Dear Maintainer,

Installing a debian in RAID 1 on two harddrives - /dev/sda and /dev/sdb
Using sw raid either during install or (migrating) on the running system. Both the same effect:
First I used
  /dev/md1 on /dev/sda1 and /dev/sdb1 as /boot (ext3)
  /dev/md5 on /dev/sda5 and /dev/sdb5 as swap (swap)
  /dev/md6 on /dev/sda6 and /dev/sdb6 as / (ext3)
  /dev/md7 on /dev/sda7 and /dev/sdb7 as LVM container
second, I eased things and dropped /boot - only using /dev/md6 as / (ext3)

When unplugging /dev/sda, grub fails to boot. It attemts to wait for /dev/md6 containing root, revealing an UUID that is inappropriate (unplugged /dev/sda).

I read about including rootdelay=90 to the grub_cmdline_linux_default option
I also tried the option GRUB_DISABLE_LINUX_UUID=true in the /etc/default/grub - same behaviour.
I also tried to assure all modules to be available through GRUB_PRELOAD_MODULES="lvm mdraid1x" - just to make sure the raid is up and running. - same behaviour.

System fails to boot on the second harddrive, which makes raid1 inefficient.

I would expect the grub2 to notice mdraid and adjust UUIDs to boot from both harddrives.