What is intelligence? I'm not talking about the spy thing, but what is the definition of intelligence in humans? This is pretty good: http://en.wikipedia.org/wiki/Intelligence#Definitions
For most people, today's self-appointed and possibly overly influential creators and thinkers working around the "One Linux" idea fit the definition of smart people - at least technically.
And his messages are quite convincing:
- Simplify multi-distro development.
- Allow faster boot times.
- Allow an architecture on-demand and event-based, similar to "modern" operating systems.
- Bring order and control to subsystems that have had as many different tools as there are distros.
All seemingly noble goals. All apparently derived from a deep desire to contribute and make things better.
Hardly anyone could argue that these clever people thought seriously about these issues, and put an enormous amount of effort into solving these problems. Unfortunately, the solution they came up with, as you may have already guessed, is "systemd".
Although it is not something new, its grotesque impact has finally reached me and I must express it publicly.
So what is systemd? Well, meet your new God. You have probably been praying at the altar [if-it-works-don't-touch-it] of simplicity, but your religion is being eradicated. That probably already happened without you realizing it during an update to your Linux machine. systemd is the all-knowing, all-controlling, all-seeing and all-supervising meta-deity. He is the new Single Master of Processes who aspires to control everything he can - and he is already doing enough. It's what init would look like if it were a transformer on steroids. Is [every geek's dream] complicated, multi-faceted, opaque, and supremely powerful.
I heard about systemd a few years ago when upstart and some others replacements to init that I can't remember were popping up on the scene. And although it looked mildly interesting, I was not in favor of using it, or any of them for that purpose. init worked just fine for me. init was simple and robust. While the configuration had its specific differences for each distro, it was often these differences that made one choose to use that distro in the first place, and stick with that distro [(and not things like which audience it was aimed at like today)]. The tools essentially were the distro. And then I just did a dist-upgrade to Jessie, and voila, PID 1 suddenly became systemd. Go shit.
In a "One Linux" world, what should distros really be? Obsolete. No longer relevant. Archaic shells of his once proud [and often dangerous] individualism. Basically now it's just a logo and a default wallpaper. Because, let's face it, there just needs to be a Unique Modern "Competitor" to the Windows / Mac domain of personal computing. A unified front to fight the malevolent empires of Redmond and Cupertino is what is needed. The various differences that created different "flavors" of Linux needed to be concentrated and brought into conformity to proceed efficiently for war. Um, what war?
For me, Linux already won that war back in 1994 when I started using it. It did so without firing a shot or trying to be like the other operating systems. He won it [always keeping 1% of the market] not giving a damn about market share. He won it by being exactly NOT like them. He won it by being simple, understandable, and configurable to be exactly like YO wanted to. It earned it by being a collection of simple, modular components that could be attached together at will to do real work. He won it by adhering to the deeply considered philosophy of keeping the user in the driver's seat, and being free to run whatever he wanted, without layers and layers of frameworks coiling their tentacles inside all those things they shouldn't touch. He won it without all those various "ShitKits" that have begun to manifest within the heart of my operating system. He won it without being too complex and unintelligible. That kind of opacity is the heart of Windows and Mac, and it's exactly what I hate about them, and exactly why I chose to use Linux in the first damn place. systemd is hugging All what I hate about Windows and Mac, and doing it in the name of "modernity" and "simplification" of the developers' work.
So why do people Very intelligent which love and use Linux would they want to create or hug such a demon "Master of All"? Obviously, it is for the reasons they say, as I mentioned at the beginning. But partially I think it is due to lack of experience. I do not mean a lack of programming hours, but a lack of time on the planet. Intelligence alone is not a substitute for life experiences and, yes, I am going to say it, neither is it for wisdom. There is no manual for wisdom. Implementing systemd on distros is not a wise move for them in the long run. That will, in fact, bring about their ultimate doom.
They are partially the larger-than-life egos of the people involved. Someone has read what does Poettering say about things? Wow. This guy is obviously convinced that he has all the answers for everyone. Traditional ideas about simplicity and freedom are colorful, but they have no real place in a "modern" operating system. Look, he's just smarter than you, so get over it and step aside. He knows what's best, and he's got it under control. How old is that guy anyway? 12 years or something? It's a fucking dick (in my humble opinion).
They are partially his shady business interests. Look, we could make money selling things to Linux users if there was a simpler and "dystro-agnostic" way to do it. Fuck the options, they'll like what they get.
Partially it could be that it is ominous and bleak by nature. One One Ring to rule them all, secure access to it would be captivating for those bent on having full awareness of the information. Confidence is not something that is really high on my list of things to concede these days.
Partially it is the belief that the Linux Community must fight against the hegemony of Windows and Mac - as if the existence of Linux depended on the defeat of the other alternatives. Those who think that Linux should care about idiots and HOYGANs they should go back to their Macs and Windows, and stop trying to "fix" Linux. It ain't broke, damn it!
Partially - and this is what I cannot accept - it is a blatant disregard and disrespect - whether on purpose or not - to the most important principles of * NIX. It is an unconscious disregard for them, and a trampling of the values that I personally hold as [old,] true and fair, and I'm not alone here. systemd is the exact opposite of what * NIX stands for. And I'm not babbling about POSIX compatibility either. It's stupid philosophy.
systemd is a coup. It is a subversive intruder designed to destroy Linux as we know it, imposed on us by the mob. we-know-him-better-than-you camelcase. And they just don't understand it as deeply as they should. systemd is not pointing in a direction it should be going. Does not support freedom. It does not incentivize choice. It does not show transparency. It does not embrace simplicity. Take control and force you to give in to him. It makes applications and major system components dependent on it, and cannot function without it. You are gaining speed by luring into the fold of naive or lazy or just clueless developers with the promise of making their lives easier. To fall into this way of thinking is to ignore the great danger that systemd represents.
Debian has always stood the line against this sort of thing in the past, and it has always earned my utmost respect and loyalty for its integrity. Debian's decision here was forced. Debian has made a serious and cowardly mistake [when Tagliamonte did not annul Garbee's goal], and they need to correct their course immediately. Incorporating systemd was not a smart move, and certainly not a well thought out [(and no, I didn't believe any of Russ Albery's reasons)]. Debian must reject systemd and its ilk, and return to the values that led Linux to this point in history, led not in a small part by Debian. They must soundly and publicly divorce GNOME, no matter how painful and annoying it may seem in the short term, and focus on the most essential values of simplicity and freedom. Place systemd and his clique in the repositories non-free where they belong if there is no choice. Let the user decide if that is what they want. Enlightenment is an excellent choice for a default desktop that doesn't have the bloated GNOME garbage. And to the leaders of Debian - after 20 years of my loyalty and evangelism, you have disappointed me and everyone. They need to have a fucking couple of eggs and do the right thing in the here and now.
Send those opportunists to shit!
GNOME. The Linux Foundation. freedesktop.org, [Figueredo] and others. They are all [a sleeve of thieves from first to last] groups with agendas. They are not those who believe in freedom. They believe in control and standardization. They believe in monotony. Who are these people? Who are these self-assigned guardians of the Linux flame? (intentional and subliminal reference to malware) What are your real agendas? Who finances these people? Why do they want so aggressively to shift the heart of Linux away from its true philosophy? Let them go and create their own "competitor" to Windows and Mac. If they did, it would be the same opaque, user-tracking, backdoor shit that Windows and Mac became. They do NOT speak for me and you should not passively allow them to speak for you either.
systemd is a Trojan. systemd is a jellyfish. systemd is Substance D. systemd is creepy - not just because its tools suck, or because it's a bloody massive hairball - but because architecturally it has a lot of concentrated power. We need to collectively kick him out of the way because he will own Linux, and by extension us and our freedoms. systemd will Linux. Sit idly and ignore this fact of our collective danger.
http://static1.gamespot.com/uploads/ignore_jpg_scale_medium/986/9866256/2540173-0237640794-Micha.gif
A while ago, the first commentator who won an all-expenses-paid trip came out. While I'm around here I'm going to be hitting the banhammer hard. Insulting comment and / or without arguments goes straight to hell, even if they make me angry that for censorship and such. It is well that the flamewar entertains but «everything with measure» ...
hahaha xD, I saw it: v, his argument was not bad at all, but he went to the grass when he generalized xD
Well, elav just resuscitated him and answered him.
HAHAHAJAJAJAJAJA, it had to be, look if he will love wasting time in pointless discussions. Well, if that's what he wants, I wash my hands and give the banhammer to him, from now on I prepare my popcorn too and I just watch ... http://i.imgur.com/YcJ0XpY.jpg
There is nothing wrong with that comment, it is a simple opinion regarding I don't know what ass ... XDD
That @tete is all crazy 😀
By the way I already have my system free from Consolekit, Policykit, and with automount without problems 🙂
good! Gentoo xD's versatility is impressive
Umm I don't know about your particular case, but my systemd gave me a lot of problems, especially with PAM, I would wait a little longer to change, for the moment, as I also use BSDs, I prefer to use OpenRC.
@roader, yuki uses OpenRC without Consolekit or Policykit
@roader, I don't like systemd, I stick with OpenRC because it's all I need, as well as being easy to handle. The * Kits I deleted from my system, and the automount I do it using IgnorantGuru's udevil and it goes perfect. About PAM, this can be eliminated, in Gentoo you have that capacity, I don't know how it will be in BSD.
I had read it, and in general I agree with what it says. Although such an email in the kernel list ... it would seem that it is just looking to do a flamewar. But as I said before, it does not mean that it mentions several valid points, but I do not think it is the place (in my opinion).
It is worth saying that three days ago I made a tweet that said: «gnome, pulseaudio and systemd have something in common: the unhappy way in which they relate to (t) the #Linux community: - /» https://twitter.com/gorlok/status/500413760595849216
It was in reference to the constant conflicts of these projects, in particular their developers / leaders how they have dealt with the community, with continuous controversies. It is clear that it is impossible to make everyone happy, all the time, and at the same time meet the goals. There is no formula for a project to leave everyone happy singing the Kumbaya, but from there to the continuous level of conflicts and constant clashes that exist in those 3 example projects, there is a stretch.
In particular with systemd, despite the conflicts and its design philosophy, it works, and solves several real and pressing problems that needed to be solved. Is the best solution? I doubt it. That is why it has gained space despite the conflicts. There is no doubt that there is room for other projects to take their place. That is why it is important to minimize the dependencies with systemd (or pulseaudio for another example), and in that I agree with the user, so we facilitate that path that will undoubtedly be traveled eventually. The other realistic alternative to systemd was upstart ... which was also not without question or conflict, almost more of the same.
On the other hand ... these conflicts are not so serious, they are more the norm than the exception. There have always been and always will be. It is worth mentioning many historical cases with GCC, for example, that generated alternative projects by testing other paths, which in the end ended up enriching the community.
I bet eventually another project will take over here as well, unless systemd eventually works through your questions and solves them. For now ... it is what it is, and I think many understand it that way, while preferring to devote their efforts to other more urgent issues.
The point is that there are many questions and problems about systemd that Leannart simply cares little about solving even though they are compelling at times, or are simply hidden by making bad modifications to systemd-related software (Kay case). The forum had already started to talk about the issue and I put on the table some things that I have had to happen with systemd and bugs that are simply not looking for a solution (for now).
link: http://foro.desdelinux.net/viewtopic.php?pid=21896#p21896
Reminds me of OpenSSL ...
The case is more or less similar at least in the background. It is something like: «Let's do many lines of code, don't comment on them, don't document them, anyway nobody checks the code well, and if an error appears, well we'll see what we do with it»
There is nothing that suits Linux users, if each distro does what it is sung about, they discuss what is wrong, that fragmentation and that the sea by car, and if they unify it is that they take away freedoms, if they do not like it, they feel the ass and program ...
I did not expect another response from a Windosero? (I don't even know if that term is well used). Windows and Mac users are used to using what they say, YES or YES. This has never happened with GNU / Linux, that's why all this anger, and believe me it is not due to lack of programming, or of asses.
Well… http://boycottsystemd.org/.
I agree in many aspects of what has been said here.
I miss with nostalgia those Arch days when I did NOT have systemd, when rc.conf existed and so on, because 90% of my system configuration was specified in that file.
In other distros (like Debian) it was not so, it was not all concentrated in a single file, but I did not care, init was pleasant to me, I understood it thoroughly, I knew how it worked and how to make it work as I wanted, the runlevels , rcconf, etc.
Now with systemd ... let's say the first thing is that it's different. journal, systemctl, and other tools, yes, they allow you to do the same but I keep asking myself, was this change really necessary? … U_U… I know my answer, you will know yours.
I don't regret using systemd on Arch now (well, it's not that I can or won't regret it ... or use it, or I forget Arch), but I do wish it had been otherwise.
Excellent article. I do not share the extreme »crusade» against SystemD, but I do agree that the simple fact that systemd is «one ring to rule them all» is not the best option, it should not be generalized and that now all distros adopt this philosophy ... well, those who have seen The Lord of the Ring already know how that phrase ends 😉
You can always use another initmanager, it is not that difficult, the only thing is that the init scripts you must download (or write) yourself.
Strange rc.conf, I don't understand systemd, it's the reason why I left Arch, don't get used to it, it's a real riot
Same and interested
http://scriptogr.am/x_fausto/post/arch-linux-sin-systemd
Thank you!
regards
Well, without the intention of offending, it seems to me a very poor article. There are very few technical questions, and the author seems to place great importance on feeling special. In my opinion, the standardization of basic parts of the system does not seem bad to me.
correction: It seems to me a rather poor letter.
At the beginning of the article it is specified that it is not an article by the author but a free translation of a letter made to the Kernel list ... it seems bad that it is commented just because.
This battle will be long, the question is whether this will result in a single init for all distros (a Ring to control them all) or someone else will have the enlightenment to make a new init with technical and design improvements with respect to the current init , that's something that remains to be seen.
Personally, I personally used systemd for a couple of years (Archlinux and Debian Testing, and I have client machines with that init) and I must say that on many occasions it was a blessing and on others the very face of the devil trolling you as much as possible . Strange bugs and even some quite dangerous, dependencies and loops in services, in short, things that are accompanied by a software that is not yet properly stable and which in turn is accompanied by a programming complexity quite unusual.
For the development section, I was reading the systemd code, to see how documented and commented it was and in my life I have seen something more austere, except perhaps, for the Hello World that they send you to do in your first class of programming.
Well, as a flamewar, it's pretty cool. But regarding the article I will explain my point of view:
Init is obsolete. Why? All the tasks you have to perform in a single thread (I think that everyone likes things to go faster and use those multicore processors that we have in our computers). It does not allow dependencies between processes so that they do not start if others are not (unless we search with the ps command, but this seems like a joke). To users so angry that they are not developers or maintainers of distribution packages, tell them that if we are maintainers and developers, these alternatives at the beginning (upstart, systemd, ...) give us life to be able to make things work easier, especially in distributions like Debian that when you install a package everything is already working (you install Apache and from that moment it is already started and it always will). Things with init can become easier, but there comes a point where it does not offer you a solution (a pick is very good and it is very easy, but I don't see many workers using a pick to dig trenches in the asphalt) . Thus, it is necessary to use a more modern system than Init. Which to choose ? It was already being debated if SystemD, if upstart, if the Init alternative (now I don't remember what some init was called with vr or hr), I have used upstart a lot to say that it is very green for the needs of developers and what I have cursed a thousand and one times. With upstart I have not had the pleasure of being able to touch it so much, but if so many distributions (Arch, Fedora, Suse, Debian, Ubuntu ...) have opted for it, it will be something, right? From what I've seen it may be a bit more complicated to use (it doesn't seem like a lot of complicated either), but it's not something that ordinary Linux users should be concerned about. And if you are a developer / maintainer / sysadmin, well, hey, in computing you always have to be recycling and learning new things; What is good yesterday, today no longer meets the needs and you have to change to improve. Regarding whether it is closed or not, if the SystemD developers are lazy and don't want to fix things, there are two options. or create another startup system that does the things that are needed or learn how systemd works and assist with development.
And if you don't like Debian's decision, you can always build your own distro and do it however you want.
sir, I take my hat off to your comment
Or install openrc (or any other manager) the init scripts you need and modify the init = option in grub. As a fan of systemd that. For reasons of practicality, use openrc instead of this I have to say that, although it is not so developed, it has nothing to envy, it has support for parallelization, it is more compatible. I just want you to change the old consolekit to logind.
Openrc, that was the alternative that was shuffled when changing Init and that I did not remember. Let's see, much of the debate that was generated came from the fact that due to the dependency systemd had a dependency (I think I remember that with the kernel) it produced that it could not be ported to other architectures, so that it would be less free. Even so, I don't think upstart would be a very valid alternative either, and openrc I haven't seen it at all and I can't comment.
Without being inside it is very easy to say what to do and what not to do, on paper it is very nice, but when you are there and the new KDE, GNOME, Wayland and other things that are increasingly related / dependent on systemd arrive. to do?
@roader, it is possible to use OpenRC without having Consolekit or Policykit installed 🙂
@raul, because OpenRC is a very simple init, the amount of code in it is less than 40 thousand lines, compared to the more than 400 thousand lines of systemd (and increasing), while in turn it has fewer dependencies and is more modular even than systemd. OpenRC has parallelization, it is portable to other * NIXs, it is capable of limiting resources by services, it has integration with CGroups and SELinux (these two features are something like the systemd flags).
I use it in Gentoo, and without using parallelization it boots my system in 22,3 seconds while systemd did it in Debian in just over 18 seconds (I don't remember the milliseconds hehe) using the same services.
@Yukiteru, the problem that I see is that with openrc you would have to cook all the startup scripts, and it is a bummer to have to do it, especially if you have services that depend on each other (for example glusterfs has to start after the network boots, but not much later than the filesystem), that's why I'm not going to make a titanic effort to rewrite all the startup scripts for another system. If the one that triumphs and the systemD community decides, well we will accept it, which is Upstart, then go ahead with it.
@Staff, if a project decides to take a path I cannot do anything other than follow it or take another alternative. My Gnome didn't like it when I switched to Gnome3 and I switched to KDE. There were people who didn't like Gnome3 either and decided to create Mate. I have not read the SystemD license, but because of how little I know LGPL 2.1 and the only thing is that it seems to use something that has a proprietary license, therefore I do not think (I speak a little with ignorance of the complexity) that it is too much difficult to fork if the time comes that everything depends on SystemD. I am sure that you will not be comfortable with everything you use in your desktop environment, but if there is an alternative you can replace it, and if not, I suppose you will accept it because there is nothing better, right?
@raul With OpenRC do you have to rewrite startup scripts? Well at least in Gentoo I have not had to do it even to use parallelization, and if necessary it has to be used in another distro, the same is the case as in Debian or any other distro that is migrating to SystemD, you have to rewrite many of the scripts to adapt them to the PATHs among other details that come with that distro in question.
Well, I think systemd is wonderful, I have tested it on distros like archlinux and it works like a bullet, and although it is implemented in debian testing there are still certain old services that have scripts in /etc/init.dy that have to be started by compatibility like "LSB", so at a certain time debian has already ported everything to systemd. The issue is not only the speed it offers, but the service management paradigm is very well structured logically and in an elegant way. In fact, looking at / usr / lib / systemd / system and / etc / systemd / system you can make a mental tree of services during startup, I have actually written on physical paper a tree extracted from the content of the files of systemd and it is clear what happens. For example, systemd by default launches default.target, which is a link to graphical.target (even if there is no graphical environment installed), and you can modify default.target in 3 ways.
1) Specify in the kernel line a value for the default target of systemd
2) In / usr / lib / systemd / system modify the default.target link
3) systemctl enable multi-user.target || systemctl enable graphical.target or whatever you want by default. Since these targets have the name default.target in the install section, installing them is equivalent to creating a link to the default target in /etc/systemd/system/default.target.
In addition, it is seen that session managers like slim, for example under archlinux, it can be seen that they are installed with the alias session-manager.service (or a similar name), and graphical.target has wants (optional dependency) to session-manager.service and as a strict dependency on multi-user.target. In short, it is very well built ...
Greetings!
And what does he mean after SystemD comes to Debian? At least that's what I think, I think he must have posted this before that happened.
On the other hand SystemD is better in all areas (except (apparently) in what is mentioned above) than the rest of the alternatives, after all it is a point of view what it exposes there.
Not that it would be so complicated to switch to OpenRC. Of course you would have to develop a compatibility layer (or simply adopt) logind.
Still with this? People don't seem to know that the init wars have already ended with systemd as the victor, because there are simply no more competitors.
What were / are the options?
-Wait
Definitely not.
-Upstart, OpenRC….
No, none of the existing options has the technical advantage or the community and business support that systemd has, we believe that the creator of Upstart (the strongest adversary) recognizes that and applauded Debian's decision to migrate to systemd.
-Create something new.
This does not need to be explained.
So if we have a winner who gives us free software without abusive CLA, then let's take advantage of it, if it is "a ring to control them all" then we have to get involved and find a way that we can all use the ring at the same time.
Regarding the ring thing, why is nobody complaining about the kernel? it's also a single kernel to dominate them all, and no one defends Hurd (and with good reason).
and I take the daring, now with all this "despiole" (attention, I am going to use a wrong generalization to emphasize my opinion) everyone puts on the BSD shirt ... suddenly they are all BSD users, and they advocate that it is available all software for BSD, AmigaOS, BeOS, Solaris etc ... let's go guys ... if you are so concerned about those systems, send patches to make systemd compatible ... or failing that, patch the applications that depend on systemd to be compatible with your YOU..
Exactly, there is a terrible double talk.
If that kernel is used, it is because there are no options with the same level of maturity, but it is not a problem, it has a free license, we can use it, get involved so that it works AS WE WANT and when the time comes to change it (What will come ) choose an option (of those that are already being developed) that continues to give us the same freedom.
@ x11tete11x
In * BSD it is not so easy, due to the way it handles projects.
GUN / Linux is by nature the paradise of workarounds, patches and more patches, many improvised, poorly documented and that in the end require more patches, that require patches, that require patches ...
* And beware * I'm not saying that one of the ways is better than the other, they are simply different.
As I responded to raul, there is a double standard (and I add apathy) in all this, which also affects what you mention.
Most of those applaud * BDS are the same ones who attack the ideological position of Free Software on things like proprietary drivers and fill their mouths saying they prefer Open Source, without seeing that for example, OpenBSD and its main spokespersons are, at least as strict as Stallman himself and less tractable.
So, those people in the end do not contribute to * BSD or GNU, as they say in baseball slang: Neither pichan, nor catch, nor do they let bat.
The article is fine, only instead of explaining why systemd is a cancer (which it is) it rants without giving explanations of why systemd is bad.
The bad thing about systemd is the small group (well paid) that runs its development, they don't give a quarter of a ball to people, it alters transparent mechanisms like syslog-ng with journald ... everything gets murky under systemd.
Unifying is fine, but for the entire community to do it, not a well-paid rubber dick for business and imperial interests.
Luckily gentoo and funtoo have said no, openrc will not be changed as the default system, we are also lucky that gentoo forked udev and eudev was created to remove syphilis from udev / systemd.
$eix -Ic eudev
[I] sys-fs / eudev (1.9-r2@27/07/14): Linux dynamic and persistent device naming support (aka userspace devfs)
$eix -Ic openrc
[I] sys-apps / openrc (0.12.4@20/04/14): OpenRC manages the services, startup and shutdown of a host
It is not enough to appear to be good, you have to prove it and that is where systemd fails, what is missing is that they fork systemd, and change its leaders.
Because it is a translation of a post to the Kernel mailing list, not a personal opinion of the author of the note.
Lennart knows how to torpedo eudev. Plan to move udev to kdbus
http://lists.freedesktop.org/archives/systemd-devel/2014-May/019657.html
I wonder why torpedo a fork project like eudev? Is Lennart not interested at all that others can create viable alternatives to the software that comes from Red Hat?
Of course this does not mean that it is impossible, but it will certainly make things more difficult and projects like eudev may collapse because they simply cannot keep up with udev / systemd / kdbus and the amount of resources that are dedicated to it.
I do not know but at the rate we are going at some point we will have the following line in our releases: «Based in Red Hat Linux OS. All rights reserved © 20XX »
I don't know how the boys are going to solve with eudev, but KDbus is a project that I bank, it is insanely ridiculous to have to change context 2 or 3 times to communicate processes via the current Dbus, it's crazy ...
if it is not under systemd, it will be under one of the init, but KDbus should not be something taken lightly ...
Certainly the point is that Lennart will do whatever he wants with that code, that's where the point is, Lennart and company already have many pissed off with their excesses, and even Linus has sometimes thrown his one or another stone for his way of doing things.
Definitely there the mask fell off, it never ceases to amaze me the power that money has.
Since it's going to affect gentoo people if this drone tries to put kdbus inside the core, he's going to have to convince Linus to put that rot inside the core first.
Besides in gentoo there are developers that can carry the changes from udev and kdbus to eudev, not difficult at all considering the size of the udev / systemd code.
There is an urgent need to fork systemd and let the entire community handle the decisions of how things are done.
What annoys me the most about systemd are the dependencies of one service on another, that control must be manual, if you lose control of which daemon is active or inactive, then you lose control of your system ... at that step they will propose to remove root and install an unalterable image on the disk… ..all running the same, a massive way in case of "an inadvertent bug" to be able to have control of all the linux pcs.
give me a reason why not implementing kdbus is beneficial (and I am not saying bring all systemd, only KDbus), it is insanely ridiculous to have to switch context 2 or 3 times to communicate processes via the current Dbus, it's crazy ...
I am not saying that it is bad to implement a desktop bus in the kernel space, what I am saying is that your kdbus implementation (Lennart's) will not be reliable because you will change it to your liking to screw up the rest of the world. of the proyects.
I have already read some Linus opinions on systemd and I don't think I am so foolish to give in.
Who allowed Lennart to stick his nose in the system logs? affecting the syslog-ng package with its journald in binary format, a simple # cat / var / log / message | grep -i something
It allowed from anywhere to see the filtered log without using anything that is not in the base system, it turns out that journald uses a binary format and only with the jpurnald executable you can see the content and filtering it with its parameters is not exactly easy.
How can you use systemd and syslog-ng without installing journald? ... that's why it is a cancer, it spreads through all parts of the system, covering, blocking and canceling other packets.
Binary log, sounds like Windows to me.
@emiliano, and do not miss the detail of having to restart the PC in case systemd and several services update at the same time, in those cases not even doing a systemctl daemon-reload will save you from doing a complete reboot to your system if you want it to work as it is. Windows? Where?
I have always used .deb distributions because for me they have always been easier to use than those of the .rpm family.
That ease I have always attributed to upstart. I was hurt the day I learned that Debian opted for systemD
Honestly the source code will still be there, make your own distro. I didn't see any explanation of what was wrong with systemd apart from off-center comments about something technical, a sea of insulting words with practically no explanation, ok apart from the lack of transparency (resistance to change), you can still use BSD, also this haiku os ... and there are more free operating systems in case you haven't noticed 🙂 I recommend haiku os. I would have liked to know what the satanic thing is about the whole thing but I smoked the whole comment and I did not see anything relevant apart from a somewhat rigged philosophical position (common in the debianeros) I guess you will not be the only one to disagree with this, forkeen should have and they make it perfect. Greetings, I will start to worry when the source code is not accessible, for now neeee.
At the top of the article, it is clearly written that it is a letter from a Debianite on the lkml.org list, but not a criticism from the author (@diazepan).
All init "problems" are related to desktop systems, where GNU / Linux does not exist. In the field of servers, where GNU / Linux is # 1, init is the best. If you see a user kicking against SystemD, it sure is because it manages many servers.
Greetings.
Hadn't Linus Torvalds kicked out the developer?
http://gutl.jovenclub.cu/linus-torvalds-desconecta-a-uno-de-los-programadores-del-kernel-de-linux
I use linux but I have never dedicated myself to knowing it internally so these discussions of upstart systemd and others sound like Chinese to me. So I ask something why they say that systemd is a cancer?
how much could affect the development of the linux kernel. Could they become inseparable things in a distribution? thanks 🙂
He was thrown out for being the one who generated bugs without knowing how to solve them. Hence the resentment.
And if that wasn't enough, here are a couple of details why Linux is gradually moving away from the UNIX philosophy:
1.- The FSF is against UNIX. Therefore, since the kernel was created, the FSF has been the one that has influenced its distance between the Unixeros.
2.- UNIX on servers is already seen as Windows XP on desktop PCs. Hence, why supercomputers are -for the most part- installed with a GNU / Linux distro.
If it weren't for those two points, Christopher Barry would have been accepted by Linux users in general. But since we have forgotten that UNIX no longer has the same relevance as before (although it must be admitted that without it, there would have been neither the internet nor computing as we know it), and that the business sector depends completely on Linux ( take a look at which distro the New York Stock Exchange trades with).
That was Kay Sievers, and Linus forbade him to add code to the kernel because it gave more problems than it solved, in addition to wanting to shut up the kernel debug to hide the flaws in the code he writes for systemd. The original report is over here:
https://bugs.freedesktop.org/show_bug.cgi?id=76935
In the end, it would be like using Windows ...
And speaking of Rome ... they just released new versions of udev-init-scripts and openrc
https://packages.gentoo.org/package/sys-fs/udev-init-scripts
https://packages.gentoo.org/package/sys-apps/openrc
I remember when I change from udev to eudev, all a risk because you have to first remove udev and then install eudev, meanwhile you are working in the air, if something goes wrong you have no other choice but to chroot hehe.
Hahahaha if you walk on the edge of the cliff, the good thing is that it has a chroot solution and nothing else 🙂
Theo de Raadt would be wallowing with laughter at this flamewar (and to think that LibreSSL was born out of the problem of maintaining OpenSSL and its ineffectiveness of the software itself in its later years).
The same would happen if SystemD would have the interest of Theo and the OpenBSD hooligans (they would call it SystemX alluding to UNIX and family, in addition they would make it as compatible as possible, and -like LibreSSL- with a web page with Comic Sans typography).
Anyway, it seems that Slackware and Gentoo would gain more followers thanks to flamewar (seriously, Slackware is the most attached to UNIX and the KISS philosophy itself).
A little fanatic all this sounds to me
I would tell you that your time in debian is over, if you want to keep your UNIX habits, you will feel very comfortable with OpenRC on Gentoo - you can install KDE in peace and you won't get Gnome hives. Systemd and Gnome go hand in hand, it would be a headache to patch each package, as I wanted ubuntu with upstart. In debian popcon keeps saying that Gnome is used, resources are limited, and they are in the second beta of the release.
Can someone explain to me what is systemd and what is the problem with it (although I imagine where the shots go)? It is that although I use Linux I do not know very well its internal workings.
systemd is a program that is responsible for starting your operating system (init) and all the services that you will use on your computer, from mounting the hard disk and usb, to the daemons that will allow you to use the internet, as well as shutting down the system and the track of user sessions, among many other things.
No way, to find Hardware that supports FreeBSD well 😛
That is false eliotime3000
1. The FSF doesn't hate UNIX. They are only interested in promoting free software. If AIX or HP-UX were free they would be recommended by the foundation. Richard Stallman has said that he chose to create a UNIX-like system because of its great portability and cleanliness, although he had some general criticisms.
And the actions of the FSF have nothing to do with the separation of UNIX and GNU / Linux, but rather this derives from their inherited or acquired development model. While the first is more centralized, stable and tested, the second incorporates very fast functionalities and features, without performing many rigorous tests.
In gentoo you can have linux kernel or freebsd kernel, that is the main reason that gentoo has had for rejecting systemd, as you will know, systemd is tied to the cgroups that only the linux kernel has and not unix kernels like freebsd.
$eix -c freebsd-sources
[N] sys-freebsd / freebsd-sources (-): FreeBSD kernel sources
Today's hardware is fully compatible in 95% of the cases, some other problem with unusual asics chips from netbooks or notebooks, but in general the rest is not problematic, some bios out there with the ACPI tables not very well finished , but nothing that can't be fixed by compiling the kernel or passing a kernel parameter in / etc / default / grub
Do the words are gone with the wind
They are only words without arguments, concrete data, comparative ...
is a philosophical speech
I would like to find documented reasoning in all that speech, but I can't find anything
"Those who think Linux should worry about idiots and HOYGANs should go back to their Macs and Windows, and stop trying to" fix "Linux."
This is the clearest example of why the vast majority of Windows users perceive us as a bunch of geeks who feel like the navel of the world ... for my part, Barry can roll up his little letter and put it there, in that part of the body where the sun never shines.
Well, my systemd does not displease me, it is no longer as unstable as it was at first, and the truth is I do not care if it is openrc, upstart systemd, as long as my pc starts quickly.
Tina, the majority of users are those who have never entered a forum to watch discussions, who do not even know about the existence of Linux and if they do, they know that there is something strange and brown called Ubuntu. Therefore these discussions, the average user does not reach or find out, in fact most still do not know how to use Windows well.
Pandev… you are absolutely right in the world; the vast majority of Windows users do not read or participate in such discussions, and that is precisely why they have a distorted image of the GNU / Linux user. And with such stupid ideas, like the one Barry wrote, they are even less interested. Who wants to be treated like "idiot and Hoygan"? Who does Christopher Barry think himself to refer to others in that way? So GNU / Linux was, is and always will be a world only suitable for "enlightened ones" who understand the technical aspects and the famous "freedoms"? The simple ordinary users who want GNU / Linux to generate more friendly, easy-to-use distros and we are very little or not interested in the programming language and Stallman's doctrine, do we not count because according to «Mr. Barry »are we idiots?
tina, pandev. Notice that the link that I put in the part of the hoygans I put it in a bad way.
Hello Diazepan. Thanks for the clarification, but I had already noticed. In fact, before making my comment I read the original letter to be clear about what Barry is trying to say and it was very true that you criticize the background -not the form- with those ironic additions, and very accurate indeed.
It is the same context of "bad milk" that I use to refer to what "Mr." Christopher Barry expressed as inadvertently wanting.
I understand, like you Diazepan, what Barry said between the lines and I totally agree with your criticism.
I think you have to place yourself in the context, who was the letter sent to? yes, from this Berry to LKML.ORG, that is, he is possibly an advanced programmer user who addresses other advanced developers ... not normal users.
GNU / Linux has arrived where it is thanks to the programmers, it is still the programmers who improve and advance all this.
We are facing an operating system and programs made by people and not by companies, in a company the manager says this is done that way because I say and for that we pay them, full stop, that way the user will never compromise with the creator of the program or system.
Now see what happens in GNU / Linux, users without knowing much get into conversations with programmers….
I do not want to raise controversy with this comment, I am not a programmer either, just to say that you have to place yourself in the context of how it happened.
Quoting the release in question:
By marking Windows users as idiots and retards is more than enough to consider Barry a idiot and HOYGAN (There are many who know how to get the most out of the Adobe suite, Autodesk programs {such as video game studios, including Valve and Bandai Namco}, modelers using SolidWorks with their 3D printers, and other technicians).
Without a doubt, this is one more reason not to consider the opinion of this fanboy that he would be better off working with Theo de Raadt on the development of LibreSSL and OpenBSD.
The systemd thing has become a fight of egos, which bigger believing gods. The idea of systemd was not bad at first but it seems to me something too big and complex for a Unix-like system, it still works half-baked (problems with the kernel debug for example). Although if the one who made the pulseaudio disaster is into the garlic (from time to time strange things happen to me with him), it does not need to say much more. When I started with linux upstart it didn't even exist, it was init and it worked.
Anonymous:
The context for this is an open letter addressed to the entire community:
Date Tue, 12 Aug 2014 15:38:12 -0400
From Christopher Barry
Subject OT: Open letter to the Linux World
"Now see what happens in GNU / Linux, users without knowing much get into conversations with programmers…." equals "Those who think Linux should cater to idiots and droolers should go back to their Macs and Windoze boxen, and stop trying to 'fix' Linux."
What do you suggest that I do then Anonymous? That I go back to my Mac or my Windows? It seems to me that part of the GNU / Linux community is made up of people who use GNU / Linux because it offers us an operating system that we like. That group, to which I belong, we are not programmers and, as is my case, we do not find Stallman's proposals at all practical. There are users, like me, who do not care and we are not interested in knowing if systemd violates freedoms - I suppose that Barry refers to those that Richard Stallman proclaims, although he does not say so - because we see in GNU / Linux a tool with a future ... although the way things are going it seems to me the future will be without the GNU.
@Tina - The letter was sent to the core email list, it is understood by its subject that it is addressed to all core developers, or at least that is what I understand.
Systemd is criticized because it started as a replacement system for init and it turns out that it is now manipulating the network and dns. As much as you say that you are not interested in knowing how it works, if it will affect you to continue as it goes ... this will surely go to the external remote control and consented by users who do not know a cucumber and think as if they did.
Security is the foundation of GNU / Linux, the security that the information is not accessed or manipulated, without the user noticing (in the best windows and mac style).
What can someone who knows something about security think by looking at the course systemd is taking? speed is a pretext for perverting security.
I'm tired of saying it, logind, consolekit, policykit, selinux, multiseat and now systemd are Trojans for the innate security that GNU / Linux always had, from outside companies and certain countries are paying very well the leaders of certain key projects to turn the security of GNU / Linux into something manipulable… .no one says it like that, I dare to say it to see if they find out once and for all.
Other attempts in that direction are in kde, kdelibs nepomuk strigi called semantic desktop.
I don't want my only trusted operating system to be ruined, but it turns out that systemd already has PID 1… and has messed with the network and system logs.
You can argue long and evenly for days, but look at the responses to that email and you will see that no one came to say that they had no reason to say what they said.
@Tina much of what Linux is today, it is thanks to the fact that the ideas of Stallman and many others have been closely followed, who advocate respecting the freedoms of software and the user to choose what they want, from according to their needs, and that has been the case not only at the user level but also at the institutional, scientific and business level. Giving up and losing these freedoms not only means being left without the ability to choose, but it would also transform Linux into something that had been fought against from the beginning. That projects like KDE, GNOME depend on systemd, will make things more difficult in this context, and due to the current trend it is possible that that will not be there at all, extending the dependency to systemd. Just to give an example, FreeBSD installs KDE by default, but if KDE depends on systemd to make it work on that system it will be a problem since the code will have to be patched to make it work, even so, Martin Gräßlin, I assure that it will keep the old KDE init for systems using X11 only, so this won't happen at least for now.
And much of what Linux is today -without GNU- has been achieved thanks to those who have not followed Stallman's rigid doctrine: the most used Linux OS in the world is Android and GNU has nothing, the most popular distros and that have come closest to the ordinary user like Ubuntu and its derivatives Linux Mint, Elementary and more recently Deepin by adapting them to our needs, it turns out that in the end they are no longer very GNU.
I have always said, here and elsewhere, that I have a lot of respect for those who believe and practice Stallman's philosophy, although I do not find it useful -at least for me- however it must also be recognized that there are many GNU / Linux users that we use their distros for reasons other than the freedoms that Don Richad promotes and that what matters least is to know the "guts" of operating systems because we are not technicians or system engineers. That group - large or small is the least of it - deserves respect, the fact that we do not know programming does not make us more or less intelligent and neither more nor less interested in improving ourselves as people. We know what we need from an operating system - be it GNU / Linux, Windows or MacOS - and we have every right in the world to present our ideas. What happens if my needs - which can also be the needs of many others - collide with Stalinian freedoms? The truth is that I don't agree much with that if we don't all get warm, then let's throw the stove into the river.
Let's agree that those users you describe do not care about systemd, upstart, openrc, etc etc 😛
That's right @ x11tete11x. When I turn on my PC, unlike Neo, I do not see the codes of the Matrix on my monitor, much less I know how to interpret them. I do not care if the OS that I use has SystemD or any other because for practical purposes what I want is that the OS works ... and that it works well. That SystemD will end up being ubiquitous and does that mean losing freedoms? Jeez, that's nonsense!
That SystemD is complicated, that it has bugs, and that its developers are so lazy that they do things by halves? Raul has said it so clearly that it is not necessary to shout it; Instead of complaining, let those who know about it, if they have time, help patch it up and correct it. Well, that's not what free software is about?
@Tub:
Repeating for the umpteenth time what I said in previous comments: learn from Theo de Raadt, for had the guts to fork OpenSSL when everyone else was just complaining.
I'm not saying that implementing SystemD is a bad idea. In fact, when I use Chromium / Chrome / Opera Blink Developer on Linux, it usually takes SysVinit on Debian forever to shut down the system completely, which normally does not happen with SystemD. Also, SystemD, it seems, would have its days numbered if it is forked (and to top it all, with the improvements and similarities with OpenRC, SysVinit and the like that many sysadmins wish they had).
My dear @ eliotime3000… you are absolutely right in the world. Although I don't see the reason why it would have to be forked. Wow, I understand why, but I don't see a justification for it.
@ eliotime3000 what you say about forking it I see it as complicated at least in BSD, in Linux it is another matter, since you can make a system that is capable of making compatible applications dependent on systemd with an init like OpenRC, by means of a layer intermediate between systemd-dependent applications and the system itself. Something like it was being planned with OpenRC-settingsd, although this project seems to have been stopped since 2013. The other question is: If you do something like that, will Lennart not do the same thing that he is doing now with eudev torpedoing the project by carrying to udev / kdbus and systemd to stronger kernel integration?
@Tina, the one that you turn on your computer and do not understand the matrix code that appears during boot, makes you an average Linux user, who are the vast majority of * users * of this system, those who come from the lands of Windows and OS X, looking for an alternative, testing or experimenting, and they saw something in Linux that just hooked them, whatever that reason, and that's respected. What is discussed and is the main source of friction between the pro and against systemd at least in my understanding, is the fact that this init becomes a kind of gravitational center from which nothing can escape, doing things that an init normally it shouldn't do, and becoming a hard dependency on everything Linux does (daemons, management software, DE). That in the end means one thing: It will make us dependent on an application that may have design and security problems (now or in the future), besides that it does not accept anything other than its own premises (Lennart is not very collaborative, sarcastic). There are many reasons to avoid that, and it does not take a lot of memory to find one, the case of OpenSSL is quite recent, and is a vivid example that depending on software itself is a bad idea (OpenSSL is the largest software in cryptography used in almost all distros) and gives us a clear example of what blindly relying on software can bring to the community and those who use the OS. And there are more reasons, but I leave it here for not extending myself.
regards
Now look where he's intruded on, yeah, with the VT core virtual terminals, it looks like he wants to replace them… ..someone has to stop this front man or things are going to go wrong.
FORK urgent please!
Systemd 216 Piles On More Features, Aims For New User-Space VT
http://www.phoronix.com/scan.php?page=news_item&px=MTc2Nzk
@ anonymous and it doesn't end there now you even have a DNS Cache. What does an init do with a built-in DNS service? I don't have the slightest idea actually, but the * monster * will even have DNSSEC and mDNS support. I just hope it doesn't get * plastic * with some other daemon like bind, unbound or pdnsd.
The systemd-terminal thing has to do with KMSCon (KMS Console), an effort to get the VTs out of the kernel-space and into user-space, but also to fix many other things, especially with Mesa, KMS, DRM, in addition to simplifying a little kernel development. The effort is good, but the implementation will be problematic after a while because the KMSCon dependencies are: table (libdrm, libgbm, libEGL, libGLESv2), udev (systemd), xproto (runtime dependency), libxkbcommon, freetype2 and glib (unicode support), in short, another lunge for init not compatible with systemd.
this is because systemd has long since stopped being an init only. init is just one of the services it brings. the other is a series of secondary daemons (like logind, and everything that ends in d) that are separate binaries and that focus on a single specific function. and in fact you can use systemd without most of these secondary daemons or have them disabled. what happens is that these daemons are, in many cases, only compatible with systemd and cannot be used outside of this (logind case since version 205) this is because they are designed to work only with it. this brings advantages and disadvantages.
and advantages
-systemd is highly modular, you can use it only with the services you need (in fact this is why systemd is mostly used in embedded systems)
- being designed to work only with one another. gives the possibility of better integration in the system. making systemd scripts practically universal (this was one of the main problems of sysvinit, scripts from one could not be ported to other distros, it was easier to create one from 0 (said by several debian developers)
-standardization: this must be explained.
- by limiting your options, you can keep a smaller and cleaner code base, thus improving security (because remember that each systemd service is a separate daemon in a separate binary) many of the most serious security problems are due to code old that is still there because of backwards compatibility or because you have to endure too many options.
disadvantages
-Lack of portability: since it is designed to work only with certain pieces of software and nothing else, if the requirements are not met, things do not work or require a lot of extra work to make it work.
lastly, being a critical dependency of all systems is not bad. we have depended on X11 for 20 years in linux for graphical interfaces and DEs. This has brought us universal security problems (the x11 problems in debian are the same as in arch for example) but the solution is not to create 20 different graphic servers, one for each distribution, that would make developing for linux a nightmare higher. It seems to me that the advantages of "imposing order in the house" are better. the linux world has been asking for years to end fragmentation and systemd is the closest piece of software that has brought this long-standing desire
Hello,
If you have really used systemd I don't see why so many complaints, I have been using Gentoo for years, which has given me the freedom to choose many things and accommodate them to my own needs.
I remember migrating my Gentoo with OpenRC and the boot speed improvement was amazing. Then SystemD appeared and after waiting a while for it to stabilize in the portage tree I decided to migrate and test it. I just kept it because it gives me what I need on my laptop, a system that boots quickly, yes, faster than OpenRC, without having made any hardware changes!
You cannot simply be a geek and want everything custom-made, we use computers because we need them and they always need to start quickly, simply and short the matter.
And well, this is linux, there will always be different opinions and alternatives for all tastes.
Beyond criticizing just for the love of being a geek and having everything tuned into our OS, better to explain solutions and alternatives ... help others who feel that the new does not accommodate them with howtos and others that contribute more than anything else in the world Linux!
Regards!
@CarlosMC
... .. and they always need to start quickly, simply and short the matter.
Look, I don't have anything mobile and my daily uptime is usually about 14 hours, with openrc it takes about 20 seconds to get to the slim graphical login, in sistemd it can do it in 12 or 14 seconds, we are talking about 8 seconds in about 14 hours. I don't need booting speed, I need security and services handled by hand as easy as it always was.
If systemd were to be forked, unfortunately many things would have to be totally discarded, such as using compressed binary formats ... really a shame to see so much useless code.
The more I read about SystemD the more afraid I am of it. Just in case I already did a 'sudo apt-get install sysvinit sysvinit-core sysvinit-utils'.
The message is harsh, but the background is totally true, many of the programmers who are helping to program the kernel are programmers paid by companies that would do anything to get more money, and one of the most aggressive companies in this regard is Red Hat that has its programmers placed there with the intention of exclusively defending their commercial interests, creating technological developments that make the use of free software incompatible in other Unix-Like systems such as BSD systems and others.
The problem is that to cause this incompatible use, they create technologies such as systemd that are totally invasive and harmful to the interests of the community, and with their extension and dissemination they will cause the system to end up not being Unix-Like, and stop doing just one thing and do it right, to do many things and do them wrong, which will result in a constant source of errors and absurd dependencies for no other reason than to avoid the proliferation of solutions by other companies, in short, another case like that of pulseaudio.
And furthermore, now with systemd after installing updates it will be necessary to restart the system, which has less and less to do with the Unix philosophy, and by making everything depend on systemd they prevent each distro from using other solutions such as openRC, Init, etc…
Let's hope Linus Torvald still has something more to say about systemd, a piece of software that is clearly not compatible with the Unix philosophy or free software.
http://www.phoronix.com/scan.php?page=news_item&px=MTY1MzA
and if this is successful, BSD systems will have to be used more and more.
It will be that linus does not care about the unix philosophy that is not religion. linus already said it. Linux is not about freedom or cheap philosophy, it is about providing effective solutions to real problems. systemd effectively solves many current real problems better than all current alternatives.
Finally. I will say that that that systemd is not unix-like is false. systemd is not a giant binary that does everything by itself. there are 2 systemd, the first is the binary for init, this is quite small. in fact it is smaller than upsart and sysvinit. the second is the project as such and all projects under the same umbrella. each of these are separate binaries (that's right, each systemd service is a separate binary from the rest, dedicated to handling only and exclusively said service and no other. This sounds to me like "doing one thing and doing it well") the thing that bothers many is that these individual binaries and daemons can only be used (some, because not all) without having systemd as PID 1. finally. Those other projects that systemd has "devoured" is because the creators of these themselves have decided to join together of their own free will. i doubt lennart put a gun to syslog or udev developers' heads. they themselves decided to join.
You said it with the exact words… .amen.
No need to worry, anyone can create a distro using openrc, nobody will be offended ... and for those who want their own, they already have gentoo / funtoo, I think sabayon and slackware also come in.
@mirage
disadvantages
-Lack of portability: being designed to work only with certain pieces of software and nothing else, if the requirements are not met, things do not work or require a lot of extra work to make it work.
Total failure, it can have all the advantages you want, but if in the end it is only compatible with itself, it is forcing everyone to use it or having to start from scratch for not wanting to use it ... if this is not imposition, what is it?
A script can be read by any human eye and is interpreted by bash ... of course it is slower to execute, a binary cannot be read by a human eye, you have to read the source code and compile it yourself to make sure ... that is, before you had to know how to read scripts in bashy now you have to know how to program in C language ... or entrust your soul to Lennart.
Regarding whether Lennart uses a gun, I don't think so, he's a RedHat figurehead to begin with and if he does use something, in any case it must be a few bundles of greenbacks ... they don't kill anyone, but they convince in the making of decisions. .
I am not a developer, I just stepped into an institute, but I understand what you say, open code and the freedom to learn from it, and modifying it is part of that "Freedom" philosophy that gnu / linux spreads and, unfortunately; framework, is today a major wound in the heart of GNU / LINUX. I suppose that, for those who have academic training, it should not matter much, since always, they will be able to do without the distros that take systemd and; build your own gnu / linux system, that is, make your gnu / linux system from scratch, without having to resort to systemd.
Archlinux, use systemd and believe me I am using it and it certainly works very well, but personally I will try to leave it to get to gentoo and from there to linux from scratch.
At least they started off with a good boycott, which will later surely morph into a FORK sovereign to systemd.
New Group Calls For Boycotting Systemd
http://www.phoronix.com/scan.php?page=news_item&px=MTc3OTM
In the comments someone passed two links where they describe in detail the design horrors (it is not known if the purpose) of systemd.
http://ewontfix.com/14/
http://ewontfix.com/15/
And where the hell is Richard Stallman and his opinions on these things?
Binary log, all controlled by systemd, not listening to the community ...
I'm neither for nor against systemd but I AM SURPRISED that Richard Stalman didn't say a word.
Brrrp!
I thought I would read a technical post in Spanish.
regards
Systemd is a very good idea flawlessly implemented. Version 217, already on Arch and in a few years on Debian - LOL! - contains security features that can only be applicable to a PID1, for example, complete isolation of services and processes, something that neither SELinux nor any other MAC could ever dream of doing with implementations other than systemd.
For my part I do not know anyone who uses systemd and wants to go back to the old system of startup and system configuration; In fact, every archero that I come across and to whom I comment on the subject, because we were all once nostalgic for /etc/rc.conf, agrees that systemd is simply impeccable from a system administration point of view and that no matter how much At one time you might have loved the old service configuration paradigm today. NEITHER CRAZY go back to that after using systemd.
People, stop breaking your balls, it's already a fact, it's here and if you don't like it, use other things like Devuan (hahahahahahahaha) or Gentoo, or Slack or even Manajaro that now supports OpenRC.
And then, while licking their wounds using anything anti-systemd, take a look at openSUSE, Fedora or Arch and see how they are hopelessly lost in time living an anachronistic reality, all for the simple itch to reject what apparently they do not know and judge according to the rest of the opinions instead of forming their own.
Whore, if even the core devs of FreeBSD are already talking about designing a system similar to systemd to face the new needs that are seen on the immediate horizon, including massive virtualization, multiplicity of connected devices, portability to mobile gadgets, security and simplified and optimized administration among others ...
Before you answer nonsense born of visceral sentimentality, read this short transcript of Lennart's recent talk about the new security features discussed above:
All the features are opt-in through the service units and are intended to isolate or compartmentalize access to system resources that daemons have and each other using kernel namespaces:
- PrivateTmp: assign a specific / tmp and / var / tmp for the daemon
- PrivateDevices: create a specific / dev for the daemon with everything necessary: null, random, urandom, etc., only what is necessary for it to work correctly. This way you prevent the daemon from having access to the disk or any connected device (block or character). AMAZING.
- PrivateNetwork: create an interface that is completely isolated from the real thing (so that you cannot access it from the host pointing to 127.0.0.1). The VERY great thing about this is that you can share the same namespace between different daemons so, for example, you can have a MariaDB connected only to the backend network that uses it and these two in turn to an Nginx that is the one that has access to the network interface.
- PrivateSystem: one step further, if you activate it, it allows the daemon to access only / etc and / usr (/ usr in R / O mode); if you set it as full, you can access / etc only in R / O mode
- PrivateHome: overlap an empty / home in the daemon's namespace so that it cannot access your information
- ReadOnlyDirectories: as its name indicates it makes those directories read-only
- InaccessibleDirectories: as is the name 🙂
- MountFlags = slave: when activating any of the previous options for a given daemon, the mounting arguments are automatically masked as slaves, that is, the general mounting table of the system is not altered, but the parameters exclusively affect the mounting performed by the devil (moi groso!).
You can also use this option stand-alone.
- CapabilityBoundingSet: instructs a process so that all threads that are created dependent on it use only the capabilities defined by the admin (http://man7.org/linux/man-pages/man7/capabilities.7.html) and thus avoid permission escalation vulnerabilities - sexy, huh !? ^ _ ^
https://drive.google.com/open?id=0Bw3eycS9_SPaQnFHZjU5TURxdmc&authuser=0
FreeBSD core developer making a projection for the next 10 years of the operating system and talking about the need to implement a solution similar to systemd:
https://www.youtube.com/watch?v=Mri66Uz6-8Y
* throw
I have been a faithful user of Debian for years and it saddens me to see the path it is taking. Only some developers are allowed to vote, but the end user has no choice but to accept that spawn of hell (systemd). I just hope that the devuan project goes ahead and receives a lot of support.
There are few distros faithful to SysV like Gentoo and pclinuxos and of course… slackware!
regards
I honestly think that for those who make money with Linux, this system is light years away from Windows and Mac and therefore all the systemd blowjobs, kits and cocks. I dare say that in less than 8 years, Nod, Avast, Kaspr and all their entourage will enter the Linux world, with them of course all their viruses. It is business and here is not going to get rid of neither god. Every day more user Linux? a cock come paca I need your money. That is so unfortunately. Now they are just putting the tip and with Vaseline .. wait a few years you'll see hahaha
I think what the author of the letter is saying is stupid. While there are several objections to systemd, whenever there were big changes, there were big objections.
Also, being a debian user (so am I) doesn't seem like a valid objection to me. If you don't like them deciding things for you, just use another "harder" distro and compile everything by hand, including the startup system.
You cannot pretend to have all the advantages of running an apt that solves all the dependencies, without any cons.
To a user of slack or gentoo, it would not have happened that, "oops ... I updated and there are things I don't like." Because to update a system by hand, you have to spend a lot of ass hours watching what you are doing.
I'm sorry if the developers of the distro you like did something that you don't like. I have an idea, why don't you do the same as Linus and build a separate OS?
Hello, I am quite new to linux and the truth is that this article seems scary and not really exempt. I who had joined Debian with all the illusion of what free software means (freedom, altruistic help to all who need it, free, simplicity) in short, that seemed to me a way of understanding fantastic and good computing ( and how I think computing should be from the beginning). In fact, I ended up on this website because I was starting to document Debian to learn how to handle myself and do things on my own, I am realizing that everything I want to target is disappearing.
We will have to create a new partition in Devuan, let's see how it ...
Thanks for this article contribution.
Greetings, Jorge.
Either way, time will prove whoever has it right.
What we as users can do is try several alternatives and hope which one ends up being better.
Regarding Devuan, I would love to give it a try, but it is currently in its test version. As soon as a stable version comes out to see if I give it a try, to see how it is.
How strange to read about such an unusual topic and agree on everything. I recently installed Debian Jessie, I already had to format 2 times for total system crashes in relatively easy updates. I wondered how nothing like this had ever happened to me and suddenly everything started to get worse. I was using wheezy and since the Debian jessie cd's downloaded the page (either recording to dvd or pendrive) they don't work for me, I install wheezy and update. As it takes time to download things is not something I do immediately so I use Wheezy for a few days and I noticed differences in terms of operation with Jessie that I had not noticed in previous updates. By chance reading another article I found out about systemd and I realized that it was not a false paranoia that I could notice with Jessie (the feeling of being in the instability of windows).
Thanks to your article and the others that I read who informed me about this topic, I decided to change the distribution.