Dear Maintainer,
update-grub installs only onto the first EFI boot partition it finds. Other partitions are ignored.
My system has 4 drives configured with 3 RAID sets. The first two drives are a single RAID set under Intel RST used for Win 7.1 Pro. The drives are configured:
Drive 1 & 2 under Intel RST RAID 1 (/dev/md126p):
100MB EFI Boot
128MB MS Reserved
Rest of drive partitioned for windows drives.
Disk 3 /dev/sdc
sdc1 - 250M - EFI Boot
sdc2 - 750M - RAID 1 (/dev/md0) for /boot on ext4
sdc3 - rest of drive - RAID 1 (/dev/md1) for LVM
Disk 4 /dev/sdd
sdd1 - 250M - EFI Boot
sdd2 - 750M - RAID 1 (/dev/md0) for /boot on ext4
sdd3 - rest of drive - RAID 1 (/dev/md1) for LVM
The installer only configures the EFI boot on the raid set made up of disks 1 or 2. So if disk 1 fails or is removed the system will continue to boot. However, loss or removal of both disks 1 & 2 should leave the system operable as a multi-os boot system has meant the Debian systems are (and should be) completely unaffected by the removal/failure of disks 1 & 2. However, the current situation renders the system unbootable.
For reliable raid the installer should ALSO populate the EFI Boot partitions on disks 3 & 4. Thus ensuring that the o/s is visible to the BIOS and will actually boot.
There is a manual work around for this which I shall add to this report for those that are interested.