No plans for any of that nonsense. ISO files work perfectly fine 
Sure, somebody can go download the various open source components (minus any enterprise bits) and build their own from our repos. Up to them to build / maintain it that way, and of course comply with terms of the GPL.
Yea, that readme instruction needs updating. The repo has debian packaging files already included, which is how the build scripts actually do the build in the first place.
100%. AI is industrialized theft with a thin layer of “progress” slapped on. Running other people’s work through a fancy laundering program doesn’t suddenly mean you can call it yours.
Thats an interesting thought
Thats an interesting thought experiment to run through. Might need to talk about that on the podcast next time
Very interesting. Open source no more.
I have been following the TN succession for the last 10+ years and did notice the changes since 2024.
It remains to be seen what they will really do with TN free, but for the first time, the idea of starting to look for an alternative came up in my mind.
If it is really true that most of the testing will be done internally, instead of needing the users of TN free to do the final testing, this could be good, but it does signal a desire to leave those users behind, with minimal features and have a new “pay/subscription” for homelabs…
Let’s see what happens in the next year to see what their real plans are.
Hi Kris,
I am a fairly new member of the open-source community who’s been speed running the learning curve due to the global shift in the world of tech in recent years with a name tag we are all aware of. I am part of an entire army of individuals trying to take back control over our data sovereignty that is rapidly eroding away day by day. I know I am not alone on this trajectory and I won’t be the last one either. The signaling of this closed-sourcing step is worrisome and I’d like to bring attention to something that might not be taken into consideration.
I feel confident that there is a market in this community iXsystems should map out thoroughly before making further decisions towards a future that would inevitably result in distancing TrueNAS from people like myself who have jumped ships several times before.
The “mountain of self-hosting” is very steep, I have no professional background in IT but I have come a very long way to understand and practice the management of data and self-hosted services. I chose TrueNAS to be part of my software stack as part of an effort to move away from other popular NAS solutions (like many others did) over the last couple of years, just like I moved away from regular cloud providers before that.
TrueNAS is not a perfect solution for me but I see great potential and value in it where I stand currently. I am willing to support the project financially (so long I can trust it) but I don’t see any option to do so without the requirement of enterprise hardware purchases, which is beyond of what I intend to do. I may not represent a revenue stream that would compete with enterprise clients but I think there should be a tech company out there to recognize the power of its supporting community before it’s too late. I hope iXsystems is going to be one of those companies. Thank you for reading all this.
How can we be sure “stuff”, such as additional tracking/telemetry, isn’t getting added to TrueNAS as part of the new closed build system?
You’ve said the free version of TrueNAS will stay free. In my experience, when something is free, the consumer is the product. A cynic might see this move as the first step towards that kind of model.
Do you want the real answer? You have to trust your vendor who does the build. How do you know we’ve been using those build scripts for the past 15 years? You don’t have a way to really know for sure. Now of course you can validate the code repos (and you should) but network traffic is probably also a great place to monitor. Which I know we’ve had people doing since almost the beginning of FreeNAS. That said we (the developers) run TrueNAS at home as well and we value our privacy just as much as anybody else here, so you can say we’re pretty heavily incentivized to do the right things. Even if only selfishly for ourselves.
I mean, it’s been free for going on 20 years now, so what’s really changed? We’ve been pretty open that the real value we get out of the massive community is the largest test lab in the world. Users tend to find the most esoteric bugs in the wild that it would nearly impossible for a company of any size to find entirely in house. Hard to put a price tag on that kind of help
For me, that’s the bottom line.
I do see a bunch of people who are trustworthy. Full stop.
.
While pondering ways to keep “free” free and “paid-for” running the business,
Has iXSystems considered some sort of entry-level or middle-ground for those who need/want some form of support or at the very least a “once-over” with a technician to ensure their deployment is sound? I’m sure there are enough home-lab and small business users that would pay $100-500 for an hour or two of an engineer’s time for a sanity check (particularly in an emergency), or for “the most stable, feature-complete” version of the software they can get, even without support. Honestly, once it’s up and running, except for tinkerers, most people would set it and forget it (until a drive fails….).
As for the hardware mish-mash nightmare, I do see how that could be a problem. Perhaps for someone to quality for the various levels of support, they could buy the hardware themselves, but they’d have to adhere to an HCL (hardware compatibility list). For VM deployments, which could be another Pandora’s box, they would have to meet super strict requirements (pass-through HBA, NIC, ample cores, RAM, etc.) to qualify for support (if at all).
Or, for supported home-lab/micro deployments, iX could create/refer people to recommended hardware combinations, like a Minisforum MS-01 with 3x NVMe drives (one boot, two mirrored).
I have no idea how big their shop is in terms of the manpower, but I doubt that on-demand supporting individual users with no steady stream of money from them is sustainable. Even providing the HCL means that someone has to actually test the recommended configurations to both verify the functionality and to know their limits. I know from my own experience - albeit in the embedded space - how hard is to answer the “will your software run on this platform for this workload” question
Having run this exact consulting model for years, I can tell you: it’s not impossible,but it is a challenging model to offer.
The core issue is the massive gap between the work and the workforce. User demand is ‘bursty’ and unpredictable; a ticket starts as ‘TrueNAS is slow’ and often ends up being an audit of their network topology or hypervisor settings. You can’t predict when those fires will start, but you can predict that a high-level engineer wants a steady, reliable paycheck.
You can’t staff a ‘help desk’ with the senior talent required to troubleshoot ZFS metadata bottlenecks or VLAN mismatches on a ‘pay-per-incident’ whim. The overhead of keeping that level of expertise on standby during the quiet periods is what eats the margin. In this business, the cost of readiness is almost always higher than the cost of the actual fix.
We are not paywalling existing free features. Period. If it’s free today, it stays free.
It might help if you could give some examples of what exactly you mean by “features” here.
For example, I’m assuming that updates to newer versions of OpenZFS (once you’re satisfied they’re production ready) will continue as normal. However, if OpenZFS adds a new feature that will require GUI support, that requires developer time, and may be deemed Enterprise only leaving the rest of us to use the command line if we want to use that feature anyway?
I believe this is basically how it already works? I guess I (and probably others) just want to be sure that’s still the deal moving forward, because “features” is a bit of a vague term so it helps to reminded where the line is drawn or what you mean exactly.
For a community, transparency and engagement are always paramount.
Sufficient transparency allows users to use the product with greater confidence, clearly understand its development roadmap, and eliminate concerns about malicious code and information leaks. Engagement makes users feel respected, thus binding them to the specific product and brand. Frankly, TrueNAS falls short in both of these areas.
Once TrueNAS gains sufficient brand recognition, it can monetize its community version by offering features such as expansion packs, cloud services, and remote management, allowing users to save time by paying a fee. For tech enthusiasts, similar functionality can be achieved through open-source solutions, but at the cost of spending dozens of hours resolving minor issues. I’ve noticed TrueNAS is already making these commercialization attempts.
TrueNAS has always been a professional storage solution. Someone proficient in Windows or Mac for office work will find it difficult to quickly grasp the concepts of Linux-based TrueNAS and OpenZFS, and to complete hardware selection, system installation, parameter configuration, and performance optimization. This means TrueNAS is lacking in this area. I’ve noticed that TrueNAS is collaborating with third parties to explore ways to simplify usage, but this doesn’t seem to be part of TrueNAS’s overall development strategy.
For community users, the best approach for TrueNAS is to offer sufficient flexibility to attract users as much as possible, making novice users willing to invest time and money, while allowing tech-savvy users the freedom to build their own home labs.
We gave a pretty complete answer to this for tomorrows T3 podcast, but if you want the short spoilers first, the answer is “it depends”. Its hard to define a firm line in the sand, because some features straddle a lot of different layers.
Examples:
RAIDZ expansion: Little to no enterprise use case. We sponsored that work in OpenZFS, did the UI / Middleware work to expose it. Pretty clean cut, not a lot of maintaining burden going forward. Went into the free bucket.
Tiering: Much larger enterprise use case. We provided and upstreamed the building blocks of it to OpenZFS (zfs rewrite). But a very heavy and active area for us to build / maintain middleware and UI’s for, especially since it’s going to evolve in TrueNAS 27 to full featured file-based tiering, policy engines, monitoring, auditing, etc. Much higher upkeep cost for us with a very relevant enterprise-use case. That one is going to be put into the enterprise bucket.
Bottom line. We take each feature addition on a case-by-case basis to make the determination of which bucket it should drop into.
Would be nicer if I had some straight-forward pat-answer formula to spit out for this one, but reality is a bit more complex than that
When I first started composing my post, I had small business IT guys in mind (the one guy who has to do it all, or the IT consultant who’s running a dozen or more little companies’ systems). For comparative business models, I was thinking of Proxmox; somehow they’re managing to do it.
My $100-$500 thought has fleshed out to be more of a one-time certification as it were: “Yes your config looks good. Don’t change it,” with a possible one-or-two support instances. Unfortunately, that lower price is probably not worth it unless the “certification” was largely automated and support was limited to common failures (i.e. things that are easier to fix).
Knowing what I know now and after the time I’ve recently spent trying to figure out the best combinations of my new systems, I’d happily pay for a feature that could automate tests, simulate real-world workloads, run a burn-in test for a few days, and spit out a report showing me how my configured system(s) will be expected to perform and make any suggestions for improvement. Something like that might already exist within the TN realm, but I am not familiar with such a tool.
As for HCL, half of it’s done already. They know what drivers are included with the FreeBSD or Linux distribution they’re built on, so pick some of the “best” on the list and call it a “support contract-compatible” list. I certainly wouldn’t make it exhaustive to be “supportable.” It could end up only being 5-10 of any one thing (HBA’s, CPU’s, motherboards) or prebuilt enterprise configurations that are common (and easily sourced second hand). The latest white-box build some guy in his basement converted from a gaming rig? Not gonna touch that. A business wants to repurpose and certify an HP ProLiant or SuperMicro box? Let’s talk. iXSystems-recommended homelab kit? Good to go.
All that said, yes, it would certainly have to make financial sense for the TN guys to do any of this. Perhaps some of that low-hanging support work gets farmed out to trusted consultants, keeping the iX guys’ focus on enterprise tickets.
This is way harder than you think. Finding good consistent talent and validating that talent is no easy task. I am not saying it’s impossible but it also has a cost to setup and maintain.
I’m waiting for iX to say when they are going to sell the repackaged Enterprise edition (ie 1 download plus 2 years of updates) as a homelabber special. And at what cost, given it’s essentially free money.
All the code, none of the support (maybe not the best tag line).
I’d like that too. It would be a nice way to test enterprise features or to use it in enterprise without having to rely on iX for the hardware. I think there would absolutely be a market for that.
Not necessarily, if it drives enough people away from buying their hardware. 
We don’t really sell hardware.. we sell an appliance and the experience that goes with it.
