systemD is coming to a port near you!

mark_j

Daemon

Reaction score: 704
Messages: 1,229

If spare time is available, I like to read vermaden 's valuable news. Specifically, this week's installment here.

What got my attention was the subject InitWare (systemd fork) Runs on OpenBSD for First Time.

A quick look at the github site gives no reasoning behind this. Further investigation yields another example here. There's probably more out there.

Of course, people are free to spend their time doing whatever they please. I am not disputing this. I am just questioning the reason/motivation for such endeavours, given what I know of systemD, it's designed to not be portable and it's inherently poor in design anyway (you know, discards the Unix principle of simplicity and minimalism [subjective]).

The best chance these have are as ports because of their licence, so again, what's the motivation?

I also wonder, once they mature, if anyone would even use them?

I did feel like posing the question to their authors but I am not on github and don't intend to be.

Hopefully they might pass this way and reply. Either way I think it's an interesting topic (in futility, if you ask me :confused:)
 

Vull

Aspiring Daemon

Reaction score: 443
Messages: 734

Maybe they want it for Wayland or KDE, or both. Who knows? Not me. I do like KDE, but it's only of peripheral interest to me, and, I hope and would certainly like to believe, to the core FreeBSD team.

It's just a port. I like the init system we already have. No need for systemd or its spawn to ever invade our base system here.
 

Trihexagonal

Son of Beastie

Reaction score: 2,405
Messages: 2,930

That's one of the things I addressed at DW this morning. I explained that just because it makes it into the ports tree doesn't mean it's going to make it into the Base System install. They weren't clear on the issue so I detailed it out for them. And why Xorg, DE or a WM are not included in the Base System.

That we like it that way and no matter how many people complain about "Why can't FreeBSD be more like that other OS" it's going to stay that way.


It's a dirty job but somebody has to get out there and do it. I'm already infamous. It can only add to it, take me closer to being famous for my infamy and achieving notoriety. :)
 

gpw928

Aspiring Daemon

Reaction score: 240
Messages: 553

If spare time is available, I like to read vermaden 's valuable news.
vermaden 's valuable news something that I also appreciate greatly.

Viewed solely as middleware, with the listed characteristics, namely:
  • InitWare has a high level of compatibility with the core interfaces of systemd;
  • InitWare is highly portable;
  • InitWare aims to be significantly more modular; and
  • InitWare is of significantly smaller scope, concerning itself only with system, service, and session management, and matters ancillary to these, such as event log management.
I think it may find favour with those who want to port Linux applications *BSD platforms, and maybe even some systemd-free Linux platforms.

I say this because of the ongoing struggle I see in the Debian community -- the effort required to maintain (systemd-free) Devuan fork is approaching impossible as systemd ingratiates itself into every nook and cranny of Linux. Getting the middleware functions, without the associated creep into the kernel, is going to make life a lot easier for some people.
 

Trihexagonal

Son of Beastie

Reaction score: 2,405
Messages: 2,930

I've got a Kali 2021.2 rolling release box sitting next to this one and unless I missed something it's still based on Debian. I don't notice any difference in the way it runs from the Debian boxen I've had in the past.

It's an Intel Core 2 Duo T7700 @ 2.4GHz with 4GB RAM at 28 days uptime. CPU usage at 2%, free memory 3449M, and apt-get works for me just like it always did.
FreeBSD 12.2 is what I use though.
 

bobmc

Member

Reaction score: 47
Messages: 90

The big problem for software systems these days is security. Google "security challenge" = 1,100,000,000 hits.

So when considering the merit of a software system change, one should ask "how does this improve/degrade security".
 

Beastie7

Aspiring Daemon

Reaction score: 590
Messages: 709

I don't know what apps would necessitate systemd. Doesn't the Linuxulator solve this issue for apps?
 

sidetone

Daemon

Reaction score: 907
Messages: 1,856

I didn't understand SystemD or whatever init it was when I tried Arch Linux. It had 3 or 4 layers represented by files for starting up programs, which, I didn't know what to put first, and why one should be in the first init. Maybe if they had an init file for internet services, instead of 1. Or if they had 2 or 3 inits. 4 left me confused on how to order them. Maybe if I left the last init blank, if I didn't need it, or didn't have enough services to need its categorization for start up order.
 

Trihexagonal

Son of Beastie

Reaction score: 2,405
Messages: 2,930

In my Kali home directory I have an .Xauthority, Xdefaults, .xsession-errors, .zshrc, .zshhistory, .ICEauthority (blank), .dmrc (desktop lightdm-xsession), .bashrc, .bash_logout, .bashrc_original and no sign of an .xinitrc or anything like it. I don't pay much attention to how long it takes to start, it's been up 28 days.

geezer, everyone who is famous has fame, but not everyone who has fame has infamy
When you're famous for your infamy, you're notorious.

When you pour a married woman a glass of Ocean Spray Mango Juice in Church during lunch break, go back upstairs when the Service begins again, are looking down with your eyes closed, look to your left and she's standing right beside you in the pew, her husband is on the other side of her trying to pull her back to where she'd been sitting by her hand and she is resisting, then they turn the Sermon to focus on Stealing a mans wife and you realize they're talking about you...

You're me.
 

Minbari

Aspiring Daemon

Reaction score: 307
Messages: 569

It's just a port. I like the init system we already have. No need for systemd or its spawn to ever invade our base system here.
I like the existing init too, yet he has its limits. Just because something exist for a long time doesn't means is good now or in the future. I think we should be more open to new ideas and changes or we will end just like Solaris & Co.
Benno Rice has a good view about systemd and why we (FreeBSD) should have a systemd like init.
 

Trihexagonal

Son of Beastie

Reaction score: 2,405
Messages: 2,930

Please tell me what's wrong with the current ~/.xinitrc / startx paradigm.

It's been working fine for me and I prefer it to a GUI.

What have I to gain from the change?
(Change Bad) (Survival Good)
 

hardworkingnewbie

Active Member

Reaction score: 161
Messages: 162

What got my attention was the subject InitWare (systemd fork) Runs on OpenBSD for First Time.

A quick look at the github site gives no reasoning behind this. Further investigation yields another example here. There's probably more out there.
Well when having a closer look then it becomes quite clear that InitWare really is an init system replacement. Probably the author thought it might be a good idea to port systemd minus some parts over to the BSDs, because then many Linux docs can be used without much change for the BSDs as well, so he did.

I mean, what were always the main selling points of systemd? Parallel startup of processes and service monitoring. The problem is what it does aside that, its creeping featurism. Also quite often usurping formerly independent projects, which have an impact for X11 and such, like elogind.

Looking at this nifty init system feature comparison table BSD's rc.d already has support for parallel startup of services, and can keep daemons alive as well. So what's the rationale then behind porting systemd is way beyond me.

Many would always argue then: bootup time is not so important as the people always tell. Truth that for both targets there are different solutions in place, which can do the same but with less luggage: OpenRC and runit. And if you think that service monitoring should not be done by the init system, there are dedicated tools for that like monit.

Complexd on the other hand is a different thing, it is a reimplementation of some systemd APIs, which might be needed by some DEs to get out the full functionality. GNOME especially had a long time a really nasty dependency on systemd stuff, meaning you could not use it properly without it.
 

Menelkir

Well-Known Member

Reaction score: 333
Messages: 272

I like the existing init too, yet he has its limits. Just because something exist for a long time doesn't means is good now or in the future. I think we should be more open to new ideas and changes or we will end just like Solaris & Co.
Benno Rice has a good view about systemd and why we (FreeBSD) should have a systemd like init.
Just because something is new doesn't means it should be changed for the sake of changing. Systemd doesn't bring something new, it just make things depend on it and try to incorporate everything inside it. There's not a single thing that systemd do that isn't possible with other less cumbersome init systems can do (openrc, runit, 66, s6, rc, you name it). And as hardworkingnewbie already stated, what is the main selling point of systemd? I think I can answer that: EEE.
Good that you pointed the video of Benno Rice, because he said to get what you like from systemd and try to port it yourself. My answer is: I don't need it, everything I like in systemd exists by at least 20 years as a standalone application.
 
OP
M

mark_j

Daemon

Reaction score: 704
Messages: 1,229

I like the existing init too, yet he has its limits. Just because something exist for a long time doesn't means is good now or in the future. I think we should be more open to new ideas and changes or we will end just like Solaris & Co.
Benno Rice has a good view about systemd and why we (FreeBSD) should have a systemd like init.
I watched that a while back and if you follow the talk through he eventually gets to the faults of systemD: the scope creep as it attempts to solve the problems its solutions have created. Just ponder that for a moment, because systemD does that.

It seems every month a new scope-creep 'feature' is offered up by the redhat boys. In that regard alone, imagine trying to port such an ever expanding piece of software written explictly to be NON-portable? Good work for a masochist.:p

Systemd is a solution in search of a problem; it may have some rationale to linux-based people but to freebsd it makes no sense. I mean apart from saying you have ported some of it to FreeBSD what problem does it solve?

I think you're wrong when you say we should be more open to ideas, thus implying a close-minded attitude. If your argument was with linux then you'd have a case because of their inherent NIH syndrome. However, FreeBSD has adopted linux emulation, ext file handling, dtrace, zfs, openbsm to name but a few. All of these can be said to enhance the OS or the user experience. I'm still looking for such from systemD.

But you are correct; FreeBSD can't afford to be an island; however it also can't berth the Titanic (aka systemD)
 
OP
M

mark_j

Daemon

Reaction score: 704
Messages: 1,229

Just because something is new doesn't means it should be changed for the sake of changing. Systemd doesn't bring something new, it just make things depend on it and try to incorporate everything inside it. There's not a single thing that systemd do that isn't possible with other less cumbersome init systems can do (openrc, runit, 66, s6, rc, you name it). And as hardworkingnewbie already stated, what is the main selling point of systemd? I think I can answer that: EEE.
Good that you pointed the video of Benno Rice, because he said to get what you like from systemd and try to port it yourself. My answer is: I don't need it, everything I like in systemd exists by at least 20 years as a standalone application.
Your email/quote from Poettering was very illuminating because that's what I've suspected all along (clearly I'm a genius :rolleyes:). This is why it makes sense for linux-based systems. There's a kernel developed totally separate from userland functionality. The only guiding light is Torvald's directive that nothing breaks userland. They are basically attempting to copy HP/UX, Solaris and BSD: systems designed as one.

However the severe synic and skeptic in me sees the darker side where all these distributions eventually are doomed by the ever-growing reach of systemD; they will all become the same (if that isn't already the case now?) They are in fact killing themselves off by adopting systemD. Soon there will be only one and it will be forever known as...Redhat.
 

Vull

Aspiring Daemon

Reaction score: 443
Messages: 734

I like the existing init too, yet he has its limits. Just because something exist for a long time doesn't means is good now or in the future. I think we should be more open to new ideas and changes or we will end just like Solaris & Co.
Benno Rice has a good view about systemd and why we (FreeBSD) should have a systemd like init.
Thanks for the Benno Rice link. Rice says systemd is like launchd except for the point that, whereas launchd is Apple-specific, systemd is Linux-specific. The FreeBSD init system is generic enough that any Un*x-like system can use it.

Being open to change and new ideas is a good thing, but breaking interoperability standards is not. In this case, and admittedly, without knowing all the minute granular details, I tend to view these types of interoperability barriers as anti-competitive practices. Whether this is being done deliberately or coincidentally is debatable, and in some situations, can even be prosecutable, to wit:

 

sidetone

Daemon

Reaction score: 907
Messages: 1,856

However the severe synic and skeptic in me sees the darker side where all these distributions eventually are doomed by the ever-growing reach of systemD; they will all become the same (if that isn't already the case now?) They are in fact killing themselves off by adopting systemD. Soon there will be only one and it will be forever known as...Redhat.
Could the theoretical outcome of SystemD be anything like Vanishing on 7th Street?
 
OP
M

mark_j

Daemon

Reaction score: 704
Messages: 1,229

Thanks for the Benno Rice link. Rice says systemd is like launchd except for the point that, whereas launchd is Apple-specific, systemd is Linux-specific. The FreeBSD init system is generic enough that any Un*x-like system can use it.

Being open to change and new ideas is a good thing, but breaking interoperability standards is not. In this case, and admittedly, without knowing all the minute granular details, I tend to view these types of interoperability barriers as anti-competitive practices. Whether this is being done deliberately or coincidentally is debatable, and in some situations, can even be prosecutable, to wit:

If you read the link Menelkir posted, https://lists.freedesktop.org/archives/systemd-devel/2010-September/000391.html, the embrace, extend, extinguish concept/philosophy is precisely what the end-goal of ibm redhat's systemd is.
You're spot on!
 
OP
M

mark_j

Daemon

Reaction score: 704
Messages: 1,229

Could the theoretical outcome of SystemD be anything like Vanishing on 7th Street?
If by that you mean:
"The population of Detroit [linux distributions] has almost completely disappeared, but a few remain [Redhat, ubuntu/debian]. As daylight disappears they realize that the Dark[systemD] is coming for them."? Then , yes you're movie selection is apt! :)

Off-topic in an off-topic: is the movie any good?
 

mer

Aspiring Daemon

Reaction score: 334
Messages: 545

I watched that a while back and if you follow the talk through he eventually gets to the faults of systemD: the scope creep as it attempts to solve the problems its solutions have created. Just ponder that for a moment, because systemD does that.
"Hi I'm from the government and I'm here to help you with the problems that the government created. The only way I can do that is if the government gets more involved in everything".

:)
 

sidetone

Daemon

Reaction score: 907
Messages: 1,856

If by that you mean:
"The population of Detroit [linux distributions] has almost completely disappeared, but a few remain [Redhat, ubuntu/debian].
Like, "The [variety of open source is disappearing], but a few [non Redhat, non Ubuntu/Debian users and environments] remain."
As daylight disappears they realize that the Dark[systemD] is coming for them."? Then , yes you're movie selection is apt! :)
Comparing SystemD to a spreading void taking over what it touches.
Off-topic in an off-topic: is the movie any good?
The paranormal aspect of it is somewhat of a stretch compared to somewhat realistic in theory sci-fi's. The acting, sequences and scenery are decent. It's good if you're a fan of Leguizamo, or if this is your kind of sci-fi. It's accurate with the trailer, with a little less suspense.

"Hi I'm from [insert bank name here] and I'm here to help you with the problems that [we] created. The only way I can do that is if [we get] more involved in everything"
Sounds more like a bank commercial to me.
 
Top