I'm still fairly new to Linux Mint Debian and recently started seeing a "boot drive is full" warning whenever I start the computer. From what I've read, old kernel packages are probably filling the /boot partition. I enabled the Update Manager option to remove obsolete kernels and dependencies automatically, but several weeks later the warning still appears.
I'm considering using apt commands to remove older kernels, but I'm not very experienced with the command line and want to avoid making the system unbootable. Is it safe to remove old kernels, and what precautions should I take? Should I keep the currently running kernel plus at least one older fallback? Would creating a Timeshift snapshot help, and are there other checks I should perform first, such as determining whether only /boot or the entire Linux partition is full?
3 Answers
Removing old kernels is generally safe when it’s done through the package manager, since it understands which files belong to which packages. The main risks are removing the kernel currently in use, deleting every fallback kernel, or removing files manually and leaving the package database inconsistent. Keep the active kernel and one known-good older version, then use the distribution’s package tools to clean up the rest. If the system can no longer boot, the GRUB menu may let you select another installed kernel; otherwise, a live USB can be used for repair.
A normal first step is to run `sudo apt autoremove --purge`, which can remove obsolete kernel packages and their unused dependencies. Before doing that, check which kernel is currently running with `uname -r`, and do not remove that one. It’s also sensible to keep at least one older kernel as a fallback. A Timeshift snapshot can help with some system changes, but it may not restore a machine if the boot files themselves are removed incorrectly.
The automatic cleanup may only happen after a package update or upgrade completes successfully. If /boot is already too full, the system might not have enough room to finish the update, so the cleanup never gets a chance to run. First determine whether only `/boot` is a separate, full partition or whether the whole filesystem is low on space. You can check with `df -h`. If necessary, boot an older kernel from the GRUB menu or use a live USB to recover enough space, rather than manually deleting random files from `/boot`.
That makes sense—there may not be enough free space for the update process to complete. The files in /boot mostly look important, so I’ll check the partition layout and running kernel before touching anything.