reassign 403299 mdadm
thanks
On Saturday 16 December 2006 12:17, Daniel Dent wrote:
> I'm quite certain it had time to fully sync. Furthermore,
> by putting a sleep into the script that remounts root RO I
> was able to see it say that it hadn't shutdown the volume
> group because the volume group was in use. I also added
> some mdadm commands to have it shutdown the raid and those
> would refuse to do it because it was still in use.
I've tried to reproduce this in vmware, but cannot.
I do admit that the messages on shutdown [1] with / on lvm on raid look
quite scary, but everything mounted cleanly for me on several reboots.
What the init scripts should be doing on shutdown is remounting /
read-only and making sure raid disks are synced before finally shutting
down things. That an array that has / on it is still in use is expected.
Reassigning this report to mdadm so its maintainer can take a final look
and see if the console messages on shutdown are correct and if maybe they
can be improved.
Cheers,
FJP
[1] http://people.debian.org/~fjp/mdadm_reboot.png
This screen was captured with a long sleep in /etc/rs6.d/S90reboot.
Attachment:
pgpK7I3eYaIYD.pgp
Description: PGP signature