At some point it’s pretty clear that the plan for that path was to be Incus. If not…you wouldn’t have integrated it.
But what reason do we have to make that migration rather than just migrate to something else?
Incorporating Incus represented a clear (and desirable) direction. Moving backwards represents…something else. A lack of direction. Chaos. Massive breaking changes with every release. I’ve been a FreeNAS/TrueNAS user since ZFS was incorporated. I was fortunate to skip the release-that-shall-not-be-named. I’ve been through the various other technical transitions…and they’ve been fine. Maybe not always ideal - but the world isn’t ideal and iX doesn’t control everything.
This feels more like pointless, wheel-spinning churn for churn’s sake without a driving external force or change in circumstance to dictate it. Just catastrophic for confidence in the product, its direction, and its ability to continue the reliable experience that I’ve known for…a long time.
The value proposition of “trade off a bit of flexibility in favor of an appliance that puts a convenient UI around what I really need, with updates handled, no temptation to muck up the base system with cruft, etc. so that I don’t need to deal with fixing/migrating/etc. stuff as much”…is just…gone. Instead there’s more churn to manage - including churn that I just don’t want or understand, and which is poorly communicated from management.
I feel like I’m not unlikely to be hitting the exits soon as well. I don’t know how to square the circle between what TrueNAS is, what management claims it is, and what management claims that they want it to be…and that’s before I even get to considering my own requirements.
I fully agree with this - anyone who was aware of the risks, and who read the Software Status Page and the Release Notes and who had patience, will indeed have waited for 25.04.2 and should indeed have an easy upgrade.
But anyone who decided to help out by being an early adopter or who wasn’t clued in enough to take the T3 “upgrade now” comments with a pinch of salt…
Also note that “people who were patient” had already been waiting years for containers in TrueNAS Linux before they rolled out a solution and then decided to completely rewrite the implementation within…weeks? of release.
What reasonable indication is there that “this time we’re really set”? There’s nothing to trust other than just not using most functionality…and at that point what’s the product even offering as a value proposition?
Exactly - once users lose trust in you, it can be very difficult to win it back.
You can call it that if you want - personally I call it an attempted wake-up call (from someone who has genuine relevant professional expertise and been around long enough to have seen similar things and to understand the consequences) to avoid further bad consequences, particularly since I give a lot of positive suggestions for a better way forward and also make positive comments where appropriate too.
I’d suggest that the reasonable range of risks extends to things like “some functionality is missing”, “migration tooling is incomplete”, “possible stability issues” and things of that nature. It’s not unreasonable (with proper disclosure and the ability to run in parallel with the old, of course…) to launch new functionality with some rough edges. It is, however, to clean-sheet rewrite with a completely different framework immediately after release, with no changes in the status of the initial underlying tech that was chosen.
Did Broadcom acquire VMware for approximately $69 billion, and get an OS that isn’t even ready for release – or is this invented criteria (and its use here for criticism) instead just an insulting exaggeration? Is there a more positive way to express a desire for help, without insulting someone because they haven’t provided it already?
For reference, here’s non-hysterical people responding to a feature being marked “experimental”. Not the end of the world again. Not even a biggie at all.
“where is the disk… I know what you guys are thinking where is the disc. Currently, one of the drawbacks of Incas being so experimental is there’s not a lot of customization. It’s going to happen in the future but for right now [the disk] just is where it is on the system…”
– https://www.youtube.com/watch?v=z2UUBW1eHsk&t=273s
… that’s kind of weird but as I said it’s still experimental. I expect a bunch more features and polish will go into that. It’s definitely a pretty good experience for anyone who just wants to quickly run a bunch of containers. It will work just fine for anyone who just wants to runs a bunch of VMs – that will be just fine too.
it looks like you can upload ISOs. You can upload raw disk images. So, that’s going to be convenient for things like Windows.
– https://www.youtube.com/watch?v=ENrQ1RWKZuI&t=1592s
But, you’ve always had that power, my dear. You’ve had it all along. Just click your ruby slippers and repeat, “There’s no place like the version I’m already running. There’s no place like the version I’m already running.”
Aggressively upgrading to the latest version, and then complaining about the loss of “stability” is high up there in the list of unreasonable made-up criteria then used to criticize others.
You know you have a perpetual free license to run TrueNAS Core or Scale Electric Eel until the end of your days, right?
Every standard of system administration strongly emphasizes a risk-managed approach to change. Any significant change to an information system, especially an upgrade, should be preceded by a thorough understanding of its potential impact on security, privacy, and functionality. This understanding is primarily gained through research, planning, impact analysis, and rigorous testing in a non-production environment before deployment to production.
If your criteria for upgrades are instead that they should be as carefree as chain-watching a TV series on NetFlix and just hitting, “play next episode”, you’re probably indulging in an unreasonable expectation.
Interesting strategy. Did endlessly insulting and berating your ex-husband get him to finally fly right and do all the things you were nagging him to do? Or was it actually demotivating, and he stopped caring about your complaints entirely?
I registered mainly to add to this discussion and also to report issues with the new VM mess. I really, really think that this was a very poor leadership decision that shouldn’t have happened in the first place because it was easily avoidable.
I can understand a bug in a 3rd party sub-component and some remote-possibility edge scenario. But this was such a glaring mistake, that I am not sure how it got past people in charge. The obvious solution? Leave “classic” VMs as they were and add experimental Incus side-by-side. That’s it. An intern could have come up with that solution and I bet there was such idea. But someone thought otherwise, and someone else above them green-lit that decision. Not sure I’d like to have those people in my team.
I too work in a software development company. We serve single end-users, enterprises and everything in between. I understand that there is some leeway and new features and fresh releases are “tested” on end-users first before they are deployed to enterprises where stability matters (in addition to staggered release). However, never ever have I seen removing a basic working functionality, and putting a half-cooked semi-solution that doesn’t really work in its place.
Up until now, I have been through the following:
Read release documentation, prepared for manual VM re-creation, saved all necessary configuration information. It gave me idea that aside from the manual steps, everything will work.
Updated 24.10 to 25.04.1. Tried to recreate VM in Incus… no go. It was a total mess. Found a YouTube video with the same issue (Windows guest VM) but without any solution.
After some tinkering, I decided to wait for 25.04.2. I deleted the Incus VM (while keeping block volume on disk) to prepare for new release.
The day finally comes. I update to 25.04.2 only to find more errors in the Virtual Machines section and inability to recreate a VM with the same name.
The waiting continues, trust diminishes…
The problem with all this is that you are in the trust business. You are not here to have the fastest networking, the ability to handle the most number of disks, the most performant virtualization, the prettiest user interface, etc. You are here to provide a solution that people can trust with their data. I believe that trust was… not lost (far from it), but weakened by this whole mess.
Over the years, I learned that a technical issue is much (much!) more understandable than poor managerial decisions. Even if it leads to the same consequence (e.g., VMs not working). Usually, that is also fixed quickly (if possible). What happened here was plain and simple a very poor descision, in my opinion. I’ve travelled some path (NAS4free > XigmaNAS > FreeNAS > TrueNAS Core > TrueNAS Scale). This is the first time I am thinking about what else is out there…
TrueNAS 24.10 had what iX called a production-ready virtualization environment. 25.04 took it away and replaced it with one they called “experimental.” Is it your position that that’s an invalid criticism? That the expectation that, before implementing the new system, iX would have looked into it to determine that it was worth using, is unreasonable?
When iX actively markets this feature, no, it isn’t the end of the world, but it’s at least a bit of a failure on iX’ part. To fail to even document how to migrate? Now it’s a major failure.
Note that even the responses you quote as illustrative of “non-hysterical people” are quite simply incorrect. No, that customization isn’t going to happen in the futire. No, “a bunch more features and polish” will not go into it.
What do you propose? Asking nicely? We’ve tried that, a lot, for a long time. The frustration you’re seeing now and reacting so hysterically to is, in part, because we’re fed up. Will iX get the message this way? I honestly doubt it, but we don’t really have a lot of other options–other than just to drop TrueNAS entirely.
In short, there were 2 bad decisions which amplify each other.
iX shouldn’t have ripped out the old framework for Incus cold turkey with no developed migration path in a release.
iX ALSO shouldn’t have gone from launching and promoting new capabilities and framework to ripping them out without a disclosed migration path in 3 months.
These are 2 rash decisions that give the impression of chickens with no heads, running in circles, expending a lot of energy, and creating rather than solving customer problems. It’s not sustainable, and doesn’t foster trust (or quality software). It’s ready fire aim that directly undermines the core value proposition on-offer.
Clone and Promote the zvol out to somewhere else, re-enable the cache settings and set the volmode to default, then recreate the vm “classically” and point at the zvol.
You should then be able to delete the old incus vm in the UI.
For reference, its similar to this, but in reverse:
Correction. Having to manually migrate every VM right after the update. With no opportunity to migrate gradually. And as was stated up in the discussion, that was a possible approach.
Also, your logic looks like:
IF “Removing working features is ok” AND “Adding experimental features is ok” THEN “Replacing working features with experimental features is ok”. But it is not. It’s not ok. Especially when there was an opportunity to not do so.
Nice one! Your advice to every single truenas user is to do research, planning, impact analysis, and even testing in non-prod before every update. Very debatable, but understandable. What about iX? Should they have done an impact analysis in particular before rolling out the update?
This truly is baffling. Either they didn’t do any homework on Incus before deciding to integrate it (and therefore failed to realize that it couldn’t be suitably integrated), or they haven’t done any homework on it since semi-integrating it, to determine how to completely integrate it. Either way, most of the 25.04 development cycle is completely wasted effort they’re just going to chuck in the bin. Leaving aside the user frustration seen here and elsewhere, that has to be demoralizing to the devs.
I’m kinda in the same camp as you, but I think some of the others have been burned a few too many times as this is now the second or third such weird update I’ve seen during my (compared to some of the old heads) short time on the forums.
I think they are actually fairly upset & frustrated because they’ve seen even more than that. Which I guess makes sense. Originally I had thought this was just a bash & hate thread, but I’ve been reading through & emotions seem high because there is a lot of care on the community side, and quite some history of such descisions by IX…
For example, TrueNAS hosts a Pi-Hole installation guide but there is no link to it from my NAS when I select it in my Apps catalog. Instead, my NAS only links to the pi-hole.net homepage, which is useful but not as useful as the awesome start of a guide your team is hosting at apps.truenas.com/catalog/pihole/.
Could future revisions of the Apps catalog point to these App installer pages that otherwise might go un-noticed since they are not referenced in the GUI?
Also, in this particular instance, I hope folk can contribute further to the above to allow for things like a individual IP address for pi-hole so it can behave like a RasPi install? Etc.
Even the release that shall not be named was more comprehensible - an ambitious reimagining that relied on external stuff that just wasn’t quite getting where it needed to be.
