I have multiple systems running manjaro stable, and few of them are set up to update automatically every week and install on shutdown. On next boot they either failed to boot or had no network/audio etc.
I came to the conclusion that the kernel update didnt go thru properly (all systems were up to date), since reinstalling linux66 kernel fixed the problem everywhere.
I rly hope this wont happen next update, but if it does, where do i look to see what causes the error? The systems are completely unrelated in hardware.
youāre on a rolling release, not to say a bleeding edge distro. hope is good but constant backups are better.
especially if you run a āunusualā configuration.
In addition, Iād not recommend unattended upgrades, that youāve had done there.
Chances are, if you didnāt rely on magic, this wouldnāt have happened.
Manjaro is stable, rock-friggin-solid. It can be used as a server even. That doesnāt mean it should be done. Itād be a much better idea to rather use Ubuntu LTS. Like I have on my Raspberry PI Home server here.
But Iāll stick to my PC for Manjaro. I didnāt even recommend it for my brother when he asked, purely because he wouldnāt be able to constantly keep it up-to-date like it needs to be.
I have timeshift snapshots set up so yea, but still its a pretty common thing to expect on a modern distro, far from unusual, most ppl dont wanna bother manually updating system every single time.
For home server i run ubuntu lts aswell, never had a problem w unattended upgrades over many years, but for desktop i prefer a rolling distro, but not bleeding edge like arch or debian unstable. I dont think having automatic updates is relying on magic, should be a fairly common option among distros, not many ppl would bother updating manually every single time.
pamac itself has specific problems with memory, full cache, crashing system ā¦ possibly related to AUR packages and auto-update.
See, ex:
(theres also a dozen or so similar threads if you use the )
To clarify a bit - just so you know ā¦ it wasnt the kernel update, or the update itself in general, that caused this. It would appear to me that what caused this was an interrupted upgrade. Which in turn was probably caused by a situation similar to the link above; ie - pamac running out the memory, causing a system hang during upgrade.
Different issues after updating suggest there is possibly more than one issue
Failure to boot might be issue from plymouth splash screen
Failure of network/audio devices could be failure to load a working device driver
But I prefer to look at the data before speculating on possible causes
where do i look to see what causes the error?
I suggest looking in /var/log/pacman.log to see:
what packages were updated
if any .pacnew files were creaed and need merging
any error messages during update
To see error messages after system reboot check system Journal
I donāt know if Raspbian comes with automatic updates or the image I have has it installed but it screwed me over so many times I had to get rid of it and update āmanuallyā. Of course I donāt do it manually as such, I use Ansible. The amount of times Iāve had to SSH to one of the RPIs went down drastically. Raspbian is based on Debian.
I donāt expect or trust any āautomaticā program to do updates properly.
Manjaro requires more. I would say you are expected to read release notes on every update and if there are steps to do beforehand, do it.
Now, I donāt know how OpenSUSE Tumbleweed is over a longer time but maybe that is an option? Still rolling-release but less active management with updates, it seems. I just donāt like TW because it has massive updates constantly. And people praise their automated build system for packages. Makes me trust it even less.
The thing is, automatic updates rely on everything going hunky-dory. a Machine canāt do something on itās own, something it hasnāt been programmed for by a human. And humans can only compensate, or handle any problems he/she/it knows about.
So in order for an āautomaticā update to work flawlessly, there are a lot of stars to align, a lot of variables. And thereās absolutely no way for something that has so many variables to be flawless. Less error-prone? Sure. Flawless? Not going to happen. Even more so with a rolling-release that pushes the updates quicker than, and with less testing than a point-release system.
Before I switched to Linux and when I still had an IT support job, I witnessed even Windows update problems. It can happen. As you just experienced.
)