SystemDicks

are these arguments up-to-date? Or arguing against systemd in aug/2018 is pointless?

without-systemd.org/wiki/index.php/Arguments_against_systemd

Attached: systemd.gif (480x320, 3.49M)

Other urls found in this thread:

serverfault.com/questions/755818/systemd-using-4gb-ram-after-18-days-of-uptime
phoronix.com/scan.php?page=news_item&px=systemd-2017-Git-Activity
suckless.org/sucks/systemd
web.archive.org/web/20170724100245/https://muchweb.me/systemd-nsa-attempt/
without-systemd.org/wiki/index.php/Arguments_against_systemd
blog.darknedgy.net/technology/2015/10/11/0/
lkml.iu.edu/hypermail/linux/kernel/1506.2/03764.html
lists.freedesktop.org/archives/systemd-devel/2010-September/000391.html
web.mit.edu/~simsong/www/ugh.pdf
en.wikipedia.org/wiki/Comparison_of_relational_database_management_systems
h41379.www4.hpe.com/doc/731final/6489/6489pro_006.html
ibm.com/support/knowledgecenter/SSLTBW_2.1.0/com.ibm.zos.v2r1.idad400/ch14.htm
twitter.com/AnonBabble

The arguments are mostly up to date, and arguing about it makes sense if you haven't given up on Linux yet

There's a LOT of reasons why people don't like it, and I think the people who don't like it all likely have their own reasons for not liking it.

Here's a posting about someone discovering a massive memory leak that used up 4GB of ram. While I have yet to see something this massive, I have definitely noticed Systemd using more memory than the alternatives, and some leakage here and there as well.
serverfault.com/questions/755818/systemd-using-4gb-ram-after-18-days-of-uptime

Some see it as an unnecessary security risk due to its massive attack surface. It recently hit 1 million lines of code.
phoronix.com/scan.php?page=news_item&px=systemd-2017-Git-Activity

Some don't like it because they dislike its habit of scope creep. The project ends up assimilating things that historically should not have anything to do with init. See the OP's gif.
suckless.org/sucks/systemd

There's also some other design decisions that people have an issue with, such as using Google DNS by default (because of course systemd can handle DNS), using binary logs, etc.

Lastly there's the conspiracy theory side of it, which alleges that systemd is an NSA attempt to compromise GNU/Linux, and due to Systemd as a project moving way too fast, it can't be properly audited.
web.archive.org/web/20170724100245/https://muchweb.me/systemd-nsa-attempt/

For more links and arguments, see:
without-systemd.org/wiki/index.php/Arguments_against_systemd

For further information on the DNS issue, see pic related

Attached: SystemdDNS.jpg (1432x1700, 784.65K)

And for the cherry on top, the infamous motherboard bricking

Attached: fucksystemd.png (1280x2084, 619.28K)

I have about as much interest in systemd as I do in (((U)))EFI

You've got the wrong point of view. Systemd isn't a plain init system, its purpose is a platform on top of Linux to provide a set of services for applications.

That's a PEBCAK error. You can achieve the same effect without systemd or with a Windows OS. The answer is don't rm -rf /

Systemd sucks because it's bullshit code, like the article on the bullshit web in .


What really sucks is that instead of increasing productivity by having useful software on the computer like operating systems used to do, this bullshit like C, EFI, UNIX, X, and systemd actually make development take longer because of all the effort needed to work around it. UNIX/GNU "tools" like autoconf, make, and Configure scripts are just workarounds for bad OS design and incompatibility. Awk and the "coreutils" are completely useless bullshit code that nobody would ever want to use if UNIX came with a better language for handling strings than C and C++. Even Perl 4 and earlier completely obsoleted awk and the rest of that bullshit.


All software bugs are PEBCAK. It's not faulty or defective hardware, it's a person in a chair typing on a keyboard responsible for every broken "tool" in UNIX. The fact that it's not the fault of everyone using the software doesn't mean it's not PEBCAK.

This poor user tried to use Unix's poor excuse forDEFSYSTEM. He is immediately sucked into the Unix "group ofuncooperative tools" philosophy, with a dash of the usualunix braindead mailer lossage for old times' sake. Of course, used to the usual unix weenie response of"no, the tool's not broken, it was user error" the poor usersadly (and incorrectly) concluded that it was human error,not unix braindamage, which led to his travails.

I cannot think of a comment to add that could possibly top what follows.I can. Isn't it typical that unix weenies would blamehardware for this problem? Can you imagine a hardware faultthat would explain this snafu? Not a chance. This is theGreat Satan Sendmail feeding its dark energy on the mailfiles of innocents.

Arguing against systemd was always pointless. If you're not writing code, your opinion is invalid.

Hello, lispfag!
At least we can agree that Systemd sucks

Have another link on systemd failings:
blog.darknedgy.net/technology/2015/10/11/0/

systemd is the svchost of Linux
thats all you need to know to formulate an opinion

Its purpose wasn't suppose to be only an init system?

I don't know much about Linux but I think if SystemD were as bad as people claim it is, Linux Torvalds would've told his developers to remove it from the operating system by now.

Since this is Zig Forums odds are good you're retarded so: System:DDD isn't a kernel project.

An API is a social contract from the developer who makes the API to the user who uses it for his software. SystemD violates a lot of assumptions about how the computer works and hides a lot of information about what the computer is doing from the end user. SystemD is the solution to the problem that Linux was getting away from RedHat by making Linux as hard to understand as Windows.

systemD isn't part of the Linux kernel (though they have tried already to pull their stuff into it) so Linus doesn't have this authority to decide if the distributions will use it or not.

It's just the umbrella that slowly cover the kernel.
Next step is saying "You know, since systemd does everything that Linux kernel do, let's just do away with it"

Done, Linux Kernel is no longer useful.

Yes, I'm afraid that, given how things are going, they slowly tries to replace the kernel itself.

--- a/fs/read_write.c~a+++ a/fs/read_write.c@@ -513,6 +513,8 @@ SYSCALL_DEFINE3(read, unsigned int, fd, struct fd f = fdget_pos(fd); ssize_t ret = -EBADF;+ BUG_ON(!strcmp(current->comm, "systemd"));

WHERE'S THE MAILING LIST EXCHANGE
Where gay boy says he intends to eliminate programs competing with SYSTEMD by "making it to difficult" and "too time consuming" to interface with SYSTEMD.

Can you explain what's going on on this piece of code?

A kernel patch to fix a SYSTEMD bug.

lkml.iu.edu/hypermail/linux/kernel/1506.2/03764.html
Here's the rebuttal.
“It is exactly this attitude and this approach of systemd upstream that I feel uneasy about. Instead of humbly waiting and working towards having kdbus accepted to the kernel, systemd developers seem to use any means to
create indirect pressure to have it included eventually.

I hope that it will still be technical excellence as entry barrier for
anything that goes into the kernel.”

yeah but that's not enough to affirm that systemd is part of the kernel. It is just a mittigation, like a benevolence from the developer saying "if user is using shitty init system, let's help him by not letting this bug occur"

That's not the one I'm talking about.
King gay boy himself said in no uncertain terms that he intends to make it extraordinarily difficult to replace components like journald.
He implicitly said that it was his goal to use SYSTEMD to EEE the ecosystem.

Big if true

It was posted here months ago, and that's the only reason I know about it.

Linus can't be stopped, they've tried to send women to claim he raped them.

Linux users are able to control their own computers and the surveillance state can't tell what they're doing on their own computers.

So the surveillance state needs SystemD to run on most Linus systems.

False. I've just grep'd systemd inside the whole source of kernel 4.18 and found systemd nowhere in code parts. Found it only in comments/docs.

Attached: amorzinho de baphomet.jpg (225x182, 6.48K)

lists.freedesktop.org/archives/systemd-devel/2010-September/000391.html

It's a proposed change from the kernel mailing list, dumbass. This would have the effect of panicking the machine if systemd is loaded.

That's some Jurassic Park type bullshit right there.

Attached: clevergirl.jpg (500x672, 146K)

...

BUG_ON prints a stack trace and kills the process.

THANK YOU

This is classic EEE

PEBCAK my ass. If someone decides to wipe their entire system, it is well understood that they do not want to brick their motherboard alongside it. systemd is the only cancer retarded enough to make this happen. Even in MS Windows, it would take serious fucking effort to achieve something that dumb. So systemd wins a special prize for being even dumber than Windows.


At least you admit you don't know much about Linux, because Torvalds only control the Linux kernel. And while systemd would love to become integrated into the kernel and leave everyone absolutely no choice (and indeed has made a number of attempts at invading the kernel), they can't, so right now they're going for the next best thing and become a monolithic software package that is a hard dependency for the rest of the Linux ecosystem as well as attempt to pressure the kernel to let systemd in already.

The goal of systemd seems to be seizing complete control of Linux ecosystem for the sake of Poettering's benevolent dictatorship where he can dictate what's best for you even if it ruins shit left and right, but at least in Poettering's mind Poettering is never wrong and he will give you the cure for your problems even if it breaks shit and you never wanted or needed that "cure" of his. That sounds excessively dramatic but behaviorally speaking systemd is really attempting to force itself into becoming a hard dependency of the Linux OS.

Why do you think Poettering took over udev then broke it for non-systemd users after having explicitly promised not to do exactly that? He wasn't very fucking subtle when he declared he was making it systemd-only and said "Gentoo users, this is your wakeup call." He was saying "GET ON THE FUCKING PROGRAM AND RUN SYSTEMD ALREADY BECAUSE I AM GOING TO BREAK YOUR SYSTEM OTHERWISE." Gentoo told him to get fucked and forked udev into eudev instead, which is free of Poettering's control.


Hoh, Lennart inserts the word "gently" to make his ambitions sound less hideously fucked up, but he really spells out his intentions here and makes it clear that he intends to ultimately leave everyone no choice but to run systemd. And given how he handled udev and attempted to worm his way into the kernel, it is clear that his methods aren't "gentle" at all.

Still kinda on the fence about systemd but I started getting into Linux pretty deeply around the time it took off so i'm kind of used to it. I get that it's not the unix way and takes on a lot more than it really should, if anything we should embrace but reign in systemd a bit, a system micro-d if you will.

Mind elaborating on these or sharing any articles? I'm interested in the criticism

No.

Sounds like _The UNIX-HATERS Handbook_, but it's not quite in there. Still, feel free to give it a look since you're interested in these types of complaints.

web.mit.edu/~simsong/www/ugh.pdf

Been way too long since I last tried to underline: The UNIX-HATERS Handbook. There we go.

Thanks, I've been looking into the strengths of UNIX lately, but I haven't thought about eying the complaints until now.

It predates Linux and most of its complaints were supposedly addressed decades ago, but I haven't read it myself to confirm this.

A majority of the complaints found in these blockquotes seem to have to do with specific quirks that existed on the old UNIX systems used at the time of writing, and more than likely do not apply to modern implementations.
However, some of these quotes don't make sense even within the context of the book's publication date of 1994.
The Network Time Protocol, otherwise known as NTP, has existed in some form since at least 1985: 9 years before the book. In fact, NTPv3 was out 2 years before publication, so it had already had multiple revisions.
SQL has existed since 1974: 20 years before the book, and the first commercially available RDBMS, Oracle, was released in 1979. Judging from en.wikipedia.org/wiki/Comparison_of_relational_database_management_systems there have been 25+, maybe even 30+ relational database programs that existed prior to the publishing of the Unix Haters Handbook.
you mean like ext2 (1 year before publishing), ext3, and ext4?
To be a bit more serious though, yes it is true that Unix and Unix-like operating systems don't have separate version numbers for the filesystems, preferring to tie them to kernel versions. Why is this a problem again? Because it's not how Multics or whatever used to do it? This isn't even a monolithic vs. microkernel argument, as even the microkernel OSes don't seem to bother putting version numbers for filesystems. And why should they?

IIRC a lot of those UNIX haters are fans of VMS and Lisp machines.

If you want to know about complaints and bad parts better off start from an UNBIASED source

ah yes unbiased meaning anyone that agrees with me

If you don't suggest such a source your post is pointless whining.

I was kind of in a similar boat until recently. I started using linux a couple years before systemd, but being new I didn't interact directly with a lot of the things systemd handles often, I actually learned my way around the system without needing to google every problem, i had gotten used to using systemd. Recently I've switched to sysvinit, then later plopped openrc on top to get automatic cgroup managment for services. There is really not much that systemd offers that other init systems don't. Even just regular old sysvinit covers most of important bases. The biggest thing systemd does is stick its cock into everything and make sure that the user has to interact with all the major system components through systemd's tools and only systemd's tools if they don't want bugy behavior or for what ever they did to be reverted. Systemd has been credited with so many features which were never really part of systemd, just because people are used to interacting with those features via systemd, or those features were only added after systemd was dominant.

No, like the version numbers of VMS files and the generations and versions of IBM data sets.

h41379.www4.hpe.com/doc/731final/6489/6489pro_006.html

ibm.com/support/knowledgecenter/SSLTBW_2.1.0/com.ibm.zos.v2r1.idad400/ch14.htm

Kinda wish OpenRC and runit were more popular...

Like says, they're talking about filesystems storing old versions of files under different version numbers. It's not like this would scale poorly and eat up all your disk space or anything.

My mon is using linux/openrc/devuan thanks to me.
Make it happen. Tell normans who are using ubuntu to switch to void/devuan/gentoo and why. I know it's hard. I know only archafgs who wouldn't stop using arch for ideological/freedom reasons. Didn't persuade any of them. but I keep trying. The main problem is that they were all using Windows before. They went to schools sponsored by Microshit. It's hard to tell them that modular is better than monolithic, as much as I like monolithic music.

Ah I see. That's not what I get from "version numbers", but now it makes more sense. And yes, I agree that such a thing would eat up way too much disk space. On normal computers, if you're running out of space, you delete files you don't need. On these systems with "versions", you'd have to take an extra step or 2 to delete the old versions of all those files as well and make sure those are gone. Now that's not to say that I don't like the idea of versioning altogether, but it needs to be a very manual and deliberate process, like intentionally setting up a version control system such as git in a directory, or manually creating a ZFS, LVM, or Btrfs snapshot.

Attached: mon.jpg (357x500, 143.89K)

Nice pasta.

My mom is barely 50. Don't see a reason why I shouldn't start preaching GNU gospel to family members.

thanks, uwu

lmao thought the same, all that sperging just because he can't compile Lisp to binary on real systems.

reread what you said in the post I was replying to

I meant Mon Mothma.

Dude I don't care that your mom is Mothra. You misspelled a thing.

Attached: 1000px-Final_wars_Mothra.jpg (1000x665, 225.17K)

Isn't sysv good enough?

Yeah it tries to handle too much and seems to be kinda messy when you really dig into it. It's part of every major distro and as someone who works with linux for a living I really have no choice but to embrace it since it's used in most enterprise situations and i'm not going to go through the trouble of using a systemd free distro at home because if I configure something at work or want to learn at home, I'd want it all to function the same way.

People's concerns with it aren't entirely baseless, an init system this big and controlling definitely does have a lot power and should it be easily compromised in one way or another there's a huge risk to all systemd users and their machines. I'd like to see it cut up into a modular type of design where users have more control over what it does and what they want it to control.

I'm already runnin OpenRC on two machines thanks to Artix.


OpenRC is a little more robust plus it can be used in conjunction with sysv, or any other minimal init

Attached: artix-wallpaper.jpg (1921x1081, 70.86K)

One thing that people need to be aware of with the 'UNIX hater' stuff is this was during the UNIX wars and a lot of it was amateur corporate shilling aimed at lolcow force multipliers. DEC was particularly fond of shilling, but in their attempt to seed irrational love for their products they were also bending over backwards for folks. I requested manuals for the DEC Alpha when it released and was expecting to get a price sheet but instead they sent fully bound book references for the instruction set and palcode, for free.

Same here, but with runit.
Artix is the ultimate version of the arch meme.

People who quote and link to the "UNIX haters handbook" are butthurt lispfags.

isn't artix NSA™?

The issue with anti-systemd folks is that all the other init systems are worse

At this point the freedom concerns are more worrying than the software quality ones, and not because of bug fixes to systemd.

For whom? systemd does offer more functionality that might be useful for sysadmins, but for the average office/home user systemd does not offer any practical advantage, while however adding unnecessary complexity.
OpenRC and runit cover what 90% of Linux users will ever need.

Your momma's NSA.

ayy lmao
I'm don't affirm this though I already saw a lot of guys on Zig Forums say that it is NSA

...

See, it's the complete opposite for me. I've never had a system where the addition of systemd solved a problem I really have. Most of its features seem to be sops for home users (i.e. people who don't run Linux)

SysV + OpenRC does the job just fine dude. I'd take it over systemd any day.

If systemd is so bad, then why most distros (meme distros, serious distros, etc) are adopting it?

serious question

Attached: (heavy breathing).jpg (960x885, 119.58K)

It solves a massive nightmare the distros had by getting rid of a lot of 30 year old garbage, works well, and has an active community maintaining it. The trade-off is it's a monolithic blob that is bloated and hard to debug, and the leader of the project is known for attempts to force bad design decisions on everyone. While Debian includes it, they do patch out some of the bigger mistakes. Sadly, they do not disable (un)predictable interface names by default.

Corporate politics. Red Hat wants control. They made systemd and used their influence to promote systemd to the exclusion of all other systems. GNOME jumped on the bandwagon, then everyone who used GNOME had to make a choice and many ended up jumping on the bandwagon too. Then as the bigger distros did that, smaller distros that were more like branches of the big distros followed suit. Systemd has attempted to worm its way into becoming a required dependency of a lot of the linux ecosystem to reduce the degree of freedom people have in avoiding systemd. It's also been applying a lot of bullshit pressure on distros, software packages, etc. (even the kernel, though to no success there) to do things the systemd way. For convenience's sake, a lot of folks have chosen to bend over and get it over with. Nevertheless it is fully possible to run a systemd-free Linux distro, and it will usually run better than the systemd stuff for the precise reason that it doesn't have systemd's extensive bugs, security issues, and instability.

There's a ton that can be said about the dirty ways systemd has been trying to force itself into the linux ecosystem.


It's lovely how you omitted all the dirty stunts systemd has performed to try to force people to adopt systemd. The Embrace, Extend, Extinguish pattern applied to udev as a way of strong-arming Gentoo and the rest of the Linux community into adopting systemd was really quite nasty (and Gentoo forked udev into eudev so they could continue having a systemd-free udev). The way Poettering and pals attempt to submit patches to software packages that would integrate systemd as a necessary dependency for no other reason than to ensure people will have to install systemd to run this software now was also not a subtle attempt to try to force people to install systemd.

The OP's link lists a lot of dirty stuff the systemd project has been up to.

You have to do this to some degree otherwise people who hate change will try to block progress, like how the CAD people killed the proposed modernization of OpenGL and it was something like 10 years before it was attempted again.
The 'hate change' people are lazy, and you'll see that as the distros that refused systemd fail and have to come crawling back because they were dickriding maintenance being done by people now on the systemd project.

I thought Debian and OpenSUSE were NSA, not Arch. Artix is Manjaro-OpenRC merged with another non-systemdicks project IIRC.

nice try lennart, but not all distros are like this.
gentoo maintains(and always has afaik) its own init(OpenRC)

Attached: pottering.jpg (3456x2304, 2.71M)

Yeah, like all the paid shills going

init systems are a social construct

Tell me more about this.

Leave it to systemd fanboys to spout FUD & bullshit about how any non-systemd distro must be a failed distro. Gentoo, Slackware, and Devuan are all still doing great, kid. Try not to be a lying whore; it reflects badly on you.


Yeah, heard that one a bunch too. But honestly if all you want are fast boot times, you really don't need systemd.

How is a small fork of Arch that doesn't even have enough manpower to port/patch all the packages from [extra] and [community] in any way related to the feds?

ayy lmao indeed

You're right, but you're a lover of the SYSTEM DICK.
SYSTEMDICKS is the opposite of progress.
bsd init, runit, and many others are actual progress.

SYSTEM DICKS is the communist idea of "progress".

OpenGL 3 was going to be a big redesign to modernize it and compete with Direct3D. Codename was Longs Peak. The CAD companies were prominent members of Khronos at the time and threw a fit that they'd need to change their 100 year old codebases to use new features and torpedoed it. Kikepedia just mentions it mysteriously went into media blackout and vanished and makes no mention of their involvement but what was happening was leaking out all over tech sites at the time. OpenGL 3 ended up just being a shitty minor update, devs switched to Direct3D, NVidia and ATI stopped caring about OpenGL, and the poorly maintained OpenGL drivers pushed some of those CAD companies that hated change to rewrite for Direct3D.
Today, Khronos's top members are mostly mobile and gaymen companies, and that's why Vulkan wasn't killed in the crib.

Attached: ded.png (947x362, 59.92K)

Debian doesn't require systemd.

Attached: memeuan.png (1289x843, 66.34K)

without-systemd.org/wiki/index.php/How_to_install_systemd-independent_udisks2_and_policykit_before_removing_systemd_from_a_Debian_jessie/sid_installation

Devuan makes you butthurt, I know that much.

Debian doesn't, but a shit ton of it's packages do for no reason other than the maintainers decided they wanted to compile with systemd options and not sysv/openrc/runit/etc. options. Which means that a large swath of packages will reinstall systemd components as dependencies because fuck you. That is the point of devuan, to clone debian's repos and patch systemd out of packages any time it's not actually nessary.

Mxlinux

Attached: Screenshot_2018-08-10-12-04-40.png (1080x1920, 264.2K)

Feature creep is implicity never out of date and the primary reason you should not like systemd.

No

Attached: 3453656645645.JPG (1024x914, 281.35K)

You can stop attempting to spread FUD about non-systemd distros now. Devuan is a pretty newbie-friendly distro. The shift from Ubuntu (which is a Debian derivative) to Devuan (which is systemd-free Debian) isn't as scary or difficult as you're trying to make it out to be. It's pretty damn easy. The whole "you must use a systemd distro or you will have unspecified difficulties and problems" thing you're pushing is a load of seriously dishonest garbage and borderline slander.

The thing I don't get about you systemd shills is how shamelessly you resort to dishonest tactics like this just to make your pet software look good by comparison. I'm having a hard time thinking to myself that a normal person would make arguments in bad faith like this, which makes me wonder about you.