A few weeks ago we talked about how iXsystems, the company behind TrueNAS CORE and SCALE, has all but confirmed that its FreeBSD-based CORE product will be put in maintenance mode, while the Linux-based SCALE product will get all the attention and focus from here on out. In an interview with Blocks & Files, the company gave more insight into this choice.
“We had a huge chunk of our engineering staff spending time improving FreeBSD as opposed to working on features and functionalities. What’s happened now with the transition to having a Debian basis, the people I used to have 90 percent of their time working on FreeBSD, they’re working on ZFS features now … That’s what I want to see; value add for everybody versus sitting around, implementing something Linux had a years ago. And trying to maintain or backport, or just deal with something that you just didn’t get out of box on FreeBSD.”
“It’s not knocking against FreeBSD. We love it. That’s our heritage. That’s our roots, I was on the CORE team elected twice. So believe me, if I felt like I could have stayed on FreeBSD for the next 20 years, I would have absolutely preferred to do that … But at some point, you gotta read the writing on the wall and say, well, all the the vendor supported-innovations are happening on the Linux side these days.”
[…]BSD aficionados don’t like this change. Moore said: “Talk is cheap and complaints are free. You know, everyone loves to complain about it. But … if people wanted to push FreeBSD forward for the last 15 years, they would have.”
↫ Chris Mellor at Blocks & Files
Above all else, my personal north star is choice, especially in technology, and as such, I want iXsystems to keep focusing on FreeBSD so that not everyone is using Linux for server- and server-like workloads. The fact that TrueNAS was a FreeBSD-based product for this long was amazing, and I would definitely have preferred if it stayed that way for many, many more years to come.
However, I don’t think the people of TrueNAS are saying anything wrong or outrageous here. They’ve got employees to feed, and the money is in Linux, not FreeBSD. If they spend more money, time, and resources on getting FreeBSD on par with features Linux has had for ages than on actually developing their own product – TrueNAS – then they’re fighting a losing battle. Honestly, I’m surprised it’s taken them this long to take this controversial step.
All we can hope for is that the things they work on, the features they develop, will make it to FreeBSD regardless.
Years ago I made the exact same decision myself for the same reasons. But I still find it sad that alternatives always seem to get crushed as dominant solution keep becoming more dominant. In the old unix world interoperability was king. But for better or worse a lot of linux development promotes linux only interfaces without regards for the broader unix ecosystem.
I hope FOSS doesn’t become a linux only mono-culture, but being the best supported FOSS platform gives it a large advantage and its easy to why most FOSS users are turning to linux at the exclusion of others. It’s not that they don’t have merit. In some ways the BSDs have more merit, but when you exist in the long tail the network effects are constantly working against you unfortunately.
I’d really like to see Haiku get some more love. I’m a greedy guy, Haiku has already surpassed the expectations of nigh on every other non-UNIX FOSS OS, but i still feel there is work to do.
Linux truly has the edge when it comes to embedded and server workloads, but it just never has quite got the edge when it comes to desktop, and that’s because the markets require very different things. Servers often require a lot of customisability, whereas desktops require “everything working, and working now”
What i liked about BSD, is the fact that at it’s core, it was standardised. Linux just doesn’t have that solid unified base, because Linux is after all just a kernel, you have to roll your own userland and even some kernel-side stuff (INIT being a prime example). FreeBSD doesn’t have that problem, as it’s a solid base already.
Maybe it’s time some guys branched BSD and made it more “linux compatible”. I don’t like it, but then there’s a lot of things that are necessary that i don’t like
The123king,
Even as a linux user, I kind of agree with this.
Yeah. This is one of the things I find a bit unfortunate even though you may be right. It’s like assuming technology doesn’t have merit or deserve our attention unless it mirrors and/or serves the leader. I find this philosophy disappointing, but for better or worse the network effects tend to dictate success or failure far more reliably than other qualities and it does seem that those who won’t follow the mainstream path often become irrelevant.
10 years from now docker or whatever containers they are using on Linux will have evolved into some mongrel … and FreeBSD will still be providing stable jails.
It has already happened before, there are already LXC containers, Kubernetes, Docker, lmctfy, rkt etc etc….
You should not mix things that are not of the same kind 🙂
Anyway, doing containers on Linux did help to refine the concept in a way that would not have been possible on a BSD or Solaris or any other Unix system with pre-existing container concept.
There are some areas it’s necessary to duplicate linux like graphics / drm for xorg and wayland. No one has ever gained marketshare from copying another guy. We need exclusives in the BSD world like freenas/truenas once was.
laffer1,
Indeed, over time the dominant APIs become the defacto standard that everyone must implement.
*caugh* linux *caugh*
Both microsoft and apple copied xerox GUI innovations too.
It’s a valid point, but if you’ll allow me to conflate metaphors: going against the stream is an uphill battle.
Edit: isn’t it weird the way we can type the wrong spelling in the moment and it only looks wrong upon re-reading: caugh->cough.
Edit2: Both spellings are crap, it should be “coff” haha.
“Maybe it’s time some guys branched BSD and made it more ‘linux compatible’. I don’t like it, but then there’s a lot of things that are necessary that i don’t like”
Well, I am sure you are aware of Linuxulator ( the ability to run Linux binaries on BSD ). You can also run Wayland on FreeBSD:
https://docs.freebsd.org/en/books/handbook/wayland/
Perhaps most interesting, you can run Podman and OCI containers on BSD:
https://www.freebsdsoftware.org/sysutils/podman.html
Or you could go the other route and use Chimera Linux which is putting the FreeBSD userland on the Linux kernel:
https://chimera-linux.org/
Probably the real problem was the bad quality of bhyve versus kvm. They now want Truenas as a virtualization host but with bhyve there were problems in performance and stability (I know because I have experienced them directly).
Bhyve has been running my pi-hole for over a year with no issues… it seems to me the main issues were expecting too much from it. And more problems originated from thier middleware than from bhyve itself…
The main issue is then that it’s possible to expect too much from bhyve in a professional setup. So, it’s probably not too much of a surprise that moving to KVM kind of start making sense at one point.
bhyve works pretty well in FreeBSD. There are areas it needs to improve, but I’ve had great luck using it as a replacement for vmware for running package build nodes for MidnightBSD.
I also was at a company faced with FreeBSD vs Linux, not just for our SASS offering, but also for embedded hardware appliance. SCO lawsuit scared the C suite to FreeBSD ( to give you an idea how long ago it was). So the SASS was required to be on FreeBSD, despite its performance loss for the application benchmarks we did. The embedded hardware was more difficult to do on FreeBSD due to drivers and third party support, even back then. We basically got it working with either, and could have launched it with FreeBSD. I left the company before it was launched, so I’m not sure how they convinced the C-suite that using linux wouldn’t cause a legal nightmare, but they apparently did. The cost of supporting one version on linux and another on freebsd would have been too high for them, and replacing one with the other in the field would have been a nightmare for us. I imagine its only gotten worse on these terms.
In an early demo one of my colleges demoed the FreeBSD version, and forgot to swap out the loading screen with the BSD Daemon and the Fortune 500 company marketing execs freaked out upon seeing it. Marketing didn’t like their brand associated with a devil, apparently.
Most of the BSDs don’t even use the BSD daemon for their logos at this point. Mine uses a cat.
Twenty years ago, Kris Moore and the rest of their team put tremendous effort in lowering the barrier to entry for FreeBSD on the desktop, and irrationally stuck with the project as a base during the down years. This while FreeBSD developed a reputation as being a side project by people enamored with Apple who only ran it in a VM on their Macbooks while basic things like Intel graphics and wifi were broken for years on baremetal installs. This is about the time that I stopped using FreeBSD.
In other words, PC-BSD and iXsystems ate their own dog food long after it expired, and somehow managed to squeak a living out of it. That’s dedication and deserves kudos.
Upstream is Linux. OpenZFS is Linux. Hardware drivers are in Linux first. The pragmatic choice is to rebase on Linux. Why waste resources duplicating work on FreeBSD at a small company when trillion dollar companies aren’t doing the work either?
Debian is a fine platform; I’ve standardized on it both professionally and on the home network.
PC-BSD was the only UNIX-like desktop system that i had zero issues with. It worked. Sure, there were flaws in compatibility and software support, but i never had it just crap out on me like the various flavours of Linux did.
Granted, this was a long time ago, Linux may have indeed improved a lot on the desktop nowadays. But i spend my working life fixing modern tech, and don’t want to have to do that when i’m at home, i have too many vintage systems i could be spending my time fixing instead.
> Granted, this was a long time ago, Linux may have indeed improved a lot on the desktop nowadays. But i spend my working life fixing modern tech, and don’t want to have to do that when i’m at home, i have too many vintage systems i could be spending my time fixing instead.
Same, except that’s the reason I am now on Debian. It just works and there are no surprises.
I made an account just to thank you for this comment. One of the best and most concise/precise summations of the saga,
Oh wow, thanks for your work!
Please stick around, OSNews is an great community of people who are opinionated but passionate about operating systems and developments in that space.
I’ve been here for over 20 years and remember the early days of PC-BSD fondly. Best wishes moving forward.
If FreeBSD is slowly dying, how’s the outlook for NetBSD, OpenBSD, DragonflyBSD and Illumos then?
j0scher,
Well, it’s only a single data point, theoretically it could just be a one off casualty and not a sign of prolonged attrition, but it is a significant project and if they’re experiencing these issues I don’t imagine they are alone. IMHO it’s important for FreeBSD and others to remain viable in order that diversity in FOSS doesn’t totally collapse.
OpenBSD has a vastly different philosophy and doesn’t suffer from the corporate pressure that FreeBSD did and still does. To my knowledge there are no corporations or other commercial entities pressuring OpenBSD or stealing away its developers’ time and resources, and having a BDFL like Theo (regardless of anyone’s opinion of his personality) pretty much ensures it will continue to grow and improve rather than stagnate and wither like FreeBSD seems to be doing.
OpenBSD diverged enough in terms of goals and funding, and did it long enough ago, that it’s probably the healthiest project of the ones listed. There are no external forces applying requirements, and with the clean code base, almost all of the *BSD momentum and innovation has been there for a while.
There were funding problems a few years ago, but enough large players make yearly contributions through the Foundation that I am not sure they are struggling with electrical bills any more. The community has invested a lot. I have been a monthly donor for a decade even if it’s not my daily driver any more. Microsoft and Google have given huge donations. There is an elephant in the room that has yet to donate, who fortifies their consumer OSes from OpenBSD code, so I stopped donating to their profit margins. But overall, the Foundation sets a yearly goal, and it is exceeded by private and corporate donations.
This is very sad to see. Longer term than iXsystems needs to worry about, we do need a vibrate OSS community as an alternative to Linux.
At some point Linus will step down. He has basically controlled the project as a single semi-benevolent-dictator. It can’t be Toooo long in the future when he chooses to step down, either for health or well deserved retirement.
When this happens I fully expect a feeding frenzy with various players trying to be come the “True” linux.
Linux Is dominant, but so was Commadore once… I like to have options as to what comes next.
my fear is that when Linus steps down, technical debt in Linux will skyrocket.
He is the one killjoy who looks past the hype around any given feature and focuses on maintainability going forward. he blocked quite a few features most people would be very eager to have, just because code was of dubious quality or features were not well thought out. or it would have become a maintenance nightmare going forward.
[email protected],
IMHO it isn’t “technical debt” that we need to worry about. Most kernel contributions are not his and many people have the technical chops to do that job. What might change more dramatically though could be management style. A change could be good or bad, we just don’t know yet and I’d argue we have more to worry about on the project management side than the technical side.
I’ve just assumed that IBM/redhat will take over linux at that point.
laffer1,
This thought crossed my mind too, but I couldn’t decide if the future of linux being taken over by a for profit company like IBM was too preposterous. It’s hypothetically possible though and a change in leadership would be the best opportunity for this type of power play.
Edit: Just to be clear, since “linux” is ambiguous, I’d like to clarify that in this context “linux” refers to the kernel itself rather than a linux distro.
Just look at the amount of projects that live and die based on Redhat money flowing in.
If every kernal maintainer suddenly starts being a Redhat employee, Redhat defacto run the Kernel.
I assume they will make a play for it.
Redhat will fork off, Ubuntu will, Amazon will, and probably a few other players. Maybe even Microsoft!
It won’t be a single big bang “I now own Linux” but there will be a kernal branch that add a new driver set fir their cloud hardware which isn’t merged into base.
They’ll be forced to make it all open source as per the licence, but as we see with distros, different orgs have different priorities. Just imagine a systemd-esq debate at kernal level…
The only thing holding Linux kernal together is it’s dictatorship. (I don’t mean that to be as negative as it can imply, but ultimately one person is in control)
Adurbe,
Yes, I shudder to think about a Poettering type taking over. the kernel. He’d be terrible for uniting the FOSS community, practically the opposite. I know Linus has been callous at times but at least he embraces community building and that’s helped linux go far.
You mentioned the license keeping linux open, but this brings up an interesting point: Say redhat took over the kernel…would they then try to apply the source code redistribution restrictions they placed on RHEL customers to the kernel itself? As far as I’m concerned, IBM/redhat’s attempt to override GPL rights by way of punishing customers who practice those rights are incompatible with FOSS principals. But putting my opinions about it aside, what would actually happen to the linux industry if ibm/redhat managed to pull this off? The formally public source code would be taken down and redhat could retaliate against customers who redistribute the source. This all seems like a work of fiction right now, but I never thought redhat would start undermining GPL rights either even though it’s actually happening.
So the disclaimer is that I am a FreeBSD developer. FreeBSD is well designed and happens to work great on many settings. Linux started being not so well designed but people managed to rewrite big chunks of it to make it do things that FreeBSD either does or doesn’t. It´’s huge and unmanageable nowadays IMHO.
Most of his happiness seems to be derived from not doing any OS development at all anymore. Most of his contributions to FreeBSD were related to having us test and support new hardware. I wonder exactly what Moore is finding so nice about Linux that he couldn’t do on FreeBSD, but if he is really happy and developing a better product than what he had, I guess I will just feel happy for him. I don’t really use anything from IXsystems, but I certainly have no hard feelings just because they started working with alternatives.
bet they don’t use FreeBSD on daily basis or use it inside vbox. hope the next one wouldn’t be replacing ZFS with Btrfs or Linux with Windows although this one was pretty equal.
Just out of interest… _Blocks & Files_ is a sister site to _the Register_ from the same publisher.
This story is in response to an earlier piece of mine for the Reg:
https://www.theregister.com/2024/03/18/truenas_abandons_freebsd/
I’ve had no official response from iXsystems. I infer from this that they are Not Happy. But I did check with them before publication and got confirmation.