Degraded RAID UI (Re: Server Team 20080812 meeting minutes)

[Kees Cook]

On Wed, Aug 13, 2008 at 11:59:53AM +0100, Matt Zimmerman wrote:
> On Tue, Aug 12, 2008 at 05:19:32PM -0700, Mathias Gug wrote:
> > ==== Boot Support for Degraded RAID ====
> > 
> > kirkland has been working on adding RAID support to grub-install. His patch is
> > waiting for sponsoring. He hopes it will make it for alpha4 scheduled to be
> > released on Thursday. If so mathiaz suggested that this new feature should be
> > documented in the Release Notes in order to get more testing.
> 
> Do you wish to continue? [y/N]
> 
> This should present no more risk than dropping to a shell, and gives the
> user the guidance to make an informed decision.

This blocks other rootfail handlers from executing.  In some magic
future the other handlers could potentially do something to solve the
md problems, resulting in a correctly booted system without any user
interaction.

However, nothing is yet attempting to use the rootfail framework beyond
MD and LVM.  Still, it seems like a bad idea to block.  At present, on a
failure, all the information needed to boot degraded is presented to the
user.  It does require a separate boot, though:

 There appears to be one or more degraded RAID devices, and your root device
 may depend on the RAID devices being online. One or more of the following RAID
 devices are degraded:
 .../proc/mdstat...
 If you want to attempt to boot with the RAID in degraded mode, add the
 following parameter to your kernel boot options:
        bootdegraded=true

As mentioned, a timeout to the handler's question could be added using
the "-t" option to "read" (which is supported in busybox, thankfully):

    ANS="unanswered"
    read -p "Boot with degraded RAID anyway? [y/N]: " -r ANS -t 15
    case "$ANS" in
        unanswered) echo "Timed out"
                ;;
        y*|Y*)  BOOT_DEGRADED=true
                ;;
       *)       BOOT_DEGRADED=false
                ;;
    esac



-- 
Kees Cook
Ubuntu Security Team