Initramfs error handling, and mdadm - How to get mdadm's mount failure hook to function.

[Scott James Remnant]

On Mon, 2008-03-10 at 16:07 -0400, Phillip Susi wrote:

> Luke Yelavich wrote:
> > The problem with mdadm in this case, is that the /dev/md device node
> > gets created, even if the array is degraded. The obvious consequence of
> > this, is that the root device node is found, and the failure hook for
> > mdadm is not run, causing the boot process to error out later on.
> 
> Seems to me that checking for the existence of the root device is not 
> sufficient.  I think this is related to the EvmsLvmRaidAgain spec and 
> the solution needs to involve a deliberate set of state information 
> indicated whether or not a device is published for use.  An offline md 
> device should not be and therefore should not be accepted as a usable 
> root device.
> 
This is why the mountroot loop in local also runs vol_id, to check not
just the existance of the device but its state.

Scott
-- 
Scott James Remnant
scott at canonical.com
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 189 bytes
Desc: This is a digitally signed message part
Url : https://lists.ubuntu.com/archives/ubuntu-devel/attachments/20080311/5f3fcacc/attachment.pgp