Ensure all servers can boot on the latest configuration deployed by morph

This does not seem to be the case for HRO-Cloud with morph deploy --on="storage003" morph/grid/<env>/grid.nix switch:

...
Executing 'switch' on matched hosts:

** storage003
updating GRUB 2 menu...
installing the GRUB 2 boot loader on /dev/disk/by-id/nvme-eui.e8238fa6bf530001001b448b463b77bf...
Installing for i386-pc platform.
/nix/store/vwv1k39h5ihms6jwrhpmv1y6k561hbjp-grub-2.06/sbin/grub-install: warning: this GPT partition label contains no BIOS Boot Partition; embedding won't be possible.
/nix/store/vwv1k39h5ihms6jwrhpmv1y6k561hbjp-grub-2.06/sbin/grub-install: warning: Embedding is not possible.  GRUB can only be installed in this setup by using blocklists.  However, blocklists are UNRELIABLE and their use is discouraged..
/nix/store/vwv1k39h5ihms6jwrhpmv1y6k561hbjp-grub-2.06/sbin/grub-install: error: will not proceed with blocklists.
/nix/store/bw8rz72if5m4chqa50a2mmlw1zwmkqmj-install-grub.pl: installation of GRUB on /dev/disk/by-id/nvme-eui.e8238fa6bf530001001b448b463b77bf failed: Inappropriate ioctl for device
Error while activating new configuration.

This makes storage003 rebooting on its legacy configuration with authorized SSH keys we'd like to get rid of!

Also, our morph documentation page only refers to test and does not mention switch or boot. It is then unclear (to me) when do we make sure the configuration will be made persistent.