I think I meant to say, that for free users iX only offers they will voluntarily allow themselves to be held accountable to some degree. No hard accountability like with paying customers.
As you correctly stated, there are different types of accountability. So what accountability does mean in relation to free users?
I had to find some breakdown what “accountability” even means because its getting confusing for me.
Responsibility (they were supposed to do something)
Answerability (they must explain or justify what they did)
Consequences (good or bad results based on their actions)
So for free users:
iX are responsible for developing TrueNAS so it works for its users. Even free users.
If iX does something “bad” they explain it here on forum or in podcast.
And consequences for them are complaining posts like this one and dissatisfied users.
iX could easily stop holding themselves accountable in this way. They could just stop explaining their actions, they could start deleting critical posts or they could shutdown forum altogether.
But iX lets themselves be accountable in this way.
You can say its not enough or something. But this is the level of accountability they are willing to accept from free users.
Higher levels of accountability are for paying users and the highest for shareholders and such as you correctly said.
So you can be dissatisfied that only accountability available are answering questions and complaining posts, but this is all you can buy for free.
This video was released over two years ago with a TN UK reseller and watching it back now I find the answers to the questions towards the end very telling about the whole TrueNAS CORE, SCALE virtualisation conversation. I’d strongly recommend anyone commenting or interested in this thread watch the video.
Actually, I would prefer it to be unnecessary - accountability is only necessary in an adversarial situation.
Were it a cooperative effort where user input was welcomed, and the views of knowledgeable users were genuinely solicited not only for feature requests but also for opinions on their draft plans, then there wouldn’t really need to be any form of accountability.
I genuinely believe that iX do welcome feedback and feature requests, and as I pointed out there is evidence that occasionally this does have an impact (6%) …
… however this is completely different from genuinely listening to these suggestions and allowing them significantly to influence your future plans (e.g. adopting the suggestion made many months ago to deliver parallel running for new technologies) …
… and that is completely different from actively involving knowledgeable community members in a genuine discussion during the early development and review of version plans. This would be a recognition that a variety of viewpoints, and independently generated ideas, and the subsequent debate about the pros and cons of these, can be extremely synergistic.
But in the end, it is iX that is funding and investing the development time, so the final decisions need to be theirs.
Fair enough… but this totally misses the point that the issue is NOT with introducing “Instances” system containers with LXC/Incus as an official alternative to community-developped “Sandboxes” with systemd-nspawn. There’s nothing wrong with a new feature being experimental.
The issue is with also introducing Incus as a manager for KVM virtual machines, introducing it as “experimental” with an incomplete feature set and an unstable work-in-progress API while removing from GUI the stable libvirt VM management which was there since Anglefish and was relied upon by a sizeable number of users…
When the Incus VM management part is stabilised, and you can hopefully provide working automatic migration, then it will be very fine to remove libvirt and rely solely on Incus for both system containers and virtual machines. Until then, the only sensible path is to retain the old and trusted libvirst system alongside the new and shiny, but still unstable/experimental, Incus system—and if it could be done relatively quickly in 25.04.2 it really should have been done in 25.04.0. Cutting that corner too early was a “saving that cost more than it saves”.
My thoughts exactly. And no, even without of the VM feature, I would still not be updating as I wait for the conservative version (at least) and that will be at least .,2. I’ve never had any trouble with any update of Scale! I just wait. Mindlessly is often the correct word, people seem to update on day 1 many times or even pre-release. I don’t mean the technical kinds.
I do really like this idea and think that would give people pause esp. less technical users. Maybe they wouldn’t update to the .0 version. Maybe they still will, but then it’s really on them.
The people who you need to make this clear to is the IX people on social media like reddit. They STILL merely say look at the great new release we are announcing, and say NOTHING about that software status page, the risks, etc. That’s on IX. By the time anyone with sense posts guys please wait, it’s the first version, thousands of people have already upgraded.
And within docker, you have the built in apps, you have create a custom app, and, you have raw yaml. And within that, you have people running dockge and portainer. All docker, but all requiring different methods for some things.
I agree that your conservative definition has been spot on. I follow it every release and simply do not encounter the hundreds if not thousands of posts of people with trouble, those do not affect me. I’ve been happily on the conservative trail and it has served me well. The issue is more the social media posts.
Social media is to generate early adopter (new user) interest, not provide technical advice to Conservative users. We have to do both to function as a business.
The middle ground is that the blogs they typically link to have “experimental” and status page references.
I would observe that new users have less issues that old users doing updates to complex setups.
We have been working on Goldeye and making that virtualization better… after resolving/testing the issues there, we backported to 25.04.2. Some of the middleware code had been maintained which made it easier.
We would never introduce the capability in .0 … it would have to be in BETA and RC1. So, our choice was really whether to use Incus for VMs or not at the start of Fangtooth.
With 20/20 hindsight we didn’t make the best choice, but with the information we had at the time I can’t say I would change my mind.
Everyone is way smarter with the benefit of hindsight. No arguing about that…
But hidsight is not required to see that replacing the KVM/libvirt system which works and which is relied upon by (some) users by an admittedly experimerimental KVM/Incus system is a regression. No change of mind about that? Really?
And then, if the only way for users to avoid the regression is to NOT update to Fangtooth (must… resist… marketing…), that means that Electric Eel should be fully maintained and updated until Goldeye is out and stable. I.e. “Goldeye .2 release deprecates Electric Eel”, not “Fangtooth .2 deprecates Electric Eel”.
We’re back to arguing for a yearly release cycle, not semestrial. Or a LTS release for conservative users.
For VM users specifically we agreed to a migration from Electric Eel to Goldeye.
That is part of making the Instances code “experimental”…
Edit: We will have to review the best update path after we have experience with 25.04.2 and the re-add of “virtualization”. The commitment is that there is an update path with is good for conservative users with VMs. I don’t yet know what the Engineers will recommend (and neither do they is my guess).
I have been hesitant to speak up in this thread because I’m not as technical as most of the folks in this forum. That said:
I’m glad to see the potshot comments have subsided today. One of the highlights of TrueNAS for me has been an inviting community and our comments to each other (and to the company that develops TrueNAS) make or break that new users decision on whether we are worth investing.
I too feel the whiplash of technology direction changes
I am able to weigh that negative fact against XI’s level of communication and interaction with the community which I have never observed with any other major software provider
TrueNAS’ primary function of serving and protecting my data had never failed me in the last 13 years
If you are looking for an admission of making a mistake on the VM piece, I’d consider 25.4.2 as that admission. I would think any other large software vendor would either give the middle finger or wait more than a year to compromise but I don’t believe any company would admit to a mistake.
Re corporates admitting to mistake - there are many good and very public examples of exactly that, including from US tech giants like Apple (Maps debacle) & Netflix (billing issues) and many others. Obviously an apology can be a strong statement, but worth nothing if it doesn’t come together with what really matters - a demonstrated willingness to learn from the past and review/take actions for the future. To my mind, that is what this, and similar recent threads, are all about. Unfortunately iX, while seemingly actively engaging and participating, are showing little sign of that (IMO) - or at least opting not to speak of it publicly - and that is clearly provocative to some.
I do believe that iX should be commended though for their investments in engaging with their user community. The roots in FreeBSD and open-source projects shines through clearly. I think part of their issue at corporate level is split-brain between engineers who want to work on and sell technology that they like, vs trying to scale (no pun intended) a long-term business with corporate enterprise customers that care only about the outcome and are completely indifferent to the actual underlying technology, as long as it works to its intended purpose and is secure. Totally different starting point and approach needed then.
It may be to encourage new adopters I suppose, but basically, post comes out that Fangtooth has been released with these great new features, and everyone posts how they are updating now. That’s the reality. It may get some new users, but generally, most are existing users. The one I have experience with is reddit.
Is it really that difficult to add a small paragraph about existing users, and link to the software status page to reference recommendations? Along with the changes to the Truenas update screen itself with some text?
