Red Hat Bugzilla – Bug 244126
f7 ks install md1 for /boot md0 for /, Error: md1 already in mdList
Last modified: 2008-02-26 09:26:30 EST
Description of problem: Kickstart installation with following disk layout works the first time. If I then re-install with the same kickstart, Anaconda complains that mdX (whereever /boot lives) is already in mdList. Swap the --device assignment of md0 and md1 between /boot and / then the install works again. Reinstall again and it fails until you swap them back. # create raid5 md0 raid pv.a0e0 --level=RAID5 --fstype="physical volume (LVM)" \ --device=md0 --spares=1 raid.a0 raid.b0 raid.c0 raid.d0 raid.e0 # create raid1 md1 raid /boot --level=RAID1 --fstype=ext3 \ --device=md1 --spares=3 raid.a1 raid.b1 raid.c1 raid.d1 raid.e1 Version-Release number of selected component (if applicable): F7 How reproducible: Every other install. Steps to Reproduce: 1. Use above disk config in kickstart (full KS attached) to install. 2. Reboot (taking bug #242107 into account) successfully. 3. Use above kickstart to install yet again. 4. Get anaconda error about mdX already being in mdList during the partition formatting step. Actual results: Traceback (most recent call first): File "/usr/lib/anaconda/raid.py", line 203, in register_raid_device raise ValueError, "%s is already in the mdList!" % (mdname,) File "/usr/lib/anaconda/fsset.py", line 2191, in setupDevice self.level, self.numDisks) File "/usr/lib/anaconda/fsset.py", line 608, in clobberDevice device = entry.device.setupDevice(chroot) File "/usr/lib/anaconda/fsset.py", line 1543, in formatEntry entry.fsystem.clobberDevice(entry, chroot) File "/usr/lib/anaconda/fsset.py", line 1621, in makeFilesystems self.formatEntry(entry, chroot) File "/usr/lib/anaconda/packages.py", line 153, in turnOnFilesystems anaconda.id.fsset.makeFilesystems (anaconda.rootPath) File "/usr/lib/anaconda/dispatch.py", line 203, in moveStep rc = stepFunc(self.anaconda) File "/usr/lib/anaconda/dispatch.py", line 126, in gotoNext self.moveStep() File "/usr/lib/anaconda/text.py", line 605, in run anaconda.dispatch.gotoNext() File "/usr/bin/anaconda", line 955, in <module> anaconda.intf.run(anaconda) ValueError: md1 is already in the mdList! Expected results: Normal re-installation. Additional info: To restate: Install / on --device=md0, /boot on --device=md1 -> works Re-Install / on --device=md0, /boot on --device=md1 -> fails Re-Install / on --device=md1, /boot on --device=md0 -> works
Created attachment 156941 [details] Kickstart Configuration
Just a quick note on your kickstart file while I'm looking at the bug. kickstart now supports multiple %ksappend directives, so your following comment no longer applies. Perhaps you can reduce some complexity. # Apparently KS will only use the first ksappend line and we want to # call modular scripts depending on the host being setup. Plus the vars # aren't expanded. So we'll serve a meta config from the web server which will # serve up a custom shell script based on the dist ver and MAC. Also, we support repo commands for pulling in packages from other repos at install time, and a chkconfig command for enabling/disabling services. Now to look at the bug itself.
Are you able to reproduce this with F8 test 1?
Should be fixed in rawhide.