Not directly related to the issue; but longtime IT Systech.
- Can you ensure that boot never fails after successful installation and usage, of course.
You never can. You can, however, prepare (for the worst). Some use BTRFS, some use timeshift, some both. Some make a whole copy of their ESPs or use immutable distros. Some just try to change their workflow, so that at least the workflow does not break everything. VMs or docker come to mind. Idk if vagrant is still alive and kicking.
www.vagrantup .com/
unix.stackexchange .com/questions/622148/rollback-root-snapshot-in-btrfs
itsfoss .com/backup-restore-linux-timeshift
Immutable OSes: for example: VanillaOS or Fedora Silverblue
- Is there any quick fix to solve this issue. (with or without boot disk). If not, is there a minimal downloadable bootable USB that can fix the issue. (just to avoid entire EOS download in a Browsing Kiosk)
I just recently booted back into my system after an issue with 2 tools. rEFInd and supergrubdisk2, which were ISOs that I just copied to my USB rescuestick (because I like to be prepared) - both tools allow you to find those sneaky EFI boot options, even if some loaders are misconfigured.
Ventoy is the tool that allows you to run multiple isos from one boot menu without permanently creating new bootable USBs (www.ventoy .net/en/index.html)
rEFInd: www.rodsbooks .com/refind/index.html
SuperGrub2disk: www.supergrubdisk .org/
(you can even replace your Boot loaders with rEFInd, but you should know what you’re doing)
I get that everyone and their mother recommend OSes that did serve them well in the past but you need to understand, that each workflow may require a different approach to selecting something that does not break for them and put their earnings at risk.
That being said;
Maybe they should look at distrobox. In that box you can install virtually everything without affecting your host system (to a degree, but for simplicity, I’ll cut it short)
Whatever the decision is: the questions will remain:
Do they like to be safe to just make a rollback? Immutable OSes may really be an option here. as soon as you see a boot failing, rolling back is an essential feature.
However: you do need more reboots between installations and that may compromise their workflow as well.
BTRFS snapshots - yea almost the same thing. But no safety net will ever be enough, as long as they are not comfortable with setting up and using it the way they are supposed to.
I hope some of the tools and ideas were helpful.
For rescue systems; almost any live system is good enough, you just need to mount your partitions and can chroot into them. Ofc it’s easiest wit an arch-based one (others do not ship arch-chroot); but you can do that as well with I would say almost any linux distro out there with a regular chroot. What live-distros are they familiar with and have a smaller footprint - is something just the both of you can answer.
(sry, can’t post as many links as would be helpful)
