Fedora Core 4… Not yet :(
For reasons I don’t entirely understand, I have been unable to try out Fedora Core 4. I have downloaded it and successfully made the install disks and I can boot off disk one with no problem. The problem arises when I select which installation to upgrade. My setup isn’t entirely normal. I have two 120GB disks running off an ATA100 controller (so they are NOT hda and hdb) that are mirrored using software mirroring. So my root device is /dev/md0. That shows up as an option and has Fedora Core 3 detected on it, but when I click next, I get a dialog that looks similar to:
+-----------------------------------------------------+ | Duplicate Labels | +-----------------------------------------------------+ | | | (*) Multiple devices on your system are labelled | | | | [ Reboot ] | +-----------------------------------------------------+
In the above ASCII art, (*) represents a symbol that looks like a critical fail icon.
What really confuses me about this dialog is that it doesn’t tell me what label is duplicated. I’m guessing this is some kind of bug because I have two physical disks that are identical in every way, but I’m not sure. I open bug 160622 on Redhat’s Bugzilla, but it got closed telling me that my RAID disks can not be labelled and suggesting that I remove the label. As far as I can tell, I have no labels of any sort, so I think I’m gonna have to reopen this bug. I’ve been poking around on the ‘net trying to find a solution, but having no luck. So for the time being, I’m unable to use Fedora Core 4… unless, of course, I install it on another machine.
UPDATE: 29-Sep-2005 – It’s installed and working!
Comments
Comment from Leonid Mamchenkov
Time June 20, 2005 at 5:04 pm
Have you tried to see/change labels with /sbin/e2label, which is a part of e2fsprogs RPM?
Comment from Jake
Time June 21, 2005 at 11:58 am
As far as I know, I’m not using LVM… I suppose anything is possible, but I don’t recall ever setting that up.
I did check using e2label and did not see any labels: https://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=160622#c4
Comment from Jeebus
Time August 28, 2005 at 11:05 am
Jake,
Are you still having trouble with FC4? I found your site when I was have the same “Duplicate Labels” problem last night during my upgrade. I have two disks but no RAID and I was still getting the message even after all sorts of fiddling with e2label.
When you get the “Duplicate Labels” reboot message and you Ctrl-Alt-Fn over to the consoles, what kind of error messages do you see? Anything to do with swap labels perchance? This is what I saw, and it led me to believe this was the real problem. I had two unlabelled swap partitions, one on each disk. After a few unnecessary missteps, I ended up doing a `mkswap -L “swap_a” /dev/` and `mkswap -L “swap_b” /dev/` and… presto! The upgrade could proceed.
So perhaps your problem could be caused by this inability of anaconda to handle multiple unlabelled swap partitions and not a RAID/LVM issue.
Just my $0.03… (includes $0.01 fuel surcharge)
Comment from <img src='http://jacob.steenhagen.us/blog/wp-content/plugins/rpx/images/openid.png'/> Jake
Time August 29, 2005 at 9:40 am
Jeebus,
Thanks for the post. I actually just firgured that out on Saturday, though I got around it by deleting one of my swap partitions not by labeling them. I updated my bug at redhat.com to say that the problem seems to be caused by the multiple swap partitions not software RAID. I haven’t made a blog posting yet to announce my success, but I did have some 
.
I also didn’t know about ‘Ctrl-Alt-Fn’, I suppose that would have come in handy…
Pingback from Jake’s Weblog » Blog Archive » Fedora Core 4 is here at last
Time August 29, 2005 at 6:11 pm
[...] I finally figured out how to get Fedora Core 4 to install on my machine. When I last posted on this topic, I thought the problem was that Anaconda didn’t like my software mirroring. As it turns out, the problem was the fact that I had two swap partitions. Once I removed one of the swap partitions (both physically and via /etc/fstab), Fedora Core sucessfully installed on my machine. I updated my bug at RedHat with this new information. The only hitch I’ve noticed so far is that Cyrus didn’t want to start because of database incompatibilities. The solution I ended up using, which was one of the ones I found on FedoraForum, was simply deleting the database files. Other than that, it seems to be working fine. [...]
Write a comment
You need to login to post comments!
Jake's Weblog
Upcoming Events via e-mail
Comment from Fred F.
Time June 20, 2005 at 5:03 pm
Are you using RAID and LVM? It may be that they are getting loaded in the wrong order. I use Gentoo so I don’t know much about Redhat but that was a problem a while back on the init scripts. I have exactly that setup and it works.