Redpill me on systemd

>inb4 'durr if you dont like systemd LoC, you can't use Linux then. muh hypocrisy'
Linux has a perfectly good reason for its size: drivers. Linux needs drivers to be able to support all of the hardware that it does. Systemd does not have that excuse.
Well technically kernels don't have that excuse either, but until we get a usable microkernel OS, they do.

CIA/Mossad backdoor

when minix is unironically best kernel

Systemd saved me from OpenRC fucking up every time I updated something.

Botnet shit made by boot-lickers enjoying their botnet funding. If you bitch about Win10 [or any other version] and use SystemD, congratulations, you're a moron.

Attached: CIA floor.jpg (512x333, 68K)

It's honestly probably the best suite of configuration tools for Linux (independent of GNU) today.
The syntax is clear for the init system, it has easily manageable services, it even standardized printers. When you use systemd you're using a product that works and is maintained by professionals, not amateur cell-dwellers that made a single startup script and posted it online.

Attached: 1548981713100.jpg (283x283, 18K)

>When you use systemd you're using a product that works and is maintained by professionals

top kek are you trying to kill me with laughter?

news.ycombinator.com/item?id=14682210

it's a fake controversy sprouted by people who love to hear themselves talk and say things like "it's bad because it's low quality" so they can promote their own operating system that will get .000000001% of the desktop market share at it's peak popularity

>it's a fake controversy
Covington was fake controversy

This is more like Bush Jr. presidency; It should be interrupted and those responsible shot.

now that we are at it, what are some good systemd-free distros? ideally with relatively up to date packages and without the need to compile stuff.
>mx linux
>devuan
>artix
>void
am i missing any gem?

I'm guessing those are the better ones, you just forgot gentoo I think. Remind you though, Devuan is even more ancient than Debian because they have to re-compile everything to work on a non systemd system. It results on an even more stable system though

nevermind, i missed the "without the need to compile stuff"

systemd was created by an SWJ faggot who works for RedHat, a company that is basically a CIA front. It was forced on all distributions almost overnight.

The arguments against systemd and why it's bad have been posted over and over, yet systemd shills continuously pop up saying no one has provided any arguments against systemd and that Lennart Poettering's cock is small so it doesn't stretch your asshole or hurt too bad while he's fucking you over with a CIA trojan that runs at PID 1.

The systemd developers are making it harder and harder to not run on systemd. Even if Debian supports not using systemd, the rest of the Linux ecosystem is moving to systemd so it will become increasingly infeasible as time runs on.

By merging in other crucial projects and taking over certain functionality, they are making it more difficult for other init systems to exist. For example, udev is part of systemd now. People are worried that in a little while, udev won’t work without systemd. Kinda hard to sell other init systems that don’t have dynamic device detection.

The concern isn’t that systemd itself isn’t following the UNIX philosophy. What’s troubling is that the systemd team is dragging in other projects or functionality, and aggressively integrating them. When those projects or functions become only available through systemd, it doesn’t matter if you can install other init systems, because they will be trash without those features.

An example, suppose a project ships with systemd timer files to handle some periodic activity. You now need systemd or some shim, or to port those periodic events to cron. Insert any other systemd unit file in this example, and it’s a problem.

There are 3 main groups of systemd haters, and they all have valid critcisms:

People who don't like the fact that systemd has massive scope creep. Specifically that it tries to reimplement many existing services instead of improving / integrating existing ones. For example user switching, network management, logging, etc.

People who don't like the idea of everything relying on systemd interfaces to work at all. For example gnome started to rely on logind and other services even though it technically didn't need to.

People who don't like the management of the project. Lennart can be a dick to people with different opinions. He also created many interesting projects which were both a bit complex and pushed before they were ready. (like pulseaudio, packagekit) Since they were forced on people via popular distros, pulseaudio became "the thing that's always broken" for a year or so. And since Lennart was the author, he became a person who breaks the system.

The only people the systemd developers had to convince to get it widely adopted were the distro maintainers and the developers of a few key software projects. The distro maintainers made systemd at least an optional (if not outright default) init system, since it (1) promised to eliminate a series of init-related problems, and (2) other critical pieces of low-level userspace functionality like udev were getting merged into systemd's codebase anyway. In theory, systemd would reduce their maintenance burden, giving them an incentive to encourage their users to adopt it.

Making it an optional or default init system wasn't enough to spur its rapid adoption, however. What helped solidify systemd's hold was a few key developers making their software depend on it to work. For example, you can't run GNOME without systemd these days, and KDE is not far behind. Other previously-unrelated daemons like upower now require systemd to work correctly. So, users who want a working desktop pretty much have to use systemd now, even if they don't have an opinion on it, or even if they want a desktop more than they want to steer clear of it.

This has created a lot of frustration within the user community. There are plenty of users who do not want systemd, but now have to either (1) go without a working desktop to do so, or (2) fork the relevant parts of the ecosystem that depend on systemd and make them so that they don't. This is exacerbated by author of systemd openly advocating that GNOME depend on it (helping to create this dilemma), having a reputation for breaking the Linux sound system (making people wonder if the same will happen to init), having a reputation for not taking criticism well, and increasingly having a reputation of being deceitful about his intentions for systemd's role in the ecosystem.

contd.

Systemd's design has also created a lot of frustration with power users and developers. While it is all well and good that every developer should maintain his/her code in the way he/she sees fit, there are two important design decisions in systemd that have questionable technical merit but have non-trivial social and political consequences for the ecosystem. First, the author of systemd has repeatedly stated that systemd will not be portable, and will reject patches from those who would make it work. This effectively locks people into Linux if they need to use software that depends on systemd, even if the software doesn't otherwise require Linux-specific features. Second, the author of systemd has publicly stated that the interfaces between the systemd components will remain unstable and undocumented for the foreseeable future. This makes the creation of alternatives to systemd components difficult, since developers have to first study the large systemd codebase to even figure out how to begin, and will need to ensure their alternatives are compatible with every version of systemd if they are to gain adoption.

contd.

Combined, these two design decisions ensure that it will be very costly and time-consuming for developers to implement alternatives, short of rewriting the whole systemd from scratch. Any work they do to address their problems with systemd can be easily undermined by the systemd author (who doesn't particularly care for alternatives). What is particularly demoralizing about this situation is that there isn't a clear technical reason why systemd had to be designed this way--it could have been written in a portable fashion (or, people would step up and make it portable if the author would accept their work), and its components could have been designed to have stable internal interfaces, so replacing parts of systemd piecemeal would be feasible. The only perceived gains from these design decisions is that the systemd author can artificially make it too costly to compete against it, leading to frustration and controversy.

Very real disadvantages of systemd:

1. systemd is tied to a specific kernel and a specific libc and specific device manager and specific journaling daemon, basically, having systemd means you're locked in to a whole lot of other things.
2. systemd is renowned for locking up during startup and boot when you have network filesystems.
3. systemd hardcodes quite a lot of the booting and shutdown process in C which other systems place in easily editable scripts.
4. systemd in practice requires quite a lot of things: ACLs, PAM, dbus, polkit, these are not hard requirements but without this the above advantages are lost so all distributions enable them at compile time.
5. logind starting to do retarded shit like user sessions and having retarded power management, in theory you can disable logind, but no distribution again does this.
6. systemd is very monolithic and comes in one configuration compared to being able to piece your system together yourself.

7. systemd appropriates the cgroup tree and takes control of it and completely messes with any other user of the cgroup tree and really wants them all to go through systemd, systemd was wirtten basically on the assumption that nothing but systemd would be using cgroups and they even tried to lobby to make cgroups a private prioperty of systemd in the kernel but that went no-where.

8. systemd's usage of cgroups for process tracking is a fundamentally broken concept, cgroups were never meant for this and it's a good way to fuck resource usage up.

9. systemd has a hard dependency on glibc for really no good reason.

10. systemd relies on DBus for IPC, as the name 'Desktop bus' implies DBus was never written with this in mind and it shows. DBus was written to facilitate IPC within a single desktop session, not as a transport during early boot. This is why systemd wanted to push kdbus heavily beause kdbus solved some of the problems inherent to DBus being used as IPC during early boot.

11. systemd's security and general code quality practices are less than stellar, a lot of security bugs pop up in systemd due to its insistence of putting quite a bit of code in pid1 and quickly adding new features and quickly changing things.

13. systemd creates dependencies and is a dependency of things for political reasons in order to encourage people to pick these things. This is not conjecture, Lennart has admitted multiple times that he creates dependencies to 'gently push' everyone to the same configuration

14. systemd is monolithic for its own sake. It's basically product tying to encourage people to pick an all-or-none deal to again gently push towards this consistency.

15. Lennart Poettering, the face of systemd and its lead dev is the biggest primadonna FOSS has ever known who continues to shift blame and demand that entire world adapt to his designs.

systemd is the new UNIX. We're gonna replace all that old, messy, broken crap with systemctl tools

Saved

gentoo is the only non systemdicked distro that's worth using
it's the only one that's mature and well established
packages are up to date and it has a great number of packages, especially if you include overlays
and besides, it comes with advantages other distros don't have
>without the need to compile stuff
literally why? compiling offers lots of advantages and gentoo makes it easy

>This is not conjecture, Lennart has admitted multiple times that he creates dependencies to 'gently push' everyone to the same configuration
here's the mailing list exchange in question, if anyone missed this
lists.freedesktop.org/archives/systemd-devel/2010-September/000391.html
Lennart himself implicitly states that his goal is to pretty much EEE the ecosystem

Attached: 3620f733c39978b74810b25fd6d2aba8de5929c9684e6e5091ebae1b9d0cbcbc.png (3896x1769, 627K)

this. also, systemd is system space, not just an init system. if you don't want the linux platform to have standardization then move over to bsd you faggots.

1. kernel: who cares, libc: anyone willing to put in the work can port it.
2. hasn't happened to me since the rocky beginnings, file a bug report.
3. they want to replace the scripts, that's the whole point of the standardized system space that they're trying to create. you can still customize it if you want, as it's open source.
4. it's modular by design, you can disable all modules except systemd (init), udevd and journald, so that's not an issue.
5. nice, standardized sessions and power management.
6. like that's a bad thing. systemd is monolithic AND modular. the key word is "modular", you can pick and choose and replace what you want. next you're going to argue that a microkern... microservice design is better. it doesn't actually matter in the real world.
7. it should go through the system space, i.e systemd.
8. go on irc and talk to them about how the design is broken and that you have a better idea and will implement it.
9. if you want a different libc and it has a hard dependency on glibc, then put in the effort.
10. d-bus has gained a lot in scope, it is no longer really limited to desktop environments, but yes, that's why they were pushing for kdbus and BUS1. meanwhile they settled on dbus-broker.
11. quality assurance could be better and they should hire more developers to work on it.
13. we'll never standardize the low level stuff if every damn distribution does everything in their own special snowflake way, so encouragement is good.
14. monolithic != bad though, because it's modular in design. standardization is a good thing.
15. who cares, he can be a drama queen all he wants. I only care about the fact that he's trying to change things for the better, even though it's not perfect.

Ask yourselves this: Why won't they package systemd components separately?

For example, even though a DE is one project, it consists of a number of separate packages, most of which can be used standalone.
You're not forced to install the whole DE, you can just install the parts you want, like the WM, the bar or the file manager, etc.
If you want the whole DE, you install the meta package that pulls all the packages that comprise it.
Why can't systemd be like this?
Why shouldn't you be able to install systemd-init, udevd, logind, networkd, journald, crond (or timerd, whatever), timedated or gummiboot (now known as systemd-boot) separately?
Never mind that they're all part of the systemd project.
There's absolutely no reason whatsoever for those things to be packaged together, nor is there any reason for any of them to depend on each other for that matter.
Just like there's no reason for a DE's WM, file manager and bar for example to be interdependent (or worse, packaged together).
They said they wanted to "fix" or replace all those "crusty old bad" system utilities or something along those lines?
Well okay, fine, they did that. They took over some half dead (?) projects and "fixed" them or wrote their own "better" alternatives.
Maybe some people would appreciate that (if there was anything that required fixing or reinnovating in the first place).
Why merge everything and bundle it all into a single package though?
Let's compare Xfce and systemd.
In here github.com/xfce-mirror we got:
* github.com/xfce-mirror/xfwm4
* github.com/xfce-mirror/thunar
* github.com/xfce-mirror/xfce4-panel
* github.com/xfce-mirror/xfce4-terminal
* ...
Mostly standalone programs, maintained separately (naturally) under the umbrella of the Xfce project.
systemd, just like Xfce consists of a lot of standalone programs (or should be standalone programs), but they're all crammed into this one git repository: github.com/systemd/systemd
Why?

If they just maintained the components separately, most, if not all of the drama could've been avoided.

This. Gentoo is the best option.

Attached: gentoo-signet.png (533x533, 107K)

>Why?
It's actually not that complicated. They want standardization of the system level components, which is a good thing.
Everyone doing low level stuff differently is a real PITA for everyone involved. Remember when we had OSS/ESD/ALSA/Jack/Pulseaudio, it was a fucking mess and my audio always needed fixing, things broke, some programs just didn't work (WINE was a PITA). Now pulseaudio is pretty much the gold standard and it just werks.
Maybe, but then it would be just another init system, just another ipc, just another cron, just another device manager and everything still wouldn't work that well together.

The whole point of systemd (not initially, but it became it), is to create a system space, a complete package that "just werks" together.

>we'll never standardize the low level stuff if every damn distribution does everything in their own special snowflake way, so encouragement is good

Attached: lennart poetering tongue.jpg (1448x1341, 225K)

>Redpill me on systemd
They've disabled comments here
youtube.com/watch?v=o_AIw9bGogo
a few hours ago, because they tended to be objective.

Powertripping instead of dialogue. systemd in a nutshell.

Attached: systemd.png (1066x621, 87K)

>uses mac complains about privacy
kek

the distributions are encouraged to replace those _fallback_ values with their own values at compile time.

>A start job is running for ______ [1 min 30s]
>A stop job is running for ______ [1 min 30s]

I complain about advertising, my lil' nigger. Google is an adcompany, my lil' nigger.
You cheeky little nigger, you.

I think they were trying to excuse their shitty code by blaming it on the language.

Can't it just unmount and mount it again when issuing that command?

>"""system level components"""
there's no such thing as """system level"""; kys, closeted wincuck
>Everyone doing low level stuff differently is a real PITA for everyone involved
every gnu + linux is the same, nobody's doing "low level stuff" differently
it's the same across distros and even across operating systems (thanks, POSIX)
>Remember when we had OSS/ESD/ALSA/Jack/Pulseaudio...
cool story, pal
>but then it would be just another init system, just another ipc, just another cron, just another device manager
which is what it is
>and everything still wouldn't work that well together
what do you even mean by that
>The whole point of systemd is to create a system space, a complete package that "just werks" together.
no thanks

Install windows LTSC

>>"""system level components"""
>there's no such thing as """system level"""; kys, closeted wincuck
there should be a reasonable amount of """system level""" components on the platform that one can reasonably expect to exist.
>>and everything still wouldn't work that well together
>what do you even mean by that
a standardized way to do specific things on the linux platform that you can reasonably expect to exist and work well together.
>>The whole point of systemd is to create a system space, a complete package that "just werks" together.
>no thanks
why aren't more people using Linux!??!?!
also you.
standardization is what is required to get everyone on board. if you want to feel special about using linux with your obscure setup that barely works then you're a faggot.

>there should be a reasonable amount of """system level""" components on the platform that one can reasonably expect to exist.
define a "system level component"
>a standardized way to do specific things on the linux platform that you can reasonably expect to exist and work well together.
>"you need a standardized way to do specific things on the _same_ operating system"

>kernel: who cares
People who use other operating systems care if their desktop environments and programs start depending on systemd (which in turn depends on Linux) for no reason.

>if you don't want the linux platform to have standardization then move over to bsd you faggots
What's the point of this standardization if it isn't cross platform?

He's acting as if before systemd, all programs were completely incompatible across distros and that the userland was more fragmented than it actually was

>without the need to compile stuff.
>you just forgot gentoo
Absolute state of this board.

artix is just arch without systemd retard, it's completely usable.

windows 10 doesn't have systemd.

No one uses systemd besides those distros that got fucked over and now they wont replace it because it would be actual work.

Google looked at it and the second o went capital

>define a "system level component"
I already did, but I'll give you some examples:
>init
>service management
>job management
>device management
>graphics management
>network management
>time management
>user management
>log management
>tui and gui
and probably quite a bit more.

here's an exercise that I encourage you to do:
>write a program that is able to run and work correctly on the following distributions: Ubuntu, Kubuntu, Debian, Devuan, Gentoo, Fedora, CentOS, RedHat, Manjaro, Arch, Slackware, OpenSuSE, SUSE
>you may only ship binaries
>use only what comes installed by the distribution by default
>this program should have a gui and console interface, be able to set the global sound level of the user, mark a specific date in the users calendar, install itself as a service onto the system with root permission by the user, request the start of a service that isn't currently running or check its status, request the system to install a specific service, update the time of the system, edit / enable or disable network interfaces, log messages to their appropriate locations, install timers that can run specific scripts specified by the user
and there's probably a dozen more I haven't thought about. something like this is hard to do on linux, but easy on other platforms like windows or mac because it's standardized. if you like it or not, we need a standardized linux platform, and systemd is the right step in the right direction.

retard

Why are you acting like programs didn't work across different distros before systemd?

before standardization programs probably had to rely on a shit ton of #ifdefs to work across different configurations, you can yell "but muh choices" as loud as you want but the lack of standardization in the linux scene has always been a pain in the ass for coders
islinuxaboutchoice.com/

>probably had to rely on a shit ton of #ifdefs to work across different configurations
No not really. You don't know how this works. Systemd isn't actually relevant in most cases either.

tell us then, how do you make sure your code work across different configs without any standardized interface?

Why is it a reference implementation's job to be cross platform? No one is stopping other free and nonfree systems to implement freedesktop standards.

TLDR of this thread: Systemd haters still peddle debunked copypastas and conspiracies trying to trick newbies into using their meme distro. Systemd is still the best choice.

What is systemd?

this is why i went to bsd and sun

every week theres a new vulnearability. dont really want to use it because of that

What is Lennarts actual response to these criticism? Links appreciated

kys lennart we all know its you the CLOSED WONTFIX attitude to the entirely legit criticism gives it away
maybe learn how to not be such a shit dev instead of shilling your garbage

>shilling
it's you and your cult who needs to stop your shilling

What, the cult of not using garbage shitpile software?

thanks for proving my point, nobody cares what you use or don't use, just shut the fuck up already

Sure, nobody cares. Except for anyone concerned about security, anyone concerned about bugs, anyone concerned about efficiency, anyone concerned about what the software they're running is actually doing, anyone concerned with ease of recoverability if something does fuck up... Oh, and the shills like you who desperately want to force everyone to use their shitty garbage for some reason instead of any of the decent options.

Arch and its derivatives are all trash. Sorry.

here we see one of the defining traits of a shill, the more you tell it to shut up the longer the replies get

Just let it go. systemd rendered their decades of esoteric initscripts knowledge obsolete overnight, let them live their dream while the rest of us have socket activation and managed containerized services.

>projecting this hard

>esoteric initscripts
Were already obsolete long before lennart thought of shitting out systemd.
>socket activation
Again, already around before systemd.
>managed containerised services
You guessed it, already around before systemd. Kindly take your false dichotomies and ignorance of init systems and service management and shove them up your ass.

>SJWs evil
This isn't a valid argument, incel.

absolutely seething

absolutely seething

how much are they paying you to do this? this campaign to deface syatemd has got to be the most persistent and well funded hit job against free software, i mean, surely nobody sane would do this for years without anyone paying them

Absolutely cope

Stop, please, my sides can only take so much. Christ, I have no idea how you managed to keep a straight face long enough to type that.

hey im just employing the same tactics you and your cult love to use, any minute now someone's gonna start accusing me of being in bed with the nsa or the cia or the fbi

This only happens on distros that misconfigure their services. The same problem would occur with a different init system.

What you most likely are is a dumb zoomer fanboy. You obviously have no idea about programming, no idea about software design, and no idea about Linux prior to 4-5 years ago. All you know is that systemd is FOSS and was adopted by some of the major distros and so by appeal to authority and post modernism it must be the Right Way(tm). Try learning and thinking for yourself for once.

easy with the projection there, loser, yelling at a stranger won't help you overcome your shortcomings

I agree with him on this one. C is an ancient language, it has so many design flaws. It was fine when it was designed, but it doesn't suite the needs of modern million-lines-of-code projects. C has no support for error handling (returning a negative number is not error handling). Multithreading is fucked up. Syntax is cumbersome (type defined before the identifier, but "being an array" after). No support for generics data structures. And while we are at it, C has a very poor type system. Strings in C are bad and utf8 is not the default encoding. etc.

t. Rust loving tranny

I am not fond of rust either. Do you have any of substance to say?

>t. don't know what t. is

>Do you have any of substance to say?
Yes, trannyism is a mental illness, now go hang yourself.

can't even reboot

All these hatred is why nobody should listen to the systemd haters, you're like a sports announcer commentating a game being played by a pack of niggers who raped and murdered your wife and children and got away with it, you've lost the ability to be objective and nobody should ever listen to whatever you have to say. If I challenge you to say one thing systemd did right you'd probably reply with some worthless nonsense like "systemd did nothing right".

t. postmodernist

you can switch between read-only and read-write on the fly in linux, just a matter of;
mount thing place -o remount,rw
so yes, you can mount it read-only, then switch to read-write only when it's needed

this. systemd is the way forward, even with its current flaws (that will get eventually get fixed). people will look back at this whole thing just like they did with pulseaudio. when pulseaudio was first introduced, it was a nightmare because it was adopted without it being ready (thanks, distribution maintainers).
now pulseaudio just werks and I wouldn't want use a linux system without it.
the only ever way they worked was by adding compile time options and specific packaging per distribution, which requires compilation per distribution.
if you don't call this model broken, then I don't know what else to tell you.
his official response is:
0pointer.de/blog/projects/the-biggest-myths.html
just because trannies and fags use the language doesn't make it any worse. judge the tool on its feature set and tooling, instead of the people that use it.

do you not use knives? trannies use knives as well, you know, better not use them then. shit they also use forks and plates, sleep on a bed and live inside a house. fuck, you're going to have to move to the jungle to escape the things that trannies use.

You're either completely right or missing the point entirely.
efivarfs should, by the laws of common sense, be mounted as read-only, and remounted as read-write only temporarily when you need to write to it. If that's what you mean, then yes, it can and should have been done this way.
(If you mean that the user who issued rm -rf / should have remounted efivarfs as read-only, then you're retarded. But surely you aren't, right?)

Judging by the wording
>Can't it just
"it" most likely refers to the systemd utility that wants to write to efivarfs

By the laws of common sense your motherboard shouldn't hardbrick from erasing efivars, and if it does take it to the chink you bought it from.