Will systemd make FreeBSD more popular?

Status
Not open for further replies.
TeamBlackFox said:
Why launchd, or why Darwin?
Click to expand...
I don't know. I am no software engineer, the only thing I can do is to hear what experts say. And by his record, Hubbard is an expert.
He did a great talk and mentioned problems which they had to get Mac OSX on mobile devices etc. and how they solved them. It could very well be that these problems can be solved without a different init system. But the experts seem to use launchd-type of init systems.
I red today that Ubuntu already 2007 wanted to use launchd for their init system but because of Apples open source licence they rejected it.
Edit: It was 2006 and not 2007 :). I looked it up again ...
Edit2: To be precise: In the Wikipedia article about launchd, they said that Ubuntu considered using launchd but rejected it because of licensing. This reads a little weaker than what I formulated above ...
 
Well, Jordan Hubbard is associated with Apple, so he's going to of course be an advocate for it. I'm a developer, not a software engineer, but I see no convincing reason to believe that a monolithic init system is the sole answer in the modern world. All of the talk about it is mostly for a usability standpoint, but usability for a GUI user translates to hell for a CLI user.

Again, I think iXSystems is a good company, but I won't stand for Jordan to have his way, I am totally serious about forking FreeBSD. If I lead, others will follow. Not all of us are Apple advocates, and now I'd like to see nothing less than Apple crumble to dust now that Steve Jobs, the only man I had any sort of respect for, is dead.
 
walterbyrd said:
I have seen lots of posts from Linux users who say they are moving to FreeBSD, or have already moved.

I have been using Linux myself for over ten years, systemd is the reason I am trying out FreeBSD.

Poettering called FreeBSD an irrelevant toy. It would be awesome if there was a surge in popularity for FreeBSD.
Click to expand...

I stopped worrying about what other people choose to run as their OS of choice long ago. I don't think about systemd a lot.
 
We seem to get somewhere :)

In the end, we have three options: Run, Fight, or Hide.
What is it we should do? Consequences, effords, gains, and so on?
 
TeamBlackFox said:
At the very least though, if I lead the charge in forking FreeBSD, I will probably ditch ZFS for HAMMER2 and improve UFS2.
Click to expand...

Out of curiosity, is there some reason why you would ditch ZFS for HAMMER2? HAMMER2 even if it were ready today would still require quite some time for it to be deployable without major show stopping bugs. ZFS has had orders of magnitude more testing.

It also seems that porting HAMMER2 won't be trivial according to Matt Dillon himself.

I understand that the BSD license (HAMMER2's presumed license) is great, but why remove what many would argue is FreeBSD's killer feature?

I would perhaps echo a previous poster's sentiments concerning contributing to DragonflyBSD instead of forking FreeBSD.

As far as the Dragonfly team not being keen on porting, I've read that they want to establish a stable scalable system on one mainstream CPU architecture first before porting. This seems to make sense.
 
Completely off the original topic but completely replacing ZFS with HAMMER (v2 of which is still mostly a research project) would be moronic. ZFS has seen 10 years of enterprise testing/use and OpenZFS is supported by many systems and the basis of many enterprise solutions. The inclusion of ZFS has been very good for FreeBSD. One thing that will intruiged Oko and TeamBlackFox though is this page that's appeared on the wiki recently. https://wiki.freebsd.org/HAMMERFS (Obviously I have no issue with HAMMER being supported as an additional file system)

Will systemd make FreeBSD more popular?
Yes. We've already seen the people coming into the forums mentioning that they've moved from Linux. That's not including the people who haven't mentioned they've moved because of systemd (even if it's the last of many reasons) or those that haven't used the forums at all.

I doubt it's actually going to be a big enough migration to actually affect the development of FreeBSD though. Most Linux shops (i.e. not just end users) will just put up with it unless it really affects their services. Linux doesn't really have any interest in maintaining UNIX philosophy. They are slowly getting to the same place as Windows/Mac OS and are developing the sort of cross-subsystem monolithic hacks that are mostly hidden away in the closed source of those other operating systems. I'm fairly sure the FreeBSD devs still do want to maintain the UNIX philosophy.

Discussions about systemd itself (and possible similar, but incredibly unlikely, developments in FreeBSD) have already been beaten to death, and wasn't the purpose of this thread, so I'm not going to comment on that.
 
Replacing ZFS would serve a purpose because while I love ZFS, the way I see forward is through other technology. With the Oracle ZFS being patent protected, none of the enhancements of later versions can make it downstream. OpenZFS is great, but I think in the at minimum of 1-2 years needed for launchd to become the default init it will be safe to assume HAMMER2 will likely be already released and well on its way to be a better alternative for ZFS. It has little to do with licensing, and everything to do with providing something well maintained and with a progressive upgrade path. You can be assured, if we have to make a move, me and the rest of Team Black Fox will definitely support ZFS until I would feel comfortable deprecating, EOLing and removing it in that order. Anyways, I'm just about done with my say in the topic. If it comes to launchd, uselessd or any other init system or change we don't agree with, we will make our move then. Talk until then is just talk
 
gofer_touch said:
As far as the Dragonfly team not being keen on porting, I've read that they want to establish a stable scalable system on one mainstream CPU architecture first before porting. This seems to make sense.
Click to expand...

I did not see this quote earlier, but this is what I understood about the DragonFlyBSD project. As I do not care to know x86_64 assembler and will not contribute to any architecture specific code involving it, I probably will not be a good fit for the DragonFlyBSD team. I learned and first debugged C on ARM32 and MIPS64 systems, so I know those platforms much better than I know x86_64. I'm not interested in learning it, and I'm also married to Nvidia, and will not use AMD or Intel integrated graphics solutions, so I am better off working on the FreeBSD project for now.

My final thoughts on the original topic are that I believe systemd in general is good for the BSDs - it puts pressure on us to hold together and pool together. Perhaps we'll be more motivated to leverage ourselves from Linux and finally get ourselves into a position to be more independent.
 
getopt said:
Some have the pleasure riding a dead horse ... :)
Click to expand...
That depends on the view. The horse is judged to be dead because it gets you nowhere, but it is also the goal that matters. If you want to stay put, so to speak, a dead horse is your best bet.

This thread is still open partly because I wanted to see where it would get, and get some consensus. If it were closed at the first mentioning of other threads which were closed, it would not get anywhere. But then you get a new thread, and another one, and none of those will go anywhere.

It looks like we have started to understand some things. As each communication trainer will tell you (about dealing with customers, f.e.), the emotional part has to come out first and the arguments come later. So when you stop listening and talking while explitives are hurled around, you will never get anywhere. After that, some cooling down is taking place and people start talking sense. I don't want to single out any example here, but if you read the last two pages of this thread, that pattern becomes obvious. So that is where we are currently heading - the high emotions are out and we can now start to deal with the issue at hand. I want to apologize to walterbyrd that this took the thread away from the original question, but it might be worth it. That is something he has to say.

So let's continue with planning and evaluations of possibilities, and also checking in what ways this will affect the future of FreeBSD.
 
Crivens said:
So let's continue with planning and evaluations of possibilities, and also checking in what ways this will affect the future of FreeBSD.
Click to expand...

I believe that systemd is actually hurting Linux, not FreeBSD. Why?

The "One Linux"/"Year Of The Desktop Linux"/"Linux Everywhere" mentality is hurting Linux more than it can ever hurt FreeBSD. Linux didn't suck when these mentalities didn't exist and when distributions focused on making themselves unique. Now, Linux really sucks, largely because of this "One Linux" mentality and the poor development decisions made as a result. (or at least in my opinion)

As a result, people who are unhappy with systemd and the "One Linux" ideology will want to switch away from Linux. I read one article that said FreeBSD is gaining interest with cloud providers because of WhatsApp's servers run FreeBSD, but I don't think it's WhatsApp driving this interest as much as the resentment of systemd and the "One Linux" ideology.
 
Crivens said:
In the end, we have three options: Run, Fight, or Hide.
Click to expand...

I'd suggest to hide. If disappointed Linux devs came to "improve" FreeBSD (or other BSDs without a definite "that won't ever happen" strategy), we can all imagine how FreeBSD will continue. Everyone can be the next Poettering.

2016, the year of the FreeBSD desktop? For some of us, indeed. I'd miss it.
 
Cthulhux said:
I'd suggest to hide. If disappointed Linux devs came to "improve" FreeBSD (or other BSDs without a definite "that won't ever happen" strategy), we can all imagine how FreeBSD will continue. Everyone can be the next Poettering.

2016, the year of the FreeBSD desktop? For some of us, indeed. I'd miss it.
Click to expand...

Jordan Hubbard is a Linux developer? WHOA
 
TeamBlackFox said:
Dealbreaker for me on using OpenBSD on the desktop is the lack of Nvidia support. I won't use AMD or Intel, the former for piss poor drivers, the latter because their GPUs are as slow as frozen molasses.

I agree though, I frankly get pissed off at the number of OS X users on here who are so uninformed about the FreeBSD desktop/workstation quality. Its obnoxious, and I wouldn't care less if people used OS X if they kept their damn opinions to themselves.
Click to expand...

Cthulhux said:
I'd suggest to hide. If disappointed Linux devs came to "improve" FreeBSD (or other BSDs without a definite "that won't ever happen" strategy), we can all imagine how FreeBSD will continue. Everyone can be the next Poettering.

2016, the year of the FreeBSD desktop? For some of us, indeed. I'd miss it.
Click to expand...

I'm glad to see some serious discussion of these issues that I myself have been wondering about for the past couple of years. I hope the sun never sets on FreeBSD, but I am preparing in the event that it does.
 
Yes, hope for the best but prepare for the worst.

FreeBSD, and I think any other *BSD, has one huuuuge advantage when it comes to those "hackers" tinkering with the codebase. The OS is one unit, kernel and user land in one. The commit priviledge needs not only to be granted, but it can also be removed again, and commits might be rolled back also. So I do not see the chance that some badly designed component would creep into the code base and stay there, together with it's comitter, for very long. Well, at least as the keepers of the gate there are keeping to the path.
 
I think Crivens made a good point earlier that was lost in the shuffle, in that if you are talking about systemd in terms of impact on FreeBSD, you really have to specify whether you mean as it exists - currently - in reality as an init system that has subsumed system services, or if you mean the systemd philosophy (which is definitely not the UNIX philosophy). Also, if you are going to discuss Jordan's presentation, it's important to separate launchd from systemd. I've been trying to educate myself on these issues, and have some useful links/videos I'd like to post here to help guide the discussion.

First, what is the systemd philosophy? As I understand it, it has not changed much since Poettering gave this presentation. The title gives you a good idea of where it's going, being "Beyond Init".

Second, the GNU Hurd developers have had some interesting discussions, along with some ranting, about the implications of systemd for OSes that are not Linux (or GNU/Linux if you will, since this affected Debian GNU/kFreeBSD too ;)). http://darnassus.sceen.net/~hurd-web/open_issues/systemd/

Third, I am not an OS X or Apple hardware user, so I'm not familiar enough with launchd to really comment on it and how it differs from systemd. But, after watching this presentation, I think I have a decent handle on it:
 
Status
Not open for further replies.
Back
Top