"Guimager"? Do you mean Guymager, a forensic disk imager program?
In that case, you've gone outside the bounds of what Users can do: a forensic disk imager needs raw access to the disks, bypassing (and possibly locking out) any active filesystem drivers accessing those same disks. This is something normally only a sysadmin has permission to do. And the extra rights bring with them extra responsibility.
I don't know about SquashFS specifically, but I do know that many filesystem drivers get unhappy if they suddenly find themselves unable to access "their" disk. Some filesystem drivers go into a persistent error state: to fix it, you must unmount the filesystem, usually run a
fsck on it, and then re-mount it.
If this happens to a root filesystem, a pragmatic sysadmin can recognize that rebooting is the fastest way to get the system back to a working state, or at least to a state that allows further fixing (= the emergency shell of initramfs, if the root filesystem is severely damaged).
Of course you could kill all user processes and most system daemons, unmount other filesystems, remount the root filesystem read-only, check it, then restore the system to a working state, but doing all that for the sake of a dogma is inefficient.
A pragmatic sysadmin can perfectly well judge that a reboot is the fastest and most reliable way to get the system working again, and then do it. But that's very different from "let's reboot and pray the error does not happen again", which is what your so-called "law of Linux philosophy" aims to avoid.