Idk if this is a hot take but βno limitβ should straight up not exist. Just set a limit and blow the whole PC up when it takes too long.
post
If anything ever makes me leave systemd. It'll be this.
Why? You can bypass it immediately if it's a one-off, or go update the unit configuration to have a short timeout if it's a regular error.
More likely though, it's something that should actually be fixed, like a misconfigured fstab. Systemd is rarely the problem here.
how do you bypass it? didn't know its a thing
If you hit Ctrl-Alt-Delete 7 times in 2 seconds it will skip immediately
I can't tell if you're fucking with us or not. π
Hard power off dont care, and neither do modern filesystems.
databases may still care
Unless the culprit is that one process permanently stuck in uninterruptible disk sleep, in which case the timer would just perpetually go up, and pressing 7x Ctrl-Alt-Del as you mentioned will simply shut down systemd and the kernel itself will start complaining (faulty hardware can cause this)
You can do systemctl edit <unit>. There you can override all options without changing the main unit file.
Ughh recently had an amd graphics kernel bug which would lock up some /proc file somehow
Youβd go βwhy are fabs at maxβ type ps or top and they would freeze. System monitor would freeze.
βOkay, Iβll just reboot, this is fuckedβ, and systemd is just there:

Luckily itβs patched now. Hopefully those hard reboots didnβt damage anything.
Hard reboots don't damage anything but the OS state, and even that has been protected against the last 2 decades.
That is good to know, because I do not want to replace a 2tb ssd.
Apart from static discharge, or a broken power supply, I don't think there's anything your computer could do to the SSD to actually break it. Worst case you just need to reformat it.
(I guess if it writes to it constantly it could wear out, but that's different than a hard shutdown)
Some computer components don't protect against voltage loss and some OS components aren't tolerant to sudden shutdown but we're 99% the way there. Any of the two which don't have protection at this point are complete garbage.
That's not really systemd's fault though. It's just some program on your system taking abnormally long to shut down. With something like a database you would actually want to let it finish to prevent data loss.
The weak must perish.
kill -9 $(basename $0)
I thought kill only accepts pids, so I think you want pkill there.
Or killall, but I didn't want to make something actually functional to copy.
How will we learn without testing? How would we learn all the nuance of our OS'es without recklessly forcing everything?
This happens, too. A burst of local brownouts killed an OPNsense firewall of mine once. (The hardware was fine. The data got wrecked.) The firewall would boot, or almost finish booting, and then another brownout would hit. After the third or fourth cycle, the poor thing was toast.
I should thank the power company for free resilience testing.
Did you learn your lesson and get a UPS?
Brownout once fried my UPS. My guess is the inverter circuit.
Better the ups than the stuff attached to it.
If only we could be so lucky. They don't make Minutemans anymore.
My house flooded once, had like an inch of water in the living room, and my Minuteman ups was just sitting there beeping cos there was no power.
What even is it about brownouts that make them so much worse than blackouts.
A blackout is just a loss of power. Nothing complicated about it. There are obvious human and societal consequences, but as far as the infrastructure is concerned, a blackout usually means that protection devices have been tripped and the system has protected itself from greater damage.
A brownout means that the electric grid (or interconnection), for some reason, can't supply enough power to satisfy the demand, but protection devices haven't tripped yet. It can result in momentary power losses, drops in voltage, drops in frequency, or some generators getting desynchronized with the rest of the grid. If reactive loads (like HVAC motors) go offline momentarily, they can cause spikes in power consumption, which can overload circuits and trip safety devices.
A UPS needs to immediately switch to battery power when there are signs of an imminent power loss. A brownout might cause a consumer-grade UPS to switch on and off repeatedly, which the inverter circuit probably doesn't like.
This channel is run by a real civil engineer, he's released some great videos about the challenges of maintaining an electric grid.
This channel is run by a real civil engineer, he's released some great videos about the challenges of maintaining an electric grid.
Practical Engineering is awesome. For those that don't want to deal with YouTube, his content is available on Nebula as well.
I did. All the important things have time to gracefully shutdown now. (Except my for my rack of 4 DL360's, which need something a little more robust than a Walmart UPS. They aren't critical and only exist to handle weird one-off projects anyway.)
Not sure about that. When I first installed Gentoo with OpenRC everything went fine, just the boot/shutdown times were a bit slower than with SystemD's when it didn't do this shenanigans. But when I started reading about the SystemD hype I decided to try it and it did this at random - and got fed up so went back to OpenRC and never experienced this again. Maybe a misconfiguration on my part or something but I didn't do something out of what the manual said about how to set it up.
If it was really the program's fault, why doesn't it happen when I switch to sysvinit? (or maybe openrc, I haven't tried that one)
possibly they ship a bad systemd unit config for their package
Not onl that but they usually remove the init script as well, so unless you go for archives you no longer have a sane mechanism for starting and stopping the service.
Fortunately Debian has a orphaned-init-scripts package that can help alleviate that pain, as you can also use those scripts (or even better, openrc scripts) as skeletons to replace systemd service units.
Because you didn't bother fixing the underlying issue but got lucky.
Or these other inits are more brutal in their shutdown procedure, but I doubt that.
That's not really systemd's fault though.
Oh, but it is. Upstart and sysV will both assassinate everything after a point. This bullshit is entirely new, either by Systemd sucking, or Systemd allowing this shit (thus, sucking), where neither thing happened beforehand.
I admit I rage when it's starting or stopping and can't win for all the losing -- especially because "startup time" was the the main (unfounded) feature on the banner they waved around.
I work with sysV and Systemd and shut each down on a regular basis, via tooling, for stuff like patching, so it's often a horse-race to offline. The difference is often negligible, for new machines or established boxes with apps, and the main difference is sometimes Systemd fails its YOU HAD ONE job from time to intermittent time.
I work in a team where Unix and Linux is ~15 very experienced admins, so don't fret over the skills and experience question or the scale question.
What I would call out systemd for is that it can make it difficult to know precisely what thing it isn't happy with. Abstracting a mishbehaving process to "user session" when it could call out the offending process by name.
Similarly, when it's trying to decide the network is 'online', and by all appearances the network interface was configured, but systemd isn't so sure, sometimes it's hard to tell why.
A lot of the bad experience with Windows similarly could be attributed to something workable, but the design obfuscates the underlying behavior. Systemd has been bringing some of that sentiment over, and that can be more aggravating.
What, not any of the other stuff?
Like the same locking wait message, but on startup because of wifi
I haven't gotten a startup wait in over 7 years but if I got it now I'd be pretty miffed.
On Windows I get a "some apps are preventing shutdown....". With a list of apps. The app that is stopping shutdown? It's called "shutting down". Wtf
I hope my installation of Mint doesn't start acting up, because I don't want to learn whatever it is half y'all are talking about.
I just want it to work, and so far it has.
The only problem I've had is nvidia driver updates breaking Blender. So, I need to restore and tell update to ignore that one.
Same, my shutdowns take about about two seconds, although suspend can sometimes take about six seconds (but normally one).
I'm on Gentoo, so it's probably a leaner install than most.
> doesnβt want to learn about Linux > runs Gentoo
So I should wish for my computer to start acting up so I can have an educational experience attempting to fix it? No thanks, there's less frustrating ways to learn.
Also, if you're trying to go for a greentext-style formatting, put two spaces at the end of your first line.
systemctl poweroff <3
I'm still using shutdown -h now, even though the -h flag has been pointless for, what, 20 years?
edit: looks like sysvinit's shutdown still defaults to single-user mode, so I guess that muscle memory will still be useful if I ever use a Devuan or antiX box.
Sometimes this just happens with systemd, it wants to manage every thing as a session and sometimes things running in the session don't exit like systemd expects.
Also, sometimes logging into a headless ssh session takes quite a while, because systemd is not super responsive to setting up the user session and ssh blocks waiting for that to happen.
Got rid of Vista, installed Mint in 2012. Never a single regret. I am so good at Mint now, it's hard to be humble.
I've been using mint for years without learning a single thing about terminal command stuff. It just works. Granted I'm a pretty light user. Light gaming, high seas stuff.
mine has started acting up quite badly,
-from the start the internet has cut off randomly, likely due to driver issues or hardwareissue. Fixed only by complete restart. There seems to be no pattern, only that machine spirit gets angry if i use internet too much or sometimes just because i woke it up.
-Sometimes when i start the pc it takes really long time to boot and just stays in black screen,
-there are always kind of worrying error messages on startup and shutdown, something about rdseed being broken, something about data being lost, something about a thing hogging the processor and maybe switch to hq_unbound,
-sometimes the pc just refuses to shutdown when shutdown and everything else goes offline but fans keep running.
All of this has just slowly accumulated overtime, without me doing anything that should have caused it.
Yet funnily enough, i'm STILL very much more satisfied than with windows that didnt have that much immidiate issues.
This actually got me to go track down the systemd service that has been responsible for being unresponsive and causing a 90-second shutdown delay way more often than it should.
Do you really want technical details?
In my case it's the MPRIS service that's supposed to provide playback control on media players and related software.
... and so now it's disabled to see what, if anything, breaks without it. So far, none of the media apps I regularly use seem to be affected (browsers, a computer emulator, DOSBox, VLC).
It looks to be Bluetooth-associated based on the package it's installed with, and I don't need Bluetooth 99.9%+ of the time, so it might be a while, if ever, before I notice anything wrong.
systemd-analyze ?
"Ask systemd what's wrong with systemd" feels like "ask Grok what is bad about child sexual abuse" lol.
What about it? I admit that I don't know a great deal about systemd and all its subcommands - it took me a while to figure out how to disable that service properly, for example - but the manual seems to suggest that systemd-analyze is for startup processes, not shutdown.
I'll gladly take pointers on how to get it to say why a particular service hung on shutdown.
Used it a few times now, so much better than holding the power button.
For users that are not aware: alt+print screen then one at a time with a few seconds inbetween press the keys above, switch the O for a B to reboot instead of shutting down
That's what kill -9 is for.
Recently I had to reboot to fix a zombie process (caused by a GPU crash, not sure if it was hardware or the driver). When shutting down systemd waited 2 minutes for all process to be killed and then another 2 minutes for the specific zombie process.
What happened when you called kill?
Absolutely nothing. kill -9, SIGCHLD to the parent which was pid 1, relogged my graphical session. After I rebooted I had the thought that I should have tried removing and reinserting the amdgpu module to avoid a reboot. I couldn't start new mpv processes so shit was pretty fucked lol.
Transcript:
Top: A woman in a parked car screaming at a child in the back seat
Caption: Why canβt you just SHUTDOWN ?
Bottom: a child buckled in the back seat screaming back
Caption: α΅’βgfβα΅’βA.cSβtβop job is running for Session c2 of user β¦ (1:30/no limit)
The transcript is appreciated, but I don't think screen reader will take kindly to mixing imgflip.com and A stop job is running.
top 98 comments