why so many people hate systemd?
 

why so many people hate systemd?

There are many threads about SystemD here at LQ. You might consider using the LQ search to find them. You'll probably find more about SystemD-hate and -love than you want to know.:)

Here's one that's currently active: http://www.linuxquestions.org/questi...ta-4175608516/

It could be that when they use it, they were not sure of the settings and everything fails. And from that time on, the hatred for systemd has sprouted.

True for developers. There were a few complaints about the quality of their work in the LKML.

For the rest of the GNU/Linux ecosystem the issue is a philosophical one. If Linux is a replacement for MacOS, Windows, or proprietary Unix for you, then you might not care as long as it works without issue. For those who care about Linux mainly because of the freedom it offers to configure your system as you see fit, to install or uninstall any piece of software, to have absolute control over every little piece of it, this is a very serious problem.

For me, personally, I refuse to be *told* what to do with my own computer. Granted, we are not there yet, but we are uncomfortably close. I would imagine many similar minded people exist within our community.

Just like with all tech-related "religious" wars, people hate systemd because if they didn't hate it, there wouldn't have a good excuse to make useless posts in internet forums.

Out of curiosity; do you disagree that artificial limits are being imposed on the Linux ecosystem?

Isn't that a new thread?

Remove systemd and it will gut your system.
Somewhat more relevant then any vi / emacs war.
I'd be fine with systemd if it didn't decimate my system if I want / need to use another init system.

Is it, now?

what a question ;) ... because it is imposed. no choice. no freedom.

just one more reason, even Linus no longer trusts it.....

https://www.linuxquestions.org/quest...se-4175610133/

I does not matter much about it, anyhow, systemd is there and it means how bad has become programming today in Linux. This is a good reason to have *BSD systems.

Doesn't it have vulnerabilities? I saw a thread here linking to a YT video about systemd vulnerabilities, and that could also be another reason why so many people hate it.

there are thousand reasons ;) Funny that it came up.

I wonder why people don't stick to programming principles sometimes. Small programs are better than an huge one.

Trust-wise, I absolutely agree with you. BSDs look to be saner systems in a lot of ways. Strong philosophical differences, but I am truly glad that they exist.

Luckily otherwise Unix would be today dead.

You could also run Slackware (or Salix) which is the closest to Unix while still being Linux.

http://www.slackware.com/getslack/

https://salixos.org/download.html

I've been on Slackware for a while now. It is all the way nice. But even venerable Slackware might have to choose between no software, or undesirable software.

unlikely, most of those systemd hard dependencies are put there by the distro maintainers.....PV and Slack team are very much against adding it. They are very familiar with avoiding unnecessary software that other distro. are quick to adopt.

Slackware is basically dictated by the main boss, if he decides that SystemD will be, it will be.

I searched back this page from DEBIAN about the "said" advantages of SysD.

To me, those arguments make really little sense:
https://wiki.debian.org/Debate/initsystem/systemd

They, Debian developers, probably forgot what is :UNIX:

True, but in all fairness, I believe the Slackware team is working very hard to stay the course despite overwhelming pressure from all sides.

Honorable mentions of Gentoo and LfS. I'm not aware of any other distros over on this side of the divide.

Ok. I don't want to read to the bottom of that page, but it seems to boil down to "we're going with the flow." I could care less about the others, but for a distro as respectable as Debian, I think that was a bit weak.

I think that it would be good for Linux that Slackware goes too for SystemD.

Debian argues that most distro i.e. fedora,... use SystemD. This is their argument. So they use SystemD.
I would be good that Slackware go for SystemD too since all Linux distributions must have something in common. SystemD will define Linux entity.

BSD is UNIX and if you want Linux, it is Linux and has nothing to do with UNIX.

The kernel is all that they need to have in common; if all distributions go systemd there will be fewer Linux (and more BSD) users*.

For the record, I don't hate systemd, and if it were small enough in scope that its replacement in a distribution were trivially easy I would almost certainly still be a systemd user. But, since it can't be easily replaced if things go pear shaped, I really can't trust it.

*Not a made up statistic, I just conducted a poll (with a sample size of 1).

The last part of your post is unclear.

If the point here is to see Linux fall and all glory to BSD, I would point out that if you are happy with your distribution the way it is, a large influx of users might be the last thing you want.

OK. Linux and BSD are luckily there for everyone, great OS, best of all time.

SystemD is a problem currently for the Linux community.
Clearly there aren't just all in favor of SysD, just looking at the poll results on LQ.

systemd blocks netflix :D
 

Interesting incendiary (noun; syn: firestarter) for the most recent SysD thread (this):

LXer news @LQ: http://www.linuxquestions.org/questi...ix-4175610482/
-> https://theregister.co.uk/2017/07/24...main_name_bug/
-> https://github.com/systemd/systemd/issues/6426

btw, OP hasn't been back since 7minutes after post#2here :eek:

The core issues for me are:
- there is no specification of what it actually is. So it keeps gobbling up parts which have no business with pid 1 at all.
- the whole mentality regarding bugs or general code quality.
- documentation? Where can I find documentation on it?
- bloatware. Bad idea for something which manages pid 1.

So far we've had 27 answers and the best anyone can come up with is "it's contrary to the spirit of Unix" or anecdotal complaints. When I hear some-one screaming that it bricked their computer, then I'll worry. If we avoided all change in Linux, we'd still use Lilo, have to do all configuration by editing scripts, and lack proper package management. Oh, that would be Slackware, wouldn't it?

This is a good enough reason I think.. Anything that's missing documentation is probably hiding something, usually backdoors of some kind.

NO_'feature'_Linux: a brutal body blow
 

No, 'that' is the distro -I- like using: mll
where BusyBox is the only executable, there's NO /etc/passwd, telnetd -l /bin/sh
and thus bb is init, sh, implements dpkg&rpm (a bit), noGUI:party:, ... :D
BYO bootloader & tinyconfig your kernel (but don't mount anything like /boot) ;)

All the arguements against systemd you can ever ask for http://without-systemd.org/wiki/inde...md#Scope_creep

And thanks Doug_G!

https://www.theregister.co.uk/2017/0...main_name_bug/

You mean that a bug might occur with it

this thread should help further the discussion:

https://news.ycombinator.com/item?id=14733255

https://plus.google.com/+TheodoreTso/posts/EJrEuxjR65J

nice link.


Ah Programming... SystemD is not very well praised.

Pottering admits systemd's true objective, much more then init:
https://lists.freedesktop.org/archiv...er/000391.html

current thread
https://www.reddit.com/r/linuxmaster...0/systemd_now/

older threads
https://www.reddit.com/r/linux/comme...d_itself_isnt/
https://www.reddit.com/r/linux/comme..._with_systemd/
https://www.reddit.com/r/linux/comme...t_controversy/

Ouch. Really looks like he's going for a One True OS. Between that goal and the aforementioned feature creep, surely SystemD should be treated as a separate Linux-based OS in its own right, like Android? At least that looks like the direction it's heading.

Personally I've had some bad experiences with large, essentially-monolithic OSes, and don't like the idea of having one Poettering along on my hardware.

casync (1st mention of it on LQ!)
 

Looks like 'He's going for' casync. FAQ@bottom: "hosted under the github systemd umbrella"

2004: PulseAudio; 2010: systemd; 2016: casync (needs to be added here:eek:); 2022: ?

What is casync ?? You forgot Wayland too!Some other thingss like Wayland?

I think that Windows'old user Linux engineers should create more stuffs. What about Start.elf and Explorer.elf? One file only, to control the Whole Gnome Linux OS System?

He could be going to be more famous that Linus making Linux/SystemD Operating System.

Originally it was for Gnome, which makes sense. It was too important to bring sysD on Gnome.

I'd just like to point out to those reading this thread who aren't in the know that the above quote was never actually said. It is a meme joke, fake news as it were, based on the words of Richard Stallman who said the same thing about GNU, not SystemD.

I joke about it, but it does look strangely familiar, this statement.

I find it rather/quite funny, but maybe, it might off-topic, I would say.

Poettering's views, agenda, objectives and paymasters have been known for years, yet the same silly threads with the same arguments from both sides, going in circles continue. Most of Linux has moved to and embraced systemd. Apparently "it works" and that's it. If you don't like it, the only choice you have if you can't write code is to vote with your feet. That's been the situation for a quite a few years and it hasn't altered much.

The argument has always been why do we need a huge monlithic program running everything and the kitchen sink that can break down and sink the whole ship faster than the Bismark sank the Hood when we can run a cluster of individual programs that serve an equal function and are interdependent yet independent of each other that can be easily restarted without the ship pretty much doesn't so much as spring a leak?

The real issue is tried and true, yet simplistic good design versus haphazard design trying to be flashy, cool, hip, and trendy. I've gone as far to say systemd is trendware, hipsterware, fadware, etc. When every big name distrubution with funding and commercial backing all to willingly dove into the pool without checking the water level, and ended up with a mess, I considered it to be as I called it.

Plenty of alternatives have existed for years. OpenRC, Runit, sysvinit combined with perp, daemontools, and any of the aforementioned. Just nobody wanted to use them for whatever reasoning. I still stand by my argument of developer laziness.

I also think Devuan was onto something and maybe they knew the DNS issue existed, laughed, and said, "let the fools figure it out if they're so all-knowing as they claim and say they know better" and just kept silent. A bug existing for so long is rather embarrassing.

This is awesome what you said.

Is that true that the DEVUAN dev team knew about the DNS issue? They excel so much in programming.

It's all very simple. Linux, GNU and before that BSD were born because someone had an idea to make something better, do it properly. simple != easy. BSD and GNU in particular, among others, were designed around the needs of academics, not businessmen, at some of the top Universities in the US. They were designed to be simple to administer troubleshoot and fix (yourself).

systemd is apparently born out of the ideology that all of this is a "mess", has always been so and needs fixing by "unification" and re-implementation of the "fragments". Software of this kind which spawns from corporate needs is not interested in being correct, secure, modular, efficient or robust. It's simply about fulfilling that specific corporate need(s) and can often be "broken by design", as well as overly complex or even deliberately obfuscated if there is a real and compelling business case for that. It is not designed to function in a "free software ecosystem", because it's developers already arrogantly assume they're replacing that. It's not designed for you to use how you want, because those are not the clearly stated goals of the software.

It seems to be evolving around the needs of Red Hat, which is all well and good. It obviously comes from a business need - a need to nurture a paid-for-support revenue stream.

Poettering's views and opinions on all of this have always been publicly on show. The objectives, the goals and the dismissive arrogant attitude towards "legacy" or "irrelevant" software can all be found if you want to search. Apologists, on this forum and others however, have adopted a "head in sand" approach, because they use it, "it works" and because they're now so heavily invested in it and their distro of choice, that it's become an inevitability and it's more reconcilable to just continue talking it up, ridiculing it's critics and closing their eyes to it's obvious defects, than it is to challenge what is actually showing itself up to be shoddy software.

Sad state of affairs.