Red Hat Bugzilla – Bug 496390
Failure to boot
Last modified: 2009-04-20 13:21:08 EDT
Created attachment 340158 [details]
second to last page of boot output
Fresh install of Rawhide onto an existing lvm parition does not boot. This happened with the Beta, and still happens if I boot into the system yum upgrade everything to the latest and re-run mkinitrd.
I'll attach photos of the last two pages of boot output - it appears that it is getting into the initrd, but not past there.
Not clear to me what other relevant information is, random notes:
* /boot is a separate ext3 partition, / is a lvm partition on a raid-1 array
* System has a Intel DQ35JOE motherboard; which is a "legacy-free" motherboard (no serial, no PS/2) with i945-ish graphics. Core 2 Duo processor.
* Hang is the same with or without nomodeset.
* System is mounted from rescue mode from the installer CD fine, looks fine as far as I can tell.
Created attachment 340160 [details]
Last page of boot output
OK, so I hacked the initrd not to run plymouth, which was eating the console messages, even with plymouth:nolog. Doing that I see:
Making device-mapper control node
mdadm: /dev/md0 not identified in config file
Scanning logical volumes
Reading all physical volumes. This may take a while...
Activating logical vlumes
Volume group "VolGroup00" not found
And continuing on with related failure to
switchroot; mount failed: no such file or directory
Booting has failed.
So that gets close to figuring out what's going wrong.
Problem was that anaconda wrote out an /etc/mdadm.conf missing the device name:
ARRAY level=raid1 num-devices=2 UUID=60c552bd:d2aade90:5a036e43:18fe0dfd
(should have /dev/md0 after ARRAY.) I hadn't tried reinstalling with a fresh Anaconda image for a few weeks, so it's possible that this problem has already been fixed. Reassigning to Anaconda.
This will be fixed in the next build of anaconda. Thanks for the bug report.