I currently have a Lenovo Yoga laptop with Fedora 43 Workstation installed on it.
I would like to set up dual boot with openSUSE because I would like to explore that distribution and how it interacts with the convertible mode and the pen. Other distributions didn’t work out of the box (Zorin OS, Debian,…), so I would like to check openSUSE (not in a VM!).
Is there something I need to keep in mind or be aware of? In the past I often had to do a fresh re-install of Fedora after having experimented with dual boot - which is something I would like to avoid since it’s my daily driver. There were always issues with the boot order or Fedora not showing anymore at all etc.
Even if issues come up, it should generally be possible to get Fedora booting again without a reinstall. Have a USB stick handy with a live Fedora image, then if necessary you can restore the bootloader by following these instructions:
Not with 2 different linux distros.
You should have separate EFI partitions, one for fedora and one for the other OS so that things don’t get overwritten. You can then always use the bios boot menu if needed to select which distro to boot even if grub fails to allow selection.
That is something you should do before any attempt to recover the boot loaders. Chances are high that they are not broken at all, as Fedora and Suse uses separate subdirectories of the ESP file system.
Also if you have secure boot enabled, remember that the Fedora shim can only boot the Fedora kernels and the Suse shim can only boot the Suse kernels. That is, unless you get the certificate from the other OS and enroll it in the MOK.
If you use the UEFI boot menu to select the OS rather than the Grub menu, this problem goes away.
How do I make sure that I have two EFI partitions? I am not using Secure Boot because I don’t know how to set that up (I know, very lame excuse but I am new to the whole Linux world and not a techie)
I don’t think it’s necessary to have two ESP; each OS will have their own subdirectory, and files won’t get overridden. In fact most UEFI bioses expect a single ESP.
I can think of a scenario where having separate EFI partitions would be beneficial.
Say an OS overwrote the partition on an update as I’ve learned MS Windows has done in the past. Then you’d have to go about the business of fixing that for your other OSes. Separate partitions would preclude that.
As @computersavvy pointed out, you can always modify the UEFI bootloader to point to the correct partition for a given OS and use that method to boot if grub fails to do the right thing. At least I think that was what he was describing.
Another scenario is where you want to dual-boot Fedora with a different distro that uses systemd-boot.
systemd-boot directories in the ESP can confuse Fedora into thinking it should install kernels as UKIs, like this user found. Keeping separate ESPs would prevent this.
I don’t think OP in that thread configured it explicitly though. They just did a fresh install onto a disk whose EFI partition contained a systemd dir from a previous different distro, and got surprising consequences.
Do the contents of /etc/kernel/install.conf get automatically created at install time based on what’s in the EFI partition?
It doesn’t. You get /usr/lib/kernel/install.conf with only comment lines. You create your own version in /etc/kernel following the systemd convention of putting distribution provided configuration files in /usr and the local user can override these files by a corresponding file in /etc.
I once installed Ubuntu on another partition (free space on the disk) and had to run Fedora via UEFI, Linux distros are not tolerant towards each other, which is why