How is FreeBSD coping with a systemd future?

Status
Not open for further replies.
tankist02 said:
PC BSD started Lumina: http://blog.pcbsd.org/2014/04/quick-lumina-desktop-faq/. Maybe this will help?

What I like about that FAQ is the answer to the question of why creating lumina:

Why create a new desktop environment? Whats wrong with KDE/GNOME/XFCE/<other>?
Answer: There are many reasons for needing a new desktop environment instead of using the existing ones, mainly because all the major existing DE’s are developed on/for Linux, not BSD. This causes all sorts of problems on BSD, and I am going to try and list a few of the big ones here:

(4-a) Porting time
Since the DE’s are written on/for linux, they have to be ported over to BSD, and this introduces a (sometimes significant) time-delay before updated versions are available (GNOME 3 anyone?).

(4-b) Porting quality
It takes quite a bit of time/effort to port a DE over to BSD, and I have to give lots of thanks to the people who volunteer their time and energy to make them available. The problem is that quite often “linuxisms” still bleed through the porting process and cause system instability, desktop/X crashes, and loss of usability on the part of the user. This is particularly true when you start looking at KDE/GNOME/XFCE because of the large number of individual pieces/applications/plugins that have to be checked during the porting process, and it gets quite difficult to check everything while doing the port.

(4-c) Linux development trends
As Linux trends continue to diverge from BSD through reliance on Linux kernel functions or Linux-specific systems/daemons, the porting process over to BSD is going to get even more difficult and take longer to accomplish. This means that if we want to have a reliable/stable desktop on BSD going forward, we have to have one designed specifically for the BSD’s.

(4-d) Linux dependency bloat.
If you look at current DE dependency lists, it is easy to see that when you install a desktop, you might be getting a lot more than you bargained for (such as additional compilers/programming languages, network libraries/daemons, audio/video daemons/applications, etc). While there might be some debate on this, my opinion is that it comes from the Linux distro mentality. Just as a Linux distribution is the Linux kernel + the distro’s favorite packages, the desktop environment is becoming the graphical interface for the system + all the favorite applications/libraries of the developers, whether or not they are actually necessary for satisfying the actual purpose of a desktop environment.
I feel like the approach on BSD is quite different because the OS is a complete entity, independent of the packages that get added later, and simply provides the framework for the user to do whatever they want with system. By this same approach, a desktop environment should simply provide the graphical framework/interface for the user to easily interact with the system, independent of what applications are actually installed on the system. Now, I understand that at this point in time a user expects that certain types of applications are expected to be available out-of-box (such as a file manager, audio/video player, pdf viewer, text editor, photo viewer, etc..), but is that really the realm of the DE to decide what the defaults are, or should it be left to the distributor of the OS? I think a point can be made that the file manager is considered essential to integrate with the DE appropriately, but I think that things like audio/video applications, text editors, pdf viewers and such are really up to the preferences of the distributor, not the DE. The DE just needs to provide a simple framework to setup those initial default applications for the distributor, not require a ton of additional applications by default. Because of this, I am taking the approach that Lumina will have a very limited number of applications included by default (there are only about 2–3 that I can think of, all written from scratch for Lumina), and will try to include basic user-level functionality within these few applications to try and cover 90% of standard user needs (at a basic level) without any additional dependencies. For example, the Lumina file manager will have basic audio/video playing and image viewing capabilities built-in because those types of abilities are available through the Qt framework without many/any additional dependencies.
 
Crivens said:
I think I see the problem we are having here in the thread.

  • Your concerns are valid
  • Our concerns are valid
  • We see your point, but do not seem to care
  • Both sides have different ways to deal with such problems

That the spread of systemd will be a problem for desktop usage is clear. BUT - many here do not use desktop environments, and thus don't care. Effort to gain ratio says you should go play with the kids instead. Most users here have an engineering background and, that is important, some experience in the field. They see systemd as one of the many many "oh look, shiny" projects out there. While Linux went from MAKEDEV to HAL to udev (now part of systemd), *BSD went to devd and was done with it. All these *Kit packages (I think of BeOS when I hear these mentioned) are also being replaced. And the same will be done with their replacement.

Standard engineering practice is to first check if a problem has to be solved. We do not have enough resources to hand this out to someone, there are more important things to do. That is the way engineers think, when they can think for themselves.

  • Can this be solved?
  • Can I solve it?
  • Why should I?
  • What would the benefits/contras be?

So before you think nobody understands what you want and run off - we do understand. We simply have a different mind set. We have some priorities, and we have some experience. I tried to point that out when I said that it might not be in the best interest of the desktop environment people to utilize systemd in any substantial way. It might save work now (short term gain) but is likely to cost them dear in the future when the makers of systemd do things to break stuff (been there, seen it).

You are welcome to work on the task, we will help you. But we will not gather around some banner for this. At least, I won't. Nothing personal, that's experience.

:beer

systemd May turn out to be the best thing that happened to FreeBSD in a long while.
 
I'm sorry, but obligatory XKCD is obligatory: http://xkcd.com/927/.

Unfortunately politics have taken over and forced systemd to almost all major Linux-based distributions, virtually overnight. Time will show that systemd is technically impossible in the land of open source where tiniest ego clashes result with forks and fragmentation and one-system-to-rule-them-all is impossible per those same open source thermo-ego-dynamics. I am actually surprised that so many distros got on board, but I think that's because of marketing and propaganda from Red Hat. I somehow think they all thought they were just voting on a new init system but got backdoored with so many replaced components. And now nobody dares to speak first.... But it'll blow. Soon, the egos will start to clash and decisions will start to get reverted, forks will be thrown around.

I declare 2015 the Year of the Linux Forks. Mark my words.

Oh, we were talking about FreeBSD... Well... *shrug*? It will continue to be FreeBSD, as awesome as ever. I think, I hope, I'm sure. :beergrin
 
One good guy in linuxquestionsdotorg once mentioned, systemd is a result of brilliant C programmer with no idea of the UNIX system administration happens, which I completely agree with.
I'm not sure if the various Linux distro maintainers didn't know that systemd is not going to be just an init system, Lennart had mentioned it clearly in the beginning that this is going to be a system management tool. We (Slackware guys) haven't still jumped the systemd bandwagon but I have been reading about the systemd architecture and I think that it's a great piece of coding and how the layers of system management has been designed around the core but the problem is, the tools it is replacing were not broken to start with.

Regards.
 
sulman said:
systemd May turn out to be the best thing that happened to FreeBSD in a long while.
Only if it stays outside.

systemd has the potential to break Linux apart in a truely epic way, I will not call 2015 the year of the forks (the prophecy is noted, @AzaShog, but IMHO the stars are aligning up even more perfect for Linux as you may think).

I predict that this movement they pin out in their slides will turn Linux into something of the Titanburg? Hindenic? You get the idea. I will try to get into the position they want it to be, the glue code to connect everything to all the rest - and then they have highjacked all the ecosystem. There will be fights, forks and arguments which will not help FOS or Linux in any way but which will be watched with glee (and some champagne) in Redmond and elsewhere. That is what I totally agree with @AzaShog, but I think that these problems will be stirred and fed quietly and efficiently by interested parties, if not now then soon. It will not stop at forks and arguments. And within one year or two, not only Linux on the desktop will be in shreds but all unix-like systems because all the applications will be in a complete state of fubar. Anyone fancy a three-year rollback to get out of that hole?

Right now I put the possibility of the above around 60%.
 
Last edited by a moderator:
Crivens said:
I will try to get into the position they want it to be, the glue code to connect everything to all the rest - and then they have highjacked all the ecosystem.

But that's the problem. systemd is not only hijacking, but providing its own versions of important system tools. It's not just a glue. If it were, then I'd have no problem with it. If it used, or patched ntpd for extra data instead of reinventing systemd-timesyncd, it would be a glue. If it did the same to mDNS responding functions, DNS resolving functions, DHCP daemons, ..., by patching for a central "hub" instead of reinventing and rewriting its own versions, I might not have a problem with it. But it goes into the hilarious part of the circus. I read it also implements a light httpd server and now implements its own terminal, vt and will act as a virtualization container/manager as well, I don't know, maybe by consuming or better yet, reinventing docker.

That's really far from being brilliant and good coding, as @CurlyTheStooge mentioned. It would be if it cleverly patched and coordinated existing tools thus leaving room for someone to truly use it as "building blocks", as the PDF linked few posts above wants us to believe, to patch in whatever daemon replacement they needed or wanted.

But, I have the freedom of choice and I've left the GNU/Linux world completely in favor of FreeBSD, both for servers some time ago and now even desktop, primarily for other reasons but systemd is good part of that. I'd have a problem with upstream limiting or removing that choice for me by hard depending on systemd in the future, and that's my only concern about systemd at the moment. Personally I don't think that will ever happen. GNOME is an exception, if it really does end up depending on it more than it's currently hinting at.
 
Last edited by a moderator:
AzaShog said:
But, I have the freedom of choice and I've left the GNU/Linux world completely in favor of FreeBSD, both for servers some time ago and now even desktop, primarily for other reasons but systemd is good part of that.

I am also in the process of migrating to FreeBSD. When the Debian initsystem debate began I completely moved to Gentoo. Now I am trying to get into FreeBSD to be prepared. Currently I am rewriting all my Ansible roles to support both Gentoo and FreeBSD. Unfortunately I don't have a choice at work where RHEL/CentOS is dominating...

To the FreeBSD folks: thank you for the clean base system, extensive documentation and good man-pages.

PS. I am also looking forward for the coming pkgng/ports (update-) improvements.
 
Just to ditto what @AzaShog and @Crivens have said. For the past year or more I've been saying Linux is no longer a Unix-like system and will become just "Linux" and incompatible with everything else to its own detriment.
 
Last edited by a moderator:
You will rather see that Linux won't be incompatible, but FreeBSD will be considered incompatible with Linux and not worth support (see also how nVidia drivers are developed for us). At the end FreeBSD will lack applications. I am on FreeBSD because I like the applications and how they are managed for me (big thanks to all in the FreeBSD ports team). But when applications go away, I will be forced to use Linux. There is no point in having a pretty system that is truly great (and undeniably better than Linux with their userland) , but on which you cannot do your basic work.

At the moment I am not dependent on KDE or Gnome, so this things can disappear without biting me (maybe not all of you agree here), but I still care that there is choice to have FreeBSD under the hood instead of Linux. For the developers of KDE/Gnome freedom of choice does not matter that much. Portable programming using well-know standards is not important, especially for developers who write on desktop software. I doubt that many reasonable developers who create daemons (services) are going to drop support for FreeBSD, but the future of desktop on FreeBSD is unsure at the moment.
 
nakal said:
You will rather see that Linux won't be incompatible, but FreeBSD will be considered incompatible with Linux and not worth support (see also how nVidia drivers are developed for us). At the end FreeBSD will lack applications. I am on FreeBSD because I like the applications and how they are managed for me (big thanks to all in the FreeBSD ports team). But when applications go away, I will be forced to use Linux. There is no point in having a pretty system that is truly great (and undeniably better than Linux with their userland) , but on which you cannot do your basic work.

^This.

People who are foreseeing a forfeiture/detriment of Linux (as a kernel) are obviously oblivious to the ground reality. A couple of Linux distros might lose some following or a small user base however that certainly will not going to be 'demise' of Linux and sudden rise of FreeBSD. How harsh it is, that seems to be the truth. If FreeBSD didn't gain momentum until now that's for a reason and it will never; it will always be a niche OS as it is. FreeBSD has already been declared incompatible by the Linux guys/hardware vendors on some parts (not in the literate terms). It's the community which is keeping its motors running, which is great.

I'm not the slightest of supporter of systemd as a system management tool, I like my text files as logs but that's not going to change the fact that the managements running the big data centers don't give that much freedom to system architects and administrators and will keep using what comes with a support contract. They don't care if the admins have to learn new things to manage the same system.

Regards.
 
Doesn't being dependent on systemd mean dropping support for Windows and Mac OSX too? I'm sure a lot of common software might not want that except maybe major desktop environments that are run only on Linux or the "irrelevant" BSDs. For example, I don't see the LAMP stack and its variants going systemd only, beyond using it as init (eg. having a unit file, and maybe dropping daemonization but that's easily patched for FreeBSD). Good portion of developers of LAMP based software are primarily Windows users. I also don't see programming languages dropping portability. Or libreoffice and that class of desktop apps. Graphics applications (gimp, inkscape, ...) are present on Windows as well. Browsers? Aren't the major ones developed on/for Windows primarily? But those are all daemon-less apps that I don't see affected by systemd anyway.

Unless someone upstream makes (g)libc integrated with systemd... which I seriously doubt will ever happen. Or if it does, I seriously doubt that'll make (g)libc-using applications unusable with FreeBSD's libc.

So really, what class of userland applications will benefit by and not lose by, becoming dependent of systemd? Major desktops environments?
 
AzaShog said:
Graphics applications (gimp, inkscape, ...) are present on Windows as well. Browsers? Aren't the major ones developed on/for Windows primarily? But those are all daemon-less apps that I don't see affected by systemd anyway.
...
So really, what class of userland applications will benefit by and not lose by, becoming dependent of systemd? Major desktops environments?
The browsers may find themselves in the position that they need to access systemd-networkd for network connections. They might decide that they need to hog your network settings like they do with cgroups, thus sitting in the position to only grant network access to those they like.

You never know. The developers of the browsers will not care, so this would make things interesting for Linux. The benefit of all this would be outside of Linux. Who would benefit from this balkanization of applications? Any ideas? Because it would not be Linux, in the end, and most other free operating systems before that.
 
Uh, no. systemd-networkd does not work that way, it's just a NetworkManager replacement on steroids (e.g. managing networking dependent devices and targets, if I'm not mistaken). I know you gave just an example, but that's exactly my point. I don't see any reason for applications to integrate with systemd, other than for desktop environments to achieve some kind of deeper integration (e.g. logind requirements). Programs might drop certain functions because systemd provides them instead (e.g. daemonization and/or process management, or virtualization container management), but I think that's easily replaceable under FreeBSD.

I think systemd is such a NIH replacement-o-rama, except of course the init functions and process control, that it can be easily shimmed and that's what OpenBSD's GSoC project is about.

As for hardware vendors dropping FreeBSD compatibility, that's an entirely another story. But that's kernel stuff unless Linus makes systemd a significant part of the kernel. I think hell will freeze over twice before that happens.
 
nakal said:
You will rather see that Linux won't be incompatible, but FreeBSD will be considered incompatible with Linux and not worth support
What I'm seeing now is what was brought up earlier, the fighting and fragmentation and complaining and splintering, etc. That is how I see as being detrimental to Linux. Already we have seen a number of Linux users talking of how they're tired of it all, here and many other forums, and switching to BSD where things seem more stable and level headed. Which leads me to another point.

The "Windows-fication" of Linux. Part of their issue is the influx of Windows users who want Linux to be a free version Windows. This has led to many of the fights between those looking for technical excellence versus "candy".

But I don't want to get off on that.
 
Even with Systemd I will probably move back to Linux if all of the basic productivity tools on FreeBSD vanish. Though that has not happened yet, and might not ever. The quality of the project was enough to convince me to switch to it. Granted I am not a normal use case, I switched because I like SCHED_ULE(4) and find it superior to cfs/bfs from Linux.
 
systemd is irrelevant to FreeBSD as far as I'm concerned. If you've been paying attention you'll have noticed systemd isn't about just improving init. It's about absorbing everything it possibly can and turning itself into a monolithic userspace kernel (MUK? Pokemon anyone?) which controls everything on the system. You will no longer have utilities which communicate with the kernel, you'll have systemd specific tools which talk dbus to systemd to do those same tasks.

The goal is clear, the systemd developers want it to be the gatekeeper of userspace. The problem with it is that it's nothing more than an undocumented, standardless, hackjob with the backing of one of the biggest players in the Linux world (RedHat) which will only benefit them, with the added long-term goal of creating vendor lock-in. How long until they decide it's going to handle package management and require RPM? How long until they start signing services and packages and locking you out from using your own unless it's signed by them? How long until they start implementing ways of preventing Oracle from basing their Linux distribution off it, and everybody else for that matter? How long until the Linux kernel itself becomes so hardwired to it that the two will be indistinguishable and we'll just have RedHat systemd/Linux and that's it, for which the motto will be "you can have any Linux you want as long as it's systemd." The possibilities they now have is scary to think about and what they'll end up doing with it, especially once more and more distributions give up trying to fight it, effectively giving RedHat even more control.

It might sound far fetched, especially 10 years ago when there was so many Linux distros and so many different people maintaining their own projects independently to provide the various pieces of userspace which systemd has now hijacked, that trying any of that would have been suicide. Now that they've gained so much influence and control over the entire ecosystem is it worth trusting them; especially after you read the lies, misinformation, and straight up bullshit on the systemd mailing lists from the lead developers? Sure it's GPL and you have the source to do what you want with, but when every piece of it is intertwined and a confusing mess of shit code with a nefarious motive the barrier to rip out the pieces you don't want, or use only the small pieces you do want elsewhere, or manage your own fork will negate the time and money to do so. I'd love to see how Stallman tries to make a GPLv4 to counter it once he realizes what's happening, but at the end of the day they're playing by the GPL rules.

Should any of the BSDs care? Nope. Let them destroy the foundation which made Linux great and turn it into a product owned and controlled by one company with the sole purpose of making a profit, and the BSDs will continue happily down their own path cherry picking the few things systemd does right, if there are any.
 
We should not hope for the demise of Linux, for a simple fact. Should that happen, a lot of contributors will look for new playgrounds. And whoever watched the mailing lists and forums associated with *buntu, for example, should consider the consequences of these coming here.

In the short run, Redhat will profit from this. But I think they will suffer in the long run. As will we.
 
Crivens said:
We should not hope for the demise of Linux, for a simple fact. Should that happen, a lot of contributors will look for new playgrounds. And whoever watched the mailing lists and forums associated with *buntu, for example, should consider the consequences of these coming here.

In the short run, Redhat will profit from this. But I think they will suffer in the long run. As will we.

Indeed, that would not be good. At the very least the quality of the project would probably drop like a rock, that would be a deal breaker for me and I imagine many others.
 
AzaShog said:
But that's the problem. systemd is not only hijacking, but providing its own versions of important system tools. It's not just a glue. If it were, then I'd have no problem with it. If it used, or patched ntpd for extra data instead of reinventing systemd-timesyncd, it would be a glue. If it did the same to mDNS responding functions, DNS resolving functions, DHCP daemons, ..., by patching for a central "hub" instead of reinventing and rewriting its own versions, I might not have a problem with it. But it goes into the hilarious part of the circus. I read it also implements a light httpd server and now implements its own terminal, vt and will act as a virtualization container/manager as well, I don't know, maybe by consuming or better yet, reinventing docker.

dpejesh said:
How long until they decide it's going to handle package management and require RPM? How long until they start signing services and packages and locking you out from using your own unless it's signed by them? How long until they start implementing ways of preventing Oracle from basing their Linux distribution off it, and everybody else for that matter? .

Think bigger. Much bigger. They want to containerise the whole caboodle.

The future? Read it and weep.

They want to eliminate the distro ecosystem in all but name. What is the point of all this otherwise? It's technically very impressive, but I cannot understand what problem it is solving.
 
I for one am very grateful to systemd. It convinced me to finally migrate my desktop to FreeBSD, which turned out to be a great decision.

I'm sure I'm not the only "systemd refugee" here.
 
bthomson said:
I for one am very grateful to systemd. It convinced me to finally migrate my desktop to FreeBSD, which turned out to be a great decision.

I'm sure I'm not the only "systemd refugee" here.
You're not. There are tons as I noted earlier.
 
Tons if you count how many FreeBSD users there were before the rise of systemd, but if you look at it globally, FreeBSD isn't even noted in Netmarketshare's statistics of desktop operating systems market share. Last I browsed http://www.netmarketshare.com/operating ... pcustomd=0, FreeBSD was at 0.01%, now it's not even in statistics. And Linux has actually risen, it was like 1% before. See e.g. here http://www.netmarketshare.com/operating ... imeframe=Y. Not only is FreeBSD in there, but also Solaris. Now they're gone.
EDIT: Now that I think about it, I've also seen more BSD topics at https://boards.4chan.org/g/, but it's hardly a measure.
 
@pkubaj, you are confusing market share with "number of people who have moved to FreeBSD". There are tons of ex-Windows users who want a free ride and have jumped on the Linux band wagon now that there are easier to install distros. Let Linux have 'em. That takes nothing away from the scores of people I see every day coming to FreeBSD boards/forums/subreddits who say, "I'm coming from Linux cause I'm sick of [insert systemd, fragmenting, etc.]"; something I had rarely seen until the past year or so which is the point I made earlier. These are generally not people who are just looking for a Windows replacement.

There are also a ton of ex-Windows users who have found that the web world has gone mobile and Microsoft does not exist in that arena so it's far more natural to switch to a Unix-like system. All their friends are on Linux so that's where they go without technical consideration. I'll pat myself on the back for giving that some thought 10 years ago when my brother-in-law, a manager of a large Microsoft shop, told me to use Linux for my business. Instead, I gave it some thought and went for what I considered to be the better technical choice and signed up for FreeBSD.

I have never regretted that choice.
 
Last edited by a moderator:
Status
Not open for further replies.
Back
Top