Love it or loathe it, systemctl is trying to do the right thing with regard to stability and data preservation.
If you really mean it, the manual offers a few levels of strength beyond the plain one: -i (don't check for busy processes, which is what's going on in the meme), -f (force, presumably asks even less nicely), and -f -f (don't even ask, just do it now, preservation be damned).
It should give you the option to abort the shutdown and sort out whatever process it is though! Or at least let you kill it manually from the shutdown terminal. I know you can technically do that with the emergency shell but I don’t like leaving that enabled. Thankfully I rarely get this issue anymore anyway
I love how this comment suggests every fucking alternative doesn't or wouldn't. That's just bloody arrogance.
Systemd's entire existence is against best coding practice. Famously, when called out just on the ability to work with others, the systemd team represented trends ably.
Never have I raged at a machine and demanded it tell me what the flying flaming fuck it was actually doing now than when systemd was trying to do what I'm charitably deciding is the right thing.
Why would be doing the right thing now? It honestly only does a thing through luck and race conditions anyway.
You can do that to Windows. They may have gotten better, but I know that my friend that ran Debian Unstable back in the late '90s-'00s swore that if he didn't properly shut down the machine every year or so, it would mess up his build.
I've recently had so many random freezes of the system, hangs on shutdown, panics on shutdown, freezes in system updates, that hard reset became a thing I did several times a day. Yet there were no systemd logs, nothing in dmesg, literally zero information on what happened.
I was skeptical in blaming Nvidia because at this point it became a Linux chiche, but then I started to switch to integrated graphics (disabling dGPU) and all of the problems miraculously went away.
Yet there were no systemd logs, nothing in dmesg, literally zero information on what happened.
this is pretty typical for hard crashes, ur system is so unbelievably fucked up that it can't even write to journal, and if it could, it wouldn't be persisted anyway (hard shutdown)
Same. I use nVidia on Wayland, and experience more crashes and panics than when using the iGPU. With older versions of the driver, I could consistently trigger a crash when exiting an app which used the discrete GPU (such as Steam), or by switching between a game and Firefox.
NVIDIA definitely has stability issues, newest drivers still kernel panic on resume from suspend. Only thing more you can do is try to capture debug logs with nvidia-bug-report.sh (I go in during a crash via SSH, usually the system is still responsive for a little while after), and post it to the NVIDIA Linux forums. They do actually seem to use the feedback there, NVIDIA reps respond from time to time and say they've submitted bug reports from the feedback. Otherwise, after that yeah you just do what you have to do for a usable system and wait...
It isn't those for me tho. It's just BTRFS that like to hang. Auto scrub going? Hang. Auto balance? Hang. Time shift/urbackup doing back up things? Let's hang!
I have a, honest to goodness breaks the electron flow, power switch for a reason, the shutdown command was a warning not a request.
Such wise words.
Love it or loathe it, systemctl is trying to do the right thing with regard to stability and data preservation.
If you really mean it, the manual offers a few levels of strength beyond the plain one:
-i(don't check for busy processes, which is what's going on in the meme),-f(force, presumably asks even less nicely), and-f -f(don't even ask, just do it now, preservation be damned).It should give you the option to abort the shutdown and sort out whatever process it is though! Or at least let you kill it manually from the shutdown terminal. I know you can technically do that with the emergency shell but I don’t like leaving that enabled. Thankfully I rarely get this issue anymore anyway
I love how this comment suggests every fucking alternative doesn't or wouldn't. That's just bloody arrogance.
Systemd's entire existence is against best coding practice. Famously, when called out just on the ability to work with others, the systemd team represented trends ably.
Never have I raged at a machine and demanded it tell me what the flying flaming fuck it was actually doing now than when systemd was trying to do what I'm charitably deciding is the right thing.
Why would be doing the right thing now? It honestly only does a thing through luck and race conditions anyway.
I'm not sure I'm a fan of systemctl either, but I think your hatred of it has caused you to read way too much into what I said.
How did you get that from their sentence, what the fuck?
You can do that to Windows. They may have gotten better, but I know that my friend that ran Debian Unstable back in the late '90s-'00s swore that if he didn't properly shut down the machine every year or so, it would mess up his build.
Runs debian unstable. Shuts down his machine every year or so.
For Debian, "unstable" just means "not running a five year old compiler".
I feel the same way when I use my turn signals. I'm not asking.
(assuming of course it's safe to follow through)
Such is life when using Linux on a laptop.
I've recently had so many random freezes of the system, hangs on shutdown, panics on shutdown, freezes in system updates, that hard reset became a thing I did several times a day. Yet there were no systemd logs, nothing in dmesg, literally zero information on what happened.
I was skeptical in blaming Nvidia because at this point it became a Linux chiche, but then I started to switch to integrated graphics (disabling dGPU) and all of the problems miraculously went away.
this is pretty typical for hard crashes, ur system is so unbelievably fucked up that it can't even write to journal, and if it could, it wouldn't be persisted anyway (hard shutdown)
Same. I use nVidia on Wayland, and experience more crashes and panics than when using the iGPU. With older versions of the driver, I could consistently trigger a crash when exiting an app which used the discrete GPU (such as Steam), or by switching between a game and Firefox.
NVIDIA definitely has stability issues, newest drivers still kernel panic on resume from suspend. Only thing more you can do is try to capture debug logs with nvidia-bug-report.sh (I go in during a crash via SSH, usually the system is still responsive for a little while after), and post it to the NVIDIA Linux forums. They do actually seem to use the feedback there, NVIDIA reps respond from time to time and say they've submitted bug reports from the feedback. Otherwise, after that yeah you just do what you have to do for a usable system and wait...
Had this issue for ages. Ditched Nvidia a month ago and now everything just works.
Yeah, lol. One of the reasons my next laptop will be one intended with linux support for the start.
Why does that even happen?
The kernel is cleaning the corpses out of the basement freezer chest before the power goes out and stinks up the place.
Leave the poor kernel out of it, it has nothing to do with this. It's Lennart, not Linus.
Unmounting CIFS , 1:30 timeout ! I just wannatgo to bed....
I've been living that for years
I prefer holding the power button for 5 seconds
has the added benefit of a satisfying POP! and turning all the lights off too...
ready for bed!
It's very energizing too
Yeah, I know. It's just Sunday morning and I'm feeling cheeky.
I was glad my server did this the other day to make sure the data Lemmy put into my database is secure.
The culprit often filesystem sync for me...
Yes. Mounted network drives.
I think a lot of the shutdown hang problems went away when i switched to systemd automount
It isn't those for me tho. It's just BTRFS that like to hang. Auto scrub going? Hang. Auto balance? Hang. Time shift/urbackup doing back up things? Let's hang!
Still faster than Windows
Meh, init 6, you coward
kill -9 1