Red Hat Bugzilla – Bug 383641
mdadm.conf incorrect, prevents boot after upgrade with 2 raid devices
Last modified: 2008-05-07 09:58:48 EDT
Description of problem:
Working F7 system was upgraded to F8. Install completed OK. Failed to boot after
upgrade. Complaint about md1 missing from mdadm.conf in startup messages
Version-Release number of selected component (if applicable):
F8 x86_64 DVD iso with updates-f8-yumloop.img
Steps to Reproduce:
1. F7 system with 4 SATA disks on AHCI controller as sda,sdb,sdc,sdd
2. /dev/md0 = RAID1 of sda1, sdb1, sdc1, sdd1
3. /dev/md1 = RAID5 of sda2, sdb2, sdc2, sdd2
4. Boot F8 x86_64 installation via USB HD with DVD iso file
5. Use updates-f8-yumloop.img from Bug 372011 to get past depsolve issue
6. Complete upgrade successfully and reboot.
Kernel & initrd loads. When it fails to start md1 and then panics when it can
not locate the root fs.
The mdadm.conf file written by anaconda was missing an entry for md1:-
# mdadm.conf written out by anaconda
ARRAY /dev/md0 level=raid1 num-devices=4 uuid=f2aa16e2:2bcef9bd:ff08a61c:f6868636
The same thing happened with f7t4 a few months ago, see Bug 240952
It can be fixed by adding the appropriate md1 line as per the previous bug.
I can confirm this bug when upgrading from FC7 to FC8. I have 3 raid devices
(md0 - raid 1, md1 - raid 5 and md2 - raid 0). anaconda somehow missed out md1
from my mdadm.conf, causing an unbootable system.
Not sure why you think this is a low priority bug - I would have thought it
should be medium or high.
this is a duplicate of 242334.
*** This bug has been marked as a duplicate of 242334 ***