Red Hat Bugzilla – Bug 487965
No boot with mdadm 3.x
Last modified: 2009-03-31 15:16:20 EDT
My /boot is on a raid-1 /dev/md0 device (everything else is lvm-ed over /dev/md1)
Since the mdadm update the boot process fails after the handover from the kernel to init. /boot can not be mounted. When I log-in at the error prompt I always find one of the two pieces of /dev/md0 mounted as /dev/md127 and the other as /dev/md0
Manually stoping /dev/md127 and /dev/md0, then re-assembling /dev/md0 and remounting pretty much everything else permits a successful init 5
Till the next reboot :(
Created attachment 333660 [details]
exactly the same here, only i can't seem to start with the newer kernel at all, I still use an f10 kernel
I have the same proplem, after upgrade f10 to f11 alpha.
As a work around try removing /etc/udev/rules.d/70-mdadm-assemble.rules .
I have been doing that since getting mdadm 3.
This still has problems if a device gets booted out of an array, as udev will try to start separate arrays with the failed devices. If this happens, you can use mdadm to add them back in and reboot.
*** This bug has been marked as a duplicate of bug 488038 ***
It seems to fixed for me in bunch of last updates.