally won't use an incomplete/damaged array. There are too many open questions there as to what led to the situation that make an upgrade in that case extremely dangerous. There's a different RFE filed about supporting the use of degraded arrays, but that's currently not a high priority item due to the risk involved. So unfortunately, the answer is "fix your array to be complete"