Basic OS upgrades, still failing in 2026.

[eeyore_syndrome@hachyderm.io]

@jorge

I am an enjoyer of:

'ujust update'

'r'

And then living my life >.>

[pfanzola@mastodon.social]

@Eeyore_Syndrome @jorge indeed, I love it!

[danathar@twit.social]

@Eeyore_Syndrome why even do that? Let the systemd timer get the updates automatically. You shouldn’t have to think about getting updates. It should just happen and the next time you reboot you’ve got them.

[whack@hachyderm.io]

@jorge brave to blog about this. Most times I talk about this problem (Linux upgrades going sideways) folks get angry that I even describe having experienced it, and why don’t I just send patches?! I don’t write about it much because the responses are truly deranged. lol

[whack@hachyderm.io]

@buckfiftyseven @jorge that’s hardly a reasonable coping strategy, but I won’t shame anyone for believing it is. “I am careful!” Is ok and I understand but like… “care” isn’t a useful risk strategy because the failures aren’t predictable (nor often recoverable)

And it models that those who end up with broken systems did so because they weren’t careful. Software needs to be less fragile 🥲
(If a user has a bad time, it’s a bug)

[buckfiftyseven@mastodon.social]

@whack @jorge really, are you going to say the people who run npm have the same risks as people who don't, and not running is just coping?

[whack@hachyderm.io]

@buckfiftyseven @jorge I’ve had many distro upgrades destroy themselves a dozen times over the past decade and none of the failures had any evidence they third party tooling like npm were a cause. I don’t understand what you’re suggesting? As example, My recent Fedora upgrade “succeeded” per the software but left the system in such a broken state I had to reinstall from scratch.

[buckfiftyseven@mastodon.social]

@whack @jorge I'm a retired guy with a deep history of software, starting in the 1980s. Environments have changed several times since then of course, but one thing has always been true. That is that software is not a uniform quantity, with a uniform risk.

If you want to say Fedora made a mistake, and should fix it, obviously. Fine.

But that does not refute the asymmetry.

[buckfiftyseven@mastodon.social]

@whack @jorge and for what it's worth, in terms of risk mitigation, I actually always do fresh installs. I love the clean slate. I have configured my methods so that it's easy and I'm back up in an hour.

[mookie@weredreaming.com]

I've been using Fedora Silverblue (the atomic edition) and it's been pretty much flawless. I've upgraded from Fedora 41 to 44 with no issues.

[jorge@hachyderm.io]

@whack Feel free to send angry #linux users my way.

I totally don't wear seatbelts either, I just drive very carefully!

[jorge@hachyderm.io]

@whack "What about the idiot that isn't paying attention and hits you?"

"It's ok, Linux is so easy to reinstall it doesn't even matter, try THAT on windows!"

[whack@hachyderm.io]

@jorge It sours me a little that so many conflate "I cope by doing X" with "It is acceptable that X is necessary in order to survive, and everyone should both know this *and* do this just like I do"

like... fuck. Were you 12 hours-deep into a video game? And you haven't saved since then? And the power went out? Baby, you gotta be more careful! Make sure you only play the first level of the game and no further, because going further risks losing progress! ?!?

[sternecker@infosec.exchange]

@jorge @whack I patch and upgrade Linux continuously as part of $dayjob. Stuff goes sideways all the time. I just attempt to keep the same weirdness from happening repeatedly. Always prefer new and unique weirdness! Whatever works for you, go with that!