Will systemd make FreeBSD more popular?

[Oko]

I think that anything Red Hat sees as a viable alternative to systemd, for most Linux users, will be gone by the end of 2015.

I honestly couldn't care less. I don't run Linux on my personal machines. Much like people who got hooked up on UNIX in late 80s early 90s I never cared for Linux or had any curiosity about it. I use Linux because somebody is paying me to do the work. I could go back and teach college mathematics again and never even touch computers for living.

 

[BSD-Kitsune]

There are many things I'd do instead of be a sysadmin in charge of systemd or launchd based systems - I'd rather be a Windows Server admin for one, because if you can't figure it out then MS brings out techs to fix the problem. Doesn't mean I enjoy it, but at least I can offload something thats too hard on someone else. With CentOS at my old job, we couldn't do that. We had to sit and isolate the problem and if we couldn't fix it then, we had to write a report describing what was needed, then the ticket would be assigned right back to you.

As a guy who grew up using Windows, hated it, went to Linux and Mac, settled on Mac with occasional BSD, Linux and commercial UNIX dabbles, then mostly BSD and commercial UNIX I have to say BSD and the commercial UNIX derivatives are the most coherent operating systems I've used, even compared to BeOS, which I love, but its an insecure dinosaur with lots of crap software. I just like working with stuff that works, and does so in a reliable and coherent manner.

 

[walterbyrd]

it is the upstream of the applications which must decide if they want to be part of that party.

As I understand it, GIMP is halfway there. It will run without systemd, but pulls in systemd libraries when it is installed.

 

[retrogamer]

walterbyrd, maybe they want to cook the frog (which does not work, by the way) - but it is the upstream of the applications which must decide if they want to be part of that party. If RedHat is killing of Linux as we know it, then the Linux users must do something about that. We can help, but not act. And if they succeed, maybe it was time for Linux to die and be replaced by something else. Maybe it is time for HURD after all. In the end, Linux is only the kernel, and the GNU userland is running amok. Time will tell who needs whom, and who wants to join the dinosaurs. In the end, we all do. But not today.

You raise a good point about the GNU userland. One thing I've found interesting is that Stallman hasn't weighed in on systemd in any meaningful way, but when GNU Guix was introduced he had praise for it. GNU Guix uses dmd as its init system, and that would presumably hold true for a release using the HURD kernel (it uses a deblobbed Linux kernel right now). There is also an effort to create a GNU distribution using a deblobbed FreeBSD kernel I posted about earlier this year. I've seen a lot of the tech media equate GNU to Linux as a 1:1 in being lockstep with regard to the direction of the OS, but I'm not so sure that's the case. I'm not personally that interested in the licensing philosophy behind the software (I'm anti-DRM, but that's about it), but I don't think it would be the worst outcome if this inadvertently created a viable FreeBSD based GNU distribution and made it easier to run some "Linux" applications on FreeBSD.

 

[Beastie7]

I doubt people here would want anything to do with GNU. I wonder though, if GTK+ will end up being swallowed by systemd also. Because AFAIK Red Hat is its biggest contributor also, and they even have paid employees dedicated to the project. If that happens I'm afraid our application support would be cut short, or would it? I hope I'm wrong. We would still have Qt/KDE but I know the Framework/App stack is heavily Linux centric, and the majority of commercial parties and devs reside of the GTK+ side I believe. Not sure.


I'd be nice if PC-BSD had more man power and momentum for support. Having our own application stack for Lumina wouldn't be a bad idea. I say use OS X and call it a day, but not everyone wants to go that route.

 

[retrogamer]

I doubt people here would want anything to do with GNU. I wonder though, if GTK+ will end up being swallowed by systemd also. Because AFAIK Red Hat is its biggest contributor also, and they even have paid employees dedicated to the project. If that happens i'm afraid our app support would be cut short, or would it? I hope i'm wrong. We would still have Qt/KDE but i know the Framework/App stack is heavily Linux centric, and the majority of commercial parties and devs reside of the GTK+ side i believe. Not sure.
.

I have to say, as someone who mostly uses FLTK applications, I'm not sure that GTK or Qt matter that much. For people who want a full on "desktop" with pretty GUIs for every task, it's something of an issue, but the ability to use /x11-wm/jwm with with /www/dillo2, /editors/ted, etc. along with command line utilities like /multimedia/livestreamer isn't going anywhere.

I don't use PC-BSD, but I thought it was already using Debian GNU/kFreeBSD for its Linux emulation, or has that been dropped?

 

[BSD-Kitsune]

I say use OS X and call it a day, but not everyone wants to go that route.

I wish we didn't have people saying stuff like this. We already have enough Apple influence as it is, we don't need to become the server extension of OS X. I'm not shooting you down Beastie, I'm merely saying people should bake and eat their own cake, or else don't bake it at all.

 

[alphaniner]

I just like working with stuff that works, and does so in a reliable and coherent manner.

From my perspective, this is exactly the direction systemd is taking Linux. I spent a few years using SysV based distros, and never learned a thing about it. Granted, I can't say I really tried. About five years ago I moved to Arch Linux, which at the time used a "BSD style init framework". It seems like I groked that much more easily, but maybe that was because in Arch you can't help but learn how things work. In any case, after getting a handle on it I felt completely lost when dealing with standard SysV systems.

And now, after the initial growing pains of the transition to systemd I can hardly imagine running Linux without it.

 

[BSD-Kitsune]

From my perspective, this is exactly the direction systemd is taking Linux. I spent a few years using SysV based distros, and never learned a thing about it. Granted, I can't say I really tried. About five years ago I moved to Arch Linux, which at the time used a "BSD style init framework". It seems like I groked that much more easily, but maybe that was because in Arch you can't help but learn how things work. In any case, after getting a handle on it I felt completely lost when dealing with standard SysV systems.

And now, after the initial growing pains of the transition to systemd I can hardly imagine running Linux without it.

Systemd isn't reliable, and isn't coherent. It works, much like Windows, but when stuff breaks, good luck getting it fixed. I had a server lock up due to systemd, rebooted, and it went into an fsck loop. It was a fucking pain in the ass to fix. With the BSD and SysV inits, this was never an issue, and if it was, I wouldn't have systemd trying to mount the volume going into single user, then try to fsck it all over again. I could, if needed on the inits, boot with the volume not mounted and do something from there to get it to fsck and mount.

Standard Linux SysV init wasn't very good, I'll admit it, but IRIX sysvinit is a dream to work with by comparison. Then again, I see IRIX as a mostly perfect OS. If it was open sourced tomorrow, I'd work to port Xsgi and the magic desktop to FreeBSD in a heart beat.

There's just as easy and usable alternatives for systemd, runit being a fantastic choice.

 

[Beastie7]

I wish we didn't have people saying stuff like this. We already have enough Apple influence as it is, we don't need to become the server extension of OS X. I'm not shooting you down Beastie, I'm merely saying people should bake and eat their own cake, or else don't bake it at all.

Oh none taken. But the reality is, is that your average Joe isn't going to care where the cake came from, or who developed it. The fact is that OS X is a good user experience and a very useful tool for end users, regardless of technical bias, licensing or whatnot. That's exactly what is it, a useful UNIX desktop based on bits of FreeBSD. Apple has successfully done that, and they understand joe blow. It's about perspective. Now I'm not implying PC-BSD is useless because it isn't for some people, but from a user experience standpoint (including mine), I can tell you I wouldn't recommend PC-BSD to anyone, not to mention third party app support. It was painful. iXsystems needs a way to attract more developers, and give users a compelling incentive to switch to PC-BSD, whether from Windows/OS X/whatever. Because users care about user experience, just look at Ubuntu. It matters. It's been six years since iXsystems acquired PC-BSD (just two years after the inception of Ubuntu), and I don't see much accomplished.

Again just my two cents, but this is the reality.

we don't need to become the server extension of OS X

I don't see how this is a bad thing, especially with awesome tools like the Casper Suite. The Mac/BSD stack is a perfect replacement for Windows/Windows Server.

 

[Beastie7]

I don't use PC-BSD, but I thought it was already using Debian GNU/kFreeBSD for its Linux emulation, or has that been dropped?

I believe Debian will drop the whole GNU/kFreeBSD project on their next release (Jessie). I read an article on it, but i can't find it right now. Maybe because of.. well, you know. lol

 

[retrogamer]

I believe Debian will drop the whole GNU/kFreeBSD project on their next release (Jessie). I read an article on it, but i can't find it right now. Maybe because of.. well, you know. lol

The project isn't being dropped (as in not developed), it just isn't going to be an official release, in the same way that Debian GNU/Hurd is not an officially supported release. I posted a link to an article about that in this very thread.

 

[BSD-Kitsune]

I don't like Apple's business model. They sell you a status symbol with overpriced, underpowered hardware, they lock their OS to the hardware and I find their interface a sickly saccharine one.

The amount of FreeBSD code in OS X isn't very big, thats a common myth. Try running XNU instead of the FreeBSD kernel and maintaining binary compatibility, it won't happen, and the subsystems are different entirely. The userland of OS X is part GNU, part BSD, and part NeXTSTEP. The XNU kernel itself consists of Mach, which is a terrible kernel (I'm opposed to the microkernel model they went with, just look at Hurd for evidence) with a NeXTSTEP-derived wrapper for certain parts of the stack. Some of the network code, most of the POSIX wrappings, and some mostly-subsystem frameworks. That's about it.

I don't want us associated with the Apple crowd, I hate the pricks at the Apple store, I think Tim Cook is pissing on Steve Job's legacy, I hate the iDevices. I hate the Apple users, I think OS X is a piece of awful dreck that needs to die and I see Apple as worse than MS. Then again, I worked for MS as a DC tech for a few months. They treat their workers like shit, sure, but the pay was good, and the work was easy. Back to the point, I won't let Linux, Red Hat, GNU, Apple or anyone else ruin UNIX for those who like it as it is.

Also, I'm one of the opponents of adding anymore Apple code to FreeBSD, to the point I have threatened to start a fork of FreeBSD if launchd comes, if you look back in the thread I've said that. And this is coming from someone with nowhere near the experience necessary to do it by myself, but I will, if I have to, and from what people here have said, I'm not alone.

I don't use PC-BSD, but I do like some of their work and have recommended it to people, who do like it. Your opinion seems to be the minority. I'd use PC-BSD if they didn't include so much antique dreck and didn't dumb down the install.

 

[Beastie7]

The project isn't being dropped (as in not developed), it just isn't going to be an official release, in the same way that Debian GNU/Hurd is not an officially supported release. I posted a link to an article about that in this very thread.

It is being support by Debian? as per this thread;

https://lists.debian.org/debian-devel-announce/2014/11/msg00005.html

they're expecting "porters" to make an unofficial release, I'm not sure who they are. Seems Debian isn't developing GNU/kFreeBSD anymore? or did I miss something else?

Regardless, I still think iXsystems should make moves with the PC-BSD project, much in the same reason why Lumina was developed instead of continuing on KDE portage from Linux dependencies.

I don't like Apple's business model. They sell you a status symbol with overpriced, underpowered hardware, they lock their OS to the hardware and I find their interface a sickly saccharine one.

The amount of FreeBSD code in OS X isn't very big, thats a common myth. Try running XNU instead of the FreeBSD kernel and maintaining binary compatibility, it won't happen, and the subsystems are different entirely. The userland of OS X is part GNU, part BSD, and part NeXTSTEP. The XNU kernel itself consists of Mach, which is a terrible kernel ( I'm opposed to the microkernel model they went with, just look at Hurd for evidence ) with a NeXTSTEP-derived wrapper for certain parts of the stack. Some of the network code, most of the POSIX wrappings, and some mostly-subsystem frameworks. That's about it.

I don't want us associated with the Apple crowd, I hate the pricks at the Apple store, I think Tim Cook is pissing on Steve Job's legacy, I hate the iDevices. I hate the Apple users, I think OS X is a piece of awful dreck that needs to die and I see Apple as worse than MS. Then again, I worked for MS as a DC tech for a few months. They treat their workers like shit, sure, but the pay was good, and the work was easy. Back to the point, I won't let Linux, Red Hat, GNU, Apple or anyone else ruin UNIX for those who like it as it is.

Also, I'm one of the opponents of adding anymore Apple code to FreeBSD, to the point I have threatened to start a fork of FreeBSD if launchd comes, if you look back in the thread I've said that. And this is coming from someone with nowhere near the experience necessary to do it by myself, but I will, if I have to, and from what people here have said, I'm not alone.

I don't use PC-BSD, but I do like some of their work and have recommended it to people, who do like it. Your opinion seems to be the minority. I'd use PC-BSD if they didn't include so much antique dreck and didn't dumb down the install.

I'm sorry you feel that way. You seem to be more emotional than objective about what I'm saying. You hate anything Apple, and I enjoy the Apple integrated experience. Let's just leave it at that.

You know, some of what you say I feel the exact same way about Microsoft and Windows, and they've done more damage to end users and national security than any other top OS vendor, not to mention Windows being a huge mess in and of itself. But I still use windows server for their apps because users (businesses) require them.

 

[retrogamer]

It is being support by Debian? as per this thread;

https://lists.debian.org/debian-devel-announce/2014/11/msg00005.html

they're expecting "porters" to make an unofficial release, i'm not sure who they are. Seems Debian isn't developing GNU/kFreeBSD anymore? or Did I miss something else?

You have misunderstood what you read. The "porters" are the developers of that port.

We discussed kfreebsd at length, but are not satisfied that a
release with Jessie will be of sufficient quality. We are dropping
it as an official release architecture, though we do hope that the
porters will be able to make a simultaneous unofficial release.

That means that the hope is for the developers to continue doing simultaneous releases with the official Debian releases.

 

[Beastie7]

Bah, alrighty then. I still think it's better to brew our own solutions instead of trailing behind updates for apps that require newer versions of kernels and graphics drivers.

 

[BSD-Kitsune]

I'm sorry you feel that way. You seem to be more emotional than objective about what I'm saying. You hate anything Apple, and I enjoy the Apple integrated experience. Let's just leave it at that.

You know, some of what you say I feel the exact same way about Microsoft and Windows, and they've done more damage to end users and national security than any other top OS vendor, not to mention Windows being a huge mess in and of itself. But I still use windows server for their apps because users (businesses) require them.

Yeah? Well Apple cheated me out of nearly $5k in the last 2 years I used their products. That's the cost of my old mini, Retina Macbook Pro and numerous accessories, along with lost revenue from terrible build quality requiring me to send off the machines for RMA 5 times collectively. Not so much as an apology from a manager from numerous botches RMAs. BBB complaint was totally stacked in their favour. I don't hate Apple's customers, rather I hate when people suggest OS X, its an upgrade from Windows, if by a knife edge. OS X really declined in quality after Leopard as well. The OS just became a side dish to iOS. I have every right to hate them as a company.

By comparison, working with MS and Dell, as a Dell tech I got prompt and professional assistance from the MS service team (I handled the hardware, they handled the OS) while working for Dell. Later was hired by MS. Guys I worked with all used *NIX at home, and just worked for MS because it paid well.

 

[TiberiusDuval]

Oh none taken. But the reality is, is that your average Joe isn't going to care where the cake came from, or who developed it. The fact is that OS X is a good user experience and a very useful tool for end users, regardless of technical bias, licensing or whatnot. That's exactly what is it, a useful UNIX desktop based on bits of FreeBSD. Apple has successfully done that, and they understand joe blow. It's about perspective. Now i'm not implying PC-BSD is useless because it isn't for some people, but from a user experience standpoint (including mine), I can tell you I wouldn't recommend PC-BSD to anyone, not to mention third party app support. It was painful. iXsystems needs a way to attract more developers, and give users a compelling incentive to switch to PC-BSD, whether from Windows/OS X/whatever. Because users care about user experience, just look at Ubuntu. It matters. It's been six years since iXsystems acquired PC-BSD (just two years after the inception of Ubuntu), and I don't see much accomplished.

Again just my two cents, but this is the reality.

I can recommend PC-BSD, it has improved quite much, especially most recent update did lot of good. Generally most problems nowadays are on growing pains side. Usually when new upgrade comes it breaks things, and seem to be a bit undertested. Now when they have automatic boot-environment creation that problem is going to be mitigated. If new update messes up your system, just boot on previous boot environment, report problems and wait for patch.

 

[walterbyrd]

I just like working with stuff that works, and does so in a reliable and coherent manner.

From my perspective, this is exactly the direction systemd is taking Linux.

Again, if you want proprietary, go with MS-Windows.

Some here think that that statement is over-emotional bashing, but it's not. I really mean it. Microsoft does proprietary better than Red Hat.

There are advantages to proprietary. An OS controlled by one group will be more coherent. Today's Linux, by contrast, is all over the map.

The downsides of a proprietary system is deliberate obfuscation, and vendor lock-in.

Red Hat is following Microsoft's playbook to the letter. Right down to the exact same propaganda, FUD, and astro-turfing. I followed MS business practices for decades, and to me, it's glaringly obvious.

Would you be surprised to learn that, according to senior Microsoft developers quoted 'on the record' and documented in O'Reilly references, one of the key design decisions was to eliminate the 'power user' class. By making the registry impenetrable to mortal humans they eliminated what they saw as a 'problem' of users thinking they could control their computer.

From another forum:

This sharper division between developers and users is also a goal of the freedesktop/systemd/gnome push. If you don't believe me, go look in /etc/udev and tell me humans are intended to touch anything in there. No line breaks, no comments, no reliable documentation other than the source. Same for dconf, although it least it, unlike the Windows Registry, has an explicit feature for help text as an option for each key... although it is pitiful how few actually have any supplied. Again, the assumption in actual use in the field is that dconf is for applications. Developers will write apps that store values in the 'registry' and those apps alone will manipulate them. If an app doesn't expose a knob to change one the user isn't supposed to manually tamper with it.

This reminds me of Microsoft paying De Icaza to attack Linux from the inside with the Mono trojan horse. Now, it is Red Hat (no doubt directed by their customer Fed Gov) directly attacking the simple, modular, do-one-thing-right Unix design philosophy and replacing it with the far-reaching, metastatizing blob that is SystemD. Why? To bake-in impossible to find, intentional backdoors and vulnerabilities as designed by Poettering and the rest of his paid-off coven.

 

[alphaniner]

Why are you on these forums if you equate "controlled by one group" with deliberate obfuscation and vendor lock-in?

Red Hat is following Microsoft's playbook to the letter. Right down to the exact same propaganda, FUD, and astro-turfing. I followed MS business practices for decades, and to me, it's glaringly obvious.

Would you be surprised to learn that, according to senior Microsoft developers quoted 'on the record' and documented in O'Reilly references, one of the key design decisions was to eliminate the 'power user' class. By making the registry impenetrable to mortal humans they eliminated what they saw as a 'problem' of users thinking they could control their computer.

I can accept all that as possible. But unless you're willing to state your case in a professional manner with evidence backing up every claim, it's just another opinion.

This sharper division between developers and users is also a goal of the freedesktop/systemd/gnome push. If you don't believe me, go look in /etc/udev and tell me humans are intended to touch anything in there.

I guess I'm not human, because I have written custom udev rules. A few months ago, I decided to see if I could use a udev rule to start a network profile when a USB ethernet device was attached. The biggest barrier I encountered was misinformation from "Tux Q. Blogger". But a mailing list message by Poettering cleared it all up in an instant.

Granted, I never considered doing anything similar without udev rules and systemd. Whether cause or effect, I can't imagine I could have accomplished anything comparable without an impractical amount of work.

 

[BSD-Kitsune]

I guess I'm not human, because I have written custom udev rules. A few months ago, I decided to see if I could use a udev rule to start a network profile when a USB ethernet device was attached. The biggest barrier I encountered was misinformation from "Tux Q. Blogger". But a mailing list message by Poettering cleared it all up in an instant.

Granted, I never considered doing anything similar without udev rules and systemd. Whether cause or effect, I can't imagine I could have accomplished anything comparable without an impractical amount of work.

Granted you like systemd apparently, what do you propose the FreeBSD do?

udev, dbus, hal, systemd, logind, consolekit, polkit, it all just reinvented the wheel in various ways. I honestly like FreeBSD's devd, its very simplistic in operation, does exactly what is needed for device node management and I've never had any issues with it. I could likely do the same thing you did in udev using a combination of devd rules and helper scripts. Do I want to? Not really. On Slackware I'm using mdev, after having to wrestle with xorg's source code. What a hassle just to use a device manager that does only what is needed.

 

[alphaniner]

Granted you like systemd apparently, what do you propose the FreeBSD do?

Keep on keepin' on. I don't propose or hope for any change in FreeBSD. For the purposes of this thread, I'm just a Linux user who is happy with systemd.

 

[gkontos]

Endless topic, short and best answer:

Will systemd make FreeBSD more popular?

NO!!!

 

[usrslashpeople]

I have been a long time supporter of FreeBSD. I have a copy of "The Design and Implementation of the FreeBSD Operating System" in my restroom on permanent loan. I have been a Linux engineer/admin for too many years. I was never excited about Linux. I grew up with IRIX and still use an Octane2 (with green skins) as my main system for doing any work in the terminal (which is most of what I do). I like UNIX or UNIX like when it comes to the choice of operating system. I currently hold two jobs as the decision maker on what we use. I was not a fan of gnome3, but it didn't matter because I could choose whatever WM I want to use (windowmaker!) I am not a fan of systemd. I really don't care what it does. The simple fact of the matter is that it makes Linux more like Windows and throws the UNIX philosophy in the trash. We have already had meetings at both of my jobs about the future direction of Linux. I will be working on a FreeBSD migration plan. I am just starting this project and I am hoping most scientific applications like R will be available. Commercial software such as SAS will run on Windows if it isn't possible to run on FreeBSD. In two years we will be on Linux without systemd/FreeBSD or FreeBSD/Windows.

If systemd is here to stay, I hope all of the hipsters in skinny jeans at starbucks really enjoy fast Linux boot times on their laptops. I will continue to work with UNIX like operating systems in the enterprise. When a server takes 10 minutes to post, I really don't care if the OS boots 3 seconds faster with systemd than with SysV init.

 

[BSD-Kitsune]

Oh hello! Another IRIX user! I have an Octane2 with blue skins, an Origin 300 server, and an Indy I want to use for N64 debugging when I can find the kit.

I agree with much of what you said, but I'm mostly OpenBox, CDE and such guy. Love the username too ^o^.