This question may seem trivial at first sight, but it has certain implications and I'm wondering what is the recommended path to take here.
Assume the following scenario: a server system. The default boot gets kicked off from a partition on the first hard drive which hosts the
/boot volume and the GRUB2 code. The
/ volume is on an
md device (RAID1, in case it matters) and GRUB2 is aware of that. Everything works fine and is nice and dandy.
Now: another system is to be set up in parallel for rescue purposes, in case something happens to the
md. The rescue shell isn't exactly too helpful, but having a full-fledged Ubuntu installation gives you a lot more power, I reckon. So the idea would be to clone the configuration - largely - which is kept under version control using
etckeeper from the default system to the rescue system in a cron job and cherry pick some pieces such as the
sshd host keys and
/etc/network/interfaces to make sure the rescue system would end up booting into a remotely accessible state resembling that of the default system (possibly locked down to only allow root logins instead - but I think I know how to take care of those parts).
How can these two installations share the same
/boot volume? It makes sense to do it, because the existing one is on the first hard disk and so will the rescue system be. However, assuming updates and eventually pruning kernels from the default system, this would leave the rescue system in an unbootable state. How can I prevent this and recycle the
/boot volume for both installations?