Q3 Command Deck Update Notification (09/29/2025) : New Apps, Curation Process and Folder updates

[Tuvar]

Any idea when VMs can be setup natively in HexOs? I know it can be done in TrueNAS but that kind of defeats the purpose of a single simple system if you need to go into the underlying OS it's built on.

[ParkyCat]

8 hours ago, jonp said:

We've discovered a handful of different issues with immich plaguing users. We are on the case and it's a priority issue for us!

Thank you, I am one of the plagued users and was wondering if I should wipe everything and start over. Mine installs ok but nothing happens if I click launch. Now I'll just wait 🙂 

Edited October 5 by ParkyCat

[Wiioscar]

8 hours ago, xaliax said:

Reinstall should work, I had the same issue and honestly didn't want to keep fighting, so I just deleted and reinstalled. It's been working and updating fine ever since.

Do you mean uninstall from the Truenas interface? Because I've tried it from Hexos and it's still the same.

[jonp]

12 hours ago, Tuvar said:

Any idea when VMs can be setup natively in HexOs? I know it can be done in TrueNAS but that kind of defeats the purpose of a single simple system if you need to go into the underlying OS it's built on.

We actually began the VM work last year but it started breaking as TrueNAS was making changes to the underlying stack.  As such, VMs were delayed as a feature. 25.10 promises even more improvements to the VM stack and API so we aren't going to resume work on them until that is stabilized. In the meantime basic start and stop functionality exists in our command deck so if you create a VM in TrueNAS it will show up for basic controls in our UI.

[Mascot]

7 hours ago, ParkyCat said:

Thank you, I am one of the plagued users and was wondering if I should wipe everything and start over. Mine installs ok but nothing happens if I click launch. Now I'll just wait 🙂 

It might be a long wait. It's been a "top priority" since June already.
 

 

[Todd Miller]

11 hours ago, Wiioscar said:

Do you mean uninstall from the Truenas interface? Because I've tried it from Hexos and it's still the same.

When you reinstall remember to remove the old dataset structure.  I forgot to do that and it made things more confusing for me as I did the new install.

[Wiioscar]

2 hours ago, Todd Miller said:

When you reinstall remember to remove the old dataset structure.  I forgot to do that and it made things more confusing for me as I did the new install.

I just tried it again as you said, but when I reinstall it, it seems to install with the same hexos structure as before.

 

[Todd Miller]

What mine did was the post_gres in an Immich dataset under applications and the rest under the Photo dataset.  Before everything was under the Applications dataset.  So if you have this structure you have the most current version and it was expected to clear up the current issues.  Please let us know if your dataset structure looks different.

First picture is my datasets.  The second is my windows SMB share.

[jonp]

13 hours ago, Mascot said:

It might be a long wait. It's been a "top priority" since June already.
 

 

To be fair, we actually did fix the issue referenced in that post.  However, additional issues have crept up especially with the recent updates to Immich in preparation for their 2.0 release.  One major change they made actually requires users to go through a migration process (a choice by the app developers).  It's important to remember that Immich is a free-to-use self-hosted solution, which means that sometimes they may need to do things like this as opposed to the slick polish you'd see on a commercial product.  That being said, we're still incredibly impressed with Immich and see it's future very bright.

[Wiioscar]

7 hours ago, Todd Miller said:

What mine did was the post_gres in an Immich dataset under applications and the rest under the Photo dataset.  Before everything was under the Applications dataset.  So if you have this structure you have the most current version and it was expected to clear up the current issues.  Please let us know if your dataset structure looks different.

First picture is my datasets.  The second is my windows SMB share.

THAT'S IT, it had the old structure, after deleting the previous one, now it works.

THANK YOU VERY MUCH

[Avari]

3 hours ago, jonp said:

To be fair, we actually did fix the issue referenced in that post.  However, additional issues have crept up especially with the recent updates to Immich in preparation for their 2.0 release.  One major change they made actually requires users to go through a migration process (a choice by the app developers).  It's important to remember that Immich is a free-to-use self-hosted solution, which means that sometimes they may need to do things like this as opposed to the slick polish you'd see on a commercial product.  That being said, we're still incredibly impressed with Immich and see it's future very bright.

All understandable, however like others said in a different thread while I don't have a problem with either waiting for an automatic migration or nuking my Immich install and re-installing I would expect some official indication (preferably in the deck) of what it is. Is an automatic migration coming or do I re-install?

[Cantos]

6 hours ago, Avari said:

All understandable, however like others said in a different thread while I don't have a problem with either waiting for an automatic migration or nuking my Immich install and re-installing I would expect some official indication (preferably in the deck) of what it is. Is an automatic migration coming or do I re-install?

They likely don't know yet, because they are still working on it.

Imagine they communicate that you need to start again from scratch and then they figure out automatic migration. Or even worse, they say that automatic migration is in the works and then they run into a huge issue that makes it unworkable. How would that be received?

[Maximus]

I'm not very technical and that Immich issue is quite scary, especially since Hexos is my main backup and the update did fail for me as well.
If there are risks and I should back everything up elsewhere to be safe, I would like to be notified.

I would also appreciate tips as to how to back things up from my existing Immich instance in such a way that reloading all my content after potentially doing a clean install would not be a complete pain.
The data is stored under thousands of folders, having to retrieve them all manually would be truly horrible 😛 

[Tuvar]

On 10/5/2025 at 10:25 AM, jonp said:

We actually began the VM work last year but it started breaking as TrueNAS was making changes to the underlying stack.  As such, VMs were delayed as a feature. 25.10 promises even more improvements to the VM stack and API so we aren't going to resume work on them until that is stabilized. In the meantime basic start and stop functionality exists in our command deck so if you create a VM in TrueNAS it will show up for basic controls in our UI.

25.10 is expected to go from Beta to General Release this month so that's good news.

[Mascot]

13 hours ago, jonp said:

To be fair, we actually did fix the issue referenced in that post.
[...]
One major change they made actually requires users to go through a migration process

My Immich was warning about the deprecated storage configuration before then, and it still is now. I've seen no new notice of migration needs since then. That was always the issue, it's not a new one. There was a separate issue with people who did not attempt to update for a long time after HexOS launched, and were then unable to do so. Is that what you are referring to? 

I get that Immich was beta, and that HexOS is still beta. But you charge money for HexOS and Immich was supposed to be one out of a whopping two applications you claimed to offer some level of support for. I didn't have expectations of extreme levels of support, but I did expect _something_ in a timeline less than several months.  But, as far as I've seen over the months, absolutely nothing was done. The only thing that's happened since that post, for me, is that after months of warning about the forthcoming deprecation, it was finally implemented and Immich will no longer update.

Over those months there have been a number of posts by people with issues both installing and updating Immich, with no indication anything had changed. No responses of "we fixed this, here's how to implement it," as far as I've spotted. Was the Q3 update even supposed to support automated migration to the new storage structure? Or are you saying you implemented something prior to Q3? How? HexOS did not support app updates before then. Was something supposed to have been deployed silently? If a solution that required an action from me was implemented, why did not HexOS notify me of this over these months? Did I miss a bunch of forum posts? The only solutions I've seen have been community posts working through how to implement the manual migration tasks as indicated in the deprecation notice.

I really have no idea what you attempted to do, or how to even have noticed it didn't work. And I have no idea what you have tried to do now with the Q3 update. All I know is that there's now an update button in HexOS, that appears to do exactly the same thing triggering an update in TrueNAS does, except HexOS takes hours to realize it has failed whereas TrueNAS reports it in a second or two.

So, from where I'm sitting, nothing was done and what you're now saying, after months of silence, seems to be "this is news to us." I hope you can excuse my less than enthusiastic response to that, as misinformed as I might be.

[Krizly]

4 hours ago, Mascot said:

...So, from where I'm sitting, nothing was done...

Not affiliated with HexOS, but here's a quick analogy to explain what's actually going on behind the curtain for a layman.

HexOS are assembling a car, which is being built on an existing Chassis (TrueNAS) that has over the past 12 months had some pretty significant structural changes. While the changes are good and will make the final car better, it means the development of the car is complicated because the way the seats bolt in, and the size of the windscreens are changing while they're assembling the car.
HexOS said 'This car is still being built, you can use it, but if you need to drive for your job you should wait until HexOS is finished before selling your old car'.
HexOS are in the process of upgrading the driving wheel, figuring out A\C, and putting in the car seats, which keeps them pretty busy. While the car drives, it's not really a production car just yet.
The radio manufacturer decided to change the wiring requirements all of a sudden with the new model they released, so HexOS need to tweak the wiring so the radio keeps working.
All of this is happening while thousands of people are driving the car, and asking when the A\C will work, or when the new seats will be available.

You can fix it by taking the radio out and putting it back in yourself (re-installing Immich)
Or you can wait until they have time to fix the wiring in the back-end, which they will do eventually. A car without a radio still drives.

Again, this is a beta. Stuff will break, and until they reach 1.0, expecting everything to work 100% and non-critical issues to be fixed quickly is just a recipe to be disappointed.

[jonp]

18 hours ago, Mascot said:

So, from where I'm sitting, nothing was done and what you're now saying, after months of silence, seems to be "this is news to us." I hope you can excuse my less than enthusiastic response to that, as misinformed as I might be.

😞 I totally get this disappointment. I don't even blame you. We are targeting a more casual user and I can see how users like yourself may have had a higher expectation closer to that of a video game beta, where it's really software that's functionally complete, but just testing integrity of the infrastructure for multiplayer and game balance. You might find a bug or two, but nothing too crazy because part of the purpose of the beta is to build hype for the game prior to release, as we just saw with Battlefield 6 Open Beta. 

All of that said, we did say this was beta and listed out all the risks in the very first post you read to get it downloaded and installed. Furthermore, we actually had a second fix in for th migration issue for Q3 but due to another immich update, it broke again literally the day before the Q3 release and we missed it because QA was the week prior. This was just a very unfortunate coincidence. 

It seems like you specifically may be very frustrated with us and not happy with your experience. If you would like, I will refund your purchase. We don't want anyone feeling raw about their transaction, so if you want to take me up on that, feel free to drop me a DM. 

We are a small but growing team and will speak to our plans to get even better about this for post 1.0. We will be building an Apps team that will be responsible for ensuring our curations remain solid and that we handle these types of issues better.

13 hours ago, Krizly said:

Not affiliated with HexOS, but here's a quick analogy to explain what's actually going on behind the curtain for a layman.

This analogy was spot on. Appreciate you taking the time to lay that all out. 

[Mascot]

13 hours ago, Krizly said:

Not affiliated with HexOS, but here's a quick analogy to explain what's actually going on behind the curtain for a layman.

I can't claim to be a layman in this, given I'm now nearing 30 years as a professional developer. I am, very intentionally, approaching this from a "regular customer" point of view, instead of as someone who is comfortable with a command line.

17 minutes ago, jonp said:

😞 I totally get this disappointment. I don't even blame you. We are targeting a more casual user and I can see how users like yourself may have had a higher expectation closer to that of a video game beta, where it's really software that's functionally complete,

That used to be what beta meant; feature complete but lacking content and polish. Then everyone started releasing alphas and calling them beta. At this point I suppose I'm quickly becoming the unreasonable one in having any expectations whatsoever of a beta, given there's a generation of developers now that have never known a beta to be anything but "some random stage between first successful compilation and full release".

I hold you at no fault at all for the unfortunate timing of the Immich 2.0 release. It's the lack of communication and/or action for the months after the deprecation warning appeared, I'm expressing disappointment in and suggesting you need to reconsider your approach to. It's not the first time. A similar laid back attitude was present when all updates broke due to a TrueNAS update being required, that no one knew if was safe to perform or not. I think everyone would benefit greatly from a more proactive approach.

My main point is that you have an ideal information channel to reach your customers: the HexOS interface. If an app is broken, make it show there. If manual action is needed, let the users know there. If there was no mechanism for doing that early on, perfectly understandable. But it should have become apparent long ago now that the information flow has not been working all that well. It does not seem unfair to say that it's not good form to leave an app broken for months without any indication in the interface that lists is as a recommended app. This should not be a thing nearly a year into selling the product, beta or otherwise.

This is very much intended as criticism of the constructive kind. I just want to see improvements. That you are responding to this kind of criticism in a positive manner, is great. But, I haven't seen much customer-facing change over the months, which I find worrisome and I genuinely feel you (as a company) should as well.

As for a refund, I appreciate the offer but it's not enough money to be worth an evening or more worth of hassle in reinstalling "just TrueNAS". I reserve the right to continue to be grumpy and disillusioned, but hopefully I'm succeeding in limiting my sharing to an endurable level. 🙂

[Maximus]

Have you considered pulling out the Immich update notification from the release? Or at the very least, posting announcements in-app an pinned on the forum about the issue, what users should  be doing, what is your plan.
I have had to scan multiple threads and posts to come to the conclusion that I should probably just wait, not try to update Immich, and that you might be sorting this out in a future release.

What I know from the team is that you are working on the issue, it's not your fault, but no guidance 😕

[jonp]

2 hours ago, Mascot said:

This is very much intended as criticism of the constructive kind. I just want to see improvements. That you are responding to this kind of criticism in a positive manner, is great. But, I haven't seen much customer-facing change over the months, which I find worrisome and I genuinely feel you (as a company) should as well.

As for a refund, I appreciate the offer but it's not enough money to be worth an evening or more worth of hassle in reinstalling "just TrueNAS". I reserve the right to continue to be grumpy and disillusioned, but hopefully I'm succeeding in limiting my sharing to an endurable level. 🙂

Appreciate the criticism and we take it in stride. I respect everything you and others post, both good and bad. Sometimes raw emotional responses can cause gut check reactions of "you're kicking my baby" but I don't let myself see it that way. You're obviously passionate about our project or you wouldn't be here pointing out all the ways in which we can do better.  I also respect not wanting a refund and I look forward to the day when I see a post from you saying, "ok, this was totally worth it and I'm glad I bought."

2 hours ago, Mascot said:

My main point is that you have an ideal information channel to reach your customers: the HexOS interface

💯 Agreed. Our CTO has a proposal for us this quarter to do exactly what you're talking about. It's premature to share but know that we are on the same page and intend to do this. We intentionally built the global HexOS messaging system for this back with the Q2 release. We now want to expand that to include notifications for app-specific issues like this. But keep in mind that was not the original plan for 1.0 so the scope is no increasing, which is fine, but we still have other projects and objectives to hit, one of which is the Local UI which is a much bigger project than I think anyone realizes. It's complicated by the fact that our actual footprint on the ISO today is very small and we do not want to bake any of our UI into that ISO either. Instead, we have to build a full container/app for this which we are doing, but it's a huge project and we have a senior dev completely focused on it right now. 

I feel like a broken record saying this, but a big part of our struggles this year is bootstrapping the business. Money doesn't solve everything, but it sure as heck would have helped to hire a lot more developers and app specialists right at the beginning of this year, but we opted for a more fiscally conservative approach to hiring to ensure we don't grow too big too fast and make commitments to employees we can't keep without proving our business model first. Expansion is a big part of our plans for 2026 and will help us get more focused on individual layers of the software. 

1 hour ago, Maximus said:

Have you considered pulling out the Immich update notification from the release?

Sure, but that comes with its own set of challenges and problems. Again, everything is still in flux and user data is not at risk. The worst anyone should have to deal with is reconfiguring immich from scratch which of course isn't ideal, but let's be honest here, it's also not the end of the world or an unrealistic expectation given the beta nature of both us and Immich itself. 

Comms of course could be better, always. But we are doing the best we can with what we have for now and we would just ask people to be patient while we work through it. This was all part and parcel of participating in the beta and if that's not going to work for anyone, waiting to deploy until more polish is applied would be the right approach. 

I've already spent hours reading and replying here and the more we try to get ahead on comms, the less time we spend on building and getting to a release. Every hour is a choice and often there are no options that will satisfy everyone. Still totally understand the frustrations, but this is the last I'll be commenting on this today so I can get back to work with the rest of the team. 

Our new docs site is a great place for us to put a guide for immich issues and resolving them in the future. Stay tuned for more, but know that we don't have the ability to always drop everything we are doing to shift focus to an issue like this when we are in the middle of other projects. There can be a lag time and that is something everyone here will have to accept while participating in the beta. Again, if that doesn't work for anyone, wait for the polish to come before deploying. 

[Cantos]

@jonp I work in software development so all of this is painfully familiar. When I saw the Immich struggles, I immediately thought "I bet they had a solution ready and it was obliterated by one of those updates".

1 hour ago, jonp said:

Sure, but that comes with its own set of challenges and problems.

That reminds me of the classic: "There are no solutions, only trade-offs." 

Any half-decent team of developers is capable of making work almost any feature and solving almost any issue. If you are thinking "couldn't they just do X?", the answer is that they probably could. However, that would always be at the cost of something else being worse, broken or delayed... or it would take too much time, require too many people or cost too much money (usually at least two out of three).

[jonp]

6 hours ago, Cantos said:

@jonp I work in software development so all of this is painfully familiar. When I saw the Immich struggles, I immediately thought "I bet they had a solution ready and it was obliterated by one of those updates".

That reminds me of the classic: "There are no solutions, only trade-offs." 

Any half-decent team of developers is capable of making work almost any feature and solving almost any issue. If you are thinking "couldn't they just do X?", the answer is that they probably could. However, that would always be at the cost of something else being worse, broken or delayed... or it would take too much time, require too many people or cost too much money (usually at least two out of three).

This man devs!

Not sure I could of said any of it better myself. 

[Norrah]

These Apps are essentially pushing the beta to a state where I’m ready to play with it. But I do have a few questions.

My end product is a low-power SSD NAS with home assistant and Jellyfin, maybe plex.

I do however need to understand how these “apps” work. Where are they stored, and how native their access is the to hardware resources?

I was considering running the OS on a 12 GB N150 from a 64gb eMMC drive, but that might differ if the apps require full VM’s behind the curtains. So what are apps exactly?

In terms of reliability… let’s assume the home assistant app crashes, how do I maintain/reboot etc. the home assistant, what are the plans there?
 

[Adam108]

Really excited to see some new apps coming into the curated list - I'd like to move my existing installs across to the "HexOS version" as and when they become available but I can't see an obvious way to do this? I can't install a second version of an app which is already installed via TrueNAS and there's no apparent way to "migrate" an app into HexOS. 

Is the current intention that we remove entirely the app versions already installed and start over via HexOS or am I missing something?

[jonp]

2 hours ago, Norrah said:

I do however need to understand how these “apps” work. Where are they stored, and how native their access is the to hardware resources?

Apps are running in docker containers. They install preferably to an SSD pool but will default to whatever you have if only HDDs are present (performance may be impacted depending on the applications storage performance requirements and other concurrent usage). 

2 hours ago, Norrah said:

I was considering running the OS on a 12 GB N150 from a 64gb eMMC drive, but that might differ if the apps require full VM’s behind the curtains. So what are apps exactly?

The OS boot drive does not store the applications themselves or their configuration data. All of that is stored on one of your storage pools. The OS boot device stores the distribution itself and it's configuration settings. That's it. 

Apps do not run in full VMs. VMs would be useful for non-containerized applications (if you have any requirements for that) or non-Linux based workloads (e.g. Windows).  Not necessary for Plex or Jellyfin

2 hours ago, Norrah said:

In terms of reliability… let’s assume the home assistant app crashes, how do I maintain/reboot etc. the home assistant, what are the plans there?

Apps work similar to a VM from a control standpoint. You can start, stop, and update them independently from one another. And if an app goes completely haywire, removing it does NOT remove the associated user data (e.g. removing Plex/Jellyfin does not delete your content).

Hope this helps!