I can't understand this reasoning. System updates obviously don't become less risky because of the OS they're updating. But going back to Win10 means having less control over when those updates happen (and much less control over, and understanding of, what is updated), and waiting much longer for them to complete.

▲

jakkos 8 hours ago | parent [-]

> System updates obviously don't become less risky because of the OS they're updating

The last time I used arch, I ran an update and it broke my bootloader, meaning the next time I restarted it wouldn't boot at all.

Sure I could make a recovery USB and fix it, but at that point I was away from home, and just really needed to do the totally crazy thing of "using my computer to actually do work".

(To be clear, I didn't and I'm not recommending going back to Windows, just a more sane Linux)

▲

ryandrake 7 hours ago | parent | next [-]

Yikes. It's 2026. "Don't break the bootloader" should be table stakes for any OS distribution's update process by now. I am not a fan of Windows or macOS, but I don't even recall the last time an operating system software broke my ability to boot--maybe during the Windows 2000 days?

Yet, when you go online to refresh your memory on how to update your Linux installation, too many of the guides still say STEP 1: Back everything up because you may not be able to boot after you do this!

▲

zahlman 7 hours ago | parent | prev [-]

My point was exactly that GGP shouldn't have expected to be able to do the system update without risk.

But the usual way to install Linux nowadays is from a live boot, so you automatically have a recovery drive anyway. It's not hard to set up regular restore points with Timeshift or similar, either.

That said, I haven't had problems like what you describe in nearly 4 years.

▲

jakkos 4 hours ago | parent [-]

I've used Windows, Ubuntu, and NixOS all for longer than I ever used arch and never had an update leave my computer inoperable.

	▲	zahlman 4 hours ago \| parent [-]
		And the other person did, as have many others on support forums. Updates generally just work, on the balance of my experience and other available evidence. It's still unwise to schedule them at a time where a failure would be more disruptive than baseline, and it's still best practice to be prepared for a failure and not have to figure things out on the fly.