I'm using a Framework 16 and want to move most of my daily work from Windows to Fedora or Nobara, but I still need Windows for a few specialized tools. I have several NVMe drives, including some used through a USB-C adapter, and my preferred setup is one operating system per drive.
When installing Windows, I disconnect the other drives so Windows gets its own EFI system partition and cannot interfere with anything else. I then choose the desired drive from the firmware boot menu whenever I want to switch systems.
I'd like to use the same approach for Fedora. Ideally, Fedora should boot independently from its own drive without GRUB detecting the Windows drive or adding it to a shared boot menu. I want to test Fedora from an external NVMe first and possibly make it my main system later.
What is the safest way to install and configure Fedora so each drive remains independent and the systems do not interfere with one another?
4 Answers
For the initial Fedora installation, physically remove the other drives or disable them in firmware if possible. In the installer, verify that the intended NVMe is selected for both the Fedora partitions and the EFI system partition. Once installation is complete, reconnect the other drives and confirm that each one appears as a separate UEFI boot option. Keeping the drives isolated during installation is safer than relying on the installer to guess correctly.
The simplest approach is to keep every operating system on its own drive with its own EFI system partition. Disconnect or temporarily disable the other NVMe drives while installing Fedora, then Fedora’s boot files will be written only to the target drive. Afterward, use the Framework’s UEFI boot menu to select which drive to start. This avoids creating a shared GRUB configuration, and the systems should remain isolated.
A separate-drive setup is generally reliable. Windows and Linux normally won’t modify another drive’s installation when each has its own EFI partition. The main thing that can occasionally change is the firmware’s default boot entry after an installation or update; if that happens, it can usually be corrected in the UEFI settings in under a minute. You can still choose any other drive manually when needed.
It’s also worth keeping a boot-repair or rescue USB around. If a firmware entry ever disappears or an installation becomes unbootable, a live Linux environment can usually recreate the UEFI entry or repair the affected drive without changing the other operating systems.
That matches what I’m trying to do. I’d rather select the drive manually than have one installation manage all the others.